<%BANNER%>

Hydrologic and Economic Impacts of Alternative Residential Land Development Methods


PAGE 1

HYDROLOGIC AND ECONOMIC IMPACT S OF ALTERNATIVE RESIDENTIAL LAND DEVELOPMENT METHODS By EVAN SHANE WILLIAMS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2003

PAGE 2

Copyright 2003 by Evan Shane Williams

PAGE 3

This document is dedicated to my family.

PAGE 4

iv ACKNOWLEDGMENTS I thank my parents for encouraging me throughout my educati on. This dissertation is as much theirs as it is mine.

PAGE 5

v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...........................................................................................................viii LIST OF FIGURES.............................................................................................................x ABSTRACT......................................................................................................................x ii CHAPTER 1 OVERVIEW OF RESEARCH.....................................................................................1 Purpose of Research.....................................................................................................1 Relevance of Work.......................................................................................................1 Morphology...........................................................................................................3 Peak Flow Characteristics and Volume.................................................................3 Groundwater..........................................................................................................3 Water Quality........................................................................................................4 Stream Warming and Biodiversity........................................................................4 Current Storm-Water Management Practices...............................................................5 Detention Ponds.....................................................................................................6 Wet Ponds..............................................................................................................7 Retention Designs..................................................................................................8 Summary..............................................................................................................10 General Discussion of Project....................................................................................11 Structure of This Dissertation.....................................................................................14 2 DEVELOPMENT ALTERNATIVES........................................................................15 Introduction.................................................................................................................15 Literature on the Development and Storm-Water Management Alternatives............15 Traditional Development.....................................................................................15 Cluster Development...........................................................................................16 A related concept: planned development district.........................................17 Purported benefits........................................................................................17 Available literature.......................................................................................18 Potential negative factors.............................................................................19 Low Impact Development (LID).........................................................................19

PAGE 6

vi Purported benefits........................................................................................21 Available literature.......................................................................................22 Potential negative factors.............................................................................24 Applicable Design Standards......................................................................................25 Design standards and metric conversion.............................................................25 Basic Development Regulations..........................................................................26 Cluster Development Regulations.......................................................................29 Planned Development District (PDD).................................................................31 Summary.....................................................................................................................32 3 SITE DEVELOPMENT PLANS................................................................................33 Introduction.................................................................................................................33 Existing Conditions....................................................................................................33 Traditional Development Plan....................................................................................35 General Design Discussion..................................................................................36 Interpretation Issues With The County Ordinance..............................................39 Design Plans........................................................................................................40 Cluster Development Plan..........................................................................................42 General Design Discussion..................................................................................43 Interpretation Issues With The County Ordinance..............................................46 Design Plans........................................................................................................46 LID Developments......................................................................................................48 General Design Discussion..................................................................................49 Interpretations of The Ordinance.........................................................................52 Design Plans........................................................................................................54 Summary.....................................................................................................................56 4 HYDROLOGIC IMPACTS........................................................................................59 Introduction.................................................................................................................59 Discussion on Modeling the Alternatives...................................................................59 Continuous Simulation Data Collection.....................................................................60 Hydrologic Model.......................................................................................................61 Experimental Methods................................................................................................63 Existing Conditions Calibration..........................................................................63 Design Alternative Models..................................................................................68 Universal model assumptions.......................................................................68 Conventional storm-water management model notes..................................71 LID storm-water management model notes.................................................73 Modeling Results........................................................................................................75 Watershed Timing...............................................................................................78 Runoff Volume....................................................................................................80 Peak Flows...........................................................................................................81 Continuous Model...............................................................................................83 Structural Controls...............................................................................................85 Model Limitations......................................................................................................85

PAGE 7

vii Hydrologic Summary and Conclusions......................................................................87 5 ECONOMIC ANALYSIS..........................................................................................89 Introduction.................................................................................................................89 Review of Literature...................................................................................................90 Methods Of Economic Analysis.................................................................................91 Construction Cost Estimation..............................................................................91 Hedonic Price Equation for Pr edicting Sale Price...............................................94 Collection of sales data................................................................................95 Collection of open space data.......................................................................96 Development of lot descriptive variables.....................................................99 Results of Economic Analysis..................................................................................101 Hedonic Price Model Results............................................................................101 Descriptive statistics...................................................................................101 Correlation results......................................................................................102 Regression results.......................................................................................103 Impact of open space frontage...................................................................107 Construction Cost Results.................................................................................108 Impact to the Developer and Potential Investors...............................................109 Limitations................................................................................................................111 Economic Summary and Conclusions......................................................................113 6 DECISION SUPPORT SYSTEMS FOR LAND DEVELOPMENT.......................116 Introduction...............................................................................................................116 Literature Review.....................................................................................................117 Construction of the DSS...........................................................................................118 Sale Price...........................................................................................................119 Construction Costs.............................................................................................119 DSS Function.....................................................................................................120 Results.......................................................................................................................1 22 Limitations................................................................................................................123 Conclusions...............................................................................................................124 7 CONCLUSION.........................................................................................................125 APPENDIX A HYDROLOGIC MODEL INPUT............................................................................127 B CONTINUOUS MODEL AND CONS TRUCTION COST RESULTS..................139 LIST OF REFERENCES.................................................................................................143 BIOGRAPHICAL SKETCH...........................................................................................149

PAGE 8

viii LIST OF TABLES Table page 2-1 Summary of lot layout standards for R1-aa (ACBCC, 2002)................................27 2-2 Yard and setback standards for R1-aa zoning (ACBCC, 2002)............................27 2-3 Selected road design standards for Alachua County (ACBCC, 2002)..................27 2-4 Comparison of full-size and cluste r lot requirements (ACBCC, 2002).................31 4-1 Adjusted depression storage for selected sub-areas of the “full” LID design.......75 4-2 Uncontrolled model results for hypothetical design storms..................................76 4-3 Controlled model results fo r hypothetical design storms......................................76 4-4 Continuous simulation selected peak flow points and total simulation volume....79 4-5 Results of the continuous simulati on for selected base flow points......................79 4-6 Final volumes of management practices................................................................82 5-1 Capital improvements included in analysis...........................................................92 5-2 Descriptive statistics of lots in sales data set.......................................................101 5-3 Open space type and proximity breakdown.........................................................102 5-4 Correlation results between independent variables..............................................103 5-5 Initial regression results.......................................................................................105 5-6 Modified regression results..................................................................................106 5-7 Construction cost estimates..................................................................................109 5-8 Estimated project return.......................................................................................110 5-9 Conceptual profit in dollars.................................................................................111 6-1 DSS Output..........................................................................................................122

PAGE 9

ix A-1 Existing conditions model areas..........................................................................127 A-2 Traditional development cover conditions...........................................................127 A-3 Cluster development cover conditions.................................................................128 A-4 “Partial” LID development cover conditions.......................................................128 A-5 “Full” LID development cover conditions...........................................................129 A-6 Initial percent of storage full and sub-area at-large groundwater routing input.....................................................................................................................129 A-7 Evapotranspiration input for all models...............................................................129 A-8 Existing conditions soil moisture units................................................................130 A-9 Traditional development soil moisture units........................................................130 A-10 Cluster development soil moisture units..............................................................131 A-11 “Partial” LID development soil moisture units....................................................131 A-12 “Full” LID development soil moisture units........................................................132 A-13 “Partial” LID additional depres sion storage for runoff control...........................132 A-14 “Full” LID additional depre ssion storage for runoff control...............................132 A-15 Existing transform input......................................................................................133 A-16 “Partial” LID transform input..............................................................................133 A-17 “Full” LID transform input..................................................................................133 A-18 Traditional transform input..................................................................................134 A-19 Cluster transform input........................................................................................134 A-20 Traditional development wet ponds 1 to 5...........................................................135 A-21 Traditional development wet ponds 6 to 8...........................................................136 A-22 Cluster development wet ponds 1 to 3.................................................................136 A-23 Cluster development wet ponds 4 to 9.................................................................137 A-24 LID development conceptual dry detention ponds..............................................138 B-1 Itemized construction costs (not adjusted for Gainesville)..................................141

PAGE 10

x LIST OF FIGURES Figure page 1-1 Conceptual diagram of solution field.......................................................................13 3-1 Existing conditions plan...........................................................................................36 3-2 The traditional de velopment plan.............................................................................41 3-3 The traditional development plan w ithout topography, soils or drainage sub-areas...................................................................................................................42 3-4 Areas of high infiltration potential...........................................................................45 3-5 The cluster development plan..................................................................................47 3-6 The cluster development plan without topography, soils or drainage sub-areas......48 3-7 “Partial” LID development plan...............................................................................55 3-8 “Full” LID development plan...................................................................................56 3-9 “Partial” LID development plan w ithout topography, soils or drainage sub-areas...................................................................................................................57 3-10 “Full” LID development plan without topography, soils or drainage sub-areas......58 4-1 Existing conditions sub-area division......................................................................64 4-2 Comparison of observed flow s versus calibrated flows...........................................67 4-3 Traditional design sub-areas and storm-water system.............................................72 4-4 Cluster design sub-areas and st orm-water management system..............................73 4-5 “Partial” LID design sub-areas.................................................................................74 4-6 “Full” LID design sub-areas.....................................................................................74 4-7 2-year, 24-hour c ontrolled hydrographs...................................................................76 4-8 2-year, 24-hour uncontrolled hydrographs...............................................................77

PAGE 11

xi 4-9 25-year, 24-hour controlled hydrographs.................................................................77 4-10 25-year, 24-hour uncontrolled hydrographs.............................................................78 6-1 Conceptual solution set revi sed from hydrologic analysis.....................................119 B-1 Traditional development continuous model results...............................................139 B-2 Cluster development continuous model results......................................................139 B-3 “Partial” LID development continuous model results............................................140 B-4 “Full” LID development continuous model results................................................140

PAGE 12

xii Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy HYDROLOGIC AND ECONOMIC IMPACT S OF ALTERNATIVE RESIDENTIAL LAND DEVELOPMENT METHODS By Evan Shane Williams December 2003 Chair: William R. Wise Major Department: Environmental Engineering Sciences This research analyzed the hydrologic and economic impacts of four alternative site planning and storm-water management designs in a hypothetical residential development in the Gainesville, Florida area. The four development options analyzed were as follows: Traditional development with full-si ze lots and conventional storm-water management. Cluster development with reduced lots for upland preservation and conventional storm-water management “Partial” Low Impact Development (LID) that implemented the LID storm-water management system on the full-size lot plan “Full” LID that implemented the LID storm-water management system on the cluster development plan. LID is an emerging method of site planni ng and storm-water management that has been presented by Prince Georges County, Ma ryland. This approach to land development combines a distributed, infiltration-based storm-water management system and reduced

PAGE 13

xiii area of disturbance. Limits on disturban ce were implemented through the use of the cluster development site plan. A hydrologic analysis was conducted. The 2and 25-year regulat ory design storms and a continuous simulation were modeled. The results were compared to the existing conditions and each other. The results showed that the distributed, infiltration-based storm-water management system provided a wa tershed response closer to the natural, existing response, particularly if combined with a site plan that limited disturbance. The economic effects of the f our alternatives were analyzed by estimating selected construction costs and the impact of lot si ze, open space proximity and type on the sale price of vacant lots. The hedonic price techni que was used to analyze the impact on sale price. The economic analysis showed that reducing lot size would adversely impact the profit from development. This impact coul d be mitigated somewhat by maximizing open space frontage. Construction costs for the LID designs were lower than the designs using conventional storm-water management. The comb ined effect of construction cost savings and sales receipts indicated that the ratio of profit to cost was hi ghest for the LID stormwater management system combined with limiting disturbance.

PAGE 14

1 CHAPTER 1 OVERVIEW OF RESEARCH Purpose of Research This research evaluated the effects of a lternative methods of land development and storm-water management on the hydrologic re sponse of a first order stream and associated wetlands. In add ition to hydrology, economic conc erns for the developer and potential investors are also explored. The goal of this dissertation was to gain in sight into the applicability of alternative land development practices in Florida and th e United States in general. The specific questions considered were whether alternative land development and storm-water management improve the hydrologic response of a developed site and what economic impact, positive or negative, results from implementation alternative practices. Improvement in the hydrologic response was defi ned as a response that is closer to the natural response of the system than resulted from conventional storm-water management practices. Relevance of Work The obvious question in regards to this di ssertation is: why should alternative land development and storm-water management prac tices be considered? An excellent answer is provided by Prince Georges County, Mary land. The Department of Environmental Resources of this County is responsible for the development of one of the alternative practices, Low Impact Development (LID), disc ussed later in this dissertation. The work done by this County was based on the conclusi on that current, generally accepted, storm-

PAGE 15

2 water management practices do not adequately protect water resources (Prince Georges County, Ch.1, p. 5, 1999). The hydrologic effects of watershed de velopment on water resources are well known. Peterson (1999), in a discussion on wate rshed restoration, points out that all landforms in a watershed have hydrologic f unction and that degradation or alteration impairs natural watershed function. Leopol d (1968) described four effects of development: Water quality and amenity value may change. Changes in peak flow usually appear as in creased flow rate and occurrence at an earlier time. Total runoff changes due to increases in runoff volume from loss of infiltration capacity and natural storage (i.e., depressi on areas) for both for individual storms and in the long-term record (increase in runoff frequency). At the same time as runoff volume is increasing, baseflow volume may also decrease due to infiltration capacity losses. A similar discussion of effects is also presented in Urban Runoff Quality Management (Water Environment Federation [WEF ] and American Society of Civil Engineers [ASCE], pp. 24-25, 1998) and includ es all aquatic ecosystems (wetlands, lakes, estuaries, etc.). It is notable that this book adds changes on stream morphology as an impact of urbanization. Somewhat less publi cized but also of concern are increases in stream temperature (Schueler, pp. 26-27, 1995) and losses of habitat and biodiversity as examples of stream degradation caused by urbanization (Schueler, pp. 28-31, 1995; WEF and ASCE, p. 26, 1998). Schuler relates all of these impacts to the degree of imperviousness of the watershed.

PAGE 16

3 Morphology A relatively low degree of watershed imperv iousness, as low as 10%, can result in stream channel instability (Schueler, pp. 23-24, 1995; WEF and ASCE, p. 25, 1998). Schueler (p. 24, 1995) states that habitat degradation due to stru ctural change is a primary effect of stream morphological changes and ci tes studies showing that habitat degradation begins to at about 10% imperviousness. Peak Flow Characteristics and Volume Changes in peak flow and discharge vol ume are perhaps the best-known hydrologic impacts of urbanization. The influence of impe rvious cover on peak flow is quite well known. Changes in the peak flow characteris tics generally manifest themselves as increases in peak flow, corre sponding to a higher flow stag e. Increases in volume are caused by impervious cover but also by soil comp action in pervious areas that are altered or graded during construction, thus reduc ing infiltration poten tial (Schueler, 2000). In addition to increased peak flow from in creased volume, peak flow is also altered from changes in watershed timing effects (tim e of concentration). Peak discharge from a developed watershed may be earlier than from the same watershed in its natural condition (Viessman and Lewis, p. 349, 1996) Groundwater Groundwater may also be adversely impacted by urbanization. The general theory is that impervious cover prevents infiltrati on of rainfall. This lost infiltration is not replaced by conventional storm-water manageme nt system. This leads to reductions in base flow. The impact of development on gr oundwater, more properly base flow to the stream, appears to be an area of variab le impact. Several studies have shown contradictory impacts of urbanization on base flow (WEF and ASCE, p. 11, 1998).

PAGE 17

4 Water Quality Recently, water quality impacts of stor m water runoff have received much attention; most often focused on pollutants, typically non point-s ource, such as total suspended solids (TSS), total nitrogen (TN), total phosphorus (TP), nitrates and heavy metals. One important study characterizing po llutants in runoff was the National Urban Runoff Program Survey (NURPS) study done by the United States Environmental Protection Agency. This study was rel eased in 1983 and is summarized in Design and Construction of Urban Stor mwater Management Systems (Urban Water Resources Research Council [UWRRC], p. 101, 1992). Recen tly, Smulen et al. (1999) “updated” the initial NURPS estimates for various pollutants. The specific concern for de veloping areas is that po llutants originating from developed areas and transported by runoff will be harmful to the receiving water. Schueler (pp. 24-26, 1995) explains that imperv ious surfaces in urban development are a key part of the water quality problem. These surfaces act as collection areas for pollutants. These pollutants are susceptible to easy wash-off during rain events. Marsalek et al. (1999) concluded that r unoff from urban land use showed at least potential toxicity 40% of the time. In addition, the U.S. E nvironmental Protection Agency (USEPA, 1996) has explained the impacts of various pollutant s. The EPA has indicated that urban runoff is a source of degradation for streams, lake s, estuaries and wetlands in the U.S. with impacts to estuaries and wetlands particular ly troublesome, number 2 and 3 impacts, respectively (USEPA, 1996). Stream Warming and Biodiversity Site Planning for Urban Stream Protection (Schueler, pp. 28-31, 1995) provides a discussion and summary of key findings fr om 17 studies that have shown that

PAGE 18

5 biodiversity is impacted by ur banization. Most of the cited studies deal with benthic communities. Schueler concludes based on the cited literature that the point where negative impacts on diversity of aquatic in sects occur was indeterminate, but fish communities were affected at about the same level of imperviousness as stream morphology. A study is also cited that showed that stream temperature can also be expected to rise with the amount of im pervious cover (Schueler, pp. 26-27, 1995). Current Storm-Water Management Practices In order to appreciate the relevance of the current re search, the performance of current storm-water management practices mu st be briefly discussed. The practices described here are mitigation methods, meaning that they seek to fix a current or future problem. Typical site planni ng practices, with the maximum number of lots possible at the regulatory minimum lot area for th e zoning district, do very little to prevent runoff. The current management approach towa rds runoff management is generally directed at removal of runoff from the developed area to the receiving water as quickly as possible (CH2MHill, Sec. 3, p. 4, 1998). This reflects a long-standing engineering attitude that water in deve loped areas is bad—unless it ha s aesthetic or recreational benefit. Dion (p. 236, 1993) has an amus ing quote attributed to an unnamed civil engineering professor who stated: There are three cardinal rules to reme mber when dealing with a construction project: 1. get rid of the water, RULES 2 AND 3. Get rid of the water. In most cases of residential development, the site is graded for rapid drainage with a transport system that is usually a highly e fficient storm sewer system (Coffman et al., 1998, 2000). These provide little or no atte nuation of peaks, volume reduction or pollutant removal.

PAGE 19

6 Local regulations, however, generally do require mitigation practices to address some of the effects of development. Contro l of hydrologic effects generally focuses on peak flow control. For example, the St Johns River Water Management District (SJRWMD, Sec. 9, pp. 8-9, 2001) requires contro l of peak flows for the 2 and 25-year, 24-hour storm events. Further regulations, such as volume control, are enforced only for special regulation areas. Alac hua County has the same peak flow requirement but does not specify storm duration for the storm event (Alachua County Board of County Commissioners [ACBCC], 2002). In Alachua County and the SJRWMD (these agencies hold concurrent jurisdiction on storm-water management), water quality must also be addressed (ACBCC, 2002; SJRWMD, Sec. 8, p. 2, 2001). The intent is to cat ch the volume of r unoff associated with the “first flush” of pollutants often observ ed in runoff. These regulations, while not necessarily as representative of typical practice as peak fl ow controls, do represent a significant concern for the effects of polluta nts in runoff. In most cases, only one management practice is required—that is, a single end-of-pipe pr actice per sub-basin. The following sections discuss some comm on practices used today. There are two focuses: water quantity (volume, peak flow) control and quality control. Drawbacks and benefits in each area are discussed. Detention Ponds Control of runoff in residential projects is frequently accomplished through some form of detention pond system—one of the ol dest storm-water management practices. These ponds are designed to store runoff temporar ily and release it at a rate not exceeding the pre-development peak flow for a given storm. Sloat and Hwang (1989) showed in a study on performance of detention basins that su ch basins only control peak flows at their

PAGE 20

7 outlet and that downstream areas still suffered increased peak flows. The cause of this under-performance is from the lack of volume control provided by detention systems. Instead of a near instantaneous peak flow the increased volume results in a peak (although no higher than pre-development cond itions) that is extended over a longer time, thus causing increased erosion. This may allow for the flow peak to intersect with other peaks from tributary streams. Additionally, th e timing of the peak flow can be delayed from its natural time, which also may allow it to coincide with the peak of the flood wave from upstream areas (Sloat and Hwang, 1989). Roesner et al. (2001) also acknowledge concern for drawn out periods of high flow s and add that the design storms typically required frequently do not control peak flows fr om events with shorter return intervals. Roesner et al. (2001) conclude that design requirements are at fault for poor hydrologic performance of detention practices rather th an the concept of dete ntion storage itself. Although they acknowledge this view is not shared by other authors. Certain detention pond designs, particular ly dry ponds that do not provide “extended” detention, have a poor record for removing pollutants. If the dry pond provides “extended” detention, the performan ce is somewhat better due to sedimentation and can be quite high, but only for particul ate pollutants (Field et al., p. 192, 1993; Schueler et al., pp. 7-13, 1992). The SJ RWMD (Sec.10, p.1, 2001) does not recommend dry detention pond systems, unless no other option is feasible. Wet Ponds Wet ponds are sometimes referred to as re tention systems (Dunn et al., 1995; WEF and ASCE, p. 220, 1998), but designs in Florid a perform as detention systems with a permanent pool for water quality. Retention syst ems in Florida, as the name implies, are designed to prevent all or a certain volume of runoff from being released into the

PAGE 21

8 receiving water, instead allowing for infiltrati on into the soil or eva poration loss. If the wet pond is designed as a detention system, flood control and water quality volumes are temporarily stored above the invert of th e outlet structure (SJRWMD, Sec. 14, pp. 1-11, 2001). In a detention configuration, the effect on volume and peak flow is essentially the same as a dry pond (unless, of course, the permanent pool needs to be recharged, in which case the pond functions in a retention role). Removal of pollutants is much better th an dry detention ponds. Schueler et al. (1992) report sediment removal rates from 50-90% and phosphorous and soluble nutrients from 30-90% and 40-80%, respectively. Other sources agree with this positive assessment of wet ponds (Field et al., p. 192, 1993; WEF and ASCE, p. 220, 1998; SJRWMD, Sec. 14, p. 1, 2001). Borden et al. (1 998) concluded that pond efficiency was seasonally variable, and noted that the ma in removal process appeared not to be sedimentation, since most pollutants did not appear associated with total suspended solids, and that biological processes accounted for most of the nutrient removal. The permanent pool depth is limited in Fl orida to prevent th ermal stratification, which is assumed to decrease pollutant rem oval, specifically the release of nutrients under anaerobic conditions (SJRWMD, Sec. 14, p. 8, 2001). Borden et al. (1998), however, observed that two ponds in North Caro lina with average depths less than 2.5 m (8.2 ft), an average depth consistent with the SJRWMD guidelines, still stratified. Furthermore, the study conclude d that this stratification mi ght actually enhance pollutant removal in certain situations. Retention Designs Retention, or infiltration, systems are desi gned not to maintain a permanent pool of water, but to infiltrate all or a portion of runoff. These systems are the only systems that

PAGE 22

9 the St Johns River Water Management Di strict calls “retention.” The SJRWMD considers a variety of systems as retention including large, flat-bottomed basins, shallow landscaped areas, pervious pavements and sw ales with some form of inlet block (SJRWMD, Sec. 11, pp. 1-5, 2001). These manage ment practices provide quality control and peak flow abatement through volume reduction determined by water quality requirements and detention storage beyond the wa ter quality volume. That fact that these systems provide volume reduction suggests that they would be an improvement, hydrologically, over detention systems. Howe ver, some retention designs suffer some serious drawbacks. Infiltration basins are not universally r ecommended since they can fail rapidly, and generally cannot be restored through regular maintenance. However, this may have been directed to the Mid-Atlantic region (Schueler et al., pp. 4950, 1992). If the basin should fail, runoff quantity and quality control aspect s will, at best, be similar to a pond system. As was previously stated, these basins are permitted in Alachua County as long as certain steps are taken to reduce the possibility of failure (SJRWMD, Sec. 11, pp. 4-5, 2001). Off-line systems that allow flows to by-pass when the retention basin is full (volume is designed to capture the “first-flush”) have had greater success in Florida (UWRRC, p. 499, 1992). Assuming that the infiltrati on basin functions as inte nded, Schueler et al. (pp. 4753, 1992) suggests that pollution removal should be high for particulates and moderate to low for soluble pollutants depending on soil charac teristics. There is a slight risk of groundwater contamination. If the pond should fail, water quality treatment may still be provided, and the basin may be retrofitted to a wet pond.

PAGE 23

10 Other infiltration practices include the infiltration trench and sand filter. These are not common in residential projects in Florid a. In fact, the “exfiltration trench” design guidelines provided by the SJRWMD notes th at these are typically built in downtown areas where space is limited (SJRWMD, S ec.13, pp. 1-6, 2001). Schueler et al. (p. 49, 1992) suggested that infiltration trench performance, at least in the water quality sense, should be similar to infiltration basins. They can serve areas of 2 to 4 ha (5 to 10 Ac) (UWRRC, p. 501, 1992). Sand filters differ from infiltr ation trenches in that th e filter medium is sand, and some designs do not allow filtered runoff to be infiltrated (Claytor and Schueler, Ch. 1, pp. 4-6, 1996; CH2MHill, Sec. 8, 1998). In te rms of pollutant removal, Claytor and Schueler (Ch. 4, p. 33, 1996), provides information for various types of sand filters showing high removals for particulate matter, moderate levels for total phosphorus and nitrogen but poor removal for nitrate/nitrites. Experience with infiltration trenches has shown that while sometimes having good pollutant removal capability, th ey require frequent maintena nce, which is usually not done (Schueler et al., p 44, 1992; UWRRC, p. 501, 1992). A variety of designs exist (Claytor and Schueler, Ch. 1, pp. 4-6, 1996) with surface types accompanied by grass filter strips being the most typical if us ed in residential areas (UWRRC, p. 501, 1992). Summary The conclusion that can be drawn from th e discussion above is that typical stormwater management using conventional end-of -pipe, single practice systems does not always provide the level of protection of water resources that is desirable—although some practices are better than others. Particul ar designs may provide adequate controls in certain areas of runoff manage ment but not in others. For example, wet ponds provide

PAGE 24

11 good pollutant control but flow control suffe rs the same drawbacks as dry detention systems. There is a growing realization that changes are needed in storm-water management practices. A variety of interesti ng approaches are availa ble in literature and are discussed in the next chapter. General Discussion of Project The 246.29 ha (608.66 ac) hypothetical property under development was located in the Camp Blanding Wildlife Management Area that is part of the Camp Blanding Florida National Guard Training Center. The site is located outside of Alachua County (about 35 miles Northeast of Gainesville) because of the desire to locate a site that was relatively undisturbed and not in private ownership. St ream stage data we re collected on the upstream side of a culvert crossing an unimpr oved road. Rainfall data were also collected with a recording gage located on the site. These data were used to calibrate an existing conditions hydrologic model that was used to evaluate impacts of development. The area surrounding the stream was consid ered a property undergoing residential development. The hypothetical property wa s “developed” with a scenario based on allowable density of housing units. The s cenario chosen allowed for minimum 1,857.88 m2 (20,000 SF) lots, the R1-aa zoning classi fication of the Alachua County, Florida, Ordinance (ACBCC, 2002). Th e hypothetical property was then “developed” for design options: Traditional development (full-size lo ts, pipe/pond storm-water management) Cluster development (half-size lots pipe/pond storm-water management) “Partial” Low Impact development (LID storm-water system, full-size lots) “Full” Low Impact Development (LID storm-water management, half-size lots) Low Impact Development (LID) is a de sign strategy developed by Prince Georges County, Maryland, which emphasizes protection of vital hydrologic features and replaces

PAGE 25

12 typical storm-sewer system with a distributed, infiltration-based system. The cluster development option, focused on land preservati on only. Each of these alternatives was analyzed for compliance with hydrologic crite ria. The basic criteria for hydrologic evaluation are discussed in detail in Chapter 2, but generally follow a typical storm-water management study (i.e., meeting regulatory c oncerns). In addition, each alternative was evaluated for how closely it maintained the natural hydrologic response. Each alternative is compared against the existing conditions of the site. The hydr ologic analysis is discussed in Chapter 4. Those who are familiar with the economic s of land development will note that the alternatives that reduce lot si ze will have an economic impact to the developer. Thus, a second set of analyses was required that dea lt specifically with the economics. A market study was done to measure the impact of re duced lot size and the effect of open space preservation. In addition, m easurement of selected capital improvements was done for each alternative. The criterion for economic evaluation was that the alternative designs should have little adverse impact to the devel oper or potential invest ors. The methods of evaluating impact were a comparison of dollar receipts (sales receipts minus construction costs) and the ratio of receipts to construction costs. Altern atives 2, 3 and 4 are compared against alternative 1, which re presents traditional development. Economics are discussed in Chapter 5. Naturally, there was no guarantee that a ny of these solutions would meet the economic criterion. The alternatives that were tested essentiall y represent the corners of a solution field. Figure 1-1 illustrates this concept. The storm-water management axis represents the degree that the system tends toward a particular approach, relative to

PAGE 26

13 increasing lot size. This is “measured” by the degree that traditional features of stormwater management, such as detention ponds or storm inlets, are present. It was expected that a portion of this solu tion set would be eliminated by the hydrologic analysis. In order to determine the optimal solution ec onomically and in terms of storm-water management system, a decision support system (DSS) was constructed, and is discussed in Chapter 6. Figure 1-1. Conceptual diag ram of solution field. The plans that were developed for this re search can be best described as “sketch plans” that meet land development guideline s (setbacks, lot dimensions, etc.) but are missing some level of detail that would make them complete. The most visible example of this is grading. The lots and streets are not graded in detail. They are “rough” graded to provide runoff flow direction (i.e ., to a detention basin) only. Commercial and industrial development is no t addressed in this research. In some respects, with residential projects, it is easie r to conduct an analysis such as this since

PAGE 27

14 development often occurs over a larger site while commercial areas are generally small parcels. Structure of This Dissertation This chapter contains introductory inform ation. The chapters that follow discuss each of the three components of the pr oject starting with hydrology, followed by economics and finishing with the decision suppo rt system example. There is a certain degree of overlap among chapte rs, and certain site design de cisions, discussed in chapter 3, were made were based on a principles, for example the economic influence of open space, that are explained in later chapters, such as Chapter 5 on economics.

PAGE 28

15 CHAPTER 2 DEVELOPMENT ALTERNATIVES Introduction This chapter reviews the literature on th e four design alternatives that were introduced in the previous ch apter. Land development and storm-water management are regulated activities. Thus, it is necessar y to review the governing regulations. The applicable design guidelines in the Alachua County Code of Ordinances and SJRWMD Applicant’s Handbook: Regulation of Stormwater Management Systems are reviewed in this chapter. Literature on the Development and St orm-Water Management Alternatives Alternative methods of storm-water manage ment have been proposed to alleviate the burden placed on water resources by developm ent activities. Proposed alternatives are not limited to engineered storm-water systems, but also include changes in the way sites are designed and developed. This section re views the literature on each of the design alternatives and presents the rationa le for inclusion in this research. Traditional Development The traditional development option reflects the minimum standards required by local ordinance. Local government entities with the power to review and approve new development require minimum standards of de sign for both the site plan and storm-water management system. In the case of site desi gn and storm-water infrastructure, these are usually physically based standards, such as lot size, lot dimensions, storm sewer pipe material and allowable lot density. The runo ff control portions of the ordinance are

PAGE 29

16 performance based. Typically, these require that a certain discharge may not be exceeded for one or more theoretical design storm events. It is often the case in the land development business to design a new devel opment to meet the minimum standards and nothing more. With this in mind, the rationa le for including this approach to land development in this research is obvious. The applicable design guidelines for this alternative are the Alachua County Code of Ordinances (Alachua County Board of Count y Commissioners [ACBCC], 2002) and the St. Johns River Water Management District (SJRWMD) Applicant’s Handbook: Regulation of Stormwat er Management Systems on the regulation of storm-water management systems (SJRWMD, 2001). S ite design standards are set by Alachua County, but both entities have concurrent jurisdiction on storm-water management. Cluster Development The cluster development concept is an element of most land development ordinances, including Alachua County. Using th e cluster development option allows the developer flexibility in lot size, while still guaranteeing at least the same number of lots as would be available if th e regular zoning requirements were followed. In return, the developer provides “open sp ace” (Schueler, pp. 55-56, 1994). In the context of this research, the cluster development option is significant because it may be used to lessen the impacts of development on natural hydrologic features, which is a relatively new application of cluster development (Schueler, p. 55, 1995). Thus, cluster development design becomes a me ans of impact minimization rather than the mitigation approach of traditional stor m-water management. The acronym “BMP” is used, in this research, to refe r to structural methods of runo ff control. However, use of the cluster design option for runoff control is itself a BMP albeit a non-structural one.

PAGE 30

17 A related concept: planned development district Also important to consider in this di scussion is the planned unit development (PUD) also called the planned development district (PDD) in Al achua County (ACBCC, 2002). The Pennsylvania best management pr actice (BMP) manual states that PUD can be used to implement the cluster desi gn concept (CH2MHill, Sec. 3, p. 8, 1998). In Alachua County, these are two design options, although they are quite similar. A detailed discussion of the County zoning regulations fo llows later in this chapter; however, to summarize the PDD design option allows a greater degree of design flexibility in site planning than the cluster opti on and is apparently meant specifically for implementing creative urban planning and encouraging environmental protection (ACBCC, 2002). Purported benefits The first benefit of cluster developments id entified in the literature is reduction in impervious area, which in turn provides a reduction in runoff and runoff-borne pollution. Impervious land uses are mainly dedicated to transportation (Sc hueler, pp. 19-20, 1995). With smaller, more compact lots, the size of the interior street system, in the case of residential projects, is lessened (CH2MHill, Sec. 3, p. 5, 1998). This is the logical conclusion, since the opposite, la rge lots increasing the amount of impervious surfaces for transportation, is frequently true (Schueler, p. 38, 61, 1995). Schueler (p. 61, 1995) cites a 1989 Maryland Office of Planning study de monstrating that cluster developments can reduce impervious cover by 10 to 50%, w ith the greatest benefit coming from clustering larger lot sizes (re ductions in impervious cover decrease as initial lot size decreases). In addition, cluster developments, with their smaller overall footprint, can also reduce the amount of uncontrolled areas in a development by concentrating runoff to

PAGE 31

18 one area and providing more options for BMP s ites, while reducing the size of the BMPs (Schueler, pp. 61-66, 1995). The second benefit of clus ter developments is preservation of open space for natural resource protection. In Pennsylvania, it is recommende d that cluster developments be used to avoid impacting natural drainage features such as ephemeral swales and areas of high infiltration (CH2MHill, Sec. 3, pp. 4-8, 1998). Cluster developments can also be used to enhance or ease (because “lost” lo ts in the buffer area may be transferred to another portion of the site due to smaller lot size) the creation of buffers around water resources (Schueler, p. 61, 1995). Logical ly, these implementations of cluster developments would require preservation of at least some open space in its natural condition. Pollution control from cluster developments may be enhanced because there is less area generating pollutants. However, the actual composition of the open space is important. Pollutant reduction, particularly fo r nutrients, will be lessened if a significant proportion of the open space is fertilized lawns, su ch as ball fields or landscaped areas. The cluster design typically uses conven tional BMPs for storm-water management. Available literature Other literature has quantifie d the benefits of cluster development in general or on issues related to the particular applicati on of cluster developments (i.e., protecting hydrolgically significant areas). Zheng and Baet z (1999), in a watershed-scale analysis of development patterns, concluded that developm ent patterns with smaller impacted areas from combining reduced lots with reduced right-of-way width and mixed dwelling types with open space result in lower peak flows than more traditional patterns. Zheng and Baetz (1999) also concluded that this benefit is most visibl e at the sub-watershed level.

PAGE 32

19 Sloat and Hwang (1989), in their study of de tention basin performance, concluded that preserving a buffer, which can be easier or less contentious through the use of cluster development, of natural land around a stream reduced peak flows. An application of some of these principles was the Rock Creek Development in Colorado, a planned community versus a “true” cluster developm ent (Galuzzi and Pflaum, 1996). Here a deliberate effort was made to preserve and protect the natural drainage by placing open space and natural areas along the stream corridors. Potential negative factors Cluster-type developments have potential negative factors that can limit their use. Some of these factors are economic and are discussed Chapter 5. A discussion of drawbacks that pertain to hydrology is found in Schueler (pp. 67-69, 1995). Although cluster developments are often not used for wa ter resource protection, Schueler notes that wording of a particular ordinance can actua lly decrease usefulness of cluster design. One concern is the local concept of development density. If th e locality has no concept of “unbuildable” area, due to na tural resources etc., imperv ious area within a cluster development may actually increase because ther e are more dwelling units in the cluster design than there would be in a full-size lot design. This can also be true if “bonus” provisions exist that allow for more units. Another issue that is discussed is the fact that local ordinances do not necessari ly require that open space to be natural areas, but might even allow impervious recreation areas to be considered open space (Schueler, pp. 67-69, 1995). Low Impact Development (LID) Low impact development is a site planni ng and design strategy that has recently been advanced by Prince Georges County, Maryland. It has been presented as an

PAGE 33

20 improvement of the current practice of la nd development storm-water management. The goal of LID is a site design that repli cates pre-development hydrologic function and response of the site (Coffman et al., 2000) These goals are accomplished through a mix of site planning and engineered manageme nt practices, which are called integrated management practices (IMPs) by LID advocat es to differentiate them from more conventional practices (BMPs) (Coffman et al., 2000). The site planning concepts of LID em phasizes minimization of impacts from development. Literature advocating LID (Coffman et al., 1998, 2000; Prince Georges County, Ch. 3, pp. 2-11, 1997, Ch. 2, pp. 2-13, 1999a), recommends minimizing the area of development to avoid water resources and th eir associated buffers, and preferably also include areas of high infiltra tion potential and natural drai nage pathways. Within the developed area, further lessening of imp acts can be accomplished through “site fingerprinting”, which is meant to redu ce disturbance and reduce and disconnect imperviousness on individual lots even preser ving woodlands on the lots themselves. The LID literature cited here does not provide any specific guidance on the extent of land preservation. Additional site planning goals include reduci ng impervious area through alternative road layout, elimin ation of sidewalks/parking lanes and narrower street sections. Thus, consideration of the site’s hydrologic functio n and means of maintaining this function must begin ear ly in the design process. In addition to site-planning aspects of LID, a fundamental change in the stormwater management system is proposed. Instea d of the conventional end-of-pipe systems, LID relies on distributed, source controls th at are designed to maintain watershed timing and encourage infiltration. Perhaps the most no ticeable practices are use of swales in lieu

PAGE 34

21 of storm sewers and bioretention areas. Bior etention areas are shallow landscaped areas used to infiltrate runoff on each indivi dual lot (Coffman et al., 1998, 2000; Prince Georges County, Ch. 3, pp. 11-15, 1997, Ch. 4, pp. 1-25, 1999a). Purported benefits The main benefits of LID, the mainte nance of natural hydr ologic function and response, have already been discussed. LID is both a prevention and an enhanced mitigation strategy. This is in contrast to the cluster design concept, which may provide prevention of some runoff and runoff-borne pollutants, but st ill relies on th e conventional view of efficient storm-water removal to end-of-pipe management practices. The prevention aspects of LID are, in most ways the same as cluster developments, although nowhere in the LID literature was cluster de velopment explicitly mentioned as a means of meeting LID site planning goals. On the other hand, LID is also not presented as a rigid design approach and community, and even site-specific interpretations are encouraged (Coffman et al., 2000). A brief definition of disconnection of im pervious surfaces, which is not addressed in discussion of cluster developments, is ap propriate. Disconnecti ng impervious surfaces means that instead of routing runoff directly to the storm sewer system, runoff from these areas flows onto pervious lawn ar eas to allow for infiltration. The LID storm-water system emphasizes source control and infiltration and even reuse. Ideally, this approach will replicat e the natural hydrology of the site. Andoh and Declerck (1997) discuss a sour ce control storm-water system that mimics nature and Kaiser (1997) stresses the need to restructure urban storm-wa ter systems to behave more like natural systems, including providing opportunity for storm-water infiltration. In the United States, a similar approach to land development called Conservation Design has

PAGE 35

22 been proposed (Horner, 2000). Heaney et al. (1998) also advocate source control of runoff and note that if on-site controls can store runoff, most locations could meet their lawn evapotranspiration (ET) requirements through reuse. The distributed source contro ls also provide the advantag e that if one unit fails, the rest are still available, whereas in the c onventional system, failure of one management practice usually means that controls for the en tire site are seriously or totally degraded. Another potential benefit of LID is polluti on control. Traditional development, and even cluster designs, place cont rols at the end of the storm sewer system. This approach places a great reliance on a small number of pr actices. In contrast, LID begins treatment of runoff at the source and continues treatme nt through the swale based transport system. This results in a “treatment train” of management practi ces that provides the most effective means of runoff treatment and control (CH2MHill, Sec. 4, p. 2, 1998). Runoff, however, from the roads will onl y be treated in the swale system, while runoff from lawn areas will be treated first in bioretention units and then in the swales if the bioretention areas should fill. Land protection aspects also assist in pollution control for the same reason discussed in the cl uster development section. Available literature Much like the cluster development concept, the benefits of LID implementation are fairly intuitive to the hydrologist. Prince Georges County, Maryland has some examples of LID computation in the context of a TR55 (U.S. Department of Agriculture [USDA], 1986) based analysis (Prince Georges Count y, Appx. D, 1997). Huhn and Stecker (1997) discussed the performance of a LID-type approach in Germany, where simulations showed that a majority of runoff could be infiltrated successfully. This study also indicated that natural soil conditions impact the effectiveness of such a system. Areas

PAGE 36

23 with better infiltration c onditions will have less of a reliance on controlled surface discharge. A theoretical study by Holman-Dodds et al. (2003) conclude d that storm-water management approaches that encourage infiltra tion require distributed controls. The most effective areas for control are in upland ar eas that tend to have soils with better infiltration potential and are generally th e area where constructi on is occurring. These high infiltration areas, howe ver, are also the areas wh ere the greatest impact of development will be seen. The same study noted that disconnecting impervious surfaces provided significant benefits over traditional development, but that these benefits decreased as the rain fall depth increased. The Jordan Cove Urban Watershed projec t, currently in progress in Waterford, Connecticut, will provide an interesting “h ead-to-head” comparison between traditional development and LID. Preliminary results ar e only available for the construction period, but favor the LID-type design for runo ff quantity. The results note pollutant concentration increases for the LID-type design that were not apparent in the traditional design but were attributed to specific ev ents (one example was unstabilized swales resulting in suspended solids output). Mass export of pollutants was higher for the traditional site; which was attributed to higher runoff volume (Clausen, 2002). Other publications deal with individual aspects of LID. For example, Braune and Wood (1999) discuss a matrix developed by ot hers in Colorado that shows disconnecting impervious surfaces to encourage infiltration as “highly effective”. Interestingly, this matrix also listed some traditional end-of-p ipe practices in the same category, while other LID practices such as swales were somewhat less effective at mitigation.

PAGE 37

24 Other examples in literature deal wi th performance of individual BMPs recommended for use in LID. A study of an aquarium parking lot in Tampa, Florida evaluated the effectiveness of grass swales for runoff quantity and quality improvement. Areas with swales had less runoff than thos e without. Pollutant removals varied with pavement type. Nutrient removal was worse th an sediments and metals, particularly for phosphorus (Rushton, 1999). This variability in pollutant removal was also reflected in another summary of swale performance in Maryland, Florida a nd Virginia (USEPA, 2000). Bioretention is another LID component that has recently been examined. Most of the focus on bioretention has been on polluta nt removal. Hsieh and Davis (2003) reported that 6 different bioretention f acilities showed high removal efficiency for sediments and lead, variable removal of phosphorus and lo w removal for nitrate and ammonium. An EPA literature review (USEPA, 2000) discusse s studies on constructed systems in the field and laboratory systems. All cited studi es used synthetic runoff. Performance was generally high for metals, although one of the st udies reported results that were lower for one of the field systems. Nutrient rem oval was high for some nutrients like total phosphorus, but quite low for nitrate. Laborat ory studies showed a relationship between nutrient removal and depth in the bioretention soil matrix. Potential negative factors LID, like most development practices, has potential negatives. The LID approach is generally applicable but dependent on a number of factors (Coffman et al., 1998) including conflicts with local regulations, property use (land use rights) issues and development density allowing enough space for LID management practices. It is also possible that LID will not tota lly eliminate the need for c onventional detention practices

PAGE 38

25 (Coffman et al., 1998). LID also relies on infiltration practices One problem with infiltration practices is a high failure rate for a variety of r easons including poor estimation of infiltration rate and soils co mpaction (Livingston, 2000). Thus, it would be wise to consider lessons of th e past with regard to infiltra tion systems, which may not be reflected in local ordinances or design guide lines. Livingston (2000) discusses of several practices and the lessons learned. For example, swale infiltration design criteria set by the SJRWMD were often unattainable due to the space required for swales. The recommendation was to use check dams or a similar means to provide depression storage in the swale itself. Applicable Design Standards It is important to introduce, briefly, so me of the land development requirements that affect this research before the discu ssion of site design. The hypothetical residential development project assumes a property si ze of 246.29 ha (608.66 ac), that will undergo residential development. This does not include other areas of the wate rshed that will drain through the project si te and are included in the hydrology model. There are two governing regulations: The Alachua County Code of Ordinances (ACBCC, 2002) and the Applicant’s Handbook: Regulation of Stormwater Management Systems from the SJRWMD (2001), which summarizes rules on st orm-water management for the District and provides design guidance on management practices. Design standards and metric conversion Site design was done in metric units. Howe ver, the applicable ordinances provide standards in Imperial units. This required rounding (and therefore slight inaccuracy) of decimals when dimensions, required by ordina nce, were converted. Conversions from feet to meters, for linear dimensions, were computed with a convers ion factor at three

PAGE 39

26 decimal places (i.e., 3.281 feet per meter), a nd dimensions were r ounded to the nearest 0.1 m or, if the conversion was closer, a “comm on fraction” such as 0.25 or 0.33 (i.e., 50 feet converts to 15.24 m so 15.25 was used inst ead of 15.2). Most erro rs are on the order of inches or fractions of inches, therefore, th is error was not an issu e for this research or even site design in general. Area dimens ions are rounded to two decimal places. This conversion convention applies only to the actual layout of the site and is intended to reflect the fact that most desi gn standards in the County Ordinance do not use fractions and are frequently rounded to a simple number. For example, the R1-aa zoning can be considered “half-acre” even though act ual minimum lot areas is 20,000 SF instead of 21,780 SF. Basic Development Regulations The traditional development option allowed for a minimum 1,857.88 m2 (20,000 SF) lots and a maximum of 4,046.46 m2 (1 Ac). This is the R1aa zoning classification for Alachua County. Allowable density of lots is based on total prope rty area, including “unbuildable” areas like wetlands, streams and their associated water buffers. The buffer that is required around water resources varies with the value of the particular feature but is a minimum of 10.67 m (35 ft). Thus, the a llowable number of lots for a given zoning classification will always be greater th an what is practical (ACBCC, 2002). Minimum lot dimensions appear to be desi gned to ensure rectangular lots (lots on curves and cul-de-sacs are exempt from these dimensional requirements). Building setback lines are also defined for each z oning classification (ACBCC, 2002). These two standards are important to this research because of the assumption of a single, generic house footprint that is constant throughout the alternatives. Th is essentially sets a lower limit on lot width beyond that required by ordinance.

PAGE 40

27 Alachua County also requires a narrow buffer between subdivisions at the boundary. Single-family resident ial districts must have a 4.6 m (15 ft) buffer. No structures are permitted in the buffer area. Th ere is no requirement that this buffer should be separate from a platted lot (ACBCC, 2002). Tables 2-1 and 2-2 summarize the relevant design standards for lots. Table 2-1. Summary of lot layout standards for R1-aa (ACBCC, 2002). Area Width Depth Buffer Minimum 1,857.88 m2 33.5 m 38.1 m 4.6 m Maximum 4,046.46 m2 None None None Table 2-2. Yard and setback standards for R1-aa zoning (ACBCC, 2002). Front Yard Rear Yard Side Yard Side on Street Minimum 7.6 m 9.1 m 3.8 m 7.6 m Maximum None None None None Design standards, listed in Table 2-3, for ro ads are also of interest. Road lane width and presence of a turning lane are determined by the level of service as computed by the average daily traffic (ADT). Road lane widt hs range from 3.1 to 3.67 m (10 to 12 ft). Right-of-way (ROW) width is determined by st reet classification (A through D, with A being the lowest level of se rvice and Types C and D further divided into C-1 and C-2 etc.). Level of service is determined by th e ADT. The ROW for a street, of type B or greater service, with a swale in lieu of storm sewer is wider (ACBCC, 2002). Table 2-3. Selected road design standa rds for Alachua County (ACBCC, 2002). Type A Street Type B Street T ype C-1 Street Type C-2 Street ADT 0 – 125 126 – 1,200 1,201 – 3,200 3,201 – 7,000 Lane Width 3.1 m 3.1 – 3.4 ma 3.67 m 3.67 m ROW Curb 15.25 m 15.25 m 18.3 mb 24.4 m ROW Swale 15.25 m 18.3 m 24.4 m 30.5 m a3.4 m lane width applies where ADT is over 400. bAssumes no on-street parking.

PAGE 41

28 Requirements for the design of on-street parking lanes are al so provided, however on-street parking is rarely seen in devel opments in Alachua County. Sidewalks are only required along streets with a high level of se rvice or where certain community features are present such as schools and parks (ACBCC, 2002). The street drainage system may either be a closed system with storm sewer and curb and gutter or a swale section. Swale desi gn has a limiting velocity of 0.9 m/s (3 ft/s) for the 10-year rain event; otherwise they requi re a paved invert. Furthermore, swale flow must not encroach on the road for this design event. Swales are also prohibited where the groundwater table is within 0.9 m (3 ft) of th e ground surface. In general, all storm sewer pipes must be reinforced concrete under pa ved roads (asphalt coated corrugated metal is acceptable elsewhere) and must have a minimu m diameter of 38.1 cm (15 in) in a closed system and 45.7 cm (18 in) where pipe is used in an open swale system. Closed systems must be designed for the three-year, rain 10minute event as determined from the local IDF curve (ACBCC, 2002). Both Alachua County and SJRWMD govern general storm-water management requirements. The regulations are, in genera l, similar. One difference is the required design storms for peak flow control. SJRW MD requires design for the mean annual (2.5year) and 25-year 24-hour storms (SJRWMD, Sec. 9, p. 8, 2001), while Alachua County states only that peak rate of discharge must not exceed existing conditions for all storms up to 25-year return interval (ACBCC, 2002). Alachua County does explicitly state that the volume of any storage system must accommodate the 25-year, 24-hour event. There is no storm-water quality design event specified in either ordinance. Instead, water quality storage requirements are defined as a depth of rainfall multiplied by the

PAGE 42

29 contributing area or the impervious area (the depth to be multiplied over the impervious area is greater than the depth for the whol e contributing area). For example wet ponds must provide water quality storage for the larger of 25.4 mm (1 in) over the entire contributing area or 63.5 mm (2.5 in) multiplied by the area of impervious cover. An offline retention based sy stem with either under drains or infiltration requires half the treatment storage of a detention system. Online retention systems, which most end-ofpipe basins are, requi re 12.7 mm (0.5 in) more storage than an off-line system over the entire watershed. Swale systems must infiltra te 80% of the 3-year, 1-hour event. Note that the swale water quality requirement is in consistent with the requirement of a paved invert for velocities above 0.9 m/s (3 ft/s) (this invert re quirement is not required in SJRWMD regulations). Treatment volumes must be recovered in 72 hours (ACBCC, 2002). Cluster Development Regulations Development standards for a cluster deve lopment are also important for this research. These apply to lot layout and build ing setback lines. The minimum lot area for the zoning district may be reduced by half with dedication of open space that corresponds to the lot reduction (i.e., 1 m2 reduction in lot area requires 1 m2 dedication of open space). Natural water resources and their requi red buffering must be included in the open space, thus Alachua County indirectly re quires a certain portion, depending on the existing conditions of the site, of the open space to be natural. Recreation areas and certain storm-water management areas, su ch as wet ponds, may also be included (ACBCC, 2002). Cluster development regulations requir e that the allowable density, based on existing zoning, shall not be increased. However, because density calculations can

PAGE 43

30 include unbuildable areas, the actual numbe r of lots for a cluster development may actually exceed that of trad itional development, depending on site conditions (ACBCC, 2002). Cluster development has different sta ndards regarding lot setbacks. Setback distance is reduced from standards listed for full-size lots to 4.6 m (15 ft) for front and rear yards and 1.5 m (5 ft) for side yards, while side yards on streets require 3.1 m (10 ft). The exception to this rule, according to the wo rding of the ordinance, occurs when a lot with reduced area abuts another parcel (in othe r words, one of the lot lines is also the subdivision boundary). In this case, building se tbacks must be the same as the standard for the zoning district, though it is not stated if this applies to a ll setbacks or just the particular yard that actually abuts the nei ghboring property. In additi on, the lot must also have a 1.8 m (6 ft) high fence or a 3.1 m (10 ft) wide vegetated buffer with 75% opacity. Logically, this would replace the 4.6 m (15 ft ) buffer, without screening requirements, which is required under standard zoning. In the previous section, it was noted that the standard buffer between subdivisions is not ma ndated as different from platted lots. The wording of the ordinance would seem to imply that if an off-lot buffer was provided, then individual lots would not abut the subd ivision boundary and cluster development buffer/setback requirements would not be required (ACBCC, 2002). Cluster developments have a provision in th e County Ordinance that allows for site density to be increased if a certain per centage of the units in the subdivision are “affordable” housing (ACBCC, 2002). If this option was applied, s ite disturbance and imperviousness may increase above what woul d normally be the case for a traditional

PAGE 44

31 development. Table 2-4 provide s a comparison of the site de sign standards for standard R1-aa zoning and a cluster development applied to the R1-aa zoning district. Table 2-4. Comparison of full-size and cluster lot requirements (ACBCC, 2002). Standard Cluster Minimum Lot Area 1,857.88 m2 928.94 m2 Minimum Lot Width 33.5 m 16.75 m Minimum Lot Depth 38.1 m 19.1 m Front yard Setback 7.6 m 4.6 m Rear Yard Setback 9.1 m 4.6 m Side Yard Setback 3.8 m 1.5 m Street Side Yard Setback 7.6 m 3.1 m Buffer/Screening 4.6 m Buffer, no screening Speciala a Requirement is for screening consisting of 1.8 m high fence/wall or a 3.1 m wide buffer with vegetation that is 75% opaque. All storm-water management requirements for cluster developments are the same for standard developments. As was previous ly stated, certain storm-water management features can be counted as open space. Th e feature should provide some amenity or aesthetic value (ACBCC, 2002). Planned Developmen t District (PDD) The PDD is intended to provide the develope r a great degree of flexibility in site planning for use of the property but also to provide enhanced pr otection of natural resources. The developer does not have to follow the minimum design standards that have been previously discussed. This makes the PDD option even more flexible than the cluster development option. The design standa rds of note include avoiding development within natural areas to the greatest extent possible, ensuring that developed areas are compact and continuous and maintaining conne ctions between natu ral areas to avoid fragmentation (ACBCC, 2002).

PAGE 45

32 Summary The material presented in this chapter presents the concept of the alternative land development and storm-water management practices that undergo hydrologic and economic analysis in the following chapters. The governing regulations have also been presented. With this discussion complete, it is now possible to introduce the actual site designs in the following chapter. These designs are the “physical” ma nifestations of the concepts presented in the preceding discussion.

PAGE 46

33 CHAPTER 3 SITE DEVELOPMENT PLANS Introduction This chapter discusses the site design of the four alternatives and the development of the existing conditions model. Where pos sible, these alterna tive land development plans were designed within governing regulations However, even in real projects, there are cases where the design engineer may wish to deviate from the required or generally accepted design standards. This is usually handled by requesting a variance from the ordinance. This assumes that the design engi neer can provide proper justification for the design and that public safety is not at risk. The general design discussion for each alternative describes the vari ous decisions and methods, incl uding clear deviations from the Alachua County ordinance, which were i nvolved in the site design. Areas where the County ordinance was vague or subject to in terpretation are discussed in a separate section for each alternative. Existing Conditions The existing conditions plan is the basi s of all the other plans. Many types of information were required to conduct even a hypothetical development project. A detailed site survey was not pr actical, so existing GIS data we re used in lieu of those a Surveyor would collect. This should not be construed as an endor sement of using GIS data to replace an actu al site survey. The GIS data have errors associated with them and also may not reflect current site conditions. With the exception of topography, these data were obtained from the Florida Geographic Da ta Library (FGDL) (University of Florida

PAGE 47

34 GeoPlan Center, 2000) and are available on CD and World Wide Web download. The data included shape files for National We tland Inventory wetlands, 100-year floodplains, soils and an aerial photograph to determine cover conditions. Topography was developed from a digital elevation model (DEM) provided by Camp Blanding. This DEM required re-samplin g to a smaller grid size to ensure smooth contours. The drawback to this approach was that the resulting contours are slightly inaccurate in areas of steep slopes where they generally underestimate the slope. This is because the original 30-meter grid interval had insufficient detail to reflect the rapid change in grade correctly. It was assumed, however, that contours resulting from the DEM were of sufficient accura cy for a theoretical project. Another parameter that is important both for development in Florida and for implementation of LID is the elevation of the water table. Normally this would be determined through test wells. The only info rmation available was from the Clay County Soil Survey (Soil Conservation Service [S CS], 1990), which wa s of questionable accuracy, and a regression equation develope d by for the SJRWMD (Boniol et al., 1993), which requires user interpolati on near surface water features. The water table elevation is important for two reasons. First, it determin es the degree of difficulty in building in a certain area. While most homes in Florida do not have basements, it may still be necessary to provide a raised building pad to keep separa tion between the foundation and water table. On the other hand, this may be an advisable building t echnique regardless of water table depth and is discussed in the various LID manuals (Prince Georges County, Ch. 3, pp. 15-16, 1997; Ch. 2, pp. 16-18, 1999a). Th e second reason that water table needs to be considered is for infiltra tion management practices. For example,

PAGE 48

35 recommended design parameters for a bioreten tion area are for a 0.6-1.2 m (2-4 ft) deep area of prepared soil and a minimum 0.6 m (2 ft) clearance from facility bottom to the water table (Prince Georges County, Ch. 4, p. 4, 1999a). This research assumed that the water table was not a limiting factor, although it is acknowledged that certain areas (a small minority of the “buildable” area) are “danger” zones where infiltration practices may be unfeasible or extra fill for building pads may be required. The hypothetical property was considered to be a collection of smaller parcels that would be grouped together. The total area of the property was 246.29 ha (608.66 ac), while the existing contributi ng watershed area (to the re ference point located on the stream where it exits the property to th e north-west) was 288.77 ha (713.64 ac). The allowable density for R1-aa zoning was 1,325 lo ts. The presence of wetland area, stream and infrastructure resulted in a density that was much lower. The existing conditions plan is shown in Figure 3-1 with topography, wood line, soil type hypothetical property lines, watershed boundaries, hydrology and exis ting improvements such as roads. Traditional Development Plan Discussion of the traditional design option is divided into three parts. The general design discussion presents the overall devel opment of the site, including areas where clear deviations from the governing regulati ons were made. The next section discusses the regulations that required interpretation because they were vague or seemed to contradict other regulations.

PAGE 49

36 Figure 3-1. Existing conditions plan. General Design Discussion The traditional design alternative represents the type of development that would be seen if the local ordinance was the only guide used. It was assumed that the developer is taking advantage of two new county roads along the southern and western edges of the property. The developer is not responsible for construction of these roads or their swale drainage system except for access improvement s (deceleration and acc eleration lanes as required) and the access road running north to south for the eastern most entrance to the development. It was assumed that the deve loper had some input on the design of the swale system on the off-site ro ads, since these roads would be allowed to drain to the storm-water management system c onstructed by the developer on-site.

PAGE 50

37 A prohibition on lots in a water resour ce buffer area was assumed for this design. Lots were prohibited in the 100-year floodplain if it was not possible to ensure that the building pad would be above the floodplain. It was also assumed that streets and stormwater ponds, and their associated grading, w ould have minimal impact on the 100-year flood elevation and, therefore, were permitte d in the 100-year floodplain. Lot dimensions, while acceptable under the local ordinance, are arbitrary. Th e lot width was set at 37.2 m (122.1 ft) and the depth at 50.3 m (165.0 ft) for r ectangular lots, result ing in lot area of 1,871.16 m2 (20,142.96 SF). Lots on curvilinear street sections need not meet the width requirements, but must meet the minimum area requirement. The standard design does, inadvertently, pr ovide a large area of open space. This consists of the wetland area around the nor thern stream and a 10.67 m (35 ft) water resource buffer around the streams and wetlands Other areas that we re unbuildable due to lack of space for a lot, storm-water manage ment, site buffer etc. were also considered open space. The site design did not place any other emphasis on protecting natural resources above that required by ordinance. The road layout paid no particular atte ntion to topography to reduce grading and followed a general grid pattern. The road ROW and lane widths are dependent on the projected average daily trips (ADT). For the most part, RO W width is 15.25 m (50 ft) and lane width is 3.1 or 3.4 m (10 or 11 ft). Generally a traffi c study would be conducted to determine the ADT, but for the purposes of th is study, the ADT was determined from the number of homes on a particular street us ing trip data (10 tr ips per home) from Site Planning for Urban Stream Protection (Schueler, p. 133, 1995). When traffic from two or more streets intersected, their ADT was combined for determination of the service

PAGE 51

38 level of the collector street. Sidewalks are only required on ro ads with a C or D level of service. This design assumes however that si dewalks are provided on all internal roads. This design plan has a generic “house” footprin t associated with it. This footprint, which includes all impervious area including driveways, was used as a guide to lot dimensions. The footprint was computed at approximately 388 m2 (4,176.8 SF). All homes are assumed single-story. The homes are only relevant to the hydrologic analysis. The economic analysis assumes the lots are so ld vacant, thus removing the impact of the home on lot sale price. The conventional storm-water design option uses curb and gutter, adding 0.6 m to the width of the road, for collection of runo ff and the storm sewer transport system. The end-of-pipe management practice is a wet detention pond, a rather common practice in Florida. The eight ponds are situated at na tural low points throughout the site and were designed utilizing the example procedur e outlined by the SJRWMD (Ch. 29, pp. 1-6, 2001). Pond computations are based upon the cont ributing area to that particular pond. It is clear that many of these ponds require baffl es to prevent “short circuiting” of runoff due to the shape of the ponds and the location of discharge points from the storm sewers. There are several small, uncontrolled areas of the site that are developed. These areas are permissible provided the developer provides compensating storm-water treatment. It was assumed that since c ounty roads are allowed to drain into the development to the wet ponds, this provide d the required compensating treatment. The storm-water system is not designed hydrau lically. The placement interval of the stormwater inlets is arbitrary. The storm sewer pipe was assumed to be 61.0 cm (24 in) diameter, reinforced concrete. In reality, pi pe width would vary with a 38.1 cm (15 in)

PAGE 52

39 minimum, but this generic diameter was dete rmined to be a reasonable average for the development. Overall slope of a pipe series was assumed as the el evation of the outlet subtracted from the elevation of the upstream inlet (minus 1.2 m or 4 ft for inlet depth). Florida Department of Transportation gui delines for the period of 2000-2001 required a minimum of 15.25 cm (6 in) of cover between the top of storm sewer and the base of the road. Thus, the 1.2 m depth in the upstream inlet is sufficient. Site grading, as was previously menti oned, was rough and for runoff flow purposes only. In other words, road profiles were not smoothed. However, it was assumed that the entire lot was cleared and graded (or comp acted if no re-grading is required) from construction vehicle traffic. Interpretation Issues With The County Ordinance There were four areas where interpretati on of the County ordinance was an issue. The issue of location of the required buffers at the subdivision bounda ry was discussed in the section on the County ordinance. This si te design assumed buffers separate from the platted lots and, thus, considered open space. This approach has been used in subdivisions in the Gainesville ar ea (for example, Haile Plantation). Another issue was required maintenance access to a storm-water pond. The final decision allowed for a 3.67 m (12 ft) access way outside of the platted lots and a strip around the basin of the same width that has a slope no greater than 8:1. The next issue of concern dealt with landscaping. This is more important in terms of economics than hydrology, but it is appropriate to identify it in this section. The landscaping ordinance was unclear as to whether the 10% landscaping requirement applied to individual lots. It was assumed that this was the case. Thus, 10% of each lot was assumed to be landscaped.

PAGE 53

40 Swale geometry was another item wher e deviation from the ordinance was possible. The SJRWMD allows for swale side slopes to be a maximum of 3:1 with a top width to depth ratio at least 6:1 (SJRWM D, Ch. 15, p. 2, 2001). Alachua County states that swale side slopes are “typically” no great er than 4:1 on the side facing the road (ACBCC, 2002). In this rese arch the SJRWMD geometry standard was applied. The swale design is parabolic, w ith a 3.67 m (12 ft) top width an d depth of 60.1 cm (2 ft). The flow depth on which time of concentration cal culations are based assumes full flow at 50.9 cm (1.67 ft), resulting in an approximate top width of 3.1 m (10 ft). Swales on the off-site roads are assumed to have a conser vative design requiri ng a concrete invert. Swales with this invert were considered to function the same as pipe (tributary impervious area is directly connected). This contradicts SJRWMD swale design standards. In this design, howev er, the swales are not used fo r runoff treatment or control. The last item of interest is the access road for the eastern most entrance to the development. This road is outside the waters hed of interest and is only considered in economic calculation. Design Plans The traditional development plan is shown in Figure 3-2, and includes topographic features for the site including proposed, r ough grading for runoff flow and storm-water management wet ponds. In the pond grading, a c ontour of a slightly different color may be observed. This represents the appr oximate elevation of the permanent pool.

PAGE 54

41 Figure 3-2. The traditional development plan. The traditional development plan resulted in 719 lots, in contrast to the allowable density of 1,325 lots. This represents a 46% reduction and was the result of site infrastructure (roads, storm-water management etc.) and areas that are unbuildable, such as the wetland and associated buffer area. These areas are counted in the allowable density computation. Figure 3-3 shows the traditional developmen t plan without topographic features or drainage basins. The lot layout is clearly visible on this pla n. While this plan was deemed consistent with County design regulations, it was accepted that somewhere in the plan the County engineers might require changes.

PAGE 55

42 Figure 3-3. The traditional development plan without topography, soils or drainage subareas. Cluster Development Plan The second design alternative was cluste r development. Much of the design discussion on the traditional development option is also applicable to the cluster design. For example, road lane and ROW widths were determined by the same procedure, and sidewalks are assumed present on all internal roads. This section focuses on differences between the cluster design and the traditional plan. Also, just like the discussion on the traditional development, the areas where interpretation of the ordinance was required are discussed.

PAGE 56

43 General Design Discussion The most obvious difference between cluste r and the traditional options were the lot dimensions. These were set at a wi dth of 27.9 m (91.5 ft) and 33.5 m (109.9 ft) for depth. These dimensions result in a standard rectangular lot area of 934.65 m2 (10,061.47 SF). Again, these dimensions are arbitrary but accommodate the gene ric house footprint. The footprint, however, is somewhat smaller, 367 m2 (3,950.75 SF), due to lower amount of pavement required for the driveway. The water resource buffer was also increas ed from its minimum size. The Alachua County ordinance specifies a 22.9 m (75 ft) buffer for Outstanding Florida Waters (ACBCC, 2002). Based on this regulation and information on forested buffers in two design manuals (Schueler, p. 86, 1995; CH2MH ill, Sec. 8, 1998), a width of 30.5 m (100 ft) was selected even though streams on the site are not specifically listed as Outstanding Florida Waters. One potentially important aspect of the road design was changed in the cluster design option. The road layout for the cluster options mostly dispensed with the grid layout. Instead, curvilinear loop streets were used, and cul-de-sacs are more frequent; especially as short branches off loop streets. The cluster design option al so required some thought on appropriation of open space, which the traditional design did not Appropriation of open space was based on several goals. Natural areas, including the wetl and and all water buffers, were required to be considered open space. Beyond these areas, the goal set for this design option was to preserve areas of natural runoff concentration (i.e., natural swales) to the greatest extent possible. Generally, this was most successful in the upper areas of the watershed where, in some cases, virtually the entire time of concentration path was preserved.

PAGE 57

44 The most important goal considered in open space preservation, as far as stormwater management was concerned, was preser vation of upland areas of high infiltration potential. These areas are roughly defined in Figure 3-4. This desc ription was based on the Clay County Soil Survey (SCS, 1990) and in cludes only areas of NRCS type A soils. Because of the general nature of soil survey s, boundaries of high infiltration soils could not be determined with any certainty. Similarl y, the GIS data that were used for site features also contained a certain degree of error. Naturally, this results in errors in Figure 3-4, such as areas of high inf iltration within the boundaries of the wetland area identified by the National Inventory of Wetla nds shape file. Still, this figure was useful for planning purposes. The location targeted to be the cente r of protection is the ridge between the two streams. There are 720 platted lots in this site design. This assumes that the developer will not take advantage of the fact that the unbuildable areas can be considered open space, thus allowing an increased number of lots albeit at reduced size. This assumption, although potentially positive from a hydrologi c perspective, led to another potential problem. It is common knowledge in the land development business that area often represents most of the value of a lot and, thus, the sale price. Therefore, reducing lot size may have a negative economic impact to th e developer, and the possibility of the developer adding more units to the development to compensate for this loss of revenue is cannot be ignored. On the other hand, proxi mity to open space may have a positive impact on lot value. These are discussed in the Chapter 5. For now, it is sufficient to state only that the cluster site layout is designed to allow for “i slands” of open space between loop streets, thus preventing lots from a butting one another along the rear lot line.

PAGE 58

45 Figure 3-4. Areas of high infiltration potential. Distance from the rear lot lines to the subdivision boundary was increased from the traditional design. The distance used in this option was 4.6 m plus the difference between lot depth from the traditional and cluster op tions. These two design features detracted somewhat from the contiguous area of open space normally resulting from a cluster development. Effort was made to design the s ite so islands of open space are as large as possible to avoid the fragmentation problems discussed in the literature cited on cluster developments. The storm-water management system is designed in the same fashion as the traditional development. Sub-areas that were totally preserved were not included in any pond computations. There are nine ponds in th e cluster design option. Pond computations

PAGE 59

46 for water quality assume a depth of runoff ove r the entire contribu ting area. If a portion of this contributing area was preserved in it’s natural state, then the pond would overtreat runoff. Interpretation Issues With The County Ordinance Land-use buffers that must exist along subdivision boundaries have an additional impact on the cluster design option. If a lot of reduced size abuts a neighboring singlefamily or agricultural parcel, screening and the more restri ctive building setback of the existing zoning district must be provided. How this requirement relates to the 4.6 m (15 ft) buffer required by the base R1-aa zoning district is uncl ear. Furthermore, use of the word “abut” seems to imply that one of the lot lines of the reduced-size lot would be coincident (the same) as the subdivisi on boundary. Therefore, it was assumed that because a buffer was provided outside of the platted lots, the more restrictive building setbacks and screening w ould not be required. Design Plans Figure 3-5 presents the cluster design option. Figure 3-5 includes topographic features for the site including proposed, r ough grading for runoff flow and storm-water management wet ponds. In the pond, grading a co ntour of a slightly different color may be observed and represents the approx imate elevation of the permanent pool.

PAGE 60

47 Figure 3-5. The cluster development plan. Figure 3-6 shows the cluster development pl an without topographic features or the drainage basins. Lot layout is clearly visible on this plan. While this plan was deemed consistent with County design regulations, it was accepted that somewhere in the plan the County engineers might require changes.

PAGE 61

48 Figure 3-6. The cluster development plan wi thout topography, soils or drainage subareas. LID Developments The final two development design alternativ es incorporate concepts of Low Impact Development (LID). The literature review on LI D pointed out that LID seeks to replicate site hydrology through a two-tiered approach. There is a runoff prev ention aspect that focuses on limiting disturbance on the site be ing developed, thus l eaving a portion of the site in its natural condition. An infiltratio n-based, distributed storm-water management system is used in areas that are disturbed for development. Combination of practices, such as a primarily LID system that require s one or two ponds to meet local peak flow regulations is possible where storm-water management requirements cannot be met by

PAGE 62

49 the infiltration-based storm-water management system. This would particularly be true where the surface area of infilt ration practices required is so large that it impacts the ability to develop the property. General Design Discussion The cluster design option is consistent with LID land preservation goals. Thus, it was decided to use the cluster layout for th e LID design and add the infiltration-based storm-water management system. Examples of LID design included in the Prince Georges County design manuals show LID site preservation as keeping all development activity out of high in filtration, hydrologic group A, soils as much as possible while providing additional preservation on the platted lots that are on the moderate infiltration soils (types B and C). Development would also ta ke place on type D soils as long as these are not wetland areas. In Flor ida, soil infiltration potential is partly determined by the high water table. This leads to many dual cl assifications. For example, soil types 19, 29 and 42, on the Clay County Soil Survey (SCS, 1990), are Osier fine sands and are classified as type D or B, with the B cl assification applying when the soil is drained. Furthermore, type B and C soils are less freque nt in Florida. They are mainly transitional areas and most upland areas ar e high infiltration type A so ils. The latter comprise the majority of the buildable area on the study site. This is both positive and negative. These areas are ideally suited for infiltration practices but are also likely to see the greatest hydrologic change from development. Developm ent in these areas cannot be avoided, but impact should be minimized. It was decided th at on-lot preservation of land was not an ideal solution since there were questions as to whether these areas would truly not be impacted and of the ability to enforce, either by the County or the Homeowners Association, the necessary deed restrictions adequately. Thus the decision to forgo any

PAGE 63

50 on-lot preservation and to use the reduced lot size of the clus ter development was justified, even though use of cluster devel opment is not mentioned in LID literature. The LID storm-water management system was also applied to the traditional design, although this is not “true” LID because of the lack of land preservation that is assumed (all lot areas are considered clea red, graded and/or co mpacted). Thus, this design is referred too as “partial” LID b ecause only infiltration-based storm-water management system is present. There are no explicit requirements in any of the LID literature on how much of the site should be preserved. Since the site design for the LID options is virtually the same as the traditional and cluster options, much of the previous discussions remain valid, except for storm-water management. There are, however, some aspect s of the LID designs that need to be discussed in addition to the design aspects previously discussed on the traditional and cluster designs. Sidewalks are generally dropped from the de velopment. For most interior streets, this still remains consistent with the County codes. Howeve r, sidewalks are required to both sides of type C-1 Streets. Therefore, tota l elimination of sidewa lk is not consistent with the County ordinance on four streets, although it affects only a small proportion of the total street length. A compromise was selected that eliminated sidewalks from all C-1 streets that crossed no more than 10 lots on a side. This eliminated all sidewalks except on the street serving the centr al upland area in th e “partial” LID design, where sidewalks and curbs were used on both sides. Another area where the LID designs differ from the County ordi nance regards road right-of-way. Use of swales for drainage requ ires an increase in ROW. For example, a

PAGE 64

51 road requiring 15.25 m (50 ft), which represents all but four interior streets on both the LID site plans, with conventional drainage would now require an 18.3 m (60 ft) ROW. The four streets, all entry roads, requiring 18.3 m of RO W would now have a 24.4 m (80 ft) ROW (ACBCC, 2002). The reasons for this additional area are not stated in the ordinance, but some possibilities are to al low room for sidewalks, and anticipation of wide swales to meet infiltration requirements or safety. The 1997 LID design manual for Prince Geor ges County discusses modification of their rural road section for use in LID devel opments. Their standard interior and primary road have the same ROW require ment as are required in this project. Several sections are proposed for use in Prince Georges County. RO W widths for these proposed sections are 16.5 m (54 ft) for most internal roads and 18.3 m (60 ft) for higher traffic roads. These sections use swales for drainage (Pri nce Georges County, App. D, pp. 15-20, 1997). The variable lane width allowed in Alachua Count y, based on level of service, allows for a fairly narrow road section, similar to the pr oposed Prince Georges County roads, and was not altered. The only road design issue wa s the setting of proposed ROW width and selecting a swale design. The swale cross-sect ion section selected was the same as the section described for the off-site County road s (discussed in the tr aditional development section). In terms of capacity, this section is probably more than is required for most interior roads. There are times where flow may be transmitted under a road, and where this is anticipated, a 60.1 cm (24 in) diameter pipe is assumed. A pipe diameter of 45.7 cm (18 in) is acceptable in an open draina ge system, but the 60.1 cm diameter was selected as a conservative figur e. All pipes are reinforced co ncrete, even under driveway crossings. The proposed swale section has a top width of 3.67 m (12 ft) and a design full

PAGE 65

52 flow width of 3.1 m (10 ft). The swales on both the interior streets and the off-site roads were considered completely pervious (no conc rete invert). However, the design still did not strictly follow the SJRWMD infiltration standards because check dams were modeled to allow for depression storage. With a 16.5 m (54 ft) ROW, identical to the Prince Georges County standard, the swale can be accommodated in the ROW of even the smallest interior street. However, it was decided that the ROW width for most interi or streets (type A or B) should remain at 15.25 m (50 ft). Thus, the design maximum fl ow depth is accommodated in the ROW, but an additional 60.1 cm (2 ft) of top widt h is provided on the lot itself. This was deemed acceptable since an easement might be required on the lot for public utilities. Type C-1 streets will have the same swal e section and an 18.3 m (60 ft) ROW and are similar to the Prince Georges County proposed section for a highe r service level LID street section. A level spreading area or device is intended to be present at the terminal point of the swales to conve rt runoff to overland flow. The final road section provide s a travel lane of 3.1-3.4 m (10-11 ft) on most interior streets with a 1.2-1.5 m (4-5 ft) grass shoulde r and 3.1 m (10 ft) of swale at a depth of 60.1 cm (2 ft). On the few higher traffic interior roads at the entrances the travel lane is 3.67 m (12 ft) and the shoulder is expanded to 2.4 m (8 ft). Interpretations of The Ordinance The previous discussions on ROW and swale section are clear de viations from the County ordinance. The LID storm-water ma nagement system design, however, is not necessarily inconsistent with the storm-wa ter management regulations that would be enforced on this site. On the other hand, how to apply the cu rrent standards defied clear interpretation. In particular, the issue is co mputation of water quality treatment volume,

PAGE 66

53 which represents the minimum storage of the system and the starting point for the hydrologic analysis. LID assumes a distributed, infiltration-bas ed storm-water management system for runoff quantity control and quality control. Sw ale design, at least in theory, in Alachua County requires infiltration of 80% runoff for a 3-year, 1-hou r rain event per both County ordinance and SJRWMD guidelines. Howeve r, Livingston (2000) cited a study that determined that the swale design guidelines are often not attainable due to space concerns. In other words, a swale width is requi red that is so great as to preclude use. Thus, it was assumed that swales in this de sign would use some form of check dam or barrier to encourage ponding for infiltra tion. The SJRWMD, whose design guidelines were used for swale cross-section, states th at an open conveyance system has no control structure before discharge to receiving waters (SJRWMD, Ch. 15, p. 1, 2001). Thus, use of check dams in the swales would mean that the swales are now considered another type of storm-water management system. They are considered a retention practice. The only question is whether they are on or off-line de tention. They are considered on-line systems (SJRWMD, Ch. 11, p. 1, 2001). These swales, wh ile providing depression storage, were not considered bioretention areas. Bioretention areas appear to also be classified as retention practices (SJRWMD, Ch. 11, p. 1, 2001). Again, the question arose as to whether they were on or off-line treatment. Arguably a bioretention unit could be an off-line system since it is not directly connected to the transport ne twork of the sub-basin and ge nerally runoff would enter as sheet flow, although SJRWMD may c onsider them on-line systems.

PAGE 67

54 The storm-water management system for the LID design options was considered a mix of on-line and off-line retention practi ces. No clear standard exists for computing treatment volume for this type of system. Inst ead of collecting runoff from an entire subarea, LID collects runoff from a single lot. This leads to the question as to what is the appropriate area for the treatment volume calculation. Pond systems are designed to provide treatment for the entire contributing ar ea, including undisturbe d areas, at a single point. The on-lot nature of LID may make control of undisturbed, preserved areas impractical because of the space required for the bioretention areas. The method that was adopted for this research was to compute treatment volume based on the disturbed area. The actual volume of treatment computed a ssumed an off-line system and used the 12.7 mm (0.5 in) of runoff over the treated area standard. It was expected that the actual storage required, due to peak flow control, would be highe r, thus providing a level of water quality treatment exceeding the computed volume, though not necessarily in strict compliance with SJRWMD rules. Design Plans Figures 3-7 and Figure 3-8 show the “partial” LID, based on the 1,871.16 m2 lot, and the “full” LID design, based on the 934.65 m2 lot, respectively. Topography is shown on these plans, but any detention basins re quired due to the resu lts of the hydrologic analysis are not shown. Also, the bioretention areas are not shown. The pipe used for driveway crossings are not shown, but a ny anticipated road crossings are.

PAGE 68

55 Figure 3-7. “Partial” LID development plan.

PAGE 69

56 Figure 3-8. “Full” LID development plan. Figures 3-9 and 3-10 show the LID devel opment plan without topographic features or the drainage basins. The lot layout is clear ly visible on these plan s. The two LID plans represent a new concept in land development. As a result, it cannot be conclusively stated that these plans would be acceptable for approval by the County Engineer. Summary This chapter introduced the actual development plans used for this research. One aspect that hasn’t been discussed is whethe r these plans truly represent a possible site plan. Site layout is as much an art as a science. Therefore, it is quite reasonable to expect that another site planner may ha ve a different interpretation of site plans. Each of the four development plans represents but one “rea lization” among an in finite number of

PAGE 70

57 variations. However, they were establis hed in a manor allowing the economic and hydrologic impacts to be explored in a systematic process. Figure 3-9. “Partial” LID development plan without topography, soils or drainage subareas.

PAGE 71

58 Figure 3-10. “Full” LID development plan without topography, soils or drainage subareas.

PAGE 72

59 CHAPTER 4 HYDROLOGIC IMPACTS Introduction This chapter covers the hydrologic analys is for four, hypothetical, alternatives for residential land development and storm-water management. Details of these alternatives were presented in the last chapter. Each alternative was compared with th e hydrologic response of the existing conditions of the watershed. Three rainfall s cenarios were modeled. The first two were the 2 and 25-year, 24-hour hypothetical storms that are the design storm events for the SJRWMD. This research also modeled th e proposed alternatives under continuous simulation based on rainfall collected on the site during the summer of 2001. Discussion on Modeling the Alternatives Hydrologic modeling for land development is governed by guidelines set forth by the regulating agency. Perhaps the most widely used techniques are based on the NRCS curve number method to determine runo ff volume. Hypothetical hydrographs are generated through a variety of synthetic unit hydrograph methods. A common method for the complete modeling process is detailed in TR-55 (USDA, 1986). Two land development alternatives are cons idered: full-size lots and reduced size lots typical of those a cluster development. These are paired with two alternatives for storm-water management: conventional wet ponds and a distributed, infiltration-based storm-water management system proposed by Prince Georges County, Maryland, in their LID publications. This second type of system requires a different approach to modeling.

PAGE 73

60 An excellent discussion of the modeling concer ns for LID is containe d in Wright et al. (2000). Two interesting points, of several, from this paper are the need for storage based modeling and continuous models to capture th e impact on small, fre quent, short duration storms. Prince Georges County has presented a NRCS curve number procedure for analysis of LID. A fundamental assumption of this pr ocedure is that the time of concentration remains the same as existing conditions. The pre and post-development curve numbers are used on two sets of nomographs, one for a retention system and one for detention. The nomographs, which vary by storm depth, pr ovide the required storage distributed across the watershed (Prince Georges County, App. C, 1999b). Continuous Simulation Data Collection The continuous simulation required actual st ream flow and rainfall data. Rainfall data were collected in 10-minute increments from 19 June 2001 until 18 October 2001. This time corresponds with most of the NorthCentral Florida rainy season. Stream stage was also monitored in 10-minute increm ents. A pressure transducer (In-Situ TROLL4000) was placed immediately upstream from a large corrugated metal pipe-arch culvert and data were collect ed at 10-minute increments. This location became the reference point for discharges from the site Cross sections were taken at the level recorder and at two places downstream of the culvert with a laser level. The stream stage data at the culvert was translated into st ream flow data using the HY8 program. This program computed stream stage, upstream of a culvert, for a given flow rate. The method used by this program is an ener gy balance approach presented in Hydraulic Design of Highway Culverts (Federal Highway Administrati on [FHWA], 1985). This approach allowed a rating curve to be developed. Interpolation between points on this rating curve

PAGE 74

61 resulted in flows for each stage measurement. The rating curve was developed up to a flow of 2.83 cms (100 cfs). No statistical op erations were performed on the rainfall or stage data. The collection process for rainfall data result ed in gaps in the data due to periods where the gage could not be accessed (the site was located on a National Guard post). The stream stage data did not suffer from data gaps because of higher memory capacity in the gages. The continuous stream stage r ecord indicated that no rain of consequence was missed when rain gage memory was full. Hydrologic Model The hydrologic model selected for this research was HEC-HMS downloaded from the Army Corps of Engineers (USACE) H ydrologic Engineering Center. This model offers the user a variety of choices for pr ecipitation input, runoff volume computation and transformation. This research used the so il moisture accounting (SMA) module for runoff volume computation. The soil moisture accountin g method is a continuous model. It can be run as a lumped model or as a distributed, grid-based mo del. In this research, the lumped parameter method was utilized mainly because of an erro r in the distributed model code that was not fixed until earl y 2003. The SMA method is well documented in the Technical Reference Manual for HEC-HMS (USACE, pp. 44-50, 2000) and a thesis on the development of the model was also available (Bennett, 1998). Additional updates are discussed in release notes although documentation in thes e cases is often inadequate. The SMA approach treats the watershed as a series of vertical reservoirs, starting with canopy storage overflowi ng to depression storage, so il profile, and two groundwater layers. For a rain event to produce surface runoff, depression storage must be full. For water to enter the soil profile through infiltration, canopy storag e must first be full. Thus,

PAGE 75

62 a logical progression for generating surface r unoff is followed: interception losses take place in the canopy, then infiltration losses and finally depression storage losses. Depression storage only fills when the rate of rainfall exceeds the infiltration rate. Similarly, surface runoff only occurs after depression storage is filled. Groundwater layers can be used to simulate fast inte rflow and slow base flow. Groundwater outflow volume is routed using a linear reservoir technique, while surface volume uses a unit hydrograph method. In this research, the first groundwater layer is used to represent interflow, and the second repr esents the base flow. The base flow portion of the model may represent a more regional phenomenon. Antecedent moisture conditions at the start of the simulation can be modeled by defini ng a percent of volume full (USACE, pp. 4450, 2000). The version of HEC-HMS used in this research contained an important modification to the SMA that was added after initial release of the program. This update was the ability to define a percent impervious for the sub-area. Rainfall on this designated percent impervious is totally converted to su rface runoff. Thus, if half the sub-area is impervious and 25.4 mm of rain falls, 12.7 mm of surface runoff is immediately generated. This modeling behavior is true only for directly c onnected impervious surfaces, where the transport system does not a llow for significant loss, such as concrete pipe. For indirectly connected impervious su rfaces, such as roofs and roads draining to swales that allow for infiltration and depressi on storage loss, this method is not realistic. One final note on the model is important. At junction points, wh ere stream reaches or sub-areas are connected, hydr ographs are summed. At these points, base flow is not reported separately. Instead, a co mposite hydrograph is reported.

PAGE 76

63 Experimental Methods Design alternatives were modeled under three conditions. In addition to the continuous simulation, which was run from 19 June until 17 October, the two design storms required by the SJRWMD for storm-water design were used. The design storms had depths of 114.3 a nd 203.2 mm (4.5 and 8.0 in) for the 2 and 25-year, 24-hour, events respectively. Thes e were determined from frequency-depth maps available in Urban Hydrology for Small Watersheds TR-55: 2nd Edition (USDA, 1986). The rainfall distribution for this area of Florida is Type II. Most storms in the period of rainfall collection were typical convection events. August was a particularly dry month; howev er, significant rainfall did occur about mid month. September was an unusual month that bega n with a series of particularly intense convection events with high peak flows. At mi d month, a tropical stor m hit the area with sustained rains. The intensity from this event was rather low, but overall precipitation depth was high. Existing Conditions Calibration The first step in the analysis was to calibra te the model against the data collected in the field. Initial values for the various SMA parameters were estimated using guidelines provided by Bennett (1998). Even though the SM A model represents a physical process, the SMA parameters, such as soil storage, might not necessarily reflect actual field conditions. The exception to this was meteor ological parameter ET, which was obtained from long-term National Oceanographic and Atmospheric Ad ministration (NOAA) records (NOAA, p. 26, 1982). Initial stream path and watershed boundaries were generated in HECGeo-HMS.

PAGE 77

64 The field data showed a hydrologic res ponse after most thunderstorms. Initial calibration attempts indicated that this respon se could not be adequa tely modeled as fast groundwater flow (interflow). On the other hand, it was inconsistent with the character of the watershed to allow for surface runoff from small storms from upland areas that have soils with very low runoff potential. Instead, an unconventional sub-area definition was used. The National Inventory of Wetlands bound ary on the smaller (northern tributary) was selected as a sub-area boundary. Anothe r boundary, roughly following the line of steeper slopes, was selected on the main channel. Figure 4-1 shows these boundaries. Figure 4-1. Existing conditi ons sub-area division. These lowland areas were calibrated to al low for surface runoff from most storms, while the upland areas only produce runoff fr om the large, trop ical event in mid September 2001. The additional divisions in subareas at mid-channel or at the upstream end of the channel were added to pr ovide load points for wet pond outlets.

PAGE 78

65 Times of concentration (Tc) were computed using th e procedure and equations outlined in Urban Hydrology for Small Watersheds: TR-55 2nd Edition 5 This method was selected because it allows for division of the Tc path into separate elements for sheet flow, overland flow and channel flow. The stream channel used an estimated cross section in the Tc computations. All channel portions of the Tc are computed assuming a velocity at full capacity using Manning’s equa tion. This assumption is made because the flow along a particular channel is not known a nd, therefore, a detailed hydraulic analysis cannot be done. The Manning roughness coeffici ent that is used for the swales assumed that a level of vegetation wa s maintained to allow for th e appropriate roughness. Off-site swales along the county roads were modele d with a value of 0.045 for the conventional storm-water management designs. This value re flects the presence of a concrete invert on a portion of the off-site swales. All swales were modeled with a value of 0.11 for the LID storm-water management designs. The roughness coefficient is also assumed to include the influence of check dams and transitions for small culverts. Roughness used for uncontrolled analysis, where no depression stor age was present, is the same as for controlled analysis. Thus, the uncontrolled analysis would imply a greater level of vegetative resistance. Various synthetic unit hydrograph methods are available in HEC-HMS including the NRCS dimensionless, Snyder and Clark. Testing of various methods demonstrated that the Clark Unit Hydrograph method allo wed for the best simulation of volume distribution under the surface runoff hydrographs observed on the site. The Clark method utilizes a time-area curve: a form of unit hydrograph. A generic formula for the time-area curve is used in HEC-HMS. The time-area curv e is then routed through a linear reservoir

PAGE 79

66 with the storage coefficient R. A linear rese rvoir relates storage at any time t to the outflow at time t multiplied by the R co efficient (Viessman and Lewis, pp. 219-221, 1996). The time-area histogram is dependent on the Tc (USACE, pp. 60-63, 2000) and will remain the same if the Tc value is constant. The R coeffi cient, however, varies based on watershed cover and development conditions and can be expressed as a multiple of Tc with a range of 8-10 *Tc for forested areas in the Vancouve r, Canada area (Russell et al., 1979). A value of 8 for upland areas and 5 for lowland areas were the best fits for the calibration. The ET coefficient was generally he ld at 0.3; however, it was adjusted as needed to fine-tune the calibration. Routing along the existing stream reaches was done using the Muskingum method. This is a simple form of hydrologic channel ro uting. Input for this model is a storage time K and a weighting factor X. The X paramete r generally averages about 0.2 and K is reasonably close to the travel time of the flood wave (Viessman and Lewis, pp. 235-236, 1996). The values of X used were 0.2 except for the channel after the stream confluence where a value of 0.25 was used. The K values were determined during the calibration process. Values of 0.25 hours were determined for the main channel segments while a value of 0.34 was determined for the tributar y segment. The routing coefficients are acceptable for flows within the range observe d in the continuous model. Flows that significantly exceed these, such as the 25-year design storm, may have greater attenuation effects due to out of bank storage. Antecedent moisture condition was represen ted by designating a percent of storage full at the beginning of the simulation for the soil and groundwater portions of the SMA

PAGE 80

67 storages. These were calibrated for the con tinuous model. For hypothetical design storms the calibrated values did not appear to re flect the assumption of “average” conditions accurately. This could be seen in the in terflow groundwater layer where the positive outflow implied that a rain event response was still underwa y. This is not particularly surprising since a storm system had impacted the area before it was instrumented. The values used for the design storms were determined through observation of the model storages in late August approximately one week after the single Augus t rain event. Input data points for the existing conditions model are contained in Appendix A. The calibrated results are shown in Fi gure 4-2. The volume error between the observed and calibrated data is 13.7 mm (0.54 in ). The volume for the calibrated model is below the observed data. The focus of the ca libration was on peak flows and volume. The time of peak flow tended to be earlier than wa s observed in the field data. This could be remedied by adjusting the sub-area Tc and would have required adjusting the results of the Tc computations. This was unacceptable since the Tc results for the predictive models were computed and consistency between methods was desired. Figure 4-2. Comparison of observed flows versus calibrated flows.

PAGE 81

68 Design Alternative Models Models were constructed for each design alternative for comparison against the existing conditions model. Several issues th at arose during the development of these models require discussion. Some were unique to either the conventional storm-water management system or the LID system, while others are universal across all four models. Naturally, the existing conditions SMA units form the basis of the sub-areas in the developed conditions. Model input parameters were altered for development, with the exception of the ET parameters, from the existing conditions parameters. Universal model assumptions The Clark unit hydrograph method presente d difficulties with the predictive models. There is no easy means of determini ng the future storage coefficient, although it is clear that it will change with development. It was assu med that the future storage coefficient would reflect the composite nature of the watershed. A weighted average was used based on natural cover, which used th e existing multiple (i.e., 8), and the developed area, which used a multiple of 1.5 from Russe ll et al. (1979). The composite multiple was then multiplied by the Tc to determine the storage coeffi cient. The Clark method results in a somewhat different situation, when m odeling LID depression st orage required, than the design procedure used in the Prince Ge orges County manuals. In the manuals, the nomographs for determining depression storag e are based on the assumption that the subarea Tc remains the same, thus assuming, depending on the synthetic unit hydrograph method, that the transformation from volume to a hydrograph has no impact on the problem. Tc may remain the same in the Clark method, but the storage coefficient does not, which can result in an increase in peak flow.

PAGE 82

69 Indirectly connected impervious surfaces also presented a significant challenge. HEC-HMS applies the soil moisture unit, for a particular sub-area, on ly to the portion of the sub-area that is “pervious”. Rainfall on the percent of the sub-area designated impervious is considered surface runoff. While this is valid for impervious areas that are directly connected to storm sewers that provid e no loss, it is not valid for swales with check dams for depression storage or for roof s that drain onto lawns. The model assumes no vertical storage for areas under directly connected impervious area; therefore, the same assumption must be made for indi rectly connected impervious areas. There were only two potential solutions to this problem. The first option was to reduce watershed area by the percentage of i ndirectly connected impervious area. Then, rainfall depth for each time step would be increased by the same percentage. In other words, the volume of rainfall was the same but was distributed over a smaller area. This assumed that rain hitting indi rectly connected surfaces such as rooftops was instantly and evenly redistributed on the pervious portions of the sub-area. For example, if a sub-area has 10% indirectly connected impervious ar ea, the first proposed solution involves distributing 110% of the rain fall over 90% of the area. The first solution was, however, impractical since it required a se parate rain gage file for each sub-area. Another potential soluti on was to adjust the storage volumes in the SMA unit proportionally. For example, usi ng the same 10% indi rectly connected impervious area, 100% of the rainfall is dist ributed on 100% of the ar ea, but with 90% of vertical storage in each layer of the SMA un it. In either case, the volume of storage and rainfall is the same. This approach was se lected for this resear ch, although there are drawbacks that are discusse d in the limitations section.

PAGE 83

70 The final universal modeling assumption de alt with the impact of compaction. Soil compaction is assumed on all cleared areas. The assumption is th at these areas are impacted due to grading, clea ring activity or both. Schuele r (2000) discusses the impact of compaction in terms of changes in bulk density. In addition, a table is presented showing the relationship between soil poros ity and bulk density. Changes in porosity imply a change in storage; therefore, this research modeled soil compaction effects by reducing soil profile storage. Comparing bul k density, from the Clay County Soil Survey (SCS, 1990), of soils on the site with reported increases in bulk density and changes in porosity from Schueler (2000), a percent decr ease in storage of 25% was determined. This decrease did not apply to open space areas ; thus, the actual reduction in storage is much less. An example of such storage reductions is appropriate. Consider an upland area with an existing soil profile depth of 25 mm and 17% indirectly connected impervious area. Since impervious area has no storage, all SMA unit storages are reduced by 17%. In the case of the soil profile in this example, th e new depth would be (rounded to the nearest mm) 21 mm. The soil profile also must be reduced for compaction. The compaction reduction calculation involved only the pervious portion of the sub-area. In this case 30% of pervious area is cleared and graded. Therefore, 30% of the site sees an additional 25%, or 5mm, reduction due to soil compaction, whil e 70% of the site re mains with 21 mm of storage. On a weighted basis, the final storage depth for the sub-area is 19 mm. Changes in infiltration/percolation rates have not been discussed. These were not altered for the developed scenarios. Rate reductions for infiltration were simulated through storage reduction for compaction. Deep percolation was not impacted.

PAGE 84

71 Conventional storm-water management model notes The conventional modeling procedure requi red few assumptions other than the universal assumptions already noted. Ponds us ed in the traditional and cluster designs sometimes captured portions of two existing su b-areas with different parameters (i.e., a portion of a lowland sub-area and a portion of an upland subarea). These situations had upland area overwhelming the lowland portion. Still, storage depths and infiltration/percolation rates we re weighted to account for the composite nature of the contributing sub-area. Starting percentages of storage full, which represented antecedent moisture conditions, were not adjusted but us ed the percentage of the dominant (upland) existing sub-area. Swales, in the conventiona l designs, are assumed to have a concrete invert on a portion of their cross section. Su ch inverts are generally used to ensure transmission of flow. Roads dr aining to these swales are c onsidered directly connected. The HEC-HMS pond routing techniques had two minor problems. First, no sharpcrested weir option is available for the ou tlet structure. The se cond stage outlet would normally be one, or more, sharp-crested weirs. In order to accommodate this, the weir coefficient for the broad-crested weirs was adjusted so that the result of the broad-crested equation would have an effec tive coefficient of around 0.4, wh ich is typical of a sharpcrested weir. Second, the reservoir routines route groundwater through the reservoir, which is not typically appropriate for stormwater management systems. This was easily fixed by taking the output hydrograph for base flow and surface runoff from each contributing sub-area (to a storm-water pond) treating them as separate sources (an HEC-HMS feature that allows a user defined flow input), and allowing base flow to bypass the pond and directly discharge to th e nearest load point. Outflows from pond outlets are assumed to translate to the nearest load point instantaneously. This is

PAGE 85

72 reasonable assumption, although a few basins had to be located farther from a load point than others. Figures 4-3 and 4-4 show the subarea divisions for the tr aditional and cluster designs. Also included are pond locations. Figure 4-3. Traditional design subareas and storm-water system. Complete tables of input parameters ar e located in Appendix A. The wet ponds are treated as directly connected impervious ar ea both in terms of runoff computations and their impact on SMA units, but are not considered in runoff quality treatment computations for the pond volume design.

PAGE 86

73 Figure 4-4. Cluster design sub-areas a nd storm-water management system. LID storm-water management model notes The sub-areas for the LID designs are s hown in Figures 4-5 and 4-6. Here, the storm-water system was designe d as a distributed, infiltra tion-based system relying on depression storage. Runoff is not deliberately redirected to be concentrated at a particular pond, thus the sub-area breakdown is much closer to the existing breakdown. The pond assumptions that applied to the traditional and cluster designs, also applied to the LID designs. These detention pond s, if required, were conceptual only (not shown on plans) and were added only if the distributed, infiltratio n-based storm-water system could not meet regulatory goals. Ponds were first added to those areas, such as Area 1, where runoff was already concentrated. If this was insufficient, ponds were then added to areas where runoff was not concentrated at a single point, such as Area 2 or 3. In this case, the pond was considered repres entative of a seri es of structures.

PAGE 87

74 Figure 4-5. “Partial” LI D design sub-areas. Figure 4-6. “Full” LID design sub-areas. Bioretention units on lots and storage fr om swale blocks and check dams were considered depression storage. Depression storage was simulated by increasing surface storage in the appropriate SMA unit above the 2 mm that all lawn and open space areas

PAGE 88

75 were assumed to have naturally. Thus, depres sion storage is inhere ntly assumed to be distributed evenly across the sub-area when, in fact, it is not n ecessarily. The volume of depression storage was first based on an interpretation of runoff quality regulations. This required that 12.7 mm (0.5 in) of storage be pr ovided for the disturbed area in each subarea. Additional increases in depression stor age were allowed for storm-water quantity management purposes. The physical area requi red for constructed depression storage assumed a 152.4 mm (6 in) dept h in bioretention areas and sw ales. Table 4-1 illustrates the results for adjusting SMA unit depression storage for selected sub-areas in the full LID design. The portion of the depression storage that is provided by bioretention was considered uncompacted area. The impact of this assumption on soil storage is small. Swales, because of their presence in the RO W were considered to be compacted area. Table 4-1. Adjusted depression storage for sele cted sub-areas of the “full” LID design. Sub-area Depth (mm)a Volume (m3) Quality Vol. Area per lot (m2)b 1 7.3 2341.7 1829.3 158.4 2 10.0 4207.0 2839.8 169.4 a Not including the natu rally occurring 2 mm. b Area of depression storage management pr actices per lot based on 152.4 mm (6 in) cell depth. The depression storage parameter is not rounded to the nearest millimeter like the other SMA parameters. This is mainly for th e economic analysis and will be discussed in the next chapter. Input data tables for the developed design models are located in Appendix A. Modeling Results Tables 4-2 and 4-3 present the modeling results for each design storm. These are also shown graphically in Figures 47 through 4-10 for the design storms only.

PAGE 89

76 Table 4-2. Uncontrolled model results for hypothetical design storms. Alternative 2yr 24hr peak flow (cms) Date and time 25yr 24hr peak flow (cms) Date and time Existing 3.344 2 June 0120 9.7969 2 June 0130 Traditional 8.577 2 June 0120 23.304 2 June 0130 Cluster 5.611 2 June 0130 15.426 2 June 0130 “Partial” LID 5.732 2 June 0140 16.076 2 June 0140 “Full” LID 3.991 2 June 0140 11.046 2 June 0140 Table 4-3. Controlled model resu lts for hypothetical design storms. Alternative 2yr 24hr peak flow (cms) Date and time 25yr 24hr peak flow (cms) Date and time Existing 3.344 2 June 0120 9.7969 2 June 0130 Traditional 2.791 2 June 0100 9.815 2 June 0310 Cluster 2.858 2 June 0110 8.090 2 June 0340 “Partial” LID 3.247 2 June 0130 10.633 2 June 0120 “Full” LID 3.269 2 June 0120 9.719 2 June 0140 Figure 4-7. 2-year, 24-hour controlled hydrographs.

PAGE 90

77 Figure 4-8. 2-year, 24-hour uncontrolled hydrographs. Figure 4-9. 25-year, 24-hour controlled hydrographs

PAGE 91

78 Figure 4-10. 25-year, 24-hour uncontrolled hydrographs Although the continuous simulation does not le nd itself to graphical representation because of the long time period, hydrographs from these simulations are in Appendix B. Results of this simulation are presented in Table 4-4 for selected peak events. Base flow results are shown in Table 4-5 for several time periods during the simulation. These sample points were times wh ere the flow was relatively steady. This indicated that only the second groundwater layer was contributing flow. Watershed Timing The design storm hydrologic models reveal se veral points of interest. The timing of the watershed response, due to the devel opment itself, did not appreciably change regardless of whether a pipe or swale system is used. Preservation of entire Tc paths was quite difficult becau se the number of paths to protect increases with the num ber of sub-areas, which make s it virtually impossible to prevent development impacts on the Tc path. Preservation of the initial sheet flow portions of the Tc path was possible. Admittedly, many sh eet flow areas began in off-site areas left undeveloped.

PAGE 92

79 Table 4-4. Continuous simulation selected peak flow points and total simulation volume. Alternativea Peak (cms) Date & time Peak (cms) Date & time Peak (cms) Date & time Peak (cms) Date & time Total Vol. mmb Existing 0.389 1 July 2210 0.196 18 July 1940 1.64 2 Sept. 1920 1.76 15 Sept. 0940 524.3 Traditional 0.41 1 July 2210 0.238 18 July 1930 1.67 2 Sept. 1920 2.75 15 Sept. 1000 544.1 Cluster 0.36 1 July 2210 0.2 18 July 2030 1.67 2 Sept. 1930 2.54 15 Sept. 1000 535.5 “Partial” LID 0.245 1 July 2210 0.18 18 July 2000 1.65 2 Sept. 1930 2.95 15 Sept. 1000 520.0 “Full” LID 0.325 1 July 2210 0.182 18 July 1940 1.65 2 Sept. 1930 2.75 15 Sept. 0930 520.6 aAll alternatives are cont rolled in this simulation. b Existing watershed area is 2.88773 km2 and developed watershed is 2.89255 km2. Table 4-5. Results of the continuous simu lation for selected base flow points. Alternative Flow (cms) Date & time Flow (cms) Date & time Flow (cms) Date & time Existing 0.119 6 July 1200 0.115 26 Aug. 1200 0.113 17 Oct. 1200 Traditional 0.095 6 July 1200 0.094 26 Aug. 1200 0.089 17 Oct. 1200 Cluster 0.099 6 July 1200 0.097 26 Aug. 1200 0.093 17 Oct. 1200 “Partial” LID 0.101 6 July 1200 0.100 26 Aug. 1200 0.095 17 Oct. 1200 “Full” LID 0.103 6 July 1200 0.101 26 Aug. 1200 0.097 17 Oct. 1200 The pond storm-water system had a more pronounced impact on timing than caused by the development itself for the 25-ye ar storm. The peaks were significantly delayed, which could create unanticipated flooding problems downstream. This delay was not present in the LID systems. Tc times for the LID system were sometimes larger due to slower transport in the sw ales and the longer flow path.

PAGE 93

80 Runoff Volume Uncontrolled runoff volume was increased in each development case, although the magnitude of this increase may be understated due to limitations in the model. The actual volume increase is always less than 25.4 mm (1 in) over existing conditions for the uncontrolled design storms. Preservation of open space had a positive bene fit to storm-water management that could be seen in comparison of traditiona l and cluster design uncontrolled results. Preserving open space, by converting to cluster development from traditional development patterns, resulted in an unc ontrolled volume reduction of 3.8 and 4.2 mm for the 2 and 25-year events respectively. If the same comparison is made between the “partial” and “full” LID desi gn, the reductions are 2.3 and 3. 0 mm for the 2 and 25-year events respectively. The LID designs also showed the bene fit of disconnecting impervious surfaces from the transportation network, at least in a qualitative sense. The off-site swales allowed for infiltration only in the LI D designs. Volume re duction was around 6.0 mm for a 2-year event and 7.5 mm for a 25-year event when comparing the traditional and “partial” LID designs. Comparing the cluster and “full” LID designs, reductions of 4.8 and 6.2 mm were observed for th e 2 and 25-year events. Therefore, it would be expected that c onverting from a traditional development, with full size lots and conventional storm-wa ter management, to a “full” LID design with a cluster layout and disconnected impervious surfaces, would result in a volume decrease of 8.5 mm (0.33 in) for a 2-year rain event. The effect on the 25-year event was observed as 10.5 mm (0.41 in).

PAGE 94

81 The pond storm-water management system provi ded little volume control, at least in the sense that surface outflow volume was d ecreased. This was expected due to the fact that HEC-HMS did not model infiltration or ET losses from the ponds. The increased volume was merely delayed in the slow rel ease, water quality treatment storage of the pond. The outflow at the end of the simulation from the conventional systems used in both the traditional land development desi gn and cluster development design are approximately 0.1 cms (3.5 cfs) greater than existing base flow for both design storms. The LID systems both returned to sub-surf ace base flows quite close to the existing conditions. The volume of water infiltrated by the LID system is dependent on extent of the area controlled. The LID storm-water ma nagement system reduced outflow volume over the simulation period by 10 mm (0.39 in ) over uncontrolled conditions for the “partial” LID design and by 5 mm ( 0.2 in) for the “full” LID design. An interesting observation was made be tween the uncontrolled “partial” LID design, where the storm-water transport syst em was swale-based, and the uncontrolled cluster design. Volume distri butions under the hydrographs we re nearly identical. This indicated that the impact on the Clark storag e coefficient from large lots was, possibly, offset by the swale-based tran sport, which has a longer Tc. The uncontrolled traditional design had a storage distribution characteristi c of many developed watersheds, while the uncontrolled “full” LID design, combining th e grass swale system with open space preservation, had a distribution closer to that of the existing watershed. The response from the controlled LID simulations was similar to the existing response. Peak Flows Both increased volume discharge and ch anges in volume distribution increased uncontrolled peak flows for all development designs. The uncontrolled peak flows for the

PAGE 95

82 cluster and “partial” LID designs are similar b ecause the grass swales may help offset the volume increase and distribution changes cause d by widespread clearing and grading. As might be expected from the previous discussi on, “full” LID results ar e the closest to the existing conditions. The pond system adequately controls peak flows for the 25-year event. The 2-year event peaks are somewhat lower than the ex isting 2-year peak. Runoff entering the pond fills the volume between the ci rcular orifice and second stage weir. This volume is the water quality treatment storage. This stor age volume is released slowly through the circular orifice to ensure acceptable treatment of runoff. The LID designs required more depression storage in all sub-areas than was calculated for water quality treatment. Table 4-6 shows the final average volumes of mana gement practices required. The depression storage values are a total site volume and an average per lot of the entire site. Volume and area of the LID management practices were varied on individual sub-areas. An exact breakdown on a sub-area basis is containe d in Appendix A. The feasibility of implementing the required area of depression st orage, as bioretention, is not addressed in this research. However, the resulting areas ar e similar to bioretention unit areas shown in examples from the Prince Georges County, Maryland design manuals. Therefore, it was assumed that the area requirements determined in this research are applicable. Table 4-6. Final volumes of management practices. Alternative Depression storage (m3) Average per lot (m3) Area per lot (m2) Total pond volume (m3)a Traditional 0 0 0 220,171 Cluster 0 0 0 157,122 “Partial” LID 32,059.11 44.59 292.58 29,576 “Full” LID 17,517.82 24.33 159.65 5,497 a Traditional and cluster designs utili ze wet ponds with permanent pool while LID designs utilize dry detention ponds.

PAGE 96

83 The LID designs could not meet peak flow control requirements with the final depression storage values. The values presen ted above represent th e stopping point at which dry detention ponds were consider ed. These ponds are conceptual and not explicitly designed, although the resulting dete ntion volume is probably quite accurate for the “full” LID design since it is placed in a sub-area (A rea 1) where runoff would concentrate naturally. The “partial” LID de sign required ponds in Area 1, Area 6 and Area 2. Of these three volumes, the one in Ar ea 2 is the most dubious since outflow from Area 2 is clearly not concentrated. Even w ith addition of three conventional detention ponds, the “partial” LID option could not meet peak flow controls for the 25-year design storm. This underscores the benefit of pres erving a portion of the watershed in natural condition. Certainly, additional ponds or preser vation of natural upland area on the lots would likely improve the performance of th e “partial” LID design, but since the ponds were conceptual, it was decided to end the analys is at this point due to the fact that ponds were now required in sub-area s that did not have a natura l flow concentration point. Continuous Model The continuous models were run with a ll runoff controls implemented. Peak flow results for the continuous models showed mi xed results. Three convection-type storms are presented first. In thes e cases, pond systems tended to be very slightly over the existing peak, while the LID designs were very slightly (with one exception on 1 July for the “partial” LID design) under the existing peak, although LID performance was close to the peak for the largest conv ection event on 2 September. The LID designs were under the existing peak for the first two events becau se the depression storage controls in the lowland sub-areas, which are the main runoff producers for these events, reduce surface runoff. The fact that pond system peak flow s were slightly higher than most existing

PAGE 97

84 peaks was possibly due to increased surface runoff from development impacts in the lowland sub-areas and pond outflow. Two storm events where all designs failed to meet peak flow controls were the convection event on 2 September and the tropical storm on 14-15 September. Each design was about 1.0 cms (35.31 cfs) over existin g conditions for the tr opical event, but only slightly so for the conv ection storm. A sustained peri od of flow observed after the large convection on event 2 Septembe r indicated that the interflow (1st groundwater layer) was full. Inspection of the model re sults showed that ju st prior to the 14-15 September event, the interflow layer had a hi gher percent of storage full than existing conditions. The reason for this might be attr ibuted to runoff from indirectly connected impervious surfaces being routed over the pe rvious areas resulting in higher antecedent moisture conditions. As would be expecte d, based on the previous discussion, open space preservation in the uplands reduces the impact However, it is also clear that LID designs are less capable of dealing with this storm. Overall volume discharged through th e simulation period increased 10-20 mm (0.39-0.79 in) for the pond system while the LID options were slightly lower, about 4.0 mm, than the existing conditions. Each design al ternative resulted in slightly lower flows at the three base flow periods observed due to a smaller area contributing groundwater flow (the volume of groundwater flow is r outed through a linear reservoir to determine flow at any given point in time). Therefore, the conclusion was that any increase in total volume is due to increased surface runoff or interflow. The magnitude of base flow difference is not great; however, the cumulativ e impact would result in a lower base flow outflow volume.

PAGE 98

85 Structural Controls Table 4-6 clearly shows the impact of preserving natural open space on structural storm-water management. The reduced lot desi gns all required less structural storm-water management. If the storm-water management system failed in the future, the impact on the receiving waters may be less for the redu ced lot designs due to less developed area. Model Limitations This model and procedure had certain limitations. The procedure used for indirectly connected impervious surfaces pr esents the biggest concern. The ability to identify a percent impervious was added after the Technical Reference Manual for HECHMS was written and is not fully documented. The same option was implemented for other loss methods, such as the curve number method, from the first release. From this documentation, it was determined that this option only represents directly connected impervious surfaces and that the SMA unit would not apply in this portion of the watershed. This was confirmed by running HEC-HMS on two test sub-areas with identical SMA parameters and one sub-area with 10% imperviousness defined. One subarea (with the 10% imperviousness) is 1.0 km2 and the other is 0.9 km2. The groundwater time-series, soil moisture time-series and gr oundwater outflow volume are identical. This is logical since the SMA model represents vertical storage. Horizontal storage is modeled through the linear reservoir gr oundwater routing routine. Ther efore, indirectly connected surfaces must also have zero ve rtical storage. The procedure used in this research to model indirectly connected impervious surf aces may introduce error to the computations. The SMA percolation equations are dependent on the ratio of storage filled in the upper contributing subsurface layer to the lower rece iving layer. Percolation between one layer and the next lowest is greatest when the upper layer is full and the lower layer is empty.

PAGE 99

86 Storages fill faster no matter which modeling “trick”, previously described, is used. However, simple recreation of the spreadshee t computations presen ted by Bennett (1998) for each option yielded slightly different re sults. A HEC-HMS test run also yielded slightly different results; how ever, the overall impact was very slight: a difference of 17 m3 for a 12.7 mm storm on 1.0 km2 sub-area (0.9 km2 for the increased intensity subarea). Regardless of the method of modeling i ndirectly connected, impervious surfaces, another source of error will occur. The model assumes that the runoff from these surfaces is instantly and uniformly distributed over the re st of the sub-area. This is, of course, not generally the case. The degree of error th at is added is not known, but is common throughout each design simulation. Any hydrolog ic model that does not explicitly have routines for indirectly connected impervi ous surfaces will present this problem. A similar concern exists for depression st orage. Surface runoff can only occur after all depression storage is filled. Depression storage was assumed to be uniformly distributed across the sub-area, which is not tr ue. Furthermore, runoff from road surfaces would not normally flow into bioretention units Therefore, it is pos sible that runoff from the road will exceed the depression storage pr ovided in the swales and flow to the basin outlet before the lawn and roof areas are co ntributing runoff to basin outlet. The model does not consider this distinction. Finally, there is the matter of the Clark storage coefficient. This unit hydrograph method is not widely used because of difficulties in determining this coefficient. The calibrated coefficients are fi ne, but the method of computing the future coefficient, by weighting the proportion of the sub-area in natural condition versus developed condition

PAGE 100

87 to compute the multiplication factor used with the Tc to determine the Clark storage coefficient, cannot be validated. The assump tion is that the greater the proportion of natural open space area, the closer the co efficient will be to the existing one. A distributed parameter model could be us eful for modeling residential storm-water management systems. This w ould be particularly true for the distributed, infiltrationbased LID storm-water management syst em. The SMA module has a distributed parameter option. However, this option is based on a grid fixed to a geographic map projection. The grid lines can not be adjusted to coincide with lots, which would be the best way to model LID systems. Hydrologic Summary and Conclusions Comparison of the four design alternatives in this hydrologic study led to several conclusions. Application of the infiltrationbased, distributed storm-water management system appears to result in a developed wa tershed response that is closest to natural conditions, particularly if it is accompanie d by a program of land preservation around stream corridors and upland high infiltration areas. Land preservation also decreased reliance on storm-water management control pr actices. This, at least in a qualitative sense, reinforces the LID design philo sophy presented by Prince Georges County, Maryland. Hydrologic performance could certa inly be improved (and possibly eliminate the conceptual ponds in Sub-ar ea 2) in the “partial” LID de sign if a portion of the lots were preserved in their natural condition. Howe ver, this approach would seem to violate the LID concept of minimizing development in high infiltration soils because large areas of uplands would still be impacted while the “full” LID impact ed a smaller area. LID storm-water systems may perform wo rse than a conventional pond system when antecedent moisture conditions are above average. The land pr eservation aspect of

PAGE 101

88 LID lessens this effect, but the “full” LID design only performed about the same as the traditional design for the 14-15 September event. The “full” LID design used a rather extreme reduction in lot size. Performance of the LID storm-water management system decreased for this event as lot size increase d, eventually resulting in peak flows worse than all other designs. Again, none of the storm-water management options were adequate for this event. The economic feasibility of the design alternatives is a critical factor to consider for two reasons. First, there is the issue of whether construction co sts associated with alternative storm-water management pract ices are competitive with those of the conventional system. Second, the increased ope n space necessarily decreases lot areas. Much of the value of a vacant lot is tied up in the lot area. Therefore, land preservation may negatively impact the receipts from lot sa les. An analysis of the economics of these four design alternatives is disc ussed in the following chapter.

PAGE 102

89 CHAPTER 5 ECONOMIC ANALYSIS Introduction This chapter is the economic portion of a study on the implementation of alternative land development and storm-wate r management on a hypothetical residential development project in NorthCentral Florida. The previous chapter dealt with hydrologic aspects of the implementati on of these practices. The effect on lot sale price is particularly important in this implementation of alternative practices because a cluster deve lopment concept was used in two of the alternatives for the site design rather than full size lots allowed by zoning. The cluster design represented land preserva tion with conventional storm-water management, and the “full” LID design used the same cluster site plan with a distribut ed, infiltration-based storm-water management system. Reductions in lot size, to preserve open space, would be expected to have a negative influence on the lot sale price since much of the value of the lot is in its size. A site design that maximizes the number of lo ts that have open space frontage may reduce the loss the developer will have due to smaller lot sizes. Construction costs for some of the a lternative design options may be lower, primarily due to reduction or removal of certain capital impr ovements, such as curb and gutter and storm sewer pipe. The lower deve lopment costs would decrease the financing required by the developer.

PAGE 103

90 Review of Literature Economic data relevant to this project can be divided into two types: general cost data and market dependent. General cost da ta are simply the costs associated with a certain construction activity or material a nd are readily available from a variety of construction cost surveys, such as the annual R.S. Means (2001) publications. Basic cost information on bioretention is available on the Low Impact Development Center’s web page (Low Impact Development Center, 2003) Sample et al. (2003) provides cost functions for other storm-water management practices. Liptan (1995) provides three examples of construction projects that did, or could have, saved significant amounts of money if alternative pr actices had been used. A literature search conducted by CH2MHill (Ch. 4, p. 2, 1993) cited a study showing that “intra-neighborhood” services, in cluding storm sewers, are highly sensitive to net density and lot size. The basic conclusion was that the capital cost per dwelling unit for these services would decrease with increasingly compact developments with smaller lots. Market data reflect the “willingness to pa y” of prospective homebuyers to certain features, such as lot size, open space, wooded lots etc. Several studies evaluated the effect of open space on house sale price. Bolitzer and Netusil ( 2000) concluded that prices rose as proximity to open space rose in Portland, Oregon. Th eir results were not statistically significant at extremely close distances—possibly due to a small number of properties in this distance range or positive and negative influences netting out. This study was the basis of this por tion of the research since it was directly “on point.” Doss and Taff (1996) studied the effect of proximity to four wetland types and concluded that only forested we tlands showed negative price influence. Correll et al. (1978) concluded that sale price decrease d with distance from greenbelts. Do and

PAGE 104

91 Grudnitski (1995) concluded that houses on lots abutting golf courses sell at 7.6% higher than those that do not. Weicher and Zerbst (1973) concluded that city parks may add value to nearby houses, but the positive effect is dependent on orientation of the house to the park (facing it or backing to it) and wh ether the lot actually faces open space or a recreation structure. Conversely, Li and Br own (1980) concluded that distance from conservation land was not significant, but that distance to the ocean, rivers and recreation areas and visual quality were significant and positive with increasing proximity or quality. Methods Of Economic Analysis There were two components to this res earch: construction costs and sale price. Construction cost estimation was a fairly simp le procedure, while pr ediction of sale price required statistical analysis and the manipulation of GIS and sales data. Construction Cost Estimation Estimation of construction costs was base d on existing cost data. The R. S. Means (2001) publication on site development costs wa s the best source of cost data. However, other sources of data were required for certain aspects of the resear ch, particularly stormwater management infrastructure excluding pipe Sample et al. (2003) was referenced for swales and pond structures. Bioretention is a relatively poorly documented practice, in terms of costs, because it is a new practice. Cost data for bioretention were taken from the Low Impact Design Center (2003). Raw construction costs were totaled from each design plan. The traditional design option became the baseline for measurement. The R. S. Means publications include location modifiers (0.816 for Gainesville, Flor ida). This modifier was applied after the raw costs are totaled. The final result was the composite cost estimate for that particular

PAGE 105

92 design option. Not every possi ble capital improvement was measured. For example, sanitary sewer construction was not analy zed. Table 5-1 lists th e capital improvements included in this analysis and the source of th eir cost estimate. Full or altered assemblies (a collection of line items for a common develo pment feature such as roads) were used wherever possible. If this wa s not possible the line item costs or a collection of line items and assemblies were used. Cost data were ta ken “as is” and not adjusted for inflation. Table 5-1. Capital improvements included in analysis. Parameter Source Notes Roads R.S. Means 24’ road assembly minus curbs & lines Curb & gutter: straight R.S. Means Steel form 24” wide Curb & gutter: radius R.S. Means Steel form 24” wide Sidewalk R.S. Means Sidewalk assembly Manhole/inlet R.S. Means Concrete, pre-cast 4’ deep assembly Storm sewer: 24” R.S. Means Class 3 RCP with 4’ deep trench assembly Grading R.S. Means All cleared areas except roads/sidewalks Pipe end section R.S. Means Quoted from private corp. Site clearing R.S. Means Clear ing/chipping med. Trees, grubbing Swales Sample et al. Function coefficient: 31.15 Landscaping R.S. Means Average of various species. 10% of lot. Bioretention LID Center $3.5 per square foot Lawn grass R.S. Means Ground cover assembly. Storm-water ponds Sample et al “Retention” basin function. Note: Units in Imperial system for ease of reference. Four costs require further di scussion. Sample et al. ( 2003) listed a function based on length of swales that has a variable coefficient. It was decided to use the mean value of this coefficient. The swale cost was assu med to include costs of additional components such as check-dams and level spreading devi ces to return flow from the channel to overland flow. The pond system function also presented a concern. Three pond functions are given in Sample et al. (2003): infiltration basin, rete ntion basin and detention basin. In Florida, “retention basin” generally m eans “infiltration basin”. Howeve r, Dunn et al. (1995) states

PAGE 106

93 that wet ponds are sometimes called retention pon ds. Therefore, this re search utilized the retention function for wet ponds. Dry detention ponds also utilized th e same function as the wet ponds rather than the function for detention ponds. The same function was used because of the uncertainties associated w ith the different naming conventions of pond systems. The final two cost line items were bi oretention and landscaping. The Low Impact Development Center (2003) reports a “r ule of thumb” cost of $32.29-$43.06 per m2 ($3$4 per ft2), which includes landscaping. The mean of this range was selected for this research. The Low Impact Development Center (2003) also suggested that landscaping costs for landscaping required by local ordinanc e be subtracted from the bioretention cost since this landscaping must be done anyway and bioretention areas will count towards landscaping. In this resear ch, required landscaping cost s for two conventional stormwater management options, traditional and cluster design were included, but no landscaping cost for the LID designs were in cluded. The required landscaping was based on 10% of the lot. One large and two medium trees were assumed. This assumption is arbitrary since the developer would be under no obligation to use tr ees. Trees, however, provide an equivalent area of landscaping. This means that one tree can be used instead of an area of shrubs or ground cover (ACBCC, 2002). The balance of the landscaped area used shrubs with an average stem density si milar to that of bioretention (CH2MHill, Sec. 8, p. 7, 1998). Costs for trees and shrubs were determined by averaging the prices of species that might be used in local landscapi ng. The prices were taken from R. S. Means (2001).

PAGE 107

94 Hedonic Price Equation for Predicting Sale Price The technique used in this study, and the ot her studies referenced here, utilized a technique called hedonic pric ing. Freeman (pp. 367-420, 1993), st ates that the sale price of a home or lot is a function of a vari ety of structural (i.e., square footage), environmental (i.e., heavy traffic) and neighbor hood (i.e., location & open space) factors. The partial derivative of the price equation with respect to a particular parameter (variable) results in the impact on sale price of a marginal change in that particular parameter. The hedonic price technique requires certain assumptions that are discussed in Freeman (pp. 367-420, 1993), and include market equilibrium and the study area being one housing market. Typically, variables describing the home a nd property are collected along with sale price, and a correlation matrix is develope d. Correlation is used to determine what degree of variation in one variable can be at tributed to linear re lationship with another variable. Two variables with a high correlation coefficient ma y indicate that only one is needed in the price equation—for example num ber of bedrooms and total square footage may be highly correlated, thus allowing th e number of bedrooms to be dropped. This assumes that a relationship is a ccepted between the variables. Once the variables are finalized, multiple re gressions, using any one of a variety of functional forms, are carried out on the data The constants of each variable in the resulting equation are the dollar amount, in a linear model, associated with an increase in that variable. An alternative is to transform the data such as taking the natural log of both the dependent and independent variables. In th is case, the resulting coefficient is not the absolute dollar impact, but the elasticity of that particular variable. Elas ticity is defined as the responsiveness of one variable to another. In this case the response of the dependent

PAGE 108

95 variable to a change in one of the independent va riables. An analysis of this sort is market specific. Thus, the results should not be applied to another market area. Collection of sales data Sales data between November 1999 and October 2001 from the Alachua County Property Appraiser’s website (Alachua County, 2003) were so rted for download using a variety of constraints includi ng single-family property, vacan t lot, qualified sale and acreage (less th an ~32,371 m2 or 8 Ac). Qualified sales ap peared to indicate that the transaction was at an “arm’s length” since many unqualified sales include transactions valued at $100. Vacant lots were used in this analysis to remove the influence of the home. The home itself has little bearing on this research. Lot sales were downloaded to a delineated text file, and then saved through Excel as a database file. This file was then load ed into ArcView 3.2. Sets of parcel GIS data (ArcView shape files) were obtained from the Alachua County Property Appraiser. The sales data were joined with the parcel map. T hose parcels that were not sold in the period of interest were identified by query and remove d. New shape files were thus created that contained sales only. The sales were divided sales into two periods from November 1999 to October 2000 (referred to as 2000 sales) and from N ovember 2000 to October 2001 (referred to as 2001 sales). The reasons for this division we re based on collection of open space data, which could only be collected for a tax (calenda r) year. The tax roll for a year is based on the condition of a parcel on 1 January of th at year. Thus, sales from late 1999 through late 2000 should be compared against open space from the 2000 tax roll. There is potential for error the later the sale takes place in the year, since the actual open space conditions in the study area are clos er to the following tax year.

PAGE 109

96 The sales data were further limited to those sales that occurred within the city of Gainesville and the urban cluster identifie d by Alachua County. The result was 282 sales in 2000 and 230 sales in 2001 for a total data set of 512. One parcel was identified as a double entry. The final data set contained 511 sales samples. The final manipulation of the sales data was addition of lot area. Th is was accomplished through a custom script downloaded from the ESRI website. Collection of open space data Collection of open space data was somewhat more difficult than the collection of sales data. Open space parcel information co uld only be downloaded for a tax year. For example, downloading year 2000 tax data will re sult in a text file of parcels, as they existed on 1 January 2000. This led to the que stion of what tax year should a set of sales be compared against. In this research, sale s were compared against the tax roll of the same year. This means that sales in 2000 were compared against th e 2000 tax roll, which would represent open space parcels as of 1 January 2000. Of course, new open space parcels are likely to be formed throughout th e year. This is most likely to occur with common areas within subdivisi ons. Any sale after 31 October of a given year was considered in the next year’s sales on the assumption that by November open space conditions are nearly the same as the next year’s tax roll. This is not an assumption that is free from error. Once the open space data were set for a year, a vi sual investigation around the lots sold was conducted to see if any parcels of open space were missed. The GIS shape file that was used in th is research was dated October 2002. Thus, there was the possibility that open space parcels identified in 2000 and 2001 might be “lost” because they did not exist in the 2002 sh ape file. This was not considered a serious concern. If a parcel actually did exist, and wasn’t a typographical error, loss of this parcel

PAGE 110

97 indicated that it wasn’t truly open space or it was later combined with another parcel. Open space in this research was assumed to be a permanent feature in the study area. Open space was classified into three categ ories: golf courses, common areas in subdivisions and forest/parks. The forest/par k classification is a category that includes state and municipal parks, both recreation a nd natural preserves, along with private and public conservation land, but does not include land zoned for commercial forestry. The 26 golf course parcels presented no problems in identification. Common areas and forests/parks presented several problem s with identification. The results of the property search for common areas and forest/p arks, using the appropriate land use code, were incomplete. There were 150 parcels from the 2002 shape file id entified as probably being common area but were classified w ith a generic parcel number 50000-000-000. These “unknown” parcels were impossible to place in a specific time period because no information was available from the Property Appraiser’s website. They were added to both the 2000 and 2001 lists of open space. The total number of forest/park parcels downloaded totaled 70 for each year. It was clear, from a visual inspection of the parc els in the study area, th at a significant number of forest/park parcels were missed in the initial data download. These parcels being classified as some form of government-own ed parcel caused the problem. In addition, parcels designated as conserva tion zoning or in private c onservation trust were missed. Downloading government-owned parcels and searching the resulting text file for appropriate parcels identifi ed around 200 missing forest and parkland parcels for each year. Parcels were identified by the owner (i.e ., Florida Dept. of Natural Resources), title of the property (as in name of the park) or by checki ng the online GIS option, which

PAGE 111

98 showed the parcel with a recent aerial photograph (this GIS data could not be downloaded). City parks were included only if they contained mostly vegetated areas. Parks that were mostly developed, such as public pools, were not included. Parcels zoned as conservation were also a dded as forest parcels. These properties were rather difficult to identify. This was pr imarily because several parcels in an area may be zoned conservation, but another parcel is not. In these cases, the parcel that was not zoned conservation was also included on the assumption that it would later be rezoned. Conservation parcels may be develope d; however, this development is heavily restricted and will always result in a large amount of open space. Private conservation trusts re presented the final component of the forest/park data set. These parcels represented a minority of th e data set (57 total). They were identified by owner name (i.e., Alachua Conservation Trus t). Then, the parcel’s appraisal and sales history was checked on the Property Appraise r’s website to determine the appropriate year (2000 or 2001) this parcel first existed. The final open space data set (combined file of each type) resulted in over 4000 open space parcels, the overwhelming majority common areas, in the general study area for each year. The numbers given in this secti on were significantly more than the total of the actual downloads. The increase was due to geoprocessing operations on the shape files, which split any polygons that have b een combined. For example, a series of separate polygons, but all with the same par cel number, may represent common area in a development. The County GIS files treat th ese as one polygon until a merge or trim operation is done. This split was considered bene ficial to this research since it allowed a more accurate rendering of the open space area nearest to each lot sold.

PAGE 112

99 Development of lot descriptive variables The hedonic price technique requires a matr ix of variables be developed that describe each lot. There is no “correct” matrix or number of variable s. The variables used to describe the lots were as follows: Lot area Distance to a reference parcel near the University of Florida Campus Location on a grid based on beari ng from the reference parcel Proximity to open space based on five distance zones Area of the nearest open space parcel Type of open space Areas of all parcels were computed w ith a custom ArcView script downloaded from the ESRI website. Distance and bearing to a reference parcel was determined with a downloaded ArcView extension (titled Dist ance/Bearing: Matched Features). The reference parcel was located at the corner of University Avenue and 13th Street between the major retail areas west of the university and downtown Gainesville. The location of each parcel was evaluated by placing an imaginary grid through the reference parcel. A pair of bi nary variables (“dummy” variable s) describes the location of the lot. A variable named “North” has a valu e of 1 if the lot is located north of the reference parcel and 0 if sout h. Similarly, the “East” variable has a value of 1 if east of the reference parcel and 0 if west. Thus, a parc el northwest of the re ference parcel has a 0 for the east-west and a 1 for the north-south axes. All open space parcels were assigned a numeric code in the ArcView database, 0 for common areas, 1 for forest/p arks and 2 for golf courses. Th ese were used in Excel to assign a binary variable for each of the three type variables of open space for the correlation and regression analyses. A sec ond extension was (titled Nearest Features v3.6d) used to determine the distance to th e three nearest open space parcels from each

PAGE 113

100 lot in the sales data set. Checking for the three nearest parcels ensured that multiple parcels at the same distance (usually frontage parcels) were selected. This extension also recorded the open space type number and it’s area. These data were then transferred to an Ex cel spreadsheet. The sa le price of the lot and all independent variables with actual num eric information, such as distance or area, were transformed by natural l ogarithm, resulting in a full natural logarith mic regression. The spreadsheet was used to determine how many of the three nearest open space features to consider for distance, area and fo r the three type binary variables. In most cases, this was only one, but there were a fe w instances where two parcels of open space were equidistant from a particular lot. In these cases, the areas of all the appropriate parcels were aggregated. Five distance zones were considered : frontage (0 distance), zone 1 (0.1-30.5 m), zone 2 (30.6-121.9 m), zone 3 (122.0-243.8 m) and zone 4 (243.8-365.7 m). These zones are similar to those utili zed by Bolitzer and Netusil (2000), although their study did not include frontage on open sp ace as a separate zone. The distance of 365.7 m (1,200 ft) was considered re presentative of how far a pe rson is willing to walk to utilize an amenity. Open space beyond 365.7 m is not considered in this analysis. The open space proximity variables were bi nary with 1 indicating presence. Type variables were also binary. A valu e of 1 was assigned to the appropriate variable (“Common”, “Forest” or “Golf”) de pending on the type of the nearest open space. If a lot was equidistant from two di fferent types of open space, it received credit for both types. The fundamental assumption of this model wa s that the lot purcha ser places all of his or her value on the neares t open space parcel. The only time more than one parcel was

PAGE 114

101 considered was when two or more open space parcels were equidistant from one of the lots in the sales dataset. Results of Economic Analysis The results of the economic analysis are pres ented in three parts. The results of the individual analyses for open space impact and construction cost are presented separately. The combined impact on the developer is then discussed. Hedonic Price Model Results The set-up of the data for regression analys is was discussed in the previous section. The results of the price analysis are presente d in four sections. First, the descriptive statistics for the data set are presented. S econd, the purpose and results of the correlation analysis is discussed. Third, the actual genera tion of the price equation through regression is discussed. Finally, the impact of ma ximizing open space proximity is discussed. Descriptive statistics The descriptive statistics for the lot data are shown in Table 5-2. The hypothetical lot sizes used in this study are within the ra nge of the lot sales data set. Therefore, the price model generated by this analysis should be applicable to the hypothetical lots. Table 5-3 shows the breakdown of the first open space type and proximity to the first open space parcel. There were 10 lots that were equidistant to two types of open space and receive credit for both ty pes. 10 parcels were located east of the reference parcel and 201 were located north. Table 5-2. Descriptive statistics of lots in sales data set. Area (m2) Sale PriceDist. To Ref. Parcel (m) Open Space Area (m2) Mean 1,722.4 43,400 33,144 279,871 Std. Dev. 2,246.2 2,2376.7 8,721 713,318 Minimum 124.1 5,500 6,366 0 Maximum 25,379.9 145,000 49,736 5,532,579

PAGE 115

102 Table 5-3. Open space type and proximity breakdown. Count Frontage 240 Zone 1: 1 to 30.5 m 40 Zone 2: 30.5 to 121.9 m 155 Zone 3: 121.9 to 243.8 m 45 Zone 4: 243.8 to 365.7 m 10 No open space in 365.7 m 21 Golf course 11 Forest or parks 19 Common area 474 Total lots in study 511 Correlation results The first statistical opera tion performed was a correl ation of the independent variables, using Excel (data analysis correla tion function). Correlation is computed as the covariance of two variables divided by the product of their standa rd deviations. The correlation coefficient represents the degr ee of linear relations hip between the two variables. The purpose of the correlation, in terms of th e hedonic price analysis, is to identify situations were two descriptiv e variables are describing the sa me feature. If this is the case, the presence of both variables in the regression may adversely affect the results since the variables in question are not independe nt. An example of this situation would be a model that includes both home area and num ber of bedrooms. The number of bedrooms in a home will increase with the home area. Thus, these two variables are both measuring the size of the home. In this situation, it would be advisable to remove one of the variables from the regression analysis. Ther e must be a relationship between the two variables, however. Two variables may have a high correlation, but have no real relationship with each other. Table 5-4 shows the results of the correlation between the independent variables.

PAGE 116

103 Table 5-4. Correlation results be tween independent variables. Lot AreaGolf East FrontageZone 1Zone 2Zone 3Zone 4Forest N orth Dist. OS Area Common Lot Area 1.00 Golf -0.10 1.00 East 0.12 -0.02 1.00 Frontage -0.15 0.13 -0.10 1.00 Zone 1 0.07 -0.04 -0.04 -0.27 1.00 Zone 2 -0.11 -0.07 -0.06 -0.62 -0.19 1.00 Zone 3 0.08 -0.05 0.01 -0.29 -0.09 -0.21 1.00 Zone 4 0.13 -0.02 0.08 -0.13 -0 .04 -0.09 -0.04 1.00 Forest 0.08 -0.03 0.27 -0.08 -0.02 0.05 0.01 0.20 1.00 N orth -0.11 0.07 0.03 -0.14 0. 09 0.10 0.05 0.00 0.05 1.00 Dist. 0.14 0.08 -0.22 0.14 -0.08 -0. 14 0.07 -0.06 -0.12 -0.44 1.00 OS Area -0.16 0.22 -0.29 0.25 0.03 0.00 0.02 0.07 0.20 0.18 -0.01 1.00 Common -0.27 -0.06 -0.51 0.19 0.05 0. 04 0.03 -0.12 -0.66 0.02 0.09 0.31 1.00 There is very little correlation between these variables except in two cases. One case is between the proximity variables Fr ontage and Zone 2. This relatively high correlation means that if a parcel does not have frontage on open space, it more than likely is between 30.6-121.9 m (>100 to 400 ft) fr om the nearest open space parcel. The common area and forest/park type variables al so show a high degree of correlation. This indicated that if a parcel does not encounter common area as the nearest open space, it is likely to encounter a forest or park. Eliminati on of either one of th e zone variables would require the elimination of each following zone, so no proximity variables were eliminated. Neither the common area or forest ty pe variables were eliminated in the first regression analysis. Regression results The next analysis was generation of a price equation through regression. The estimated lot sale price is the dependent vari able, while the other parameters are a set of independent variables. Excel us es a linear, least-squares regr ession. In this analysis the data was transformed by taking the natura l logarithm of both the independent and

PAGE 117

104 dependent variable. Multiple regression results in an intercept and series of regression coefficients. Thus, the price equa tion can thus be represented as: Ln(P) = I + C1ln(x1) + C2ln(x2)….+Ciln(xi) (Eqn. 5-1) Where x is an independent variable, I is the e quation intercept, P is sale price in dollars and C is a regression coefficient. Typically, a du mmy variable is represented as 1 or 0. In natural logarithm terms, these dummy variab les are technically ~2.71 and 1, respectively (the natural log of these values result in 1 and 0). Significance of the independent variable s is measured by the t statistic. The desirable level of significance, for this an alysis, was at the 95% level (t stat = 1.96). Significance at the 90% level (t stat = 1.645) was considered the minimum acceptable level. The Adjusted R2 value is the measure of the “goodne ss of fit” of the price equation. Higher R2 values are, of course, more desirable. The results of the Excel regression are s hown in Table 5-5. Th e variables defining forest/parks parcels as the nearest open space type, distance from the reference parcel and the lot location north of the reference parcel are significant at the 90% level. The variable defining common area as the nearest type and the open space area are insignificant (t stat less than 1.645). All other va riables are signif icant at or above the 95% level. The common area variable may be insignifi cant because the forest/park variable was defining the same characteristic of the lot. Specifically, the presence of forest/park as the first open space encountered denoted the lack of common area. It is also possible that the common area variable is insignificant beca use it represents a baseline condition since most lots in the study encounter common area as the first open space parcel. Regardless,

PAGE 118

105 the results of the regression indicate that open sp ace type is only signifi cant if it is a golf course or a forest/park. Table 5-5. Initial regression results. Coefficients t stat Intercept 6.969 17.661 Lot Area 0.502 28.109 Golf 0.336 3.627 East -0.971 -9.184 Frontage 0.973 7.017 Zone 1 0.870 6.189 Zone 2 0.755 5.584 Zone 3 0.594 4.268 Zone 4 0.578 3.617 Forest -0.228 -1.788 North -0.051 -1.779 Dist. -0.070 -1.751 OS Area -0.006 -0.707 Common -0.048 -0.413 R2 0.689 F stat. 84 Table 5-6 shows the results of a regressi on if the insignificant open space area and common nearest variables are eliminated. This version of the price equation was used to evaluate the economics of the f our design alternatives discusse d in the previous chapters. The elimination of the two variables did not impact the overall fit of the data. Most variables are significant at least the 95% le vel. Distance from the reference parcel remained significant at the 90% level.

PAGE 119

106 Table 5-6. Modified regression results. Coefficients t Stat Intercept 6.986 17.809 Lot Area 0.500 28.589 Golf 0.326 3.827 East -0.955 -9.266 Frontage 0.880 12.005 Zone 1 0.781 9.470 Zone 2 0.667 8.887 Zone 3 0.505 6.297 Zone 4 0.487 4.412 Forest -0.199 -2.896 N orth -0.056 -1.998 Dist. -0.072 -1.788 R2 0.682 F stat. 100 The price equation, following the format of Equation 5-1, is shown in equation 52. The R2 value of 0.682 indicates that 31.8% of the variation in price is not represented by this equation. Ln(price) = [6.986 + 0.5ln(lot area) + 0.326(golf) -0.955(east) + 0.88(frontage) + 0.781(Z1) + 0.667(Z2 ) + 0.505(Z3) + 0.487(Z4) -0.199(forest) -0.056(north) -0.072ln(dist.)] (Eqn. 5-2) Lot area has a highly significant and positive impact on sale price. The negative coefficients on the “North” and “East” locati on values should be inte rpreted as indicating that lots north and/or east of the reference pa rcel are expected to sell for less than those south and/or west. The magnitude of th e “East” coefficient clearly shows the attractiveness, to homebuyers, of the suburba n areas west of Gainesville. The negative impact of distance from the reference parcel was expected. This could be interpreted as the negative influen ce of longer trips. The open space values must be considered together to measure the impact of open space. Open space area is insignificant so value of open space is measured by both

PAGE 120

107 proximity to the lot and the type of open space. Thus, the full impact of open space is the combination of the proximity and type effects. The negative relationship of the forest/park variable indicated that some form of negative influence resulted from these type s of open space. One possibility, since many of these parcels are neighborhood parks, is th at noise was a factor. Another possibility was that common areas, the most frequent open space type encountered, were perceived as a “private” park only availa ble to residents of the subdi vision, whereas the forest/park variable represented areas open to the public. Regardless, the impact of a lot being within 365.7 m (1,200 ft) of a forest or park was still positive, though not as positive as proximity to common areas. Proximity to a golf course had a positive impact on sale price. Stating that common areas have a neutral influence while golf courses and forest/parks have positive and negative in fluences, respectively, could summarize the type variable results. The proximity coefficients are positive in relation to price and show a lessening impact as distance increases. The proximity re sults indicated that the positive impact of open space on home sale price is less with in creasing distance. A possible explanation for this result is that lots with open space fr ontage provide a greater appeal since the owner has a view of a natural or landscaped area rather than another owner’s back yard. Impact of open space frontage Certainly, if the cluster and “full” LID f ootprints did not attempt to maximize open space frontage, the return to the developer would be less. As an example, consider the average open space proximity coefficient that was determined for the reduced lot size footprint. The development scenarios were analyzed to determin e the actual number of lots for each open space proximity zone. An average coefficient for open space proximity

PAGE 121

108 was determined for the entire development assuming that natural open space is considered common area. For the full size lo t designs, this coefficient was 0.746 and for designs with reduced lot size, cluster f ootprint 0.874. Average lot sizes of 1,871.16 m2 (~20,000 SF) and 934.65 m2 (~10,000) were considered (the actual average over each design was certainly somewhat higher). Lots we re assumed to be at an average distance of 12,191.41 m (40,000 ft or 7.5 miles) from the re ference parcel and northwest from this parcel. Full size lot designs have predicted receipts from sales of approximately $34.1 million compared to $27.4 million for the reduced lot designs. If this simple analysis is redone, with open space islands removed (most open space in now contiguous), the average open space proximity coefficient for the reduced-area lots is now 0.798. The sales receipts would now be $25.4 million, representing a reduction of $2.0 million. Thus, the loss the developer would suffer from reducing lot size would be decreased from $8.7 million to $6.7 million if open space frontage is maximized. Construction Cost Results Construction costs were measured for th e four development plans generated in AutoCAD Land Desktop 3 (developed by AutoDesk). The construction costs are measured directly form the site plans. Constr uction take-offs of this nature must also be considered estimates, particularly because th ey are priced through sources that compile average price data. The actual construc tion price will depend upon availability of materials and other market factors at the time materials are ordered. Complete results for each line item are available in Appendix B. Table 5-7 lists the total estimated costs and costs per lot for each design alternative. These costs are adjusted with the 0.816 location factor for Gainesvi lle from Means (2001). The traditional and

PAGE 122

109 “partial” LID designs are based on a 719-lot su bdivision, while the ot her alternatives are based on a 720-lot subdivision. Table 5-7. Construction cost estimates. Alternative Total: Per Unit Traditional $12,287,688$17,090 Cluster $8,279,119$11,499 Partial LID $12,038,240$16,743 Full LID $7,012,326$9,739 Construction cost estimates indicated that the greatest cost savings resulted from decreasing the area impacted by the developmen t project. The actual cost reductions from implementing the LID-type storm-water management system, including removing sidewalks, does result in cost savings less than $2,000 per unit for the “full” LID option and less than $500 for the “partial” LID opti on. This relatively small reduction was attributed to the heavy use of bioretention units in the LID designs. If smaller units were used, either by increasing ve rtical storage or by placing gr eater reliance on storm-water ponds, the costs might have been lower. On the other hand, greater use of ponds would require more storm sewer pipe, wh ich would increase the total cost. Impact to the Developer and Potential Investors This leaves the question as to whethe r the combined influence of minimizing development footprint and/or implementing an LID storm-water management system, while maximizing open space frontage, provides the same return to the developer. A simple comparison, outlined in the section di scussing the results of the open space, was done to estimate these impacts. The average co efficient for the traditional and “partial” LID developments was 0.746 and 0.874 for the cl uster and “full” LID. The total receipt from sales was computed and the constructi on costs subtracted. Av erage lot sizes of 1,871.16 m2 (~20,000 SF) and 934.65 m2 (~10,000) were considered (the actual average

PAGE 123

110 over each design was certainly somewhat higher). The lots were assumed to be at an average distance of 12,191.41 m (40,000 ft or 7.5 m iles) from the reference parcel and northwest from this parcel. Table 5-8 shows the results of this computation. The gross profit is computed as the sale s receipts minus the constructio n costs. The difference from the traditional devel opment is also shown. Table 5-8. Estimated project return. Alternative Sales receiptsGross profit a Difference b Return/cost Traditional $34,136,023 $21,848,335 ----1.78 Cluster $27,462,215 $19,183,096 $2,665,239 2.32 “Partial” LID $34,136,023 $22,097,783 $249,448 1.84 “Full” LID $27,462,215 $20,449,889 $1,398,446 2.92 a Computed as sales receipts minus construction cost. b Difference from the traditional development profit. The return on these design alternatives is conceptual and may not represent the actual return on a development project because not all costs are represented. These results could be interpreted two ways. The first, a nd obvious, interpretation is that two of the alternative designs produce less money for th e developer due to reduced lot size. The cluster design profit is $2.67 million less than the traditional design. The LID design results also clearly show the influence of construction cost. The “partial” LID profit is $0.25 million higher than the traditional desi gn and the “full” LID option is $1.4 million lower, but the difference is le ss than the cluster design. Consider the simple case where the develope r agrees to share the profits, after the loan balance is repaid, with an investor, who is financing the construction. In this case, shown in Table 5-9, the developer receive d less money from tw o alternatives. The developer’s profit may drive the choice of alternative if profit was the only factor involved, although this might not be the case if the developer is receiving outside financing.

PAGE 124

111 Table 5-9. Conceptual profit in dollars. Alternative Construction cost Gross profit To investor To developer Traditional $12,287,688 $21,848,335 $10,924,168 $10,924,168 Cluster $8,279,119 $19,183,096 $9,591,548 $9,591,548 “Partial” LID $12,038,240 $22,097,783 $11,048,892 $11,048,892 “Full” LID $7,012,326 $20,449,889 $10,224,945 $10,224,945 The second interpretation involves looking at the ratio of return, or profit, to the development costs. This view may be more appropriate. In this case the alternative practices, particularly the “full” LID option, have a higher return. In other words, the costs of development may drive the choice of alternative. Consider the investor. The investor is being asked to fi nance construction. He or she may very well favor the “full” LID option since the investment is lower, a llowing the money that would have gone to construction to be invested elsewhere, but the profit is comparable to a traditional development. The ratio of return to cost would be 2 .63 for the “full” LID and 2.08 for the cluster design if open space frontage had not been maximized. These are still higher than the traditional development option. Limitations This study, while useful, has certain lim itations. The hedonic price equation only provides an estimate of sale price. The mode l does not represent about one third of the variation in actual sale pri ce. For example, the model considers neither discounts for early sales (“pioneers”) nor premiums for fina l sales in popular developments. It does not consider the “character” of the developmen t. For example buyers may perceive the LID storm-water management system in a less desi rable light. This may also affect the loan terms from investors. Since the LID storm-water management system may be considered “new”, it may be seen as a riskier investment.

PAGE 125

112 Construction costs compared in this study were limited to those that appeared most important. Other costs, such as sanitary sewe rs and lift stations, transport of materials, taxes, re-zoning costs, land acquisition, and engineering cost s, are not included. They are assumed equal across the development scenario s. In some cases, this may be a valid assumption, but in other cases these costs may favor one alternative. This study also assumed that developmen t and sales took take place over a short time. Sales price may be lowered to provide a sufficient cash flow to cover financial obligations such as loan payments. It is possible that designs that deviate from the traditional development may require higher pr ice reductions due to buyer and investor perceptions. Cost and sales pric es are not adjusted for inflation. However, since the study period is relatively short, this ad justment was not deemed necessary. Maintenance of the LID storm-water manage ment system is not addressed in this analysis. The LID literature cited in this research provided no information on the maintenance costs. However, they describe the anticipated maintenance as being routine upkeep of the landscaping, normally the res ponsibility of the pr operty owner (Prince Georges County, Ch. 5, p. 7, 1997, Ch. 4, pp. 9,18, 1999a). The conventional system would be normally maintained by the Homeow ners Association except for inlets and pipes on streets dedicated to the County (ACBCC, 2002). It would be reasonable to assume that the County would also maintain the swales if the interior streets were dedicated. A certain level of upkeep by the hom eowner is a reasonable expectation since most owners would want to keep th e landscaping aesthetic ally pleasing. The Homeowners Association could be expected to provide more extensive maintenance per the requirements set forth by Alachua County an d ensure that drainage to the bioretention

PAGE 126

113 units is maintained, although th e issue of enforcement of ma intenance on private property remains an issue. The question is whether the long-term maintenance costs of the LID storm-water management system are higher than the costs of the conventional system. If this is the case it might translate into highe r expenses to the homeowner either through Association dues or upkeep costs. How this a ffects the decision of a prospective buyer is not known, nor can it easily be measured. Th ese long-term costs cannot be considered developer construction costs since the de veloper is usually only responsible for maintenance until a sufficient number of owners exist to form a Homeowners Association. A possible measur e of impact, character of the development, was discussed earlier in this section. Howe ver, it is possible that th e negative impact of higher Association dues or upkeep costs may be offset by the attractiveness of the development as an environmentally friendly develo pment. Marketing the development as “environmentally friendly” also would provide the developer an opportunity to educate the buyer on the required upkeep of the storm-water management system (Prince Georges County, Ch. 6, pp. 3-4, 1997). Economic Summary and Conclusions The impact of open space is certainly positive. As expected, the impact of open space decreased with distance. Also as exp ected, type of open space was important. Golf courses had the highest benefit while forest/p arks generally had a negative influence. Common areas were essentially neutral. Open space type and proximity must be considered together to understand the full e ffect of open space on lot sale price. The net effect of open space is positive for all types, but the differing influences of open space type mean that golf courses have the gr eatest positive influe nce followed by common areas and, finally, off-site park -land or forest preserves. At the same time, the hedonic

PAGE 127

114 price analysis clearly shows the impact of lot size. Maximizing the number of lots with open space frontage did return a significant dol lar amount to the developer, although not enough to mitigate the loss of money from smaller lots. The construction cost results show that significant savings c ould be realized by minimizing the development footprint. The use of an LID storm-water management system that emphasizes costly bioretention units, but minimizes storm sewer pipe, curb and gutter and sidewalks also pr ovided a smaller degree of co st savings. The construction cost savings may be the driving factor for this site, particularly if competition for financing is high. All alternatives to traditio nal development had lower overall costs. It can be concluded, for this period in th e Gainesville area, that the “full” LID alternative has the potent ial to be, at minimum, economically neutral. If the ratio of profit to cost is the statistic of importance, th e “full” LID design was the most desirable. However, in terms of dollar amount of profit, the “full” LID option resulted in less than the traditional design. The “partial” LID design resulted in the highest dollar amount (a small increase) for the developer, yet this design was not as beneficial in terms of site hydrology (assuming no preservation of upland areas). Th is presents a problem that must be addressed. The developer may want to develop the site as a traditi onal or “partial” LID design to maximize profits. In fact there is only a small financial incentive to develop with the “partial” LID design. Even if “par tial” LID profit was hi gher and site hydrology with full-size lots was improved through on-lo t preservation of la nd, the reason that the reduced lots were considered must be rememb ered. There are concerns as to whether the on-lot preservation of uplands areas could ade quately protect these areas from impact and

PAGE 128

115 whether the conservation easements could be enforced in perpetuity. Furthermore, the LID design philosophy recommends that develo pment be avoided, to the greatest extent practical, in high infi ltration, upland areas. Extensive preservation of upland areas away fr om the lots is practical with a larger lot size than was used for the cluster footprin t. The question is what lot size and site layout is the “ideal”. If open space frontage were maximized, the required lot sizes would be lower. Lower lot sizes also would result in lower construction costs, which depend on actual lot geometry. The next chapter disc usses the use of a simple decision support system that could be utilized to provide guidance on lot size, geometry and open space frontage.

PAGE 129

116 CHAPTER 6 DECISION SUPPORT SYSTEMS FOR LAND DEVELOPMENT Introduction The previous two chapters dealt with hydrologic and economic analyses of traditional and alternative land development and storm-water mana gement. A LID-type storm-water management system accompanie d by reduction in development footprint through cluster design provided a hydrologic re sponse that closely mimicked the natural response of the undeveloped watershed. Trad itional and cluster developments met hydrologic regulations, but with significant al terations in the watershed response. Full size, cleared lots with no land preservation and an LID st orm-water management system did not meet regulatory guidelines even af ter addition of three conceptual detention ponds, but the response was sufficiently close to the natural response that preservation of some upland areas on the lots could improve the hydrology. The economic analysis showed that the LID storm-water management system was less expensive than conventional systems for the construction line items analyzed. Cost savings, larger than those r ealized from implementing the LID storm-water management system, were obtained by utilizing the cluster design footprint, which reduces lot size and the need for infrastructure that is depe ndent on lot dimensions. The cluster design footprint does negatively impact the developer because much of the value of the lots is in their area. A portion of this loss can be mitigated through maximizing open space frontage.

PAGE 130

117 These results may not necessarily be repres entative of all development cases. In fact, they represent a particular “snap-shot” in time and don’t consid er all costs. They do indicate that it would be po ssible to use an optimization procedure to determine the “best” lot size and storm-water management sy stem. The term “best” is defined in this case as the same receipts as the traditional design. It can be argued, based on the results presented in the previous chap ter that receipts from the “partial” LID design should be used since they represent the maximum pr ofit. This could well be the developer’s viewpoint. However, profit from the traditiona l design is virtually the same. The purpose of this chapter is to intro duce a simple decision support system (DSS) as an example of computing the “best” site configuration. A lthough the focus in this chapter is on the developer’s profit, the DSS example could also be applied to meet a desired ratio of profits to cost. The typical land development procedure can, generally, be described as the developer hiring an engineer and planne r to produce acceptable plans for a design concept that has already been selected. The land planner or e ngineer’s goal is to produce plans that will meet regulatory goals and stay close to the de veloper’s concept. This is critical because the developer may have al ready produced a financial schedule based on the concept. If the developer is wise, an initial consultation has taken place with an engineer or planner to see wh at is possible with the site. A DSS would provide a means to evaluate alternatives to the traditional de velopment. Such a system could include the developer’s market study with statis tical data on construction costs. Literature Review Sample et al. (2001, 2003) provide two ex amples of a DSS applied to land development. Sample et al. (2001) utilized Exce l Solver to determine the least cost option

PAGE 131

118 for lots, including opportunity costs, while ma intaining the initial abstraction of the NRCS curve number method. In theory, the ch aracter of the development was analyzed as an alternative. Sample et al. (2003) present another anal ysis of the same hypothetical site used in the 2001 study. Again, th e goal is the least cost system. Construction of the DSS The first step in construction of the DSS wa s to decide what the ultimate goal of the output was. The obvious goal is to ensure th at receipts for the alternate design are identical to the traditional design. This would ensure that profits would be the same. Ideally, the DSS will converge on a so lution that is close to the 934.65 m2 (~10,000 SF) lots that were analyzed in the hydrology portion of this research. This lot size, with the LID storm-water management system, provi ded the most desirable result from a hydrologically and environmen tally perspective. The work outlined in the previous chapters was presented in Chapter 1 as a test of four corners of a solution space. The lot size axis is intuitive and easily measured. It represents the relative degree of open space pr eserved. The second axis is the measure of the character of the development, specifi cally whether the storm-water management system is an LID-type system or not. This is conceptual, but a measure of the amount of storm-water ponds required may be consider ed. A portion of this solution set was eliminated during the hydrologic analysis. Sp ecifically, since each design required stormwater ponds, the upper portion of the solution set, representing “pure” LID storm-water management using only bioretention units, is eliminated. Figure 6-1 shows this graphically.

PAGE 132

119 Sale Price The first portion of the DSS was estimation of receipts from sales. This used the results of a hedonic price analysis outlined in Chapter 5. This procedure was recently used by Bolitzer and Netusil (2000), Doss a nd Taff (1996) and Do and Grudnitski (1995) in analyses of the impact of open spaces on home sale prices. Figure 6-1. Conceptual solution set revised from hydrologic analysis. This DSS was constructed utilizing an Ex cel spreadsheet with the Solver add-on. The baseline conditions of the traditional design use an average lot size of 1,871.16 m2, an average open space proximity coefficien t of 0.746 (assuming co mmon area), location north and a distance of 12,191.41 m (40,000 ft or 7.5 miles) from a reference parcel. The resulting sale price per lot is multiplied by 719 lots. Construction Costs An estimation of construction costs was re quired. Costs for individual construction line items were determined from Means (2001) and Sample et al. (2003). Since the LID

PAGE 133

120 storm-water management system was less expe nsive than the conventional system, cost estimates for the conventional system are ir relevant. This might be the case if all assumptions used in design or in the LID c onstruction costs are correct. This, however, may not always be true. For example, the LID storm-water management system or the modified cluster footprint may require re -zoning as a Planned De velopment District. Legal costs associated with this are no t included in the res earch. Tabulation of conventional storm-water management was done even though its impact on the final outcome of this particular DSS analysis is doubtful. It was necessary to generate cost functi ons for each line item. Ideally, these would be generated from a statistical database base d on similar projects in the area. For this example, a linear relationship was assumed (r ecall Figure 1-1 and the description of this research as a test of four corners of a so lution set). Each line it em was computed using the physical parameter (lot area or lot widt h) determined to be most applicable. For example, curb is based on lot width, while th e volume of storm-water ponds or area of bioretention storage was based on lot area. Cost items not included in this analysis, such as the cost of land acquis ition, can easily be added. DSS Function The DSS function is defined as: 13 R = SR (Ci) (Eqn. 6-1) i = 1 Where R is net receipts, SR is total value of sales and C is the 13 construction line items. Each cost function is a simple linear relati onship. SR is computed from the function: SR = x ey (Eqn. 6-2)

PAGE 134

121 Where x is the number of lots in that particul ar design, and y is the sum of the products of coefficients and variables in the hedonic price equation. Three parameters were allowed to vary: open space coefficient in the hedonic price equation, lot width along th e street and lot depth. The an alysis was subject to the following constraints: The open space coefficient could vary between 0.746 and 0.874. Total lot area could be no less than 928.9 m2 and no greater than 1,700.02 m2 (less than the traditional development lot size, but the maximum lot size that was determined to allow full frontage on open space). The lot width must be less than the depth. The lot width must be at least 21.9 m, wh ich is the minimum width for standard house footprint and side lots. The maximum open space frontage was based on the cluster design. The coefficient used as a constraint reflects the fact that some lots do not have open space frontage in the cluster design footprint. The limitation on maxi mum lot size was a further reduction of a portion of the solution set (Figure 6-2). This constraint was required to prevent the model from returning a solution with the same lot size as the traditional development, but with maximum open space frontage. This is an impo ssible situation. This is not to say that a better constraint could be developed.

PAGE 135

122 Figure 6-2. Conceptual solution set w ith constraint on maximum lot area. Results The resulting lot dimensions from the D SS, under the constraints listed above, are presented in Table 6-1. Lot dimensions of th e full size and cluster footprints are also listed for comparison. Table 6-1. DSS Output. Parameter DSS output Full size lot Cluster lot Width 27.45 m 37.2 m 27.9 m Depth 39.10 m 50.3 m 33.5 m Area 1072.06 m2 1871.16 m2 934.65 m2 The cluster lot was used in two of the a lternative site designs : cluster development and “full” LID. This result calls for a slightly larger and deeper lot. This result called for most lots to have frontage on open space, ju st as the “full” LID and cluster designs provided. The larger lot size compensates for the fact that open space frontage does not fully mitigate losses due to reduced lot size. Th e final lot size was still sufficiently close to the “full” LID lot size so it can be conc luded that storm-water management goals for

PAGE 136

123 the design storms should be realized thr ough a system maximizing shallow depression storage in swales and bioretention areas. Th is solution returned the same net dollar amount, but not all development costs such as sanitary sewer or electrical distribution systems were included in the cost function. The solution cannot be considered unique Different starting c onditions will result in convergence on other answers. The starting co nditions used in this run were lot width and depth of 1 m and an open space coe fficient of 0.746 (same as traditional development). Limitations There were certain limitations in this mode l that need to be considered. The output from this model assumed rectangular lots. Ther efore, the actual costs of the development may not exactly match the output from the DSS, which should be considered an estimate. Not all construction and land development cost s were considered. These additional costs could, however, be easily added. The model was also not perfect in the sense that it does not directly attempt to maximize open space pr eservation. Additional computations in the spreadsheet might allow this. Finally, the input for costs and estimated sale price are only as good as the methods utilized to determ ine them, and their weaknesses should be considered. The DSS did not directly consider stor m-water management goals. The DSS is based on results of the hydrologic analysis fo r a particular site. However, a statistical database of storm-water controls required on similar developments might improve the confidence of the estimates. The cost functions themselves may not be the best measures of actual costs. This is particularly a concer n for parameters related to lot area. Perhaps a better measure would be to base these cost line items on the percent of the subdivision

PAGE 137

124 left in natural condition. Spatia l layout of the lots and open space are not considered. This is still up to the land planner or engineer. It is possible, depending on the actual site conditions, that the land preser vation requirements of both LID and the desired return cannot be met. Conclusions The purpose of this example was to show the potential usefulness of a DSS in the land development process. In this case, the DSS was used to determine the most advantageous lot geometry, compared to the traditional development, using estimates for lot sale price and construction costs. Admitte dly, the problem and the DSS presented here were simple and the outcome rather predicta ble. However, this simple example proves the major point that use of a DSS would gr eatly simplify the land development process for LID-type developments. The solution provided by the DSS would have required development of site plans until the developer was satisfied, if done by “trial and error”. Therefore, this DSS would have saved considerable time if it had been used in the initial planning process. A DSS could be applied to much more complex problems. For example, a DSS could be integrated into a deve loper’s financial pro forma with the target being a specific rate of return for investors. The DSS presen ted here was quite simple and meant only to show the usefulness as a la nd development tool. The DSS itself could be made more complex. For example, it could be developed to maximize open space preserved. The usefulness of these tools should not be ignored if alternative land development practices are considered.

PAGE 138

125 CHAPTER 7 CONCLUSION This research consisted of three compone nts: a hydrologic an alysis, an economic analysis and the construction of a simp le Decision Support System. An LID land development plan using reduced lot size, maximizing open space frontage, and an infiltration based storm-water management system, resulted in a hydrologic response very close to the natu ral watershed response for two desi gn storms. This was superior to the hydrologic response of a traditional development with full size lots or a cluster development with reduced lots and conventi onal storm-water management. Full size lots with the infiltration based stor m-water management provided a response closer to natural conditions, but when analysis was stopped, this plan did not meet regulatory requirements for the 25-year, 24-hour design storm. The LI D designs, however, had the potential to perform worse than a pond system under condi tions where the antecedent moisture is high, although the pond systems also perfor med worse than the natural watershed response. The “partial” LID plan provided the greates t, conceptual profit, followed by the traditional design, “full” LID design and cluster development. The ratio of net receipts to costs was highest for the “full” LID design followed by the cluster design, “partial” LID and traditional development. The third portion of this rese arch presented a simple vers ion of a DSS as a potential method for determining optimal geometry and si ze of the lots. This was done to illustrate the potential usefulness of thes e systems in the land development process, particularly

PAGE 139

126 where alternative methods of land development and storm-water management introduce questions of economic imp acts to the developer. It is hoped that this research will enc ourage use of alternativ e practices in NorthCentral Florida. Alternativ e practices often encounter opposition when first proposed. This may not be based on scientific grounds but instead on economics. This research should help dispel concerns al ong these lines. What makes this research different is that economics, from the developer’s viewpoint, were addressed. This study contributes to a growing body of literature on the subject of LID and alternative land development and storm-wate r management practices. However, more work is needed, particularly field studies on individual st orm-water management systems and actual development projects, to conclude ad equately that these pr actices are superior to the traditional methods of land development and storm-water management.

PAGE 140

127 APPENDIX A HYDROLOGIC MODEL INPUT Table A-1. Existing conditions model areas. Sub-area Area (km2) Area 1 0.31718 Area 2 0.41887 Area 3 0.49182 Area 4 0.15484 Area 5 0.16217 Area 9 0.08474 Area 8 0.07678 Area 10 0.21945 Area 11 0.19110 Area 6 0.27353 Area 7 0.34148 Area 12 0.15577 Table A-2. Traditional developmen t cover conditions (all units m2). Sub-Area Road/Sidewalk Open spaceHousesLawns Total %DCIA Basin 1 16467 40899 1901288839 165217 10% Basin 2 32486 54569 25220123150 235424 14% Basin 3 26451 5737 30070165985 228244 12% Basin 4 53249 10926 49082251162 364419 15% Basin 5 29397 24490 25220128035 207142 17% Basin 6 35437 36220 29876145112 246644 14% Basin 7 31665 39968 35114177688 284434 11% Basin 8 43585 91324 49082237409 421400 10% UnCon 1 0 6661 0 6785 13446 0% UnCon 2 0 0 1552 8232 9784 0% Uncon 4 0 1958 3492 13797 19248 0% UnCon 5 0 43638 0 496 44134 0% UnCon 6 0 25495 1940 13259 40694 0% UnCon 7 0 17754 3492 16293 37539 0% UnCon 8 0 16056 0 2582 18638 0% UnCon 9 0 72410 0 986 73397 0% UnCon 10 0 117918 3686 36467 158071 0% UnCon 11 0 154451 1358 16541 172350 0% UnCon 12 0 147550 776 3997 152323 0%

PAGE 141

128 Table A-3. Cluster development cover conditions (all units m2). Sub-Area Road/Sidewalk Open spaceHousesLawnsTotal %DCIA Basin 1 14686 95168 2148055247 1865808% Basin 2 38453 107244 4025210503329098213% Basin 3 6827 37220 7220 21020 72287 9% Basin 4 35496 104691 2743682249 24987214% Basin 5 22997 45731 2851971236 16848314% Basin 6 26092 147595 3411580394 2881959% Basin 7 23409 66420 3339381260 20448211% Basin 8 26347 152055 4097492526 3119028% Basin 9 7596 27638 1010828673 74015 10% UnCon 1 0 15431 1083 9848 26362 0% UnCon 2 0 11641 3610 8266 23516 0% UnCon 3 0 3046 1083 2076 6205 0% Uncon 4 0 154841 0 0 1548410% UnCon 5 0 65983 2166 3286 71435 0% UnCon 6 0 51657 1083 3468 56208 0% UnCon 7 0 18043 1444 12248 31734 0% UnCon 8 0 76777 0 0 76777 0% UnCon 9 0 72719 0 802 73521 0% UnCon 10 0 167311 4874 22223 1944080% UnCon 11 0 177309 1083 1903 1802950% UnCon 12 0 150449 0 0 1504490% Table A-4. “Partial” LID developm ent cover conditions (all units m2). Sub-Area Road/Sidewalk Open spaceHousesLawnsTotal Area 1 27154 69861 38024198055333094 Area 2 21101 66809 50052237627375590 Area 3 22447 132875 59752281437496511 Area 4 11428 10096 22116114764158404 Area 5 4803 62147 1241664292 143658 Area 6 27277 65108 36666186918315970 Area 7 18877 95800 38412188396341485 Area 8 2012 32596 6596 35572 76777 Area 9 0 84743 0 0 84743 Area 10 6224 149277 9894 54057 219451 Area 11 2146 153867 4268 30818 191100 Area 12 0 150564 776 4426 155767

PAGE 142

129 Table A-5. “Full” LID developmen t cover conditi ons (all units m2). Sub-Area Road/Sidewalk Open spaceHousesLawnsTotal Area 1 19323 171262 3559994600 320784 Area 2 26469 197098 59821137315420704 Area 3 22286 235458 70831150642479217 Area 4 0 154841 0 0 154841 Area 5 5735 105378 1835038614 168076 Area 6 16929 133387 3339782817 266531 Area 7 18177 206569 3926990543 354558 Area 8 0 76777 0 0 76777 Area 9 0 84743 0 0 84743 Area 10 347 214900 1468 2736 219451 Area 11 1657 170572 5505 13366 191100 Area 12 0 155767 0 0 155767 Table A-6. Initial percent of storage full and sub-area at-large groundwater routing input. GW1 Rt2 #RES GW 1 GW2 Rt2#RES GW 2%Soil full%GW1 full %GW2 full 2 2 1 1 0 0 95/91 2 2 1 1 0 0 95/91 2 2 1 1 0 0 95/91 2 2 1 1 0 0 95/91 2 2 1 1 0 0 95/91 0.5 1 1 1 30/6 50/0 97/93 2 2 1 1 0 0 95/91 0.5 1 1 1 30/6 50/0 97/93 0.5 1 1 1 30/6 50/0 97/93 2 2 1 1 0 0 95/91 2 2 1 1 0 0 95/91 0.5 1 1 1 30/6 50/0 97/93 Note: At-large groundwater input is the same for developed conditions. Coefficients are applied based upon the dominant character (u pland or lowland) of the developed subarea. Percent full lists 2 numbers. The first is for the continuous model and the second for the design storms (assumed average antecedent moisture condition). Groundwater routing coefficients have units of hr. Table A-7. Evapotranspiration input for all models. Parameter June July August September October ET Depth (mm) 193 181 170 145 126 Pan coefficient 0.3 0.3 1.0 0.3 0.7

PAGE 143

130 Table A-8. Existing conditions soil moisture units. Sub-Area Canopy (mm) Surface (mm) Infiltration (mm/hr) Soil (mm) Tension (mm) Soil Perc (mm/hr) GW1 (mm) GW1 Perc (mm/hr) GW1 Rt (hr) GW2 (mm) Deep Perc (mm/hr) GW2 Rt (hr) Area 1 1 2 304 25 4 304 25 104 125 1250 0.04 7000 Area 2 1 2 304 25 4 304 25 104 125 1250 0.04 7000 Area 3 1 2 304 25 4 304 25 104 125 1250 0.04 7000 Area 4 1 2 304 25 4 304 25 104 125 1250 0.04 7000 Area 5 1 2 304 25 4 304 25 104 125 1250 0.04 7000 Area 9 2 2 304 20 13 52 12 25 50 600 0.04 7000 Area 8 1 2 304 25 4 304 25 104 125 1250 0.04 7000 Area 10 2 0 304 20 13 52 12 25 50 600 0.04 7000 Area 11 2 0 304 20 13 52 12 25 50 600 0.04 7000 Area 6 1 2 304 25 4 304 25 104 125 1250 0.04 7000 Area 7 1 2 304 25 4 304 25 104 125 1250 0.04 7000 Area 12 2 2 304 20 13 52 12 25 50 600 0.04 7000 Table A-9. Traditional devel opment soil moisture units. SubArea Canopy (mm) Surface (mm) Infiltration (mm/hr) Soil (mm) Tension (mm) Soil Perc (mm/hr) GW1 (mm) GW1 Perc (mm/hr) GW1 Rt (hr) GW2 (mm) Deep Perc (mm/hr) GW2 Rt (hr) Basin 1 1 2 304 18 4 280 21 97 119 1029 0.04 7000 Basin 2 1 2 304 18 3 304 22 104 125 1095 0.04 7000 Basin 3 1 2 304 16 3 304 21 104 125 1064 0.04 7000 Basin 4 1 2 304 16 3 291 20 100 121 1025 0.04 7000 Basin 5 1 2 304 17 3 303 21 104 125 1071 0.04 7000 Basin 6 1 2 304 17 3 304 21 104 125 1073 0.04 7000 Basin 7 1 2 304 17 3 282 21 97 119 1030 0.04 7000 Basin 8 1 2 304 17 3 290 21 100 121 1057 0.04 7000 Ucon 1 1 2 304 22 3 304 25 104 125 1250 0.04 7000 Ucon 2 1 2 304 16 3 304 21 104 125 1052 0.04 7000 Ucon 4 1 2 304 16 3 304 20 104 125 1023 0.04 7000 Ucon 5 1 2 304 25 4 304 25 104 125 1250 0.04 7000 Ucon 6 1 2 304 22 3 304 24 104 125 1190 0.04 7000 Ucon 7 1 2 304 20 3 304 23 104 125 1134 0.04 7000 Ucon 8 1 2 304 24 4 304 25 104 125 1250 0.04 7000 Ucon 9 2 2 304 20 13 52 12 25 50 600 0.04 7000 Ucon10 2 0 304 18 12 52 12 25 50 586 0.04 7000 Ucon11 2 0 304 19 13 52 12 25 50 595 0.04 7000 Ucon12 2 2 304 20 13 52 12 25 50 597 0.04 7000

PAGE 144

131 Table A-10. Cluster development soil moisture units. SubArea Canopy (mm) Surface (mm) Infiltration (mm/hr) Soil (mm) Tension (mm) Soil Perc (mm/hr) GW1 (mm) GW1 Perc (mm/hr) GW1 Rt (hr) GW2 (mm) Deep Perc (mm/hr) GW2 Rt (hr) Basin 1 1 2 304 19 4 280 21 96 118 1031 0.04 7000 Basin 2 1 2 304 18 3 304 21 104 125 1051 0.04 7000 Basin 3 1 2 304 20 3 304 22 104 125 1112 0.04 7000 Basin 4 1 2 304 19 3 301 22 103 124 1084 0.04 7000 Basin 5 1 2 304 17 3 294 20 101 122 981 0.04 7000 Basin 6 1 2 304 20 3 304 22 104 125 1087 0.04 7000 Basin 7 1 1 304 17 3 286 19 98 120 974 0.04 7000 Basin 8 1 2 304 19 3 302 21 103 125 1067 0.04 7000 Basin 9 1 1 304 18 5 265 19 92 125 961 0.04 7000 Ucon 1 1 2 304 22 3 304 24 104 125 1199 0.04 7000 Ucon 2 1 2 304 19 3 304 21 104 125 1058 0.04 7000 Ucon 3 1 2 304 19 3 304 21 104 125 1032 0.04 7000 Ucon 4 1 2 304 25 4 304 25 104 125 1250 0.04 7000 Ucon 5 1 2 304 24 4 304 24 104 125 1212 0.04 7000 Ucon 6 1 2 304 23 4 304 24 104 125 1193 0.04 7000 Ucon 7 1 2 304 21 3 304 24 104 125 1193 0.04 7000 Ucon 8 1 2 304 25 4 304 25 104 125 1250 0.04 7000 Ucon 9 2 0 304 20 13 52 12 25 125 600 0.04 7000 Ucon10 2 0 304 19 12 52 12 25 125 585 0.04 7000 Ucon11 2 0 304 20 13 52 12 25 125 600 0.04 7000 Ucon12 2 2 304 20 13 52 12 25 125 600 0.04 7000 Table A-11. “Partial” LID development soil moisture units. Sub-Area Canopy (mm) Surface (mm) Infiltration (mm/hr) Soil (mm) Tension (mm) Soil Perc (mm/hr) GW1 (mm) GW1 Perc (mm/hr) GW1 Rt (hr) GW2 (mm) Deep Perc (mm/hr) GW2 Rt (hr) Area 1 1 2 304 17 3 304 20 104 125 1005 0.04 7000 Area 2 1 2 304 17 3 304 20 104 125 1013 0.04 7000 Area 3 1 2 304 18 3 304 21 104 125 1043 0.04 7000 Area 4 1 2 304 16 3 304 20 104 125 985 0.04 7000 Area 5 1 2 304 20 3 304 22 104 125 1100 0.04 7000 Area 6 1 2 304 17 3 304 20 104 125 997 0.04 7000 Area 7 1 2 304 18 3 304 21 104 125 1040 0.04 7000 Area 8 1 2 304 20 3 304 22 104 125 1110 0.04 7000 Area 9 2 0 304 20 13 52 12 25 50 600 0.04 7000 Area 10 2 0 304 17 11 52 11 25 50 556 0.04 7000 Area 11 2 0 304 19 12 52 12 25 50 580 0.04 7000 Area 12 2 2 304 20 13 52 12 25 50 597 0.04 7000

PAGE 145

132 Table A-12. “Full” LID development soil moisture units. Sub-Area Canopy (mm) Surface (mm) Infiltration (mm/hr) Soil (mm) Tension (mm) Soil Perc (mm/hr) GW1 (mm) GW1 Perc (mm/hr) GW1 Rt (hr) GW2 (mm) Deep Perc (mm/hr) GW2 Rt (hr) Area 1 1 2 304 19 3 304 21 104 125 1036 0.04 7000 Area 2 1 2 304 18 3 304 20 104 125 994 0.04 7000 Area 3 1 2 304 18 3 304 20 104 125 1007 0.04 7000 Area 4 1 2 304 25 4 304 25 104 125 1250 0.04 7000 Area 5 1 2 304 20 3 304 21 104 125 1071 0.04 7000 Area 6 1 2 304 19 3 304 20 104 125 1014 0.04 7000 Area 7 1 2 304 20 3 304 21 104 125 1047 0.04 7000 Area 8 1 2 304 25 4 304 25 104 125 1250 0.04 7000 Area 9 2 2 304 20 13 52 12 25 50 600 0.04 7000 Area 10 2 0 304 20 13 52 12 25 50 595 0.04 7000 Area 11 2 0 304 19 12 52 12 25 50 578 0.04 7000 Area 12 2 2 304 20 13 52 12 25 50 600 0.04 7000 Table A-13. “Partial” LID additional de pression storage for runoff control. Sub-Area Storage (mm) Total vol. (m3) Swale vol. (m3) Bioretention vol. (m3) Bio. Area/lot(m2) Area 1 13.0 4330 786 3544 237 Area 2 15.0 5634 1035 4599 234 Area 3 14.0 6951 1235 5716 244 Area 4 16.0 2534 457 2077 239 Area 5 10.0 1437 257 1180 242 Area 6 13.0 4108 758 3350 233 Area 7 13.0 4439 794 3645 242 Area 8 10.0 768 136 631 244 Area 9 0.0 0 0 0 0 Area 10 5.5 1207 205 1002 258 Area 11 3.0 573 88 485 289 Area 12 0.5 78 16 62 203 Table A-14. “Full” LID additional depr ession storage for runoff control. Sub-Area Storage (mm) Total vol. (m3) Swale vol. (m3). Bioretention vol. (m3) Bio. Area/lot(m2) Area 1 7.3 2342 526 1816 123 Area 2 10.0 4207 883 3324 134 Area 3 9.0 4313 1046 3267 111 Area 4 0.0 0 0 0 0 Area 5 6.2 1042 271 771 101 Area 6 9.0 2399 493 1906 137 Area 7 7.3 2588 580 2008 123 Area 8 0.0 0 0 0 0 Area 9 0.0 0 0 0 0 Area 10 0.5 114 22 92 151 Area 11 2.7 513 81 432 189 Area 12 0.0 0 0 0 0

PAGE 146

133 Table A-15. Existing transform input. Sub area Tc (hr) Clark R Coeff. Area 1 1.14 9.12 Area 2 1.15 9.20 Area 3 0.90 7.20 Area 4 0.77 6.16 Area 5 0.78 6.24 Area 6 1.04 8.32 Area 7 1.15 9.20 Area 8 0.45 3.60 Area 9 0.40 2.00 Area 10 0.53 2.65 Area 11 0.47 2.35 Area 12 0.66 3.30 Table A-16. “Partial” LID transform input. Sub-Area Tc (hr) Clark R coeff Area 1 1.18 3.37 Area 2 1.32 3.51 Area 3 1.00 3.25 Area 4 0.94 1.80 Area 5 0.78 3.36 Area 6 1.29 3.66 Area 7 1.37 4.56 Area 8 0.88 3.76 Area 9 0.40 2.00 Area 10 0.62 2.40 Area 11 0.47 2.03 Area 12 0.66 3.30 Table A-17. “Full” LI D transform input. Sub-Area Tc (hr) Clark R coeff Area 1 1.14 5.81 Area 2 2.18 9.92 Area 3 1.19 5.56 Area 4 0.77 6.16 Area 5 0.78 4.35 Area 6 1.03 4.90 Area 7 1.72 9.09 Area 8 0.45 3.60 Area 9 0.40 2.00 Area 10 0.53 2.61 Area 11 0.47 2.17 Area 12 0.66 3.30

PAGE 147

134 Table A-18. Traditiona l transform input. Sub-Area Tc (hr) Clark R coeff Basin 1 0.72 2.24 Basin 2 0.88 2.57 Basin 3 0.53 0.87 Basin 4 0.83 1.40 Basin 5 0.82 1.87 Basin 6 0.78 1.91 Basin 7 0.80 1.94 Basin 8 0.82 2.38 UnCon 1 0.62 2.92 UnCon 2 0.55 0.87 Uncon 4 0.53 1.15 UnCon 5 0.60 4.72 UnCon 6 0.83 4.61 UnCon 7 0.59 2.69 UnCon 8 0.50 3.54 UnCon 9 0.40 2.00 UnCon 10 0.51 2.10 UnCon 11 0.47 2.18 UnCon 12 0.66 3.30 Table A-19. Cluster transform input. Sub-Area Tc (hr) Clark R coef f Basin 1 0.74 3.56 Basin 2 0.98 3.80 Basin 3 0.85 4.14 Basin 4 1.16 4.92 Basin 5 0.68 2.22 Basin 6 1.03 4.95 Basin 7 0.84 3.01 Basin 8 0.82 3.81 Basin 9 0.54 2.10 UnCon 1 0.56 2.99 UnCon 2 0.71 3.35 UnCon 3 0.33 1.55 Uncon 4 0.77 6.16 UnCon 5 0.60 4.50 UnCon 6 1.02 7.60 UnCon 7 0.59 3.07 UnCon 8 0.45 3.60 UnCon 9 0.54 2.10 UnCon 10 0.53 2.39 UnCon 11 0.47 2.32 UnCon 12 0.66 3.30

PAGE 148

135 Table A-20. Traditional deve lopment wet ponds 1 to 5. Basin 1 Basin 2 Basin 3 Perm. pool vol (m2) 4302 Perm. pool vol (m2) 6520 Perm. pool vol (m2) 6256 Treat. vol (m2) 4197 Treat. vol (m2) 5980 Treat. vol (m2) 5797 Elev (m) Area (m2) Elev (m) Area (m2)Elev (m) Area (m2) 0 2286 0 4743 0 6404 0.5 2610 0.5 5187 0.5 6997 1 2954 1 5648 1 7609 1.5 3318 1.5 6128 1.5 8239 2 3701 2 6632 2 8888 2.5 4104 2.5 7160 2.5 9552 3 4523 3 7712 3 10232 3.5 4957 3.5 8286 3.5 10927 Orifice center (m) 1.57 Orifice center (m)1.28 4 11638 Diameter (m) 0.080 Diameter (m ) 0.11 Orifice center (m) 0.95 Weir crest (m) 1.63 Weir crest (m) 2.17 Diameter (m) 0.11 Weir length (m)3 Weir length (m)0.5 Weir crest (m) 1.63 Basin 4 Basin 5 Weir length (m) 0.4 Perm. pool vol (m2) 11181 Perm. pool vol (m2) 6167 Treat. vol (m2) 9256 Treat. vol (m2) 5261 Elev (m) Area (m2) Elev (m) Area (m2) 0 9686 0 3054 0.5 10396 0.5 3428 1 11122 1 3823 1.5 11864 1.5 4241 2 12621 2 4680 2.5 13394 2.5 5145 3 14181 3 5675 3.5 14983 3.5 6220 4 15800 Orifice center (m)1.71 Orifice center (m) 1.14 Diameter (m) 0.10 Diameter (m) 0.15 Weir crest (m) 2.74 Weir crest (m) 1.85 Weir length (m)1.75 Weir length (m)0.75

PAGE 149

136 Table A-21. Traditional deve lopment wet ponds 6 to 8. Basin 6 Basin 7 Basin 8 Perm. pool vol (m2) 7298 Perm. pool vol (m2) 7742 Perm. pool vol (m2) 11133 Treat. vol (m2) 6265 Treat. vol (m2) 7225 Treat. vol (m2) 10704 Elev (m) Area (m2) Elev (m) Area (m2)Elev (m) Area (m2) 0 4417 0 4174 0 5817 0.5 4869 0.5 4720 0.5 6523 1 5336 1 5296 1 7272 1.5 5819 1.5 5891 1.5 8039 2 6316 2 6505 2 8825 2.5 6828 2.5 7137 2.5 9629 3 7355 3 7787 3 10451 3.5 7896 3.5 8453 3.5 11289 Orifice center (m) 1.49 4 9135 4 12144 Diameter (m) 0.1 Orifice center (m)1.59 Orifice center (m) 1.66 Weir crest (m) 2.44 Diameter (m) 0.11 Diameter (m) 0.13 Weir length (m)1 Weir crest (m) 2.64 Weir crest (m) 2.76 Weir length (m)0.5 Weir length (m) 1 Table A-22. Cluster development wet ponds 1 to 3. Basin 1 Basin 2 Basin 3 Perm. pool vol (m2) 4867 Perm. pool vol (m2) 9361 Perm. pool vol (m2) 1790 Treat. vol (m2)4739 Treat. vol (m2)7391 Treat. vol(m2) 1836 Elev (m) Area (m2) Elev (m) Area (m2)Elev (m) Area (m2) 0 2286 0 5558 0 1921 0.5 2610 0.5 5997 0.5 2180 1 2954 1 6450 1 2455 1.5 3318 1.5 6917 1.5 2749 2 3701 2 7400 2 3071 2.5 4104 2.5 7897 2.5 3420 3 4523 3 8408 Orifice center (m) 0.86 3.5 4957 3.5 8935 Diameter (m) 0.06 Orifice center (m) 1.74 Orifice center (m)1.57 Weir crest (m) 1.55 Diameter (m) 0.084 Diameter (m) 0.13 Weir length (m) 2 Weir crest (m) 2.90 Weir crest (m) 2.50 Weir length (m)2 Weir length (m)1.25

PAGE 150

137 Table A-23. Cluster development wet ponds 4 to 9. Basin 4 Basin 5 Basin 6 Perm. pool vol (m2) 6555 Perm. pool vol (m2) 5750 Perm. pool vol (m2) 7914 Treat. vol (m2)6347 Treat. vol (m2)4279 Treat. vol(m2) 7320 Elev (m) Area (m2) Elev (m) Area (m2)Elev (m) Area (m2) 0 9686 0 3908 0 4142 0.5 10396 0.5 4333 0.5 4578 1 11122 1 4784 1 5030 1.5 11864 1.5 5257 1.5 5497 2 12621 2 5747 2 5978 2.5 13394 2.5 6252 2.5 6474 Orifice center (m) 0.70 3 6772 3 6985 Diameter (m) 0.12 Orifice center (m)1.33 3.5 7511 Weir crest (m) 1.22 Diameter (m ) 0.10 Orifice center (m) 1.68 Weir length (m)2 Weir crest (m) 2.07 Diameter (m) 0.11 Basin 7 Weir length (m)2 Weir crest (m) 2.80 Perm. pool vol (m2) 6555 Basin 8 Weir length (m) 1.5 Treat. vol (m2)5194 Perm. pool vol (m2) 8675 Basin 9 Elev (m) Area (m2) Treat. vol (m2)7922 Perm. pool vol (m2) 2092 0 4174 Elev (m) Area (m2)Treat. Vol (m2) 1880 0.5 4720 0 4831 Elev (m) Area (m2) 1 5296 0.5 5367 0 1705 1.5 5891 1 5913 0.5 2054 2 6505 1.5 6477 1 2424 2.5 7137 2 7059 1.5 2815 3 7787 2.5 7659 2 3224 Orifice center (m) 1.38 3 8277 2.5 3648 Diameter (m) 0.11 3.5 8914 Orifice center (m) 1.04 Weir crest (m) 2.16 Orifice cent er (m)1.59 Diameter (m) 0.06 Weir length (m)1.25 Diameter (m) 0.12 Weir crest (m) 1.70 Weir crest (m) 2.63 Weir length (m) 1.25 Weir length (m)1.25

PAGE 151

138 Table A-24. LID development c onceptual dry detention ponds. Partial LID B-1 Partial LID B-2 Elev. (m) Area (m2)Elev. (m) Area (m2) 0 5819 0 6128 0.5 6316 0.5 6632 1 6828 1 7160 1.5 7355 1.5 7712 2 7896 2 8286 Weir crest (m) 0.00 Weir crest (m) 0.00 Weir length (m) 0.25 Weir length 0.25 Partial LID B-3 Full LID B1 Elev. (m) Area (m2)Elev. (m) Area (m2) 0 11122 0 4142 0.5 11864 0.5 4578 1 12621 1 5030 1.5 13394 1.5 5497 Weir crest (m) 0.00 Weir crest (m) 0.00 Weir length (m) 0.025 Weir length (m) 0.31

PAGE 152

139 APPENDIX B CONTINUOUS MODEL AND C ONSTRUCTION COST RESULTS Figure B-1. Traditional developm ent continuous model results. Figure B-2. Cluster development continuous model results.

PAGE 153

140 Figure B-3. “Partial” LID developm ent continuous model results. Figure B-4. “Full” LID developm ent continuous model results.

PAGE 154

141 Table B-1. Itemized construction cost s (not adjusted for Gainesville). Item: Paving Item: Straight C&G Alternative Cost per SM: 17.97 Alternative Cost per M: 27.72 Amount Cost Amount Cost Traditional 120883 2172631 Traditional 30806 854066 Cluster 89940 1616492 Cluster 17651 489364 “Partial” LID 120883 2172631 “Partial” LID 722 20015 “Full” LID 89940 1616492 “Full” LID 0 0 Item: Radius C&G Item: Sidewalk Alternative Cost per M: 40.85 Alternative Cost per M: 58.89 Amount Cost Amount Cost Traditional 4802 196165 Traditional 35608 2097096 Cluster 6650 271642 Cluster 24301 1431182 “Partial” LID 161 6577 “Partial” LID 883 52003 “Full” LID 0 0 “Full” LID 0 0 Item: ManHole/Catch Basin Item: Storm Sewer Alternative Cost each: 1550 Alternative Cost per M: 154 Amount Cost Amount Cost Traditional 573 888150 Traditional 19113 2934189 Cluster 486 753300 Cluster 13381 2054224 “Partial” LID 33 51150 “Partial” LID 5640 865903 “Full” LID 12 18600 “Full” LID 4255 653219 Item: Grade Item: End Sec. Alternative Cost per SM: 0.16 Alternative Cost each: 293.00 Amount Cost Amount Cost Traditional 1828032 284248 Traditional 32 9376 Cluster 1031365 160371 Cluster 28 8204 “Partial” LID 1546698 284248 “Partial” LID 1565 458545 “Full” LID 806534 125411 “Full” LID 1516 444188 Item: Clearing Item: Swales Alternative Cost per SM: 1.29 Alternative Cost per M: 31.15 Amount Cost Amount Cost Traditional 1984523 2562519 Traditional 0 0 Cluster 1145606 1479266 Cluster 0 0 “Partial” LID 1818805 2348535 “Partial” LID 37205 1158932 “Full” LID 985797 1272912 “Full” LID 25192 784737 Item: Landscaping Item: Bio retention Alternative Cost per lot: 684.00 Alternative Cost per SM: 37.68 Amount Cost Amount Cost Traditional 719 983592 Traditional 0 0 Cluster 720 492480 Cluster 0 0 “Partial” LID 0 0 “Partial” LID 150341 5664849 “Full” LID 0 0 “Full” LID 89323 3365691

PAGE 155

142 Table B-1. Continued. Item: Grass Item: Pond Alternative Cost per SM: 0.51 Alternative Cost per CM: Variable Amount Cost Amount Cost Traditional 1398719 715215 Traditional 220171 1284320 Cluster 767125 392258 Cluster 157122 997197 “Partial” LID 1267726 648233 “Partial” LID 29576 284976 “Full” LID 452971 231620 “Full” LID 5497 80667

PAGE 156

143 LIST OF REFERENCES Alachua County Board of Count y Commissioners (ACBCC). (2002). Alachua County Code of Ordinances Tallahassee, FL: Municipal Code Corporation. Alachua County. (2003). Alachua County Property Appraiser’s Website. www.acpafl.org. Last accessed: 10/25/2003. Andoh, R. Y. G. and Declerck, C. (1997). A Cost Effective Approach to Stormwater Management? Source Control and Distributed Storage. Water Science Technology, 36 (8-9), 307-311. Bennett, T. H. (1998). Development and Application of a Continuous Soil Moisture Accounting Algorithm for the Hydrologi c Engineering Center Hydrologic Modeling System (HEC-HMS). MS Thesis. Davis, CA: Un iversity of California, Davis. Bolitzer, B. and Netusil, N. R. (2000). The Im pact of Open Spaces on Property Values in Portland, Oregon. Journal of Environmental Management. 59 (3), 185-193. Boniol, D., Williams, M. and Munch, D. (1993). Technical Publication SJ93-5 Mapping Recharge to the Floridian Aquifer Using a Geographic Information System. Palatka, FL: St Johns River Water Management District. Borden, R. C., Dorn, J. L., Stillman, J. B. and Liehr, S. K. (1998). Effect of In-Lake Quality on Pollutant Removal in Two Ponds. Journal of Environmental Engineering, 124 (8), 737-743. Braune, M. J. and Wood, A. (1999). Best Management Pr actices Applied to Urban Runoff Quantity and Quality Control. Water Science Technology, 39 (12), 117-121. CH2MHill. (1993). Cost of Providing Government Services to Alternative Residential Patterns. Washington D. C.: U.S. EPA. CH2MHill. (1998). Pennsylvania Handbook of Best Management Practices for Developing Areas. Philadelphia: CH2MHill. Claytor, R. A., and Schueler, T. R. (1996). Design of Stormwater Filtering Systems. Silver Spring, MD: Center for Watershed Protection. Clausen, J. C. (2002). Annual Report. Jordan Cove Urban Watershed Section 319 National Monitoring Program Project. Storrs, CT: University of Connecticut.

PAGE 157

144 Coffman, L., Clar, M. and Weinstein, N. ( 1998). Overview of Low Impact Development for Stormwater Management. Water Resources and the Urban Environment, Proceedings of the 25th Annual Conference on Water Resources Planning and Management. (pp.16-21). Reston, VA: American Society of Civil Engineers. Coffman, L., Clar, M. and Weinstein, N. (2000). Low Impact Development Management Strategies for Wet Weather Flow (WWF) Control. Building Partnerships: Proceedings of 2000 Joint Conference on Water Resource Engineering and Water Resources Planning and Management. (pp. 1-7). Reston, VA: American Society of Civil Engineers. Correll, M. R., Lillydahl, J. H. and Singell L. D. (1978). The Effects of Greenbelts on Residential Property Values: Some Findi ngs on the Political Economy of Open Space. Land Economics, 54 (2), 207-217. Dion, T. R. (1993). Land Development for Civil Engineers. New York: John Wiley and Sons. Do, A. Q. And Grudnitski, G.. (1995). Golf Courses and Residentia l House Prices: An Empirical Examination. Journal of Real Estate Finance and Economics, 10 261270. Doss, C. R. and Taff, S. J. (1996). Th e Influence of Wetland Type and Wetland Proximity on Residential Property Values. Journal of Agricultural and Resource Economics, 21 (1), 120-129. Dunn, C., Brown, S., Young, G. K., Stein, S. and Mistichelli, M. P. (1995). Current Water Quality Best Management Practices Design Guidance. Transportation Research Record, 1483 80-88. Federal Highway Admini stration (FHWA). (1985). Hydraulic Design of Highway Culverts. McLean, VA: Turner-Fairbank Highway research Center. Field, R., O’Shea, M. L. and Chin, K. K. (1993). Integrated Stormwater Management. Boca Raton, FL: Lewis Publishers. Freeman, A. M. (1993). The Measurement of Environm ental and Resource Values. Washington D.C.: Resources for the Future. Galuzzi, M. R. and Pflaum, J. M. (1996). In tegrating Drainage, Water Quality, Wetlands, and Habitat in a Planned Community Development. Journal of Urban Planning and Development, 122 (3), 100-108. Heaney, J. P., Wright, L. and Sample, D. (1998). Innovative Wet-Weather Flow Management Systems for Newly Urbanizing Areas. Water Resources and the Urban Environment, Proceedings of the 25th Annual Conference on Water Resources Planning and Management. (pp. 325-329). Reston, VA: American Society of Civil Engineers.

PAGE 158

145 Hollman-Dodds, J. K., Bradley, A. A. a nd Potter, K. W. ( 2003). Evaluation of Hydrologic Benefits of Infiltration Based Urban Storm Water Management. Journal of the American Wa ter Resources Association, 39 (1), 205-215. Horner, W. R. (2000). Conservation Design: Managing Stormwater Through Maximizing Preventive Nonstructural Practices. In, National Conference on Tools for Urban Water Resource Management and Protection, Proceedings. (pp. 147-157). Cincinnati: U.S. EPA. Huhn, V. and Stecker, A. (1997). Alternativ e Stormwater Management Concept for Urban and Suburban Areas. Water Science Technology, 36 (8-9), 295-300. Hsieh, C. and Davis, A. P. (2003). Evaluati on of Bioretention for Treatment of Urban Storm Water Runoff. World Water & Environmental Resources Congress. (pp. 18). Philadelphia: American Society of Civil Engineers. Kaiser, M. (1997). Requirement s and Possibilities of Best Management Practises for Storm Water Run-off from the View of Ecological Town Planning. Water Science Technology, 36 (8-9), 319-323. Leopold, L. B. (1968). Hydrology for Urban Land Pl anning: A Guidebook on the Hydrologic Effects of Urban Land use USGS Circular 554. Li, M. M. and Brown, H. J.. (1980). Mi cro-Neighborhood Externalities and Hedonic Housing Prices. Land Economics, 56 (2), 125-141. Liptan, T. (1995). A Cost Comparison of Conventional and Water Quality-Based Designs. Proceedings of the 4th Biennial Stormwater Research Conference. (pp. 222-231). Clearwater, FL: Southwest Fl orida Water Management District. Livingston, E. H. (2000). Lessons Learne d About Successfully Using Infiltration Practices. National Conference on Tools for Ur ban Water Resource Management and Protection, Proceedings. (pp. 81-96). Cincinnati: U.S. EPA. Low Impact Design Center. (2003). Low Impact Development (LID) Urban Design Tools www.lid-stormwater.net. La st Accessed: 10/25/2003. Marsalek, J., Rochfort, Q., Brownlee, B., Mayer, T. and Servos, M. (1999). An Exploratory Study of Ur ban Runoff Toxicity. Water Science and Technology, 39 (12), 33-39. Means, R. S. (2001). Site Construction Cost Data. Kingston, MA: R. S. Means Company. National Oceanographic and Atmosphe ric Administration (NOAA). (1982). NOAA Technical Report NWS 34 Mean Monthly, Seasonal, and Annual Pan Evaporation for the United States. Washington D.C.: NOAA.

PAGE 159

146 Peterson, M. M. (1999). A Natural Approach to Watershed Planning, Restoration and Management. Water Science Technology, 39 (12), 347-352. Prince Georges County. (1997). Low Impact Development Design Manual. Largo, MD: Prince Georges County Department of Environmental Resources. Prince Georges County. (1999a). Low Impact Development Design Strategies: An Integrated Approach. Largo, MD: Prince Georges County Department of Environmental Resources. Prince Georges County. (1999b). Low Impact Development Hydrologic Analysis. Largo, MD: Prince Georges County Departme nt of Environmental Resources. Roesner, L. A., Bledsoe, B. P. and Bras hear, R. W. (2001). Are Best ManagementPractice Criteria Really E nvironmentally Friendly? Journal of Water Resources Planning and Management, 127 (3), 150-154. Rushton, B. T. (1999). Low Impact Parking lot Design Reduces Runoff and Pollutant Loads. Annual Report #1. Brooksville, FL: Southwest Florida Water Management District. Russell, S. O., Kenning, B. F. I. And Sunnell, G. J. (1979). Estimating Design Flows for Urban Drainage. Journal of the Hydraulics Division, HY1 43-52. Sample, D. J., Heaney, J. P., Wright, L. a nd Koustas, R. (2001). Geographic Information Systems, Decision Support Systems and Urban Storm-Water Management. Journal of Water Resources Planning and Management, 127 (3), 155-161. Sample, D. J., Heaney, J. P., Wright, L. T., Fan, C., Lai, F. and Field, R.. (2003). Costs of Best Management Practices and Associated Land for Urban Stormwater Control. Journal of Water Resources Planning and Management, 129 (1), 59-68. Schueler, T. R. (1994). The Stream Protection Approach Washington D.C.: Metropolitan Washington Council of Governments. Schueler, T. R. (1995). Site Planning for Urban Stream Protection Washington D.C.: Metropolitan Washington Council of Governments. Schueler, T. R. (2000). The Compaction of Urban Soils. Watershed Protection Techniques, 3 (2), 661-665. Schueler, T. R., Kumble, P. A. and Heraty, M. A. (1992). A Current Assessment of Urban Best Management Practices. Washington D.C.: Metropolitan Washington Council of Governments. Sloat, M. S. and Hwang, R. B. (1989). Se nsitivity Study of Detention Basins in Urbanized Watersheds. Journal of Urban Planning and Development, 115 (3), 135155.

PAGE 160

147 Smulen, J. T., Shallcross, A. L. and Cave K. A. (1999). Updating the U.S. Nationwide Urban Runoff Quality Database. Water Science and Technology, 39 (12), 9-16. Soil Conservation Service (SCS). (1990). Soil Survey of Clay County, Florida. Washington D.C.: The Soil Conservation Service. St. Johns River Water Manageme nt District (SJRWMD). (2001). Applicant’s Handbook: Regulation of Stormwater Manageme nt Systems Chapter 40C-42, F.A.C. Palatka, FL: St Johns River Water Management District. U.S. Army Corps of Engineers (USACE). (2000). Hydrologic Modeling System, HECHMS Technical Reference Manual. Davis, CA: Hydrologic Engineering Center. U.S. Department of Ag riculture (USDA). (1986). Urban Hydrology for Small Watersheds: TR-55 2nd Edition. Washington D.C., Government Printing Office. U.S. Environmental Protection Agency (USEPA). (1996). National Water Quality Inventory Report to Congress, Section 1 Washington D.C.: United States Environmental Protection Agency. U.S. Environmental Protection Agency (USEPA). (2000). Low Impact Development (LID), A Literature Review. Washington D.C.: United States Environmental Protection Agency. University of Florida GeoPlan Center (2000), Clay County FGDL Data CD, Version 3. Gainesville, FL: University of Florida Urban Water Resources Research Council (UWR RC) of The American Society of Civil Engineers. (1992). Design and Construction of Ur ban Stormwater Management Systems. New York, NY: American Soci ety of Civil Engineers. Viessman, W. Jr. and Lewis, G. L. (1996). Introduction to Hydrology 4th Edition. New York: Harper Collins College Publishers. Water Environment Federation (WEF) and Amer ican Society of Civil Engineers (ASCE). (1998). Urban Runoff Quality Management. Alexandria, VA: Water Environment Federation. Weicher, J. C. and Zerbst, R. H. (1973). The Externalities of Neighborhood Parks: An Empirical Investigation. Land Economics, 49 99-105. Wright, L., Heaney, J. P., Weinstein, N. (2000). Micro-scale Mode ling of Low Impact Development. Building Partnerships: Proceedings of 2000 Joint Conference on Water Resource Engineering and Wa ter Resource Planning & Management. (pp. 1-8). Reston, VA: American So ciety of Civil Engineers.

PAGE 161

148 Zheng, P. Q. and Baetz, R. W. (1999). GI S-Based Analysis of Development Options From a Hydrology Perspective. Journal of Urban Planning and Development, 125 (4), 164-181.

PAGE 162

149 BIOGRAPHICAL SKETCH Evan Shane Williams is a native of Shavertown, Pennsylvania. He received his Bachelor of Civil Engineeri ng from Villanova University in 1995. He then continued his education at Villanova University, receiving a Master of Civil Engineering in 1998 under the guidance of Robert G. Traver, Ph.D., P.E. Evan enrolled at the University of Florida in late 1998, receiving a Do ctor of Philosophy in environmental engineering in 2003 under the guidance of Willia m R. Wise, Ph.D., P.E.


Permanent Link: http://ufdc.ufl.edu/UFE0002367/00001

Material Information

Title: Hydrologic and Economic Impacts of Alternative Residential Land Development Methods
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0002367:00001

Permanent Link: http://ufdc.ufl.edu/UFE0002367/00001

Material Information

Title: Hydrologic and Economic Impacts of Alternative Residential Land Development Methods
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0002367:00001


This item has the following downloads:


Full Text












HYDROLOGIC AND ECONOMIC IMPACTS OF ALTERNATIVE RESIDENTIAL
LAND DEVELOPMENT METHODS
















By

EVAN SHANE WILLIAMS


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2003

































Copyright 2003

by

Evan Shane Williams

































This document is dedicated to my family.















ACKNOWLEDGMENTS

I thank my parents for encouraging me throughout my education. This dissertation

is as much theirs as it is mine.
















TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES .............................. ........... ... .. ...... .............. viii

LIST OF FIGURES ................................. ...... ... ................. .x

ABSTRACT .............. ..................... .......... .............. xii

CHAPTER

1 O V ER V IEW O F RE SEA R CH ................................................................. ...... ....1

Purpose of R research .................................. .. .. .. ........ .... ............. .
R elevance of W ork .................. ........................... .. ...... ... .............. 1
M orphology .............. .. .................................... ....................... 3
Peak Flow Characteristics and Volume ............ ............................................3
Groundwater ............... ......................... ......... 3
Water Quality ........................................4..........4
Stream W arm ing and Biodiversity ............................................. ............... 4
Current Storm-Water Management Practices............................................................5
D detention P onds ................................................ .................... 6
W et P onds........................................................................................ 7
R detention D designs .............. .............. ............... ..... ........ 8
Su m m ary ................ .................................................................... ............... 10
G general D discussion of P project ................................................................ ...... .... 11
Structure of This D issertation ............................................. ............................ 14

2 DEVELOPMENT ALTERNATIVES.................................. ...................15

Intro du action ............... ... .. ......................................... .. ...................15
Literature on the Development and Storm-Water Management Alternatives ...........15
Traditional D evelopm ent........................................................... ............... 15
Cluster Development ...................... ............... .......... ................. 16
A related concept: planned development district .............. ..................17
P urported b benefits ............................... .... .......... .................. .............. 17
Available literature ...... ........ ...... ........................ .. .......... .... 18
Potential negative factors ........................................ ......... ............... 19
Low Im pact D evelopm ent (LID)...................................................................... 19


v









P urported b benefits ............................... .............. .................. ..............2 1
A available literature ...................................................... .... ................ .. ... 22
Potential negative factors ........................................ ......... ............... 24
A applicable D design Standards.................................. ........................................25
Design standards and m etric conversion .................................. ............... 25
Basic Development Regulations.................. ........ .................................. 26
Cluster Development Regulations................... .............. ............... .29
Planned Development District (PDD)..... .......... ....................................... 31
S u m m ary ...................................... .................................................. 3 2

3 SITE DEVELOPM EN T PLAN S.......................................... .......................... 33

In tro d u ctio n .......................................................................................3 3
E existing C conditions .......................................... ... .... ........ .... .. ... 33
Traditional D evelopm ent Plan........................................................ ............. 35
General Design Discussion......................... ................................ 36
Interpretation Issues With The County Ordinance............................................39
D design P lan s .......................................................................40
Cluster D evelopm ent Plan ................................................ .............................. 42
General Design Discussion......................... ............................... 43
Interpretation Issues With The County Ordinance.............. .. ............ 46
D design P lan s .......................................................................46
L ID D evelopm ents........ .................................................................. .. ........ ... ..48
G general D esign D iscussion............................................ ........... ............... 49
Interpretations of The O rdinance..................................... ......... ............... 52
D design P lan s .......................................................................54
Summary ............. ............................................... .... ... ...... ..56

4 H YD ROLOGIC IM PA CTS.............................................. .............................. 59

Introduction ............. ............ ..... ............ 59
Discussion on Modeling the Alternatives............................................ ..............59
Continuous Simulation Data Collection....... ................................. ...............60
Hydrologic M odel ............................................................... .. ... ..... 61
E xperim ental M ethods.............................................. .................. ............... 63
Existing Conditions Calibration ........................................ ...................... 63
D esign A alternative M odels............................................ ........... ............... 68
Universal model assumptions..................... ....... ...............68
Conventional storm-water management model notes .................................71
LID storm-water management model notes..............................................73
M modeling R results ............... ............... ................................. ............ .. .... 75
W atershed T im ing ..................... .. ...... ................... ...... .. ............... 78
R unoff V olum e ......... ............................. .......... ............. ........... 80
P e ak F lo w s.................................................................... 8 1
C ontinuou s M odel .............................. ........................ .. ...... .... ............83
Structural C ontrols........... ............................................................ .... .... ... ... 85
M odel L im stations ................. ................................ .. .......... ...... .. .. ..85









Hydrologic Summary and Conclusions........................... ........... ..............87

5 ECON OM IC AN ALY SIS ................................................ .............................. 89

In tro du ctio n ........................................... .. ....89..........
R review of L literature .............. ............................................................ .. .....90
M ethods Of Economic Analysis ................................................... ...................91
Construction Cost E stim ation...................................... .......................... ....... 91
Hedonic Price Equation for Predicting Sale Price........................................94
C collection of sales data ........................................ .......................... 95
Collection of open space data.................................... ........ ............... 96
D evelopm ent of lot descriptive variables............................... ...................99
Results of Economic Analysis ......... ................... ................. ... ............ ... 101
Hedonic Price Model Results ............................................ ...................101
D escriptive statistics........... ........................................ ............ .......... 10 1
C orrelation results ............................ .. ............ .................. .............. 102
R egression results.............. ...... .. .. .... ...... .............................. 103
Impact of open space frontage .............. ...... .....................................107
C construction C ost R results ............................................... ............... ...108
Impact to the Developer and Potential Investors..............................................109
L im stations ............... ................ ..................................... ............... 111
Economic Summary and Conclusions .......................................... ...............113

6 DECISION SUPPORT SYSTEMS FOR LAND DEVELOPMENT.......................116

Introdu action ...................................... ................................................ 116
Literature R review ...................................................... ................ 117
C construction of the D SS ........................................................... ............. 118
Sale Price ..................................................................... .........119
Construction Costs .................. .......................... ........ ................. 119
D S S F u n action ............................................... ................ 12 0
R e su lts ...................................... ...................................................... 12 2
Lim stations ................................... ................................. ......... 123
C on clu sion s .............................................. ......... ............... 124

7 C O N C L U SIO N ......... .................................................................... ........... ..... .. 125

APPENDIX

A H YDROLOGIC M OD EL INPU T ...........................................................................127

B CONTINUOUS MODEL AND CONSTRUCTION COST RESULTS ................39

L IST O F R E F E R E N C E S ...................................................................... ..................... 143

BIOGRAPHICAL SKETCH ............................................................. ............... 149
















LIST OF TABLES


Table pge

2-1 Summary of lot layout standards for R1-aa (ACBCC, 2002)..............................27

2-2 Yard and setback standards for R1-aa zoning (ACBCC, 2002). ...........................27

2-3 Selected road design standards for Alachua County (ACBCC, 2002). ................27

2-4 Comparison of full-size and cluster lot requirements (ACBCC, 2002) ...............31

4-1 Adjusted depression storage for selected sub-areas of the "full" LID design. ......75

4-2 Uncontrolled model results for hypothetical design storms. ................................76

4-3 Controlled model results for hypothetical design storms. ....................................76

4-4 Continuous simulation selected peak flow points and total simulation volume....79

4-5 Results of the continuous simulation for selected base flow points ....................79

4-6 Final volumes of management practices........................................................82

5-1 Capital improvements included in analysis. .................................. ...............92

5-2 Descriptive statistics of lots in sales data set. ................... ................... .......... 101

5-3 Open space type and proximity breakdown........................................................ 102

5-4 Correlation results between independent variables.............................................103

5-5 Initial regression results. ...... ........................... ....................................... 105

5-6 M modified regression results .... ... .............................................. .. ............... 106

5-7 Construction cost estim ates....................................................... ............... 109

5-8 E stim ated project return ......... .................................................. ............... 110

5-9 Conceptual profit in dollars. ....................................... ........ ...................111

6-1 D SS O utput. .......................................................................122









A-i Existing conditions m odel areas. .............................................. ............... 127

A-2 Traditional development cover conditions.............................. ...............127

A-3 Cluster develop ent cover conditions.............................................................. 128

A-4 "Partial" LID development cover conditions...........................................128

A-5 "Full" LID develop ent cover conditions.........................................................129

A-6 Initial percent of storage full and sub-area at-large groundwater routing
in p u t. .......................................................................... 12 9

A-7 Evapotranspiration input for all models............... ..................................129

A -8 Existing conditions soil m oisture units. ........................................... .................130

A-9 Traditional develop ent soil m oisture units....................................................... 130

A-10 Cluster development soil moisture units ....................................................131

A-11 "Partial" LID development soil moisture units..............................131

A-12 "Full" LID development soil moisture units....................................................... 132

A-13 "Partial" LID additional depression storage for runoff control .........................132

A-14 "Full" LID additional depression storage for runoff control ............................132

A -15 Existing transform input. ............................................ ............................. 133

A-16 "Partial" LID transform input. ........................................ ......................... 133

A -17 "Full" LID transform input ..................................................... ............... 133

A -18 Traditional transform input ................................. ............... ............... 134

A -19 Cluster transform input. .............................................. ............................. 134

A-20 Traditional development wet ponds 1 to 5................................... ............... 135

A-21 Traditional development wet ponds 6 to 8................................... ............... 136

A-22 Cluster development wet ponds 1 to 3 ..... ...............................................136

A-23 Cluster development wet ponds 4 to 9............................................ ..........137

A-24 LID development conceptual dry detention ponds.......... .......... ...............138

B-l Itemized construction costs (not adjusted for Gainesville) ..............................141
















LIST OF FIGURES

Figure page

1-1 Conceptual diagram of solution field. ............................................................... 13

3-1 E existing conditions plan. ........................................... ..........................................36

3-2 The traditional develop ent plan .......................................................................41

3-3 The traditional development plan without topography, soils or drainage
sub-areas ............................................................. ..... ..... ......... 42

3-4 Areas of high infiltration potential. ....................................................................... 45

3-5 The cluster develop ent plan. ............................................................................ 47

3-6 The cluster development plan without topography, soils or drainage sub-areas......48

3-7 "Partial" LID development plan................................ .......................... 55

3-8 "Full" LID develop ent plan .............. ..................................... ............... 56

3-9 "Partial" LID development plan without topography, soils or drainage
sub-areas ............ ............................................ ........... .. .. ...... .... 57

3-10 "Full" LID development plan without topography, soils or drainage sub-areas......58

4-1 Existing conditions sub-area division. ........................................ ............... 64

4-2 Comparison of observed flows versus calibrated flows .......................................67

4-3 Traditional design sub-areas and storm-water system. .........................................72

4-4 Cluster design sub-areas and storm-water management system. ..........................73

4-5 "Partial" LID design sub-areas.......................................... ........................... 74

4-6 "Full" LID design sub-areas.......................................................... ............... 74

4-7 2-year, 24-hour controlled hydrographs......... ........... ......................... 76

4-8 2-year, 24-hour uncontrolled hydrographs................. ................ ..... ....77









4-9 25-year, 24-hour controlled hydrographs.............................. ............... 77

4-10 25-year, 24-hour uncontrolled hydrographs........ .............. ............... 78

6-1 Conceptual solution set revised from hydrologic analysis................................119

B-l Traditional development continuous model results. ............................................139

B-2 Cluster development continuous model results................................................... 139

B-3 "Partial" LID development continuous model results ................. .. ...................140

B-4 "Full" LID development continuous model results ...............................................140
















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

HYDROLOGIC AND ECONOMIC IMPACTS OF ALTERNATIVE RESIDENTIAL
LAND DEVELOPMENT METHODS

By

Evan Shane Williams

December 2003

Chair: William R. Wise
Major Department: Environmental Engineering Sciences

This research analyzed the hydrologic and economic impacts of four alternative site

planning and storm-water management designs in a hypothetical residential development

in the Gainesville, Florida area.

The four development options analyzed were as follows:

* Traditional development with full-size lots and conventional storm-water
management.

* Cluster development with reduced lots for upland preservation and conventional
storm-water management

* "Partial" Low Impact Development (LID) that implemented the LID storm-water
management system on the full-size lot plan

* "Full" LID that implemented the LID storm-water management system on the
cluster development plan.

LID is an emerging method of site planning and storm-water management that has

been presented by Prince Georges County, Maryland. This approach to land development

combines a distributed, infiltration-based storm-water management system and reduced









area of disturbance. Limits on disturbance were implemented through the use of the

cluster development site plan.

A hydrologic analysis was conducted. The 2- and 25-year regulatory design storms

and a continuous simulation were modeled. The results were compared to the existing

conditions and each other. The results showed that the distributed, infiltration-based

storm-water management system provided a watershed response closer to the natural,

existing response, particularly if combined with a site plan that limited disturbance.

The economic effects of the four alternatives were analyzed by estimating selected

construction costs and the impact of lot size, open space proximity and type on the sale

price of vacant lots. The hedonic price technique was used to analyze the impact on sale

price. The economic analysis showed that reducing lot size would adversely impact the

profit from development. This impact could be mitigated somewhat by maximizing open

space frontage. Construction costs for the LID designs were lower than the designs using

conventional storm-water management. The combined effect of construction cost savings

and sales receipts indicated that the ratio of profit to cost was highest for the LID storm-

water management system combined with limiting disturbance.














CHAPTER 1
OVERVIEW OF RESEARCH

Purpose of Research

This research evaluated the effects of alternative methods of land development and

storm-water management on the hydrologic response of a first order stream and

associated wetlands. In addition to hydrology, economic concerns for the developer and

potential investors are also explored.

The goal of this dissertation was to gain insight into the applicability of alternative

land development practices in Florida and the United States in general. The specific

questions considered were whether alternative land development and storm-water

management improve the hydrologic response of a developed site and what economic

impact, positive or negative, results from implementation alternative practices.

Improvement in the hydrologic response was defined as a response that is closer to the

natural response of the system than resulted from conventional storm-water management

practices.

Relevance of Work

The obvious question in regards to this dissertation is: why should alternative land

development and storm-water management practices be considered? An excellent answer

is provided by Prince Georges County, Maryland. The Department of Environmental

Resources of this County is responsible for the development of one of the alternative

practices, Low Impact Development (LID), discussed later in this dissertation. The work

done by this County was based on the conclusion that current, generally accepted, storm-









water management practices do not adequately protect water resources (Prince Georges

County, Ch.1, p. 5, 1999).

The hydrologic effects of watershed development on water resources are well

known. Peterson (1999), in a discussion on watershed restoration, points out that all

landforms in a watershed have hydrologic function and that degradation or alteration

impairs natural watershed function. Leopold (1968) described four effects of

development:

* Water quality and amenity value may change.

* Changes in peak flow usually appear as increased flow rate and occurrence at an
earlier time.

* Total runoff changes due to increases in runoff volume from loss of infiltration
capacity and natural storage (i.e., depression areas) for both for individual storms
and in the long-term record (increase in runoff frequency).

* At the same time as runoff volume is increasing, baseflow volume may also
decrease due to infiltration capacity losses.

A similar discussion of effects is also presented in Urban Runoff Quality

Management (Water Environment Federation [WEF] and American Society of Civil

Engineers [ASCE], pp. 24-25, 1998) and includes all aquatic ecosystems (wetlands,

lakes, estuaries, etc.). It is notable that this book adds changes on stream morphology as

an impact of urbanization. Somewhat less publicized but also of concern are increases in

stream temperature (Schueler, pp. 26-27, 1995) and losses of habitat and biodiversity as

examples of stream degradation caused by urbanization (Schueler, pp. 28-31, 1995; WEF

and ASCE, p. 26, 1998). Schuler relates all of these impacts to the degree of

imperviousness of the watershed.









Morphology

A relatively low degree of watershed imperviousness, as low as 10%, can result in

stream channel instability (Schueler, pp. 23-24, 1995; WEF and ASCE, p. 25, 1998).

Schueler (p. 24, 1995) states that habitat degradation due to structural change is a primary

effect of stream morphological changes and cites studies showing that habitat degradation

begins to at about 10% imperviousness.

Peak Flow Characteristics and Volume

Changes in peak flow and discharge volume are perhaps the best-known hydrologic

impacts of urbanization. The influence of impervious cover on peak flow is quite well

known. Changes in the peak flow characteristics generally manifest themselves as

increases in peak flow, corresponding to a higher flow stage. Increases in volume are

caused by impervious cover but also by soil compaction in pervious areas that are altered

or graded during construction, thus reducing infiltration potential (Schueler, 2000).

In addition to increased peak flow from increased volume, peak flow is also altered

from changes in watershed timing effects (time of concentration). Peak discharge from a

developed watershed may be earlier than from the same watershed in its natural condition

(Viessman and Lewis, p. 349, 1996)

Groundwater

Groundwater may also be adversely impacted by urbanization. The general theory

is that impervious cover prevents infiltration of rainfall. This lost infiltration is not

replaced by conventional storm-water management system. This leads to reductions in

base flow. The impact of development on groundwater, more properly base flow to the

stream, appears to be an area of variable impact. Several studies have shown

contradictory impacts of urbanization on base flow (WEF and ASCE, p. 11, 1998).









Water Quality

Recently, water quality impacts of storm water runoff have received much

attention; most often focused on pollutants, typically non point-source, such as total

suspended solids (TSS), total nitrogen (TN), total phosphorus (TP), nitrates and heavy

metals. One important study characterizing pollutants in runoff was the National Urban

Runoff Program Survey (NURPS) study done by the United States Environmental

Protection Agency. This study was released in 1983 and is summarized in Design and

Construction of Urban Stormwater Management Systems (Urban Water Resources

Research Council [UWRRC], p. 101, 1992). Recently, Smulen et al. (1999) "updated"

the initial NURPS estimates for various pollutants.

The specific concern for developing areas is that pollutants originating from

developed areas and transported by runoff will be harmful to the receiving water.

Schueler (pp. 24-26, 1995) explains that impervious surfaces in urban development are a

key part of the water quality problem. These surfaces act as collection areas for

pollutants. These pollutants are susceptible to easy wash-off during rain events. Marsalek

et al. (1999) concluded that runoff from urban land use showed at least potential toxicity

40% of the time. In addition, the U.S. Environmental Protection Agency (USEPA, 1996)

has explained the impacts of various pollutants. The EPA has indicated that urban runoff

is a source of degradation for streams, lakes, estuaries and wetlands in the U.S. with

impacts to estuaries and wetlands particularly troublesome, number 2 and 3 impacts,

respectively (USEPA, 1996).

Stream Warming and Biodiversity

Site Planningfor Urban Stream Protection (Schueler, pp. 28-31, 1995) provides a

discussion and summary of key findings from 17 studies that have shown that









biodiversity is impacted by urbanization. Most of the cited studies deal with benthic

communities. Schueler concludes based on the cited literature that the point where

negative impacts on diversity of aquatic insects occur was indeterminate, but fish

communities were affected at about the same level of imperviousness as stream

morphology. A study is also cited that showed that stream temperature can also be

expected to rise with the amount of impervious cover (Schueler, pp. 26-27, 1995).

Current Storm-Water Management Practices

In order to appreciate the relevance of the current research, the performance of

current storm-water management practices must be briefly discussed. The practices

described here are mitigation methods, meaning that they seek to fix a current or future

problem. Typical site planning practices, with the maximum number of lots possible at

the regulatory minimum lot area for the zoning district, do very little to prevent runoff.

The current management approach towards runoff management is generally

directed at removal of runoff from the developed area to the receiving water as quickly as

possible (CH2MHill, Sec. 3, p. 4, 1998). This reflects a long-standing engineering

attitude that water in developed areas is bad-unless it has aesthetic or recreational

benefit. Dion (p. 236, 1993) has an amusing quote attributed to an unnamed civil

engineering professor who stated:

There are three cardinal rules to remember when dealing with a construction
project: 1. get rid of the water, RULES 2 AND 3. Get rid of the water.

In most cases of residential development, the site is graded for rapid drainage with

a transport system that is usually a highly efficient storm sewer system (Coffman et al.,

1998, 2000). These provide little or no attenuation of peaks, volume reduction or

pollutant removal.









Local regulations, however, generally do require mitigation practices to address

some of the effects of development. Control of hydrologic effects generally focuses on

peak flow control. For example, the St Johns River Water Management District

(SJRWMD, Sec. 9, pp. 8-9, 2001) requires control of peak flows for the 2 and 25-year,

24-hour storm events. Further regulations, such as volume control, are enforced only for

special regulation areas. Alachua County has the same peak flow requirement but does

not specify storm duration for the storm event (Alachua County Board of County

Commissioners [ACBCC], 2002).

In Alachua County and the SJRWMD (these agencies hold concurrent jurisdiction

on storm-water management), water quality must also be addressed (ACBCC, 2002;

SJRWMD, Sec. 8, p. 2, 2001). The intent is to catch the volume of runoff associated with

the "first flush" of pollutants often observed in runoff. These regulations, while not

necessarily as representative of typical practice as peak flow controls, do represent a

significant concern for the effects of pollutants in runoff. In most cases, only one

management practice is required-that is, a single end-of-pipe practice per sub-basin.

The following sections discuss some common practices used today. There are two

focuses: water quantity (volume, peak flow) control and quality control. Drawbacks and

benefits in each area are discussed.

Detention Ponds

Control of runoff in residential projects is frequently accomplished through some

form of detention pond system-one of the oldest storm-water management practices.

These ponds are designed to store runoff temporarily and release it at a rate not exceeding

the pre-development peak flow for a given storm. Sloat and Hwang (1989) showed in a

study on performance of detention basins that such basins only control peak flows at their









outlet and that downstream areas still suffered increased peak flows. The cause of this

under-performance is from the lack of volume control provided by detention systems.

Instead of a near instantaneous peak flow, the increased volume results in a peak

(although no higher than pre-development conditions) that is extended over a longer time,

thus causing increased erosion. This may allow for the flow peak to intersect with other

peaks from tributary streams. Additionally, the timing of the peak flow can be delayed

from its natural time, which also may allow it to coincide with the peak of the flood wave

from upstream areas (Sloat and Hwang, 1989). Roesner et al. (2001) also acknowledge

concern for drawn out periods of high flows and add that the design storms typically

required frequently do not control peak flows from events with shorter return intervals.

Roesner et al. (2001) conclude that design requirements are at fault for poor hydrologic

performance of detention practices rather than the concept of detention storage itself.

Although they acknowledge this view is not shared by other authors.

Certain detention pond designs, particularly dry ponds that do not provide

"extended" detention, have a poor record for removing pollutants. If the dry pond

provides "extended" detention, the performance is somewhat better due to sedimentation

and can be quite high, but only for particulate pollutants (Field et al., p. 192, 1993;

Schueler et al., pp. 7-13, 1992). The SJRWMD (Sec. 10, p.1, 2001) does not recommend

dry detention pond systems, unless no other option is feasible.

Wet Ponds

Wet ponds are sometimes referred to as retention systems (Dunn et al., 1995; WEF

and ASCE, p. 220, 1998), but designs in Florida perform as detention systems with a

permanent pool for water quality. Retention systems in Florida, as the name implies, are

designed to prevent all or a certain volume of runoff from being released into the









receiving water, instead allowing for infiltration into the soil or evaporation loss. If the

wet pond is designed as a detention system, flood control and water quality volumes are

temporarily stored above the invert of the outlet structure (SJRWMD, Sec. 14, pp. 1-11,

2001). In a detention configuration, the effect on volume and peak flow is essentially the

same as a dry pond (unless, of course, the permanent pool needs to be recharged, in

which case the pond functions in a retention role).

Removal of pollutants is much better than dry detention ponds. Schueler et al.

(1992) report sediment removal rates from 50-90% and phosphorous and soluble

nutrients from 30-90% and 40-80%, respectively. Other sources agree with this positive

assessment of wet ponds (Field et al., p. 192, 1993; WEF and ASCE, p. 220, 1998;

SJRWMD, Sec. 14, p. 1, 2001). Borden et al. (1998) concluded that pond efficiency was

seasonally variable, and noted that the main removal process appeared not to be

sedimentation, since most pollutants did not appear associated with total suspended

solids, and that biological processes accounted for most of the nutrient removal.

The permanent pool depth is limited in Florida to prevent thermal stratification,

which is assumed to decrease pollutant removal, specifically the release of nutrients

under anaerobic conditions (SJRWMD, Sec. 14, p. 8, 2001). Borden et al. (1998),

however, observed that two ponds in North Carolina with average depths less than 2.5 m

(8.2 ft), an average depth consistent with the SJRWMD guidelines, still stratified.

Furthermore, the study concluded that this stratification might actually enhance pollutant

removal in certain situations.

Retention Designs

Retention, or infiltration, systems are designed not to maintain a permanent pool of

water, but to infiltrate all or a portion of runoff. These systems are the only systems that









the St Johns River Water Management District calls "retention." The SJRWMD

considers a variety of systems as retention including large, flat-bottomed basins, shallow

landscaped areas, pervious pavements and swales with some form of inlet block

(SJRWMD, Sec. 11, pp. 1-5, 2001). These management practices provide quality control

and peak flow abatement through volume reduction determined by water quality

requirements and detention storage beyond the water quality volume. That fact that these

systems provide volume reduction suggests that they would be an improvement,

hydrologically, over detention systems. However, some retention designs suffer some

serious drawbacks.

Infiltration basins are not universally recommended since they can fail rapidly, and

generally cannot be restored through regular maintenance. However, this may have been

directed to the Mid-Atlantic region (Schueler et al., pp. 49-50, 1992). If the basin should

fail, runoff quantity and quality control aspects will, at best, be similar to a pond system.

As was previously stated, these basins are permitted in Alachua County as long as certain

steps are taken to reduce the possibility of failure (SJRWMD, Sec. 11, pp. 4-5, 2001).

Off-line systems that allow flows to by-pass when the retention basin is full (volume is

designed to capture the "first-flush") have had greater success in Florida (UWRRC, p.

499, 1992).

Assuming that the infiltration basin functions as intended, Schueler et al. (pp. 47-

53, 1992) suggests that pollution removal should be high for particulates and moderate to

low for soluble pollutants depending on soil characteristics. There is a slight risk of

groundwater contamination. If the pond should fail, water quality treatment may still be

provided, and the basin may be retrofitted to a wet pond.









Other infiltration practices include the infiltration trench and sand filter. These are

not common in residential projects in Florida. In fact, the "exfiltration trench" design

guidelines provided by the SJRWMD notes that these are typically built in downtown

areas where space is limited (SJRWMD, Sec. 13, pp. 1-6, 2001). Schueler et al. (p. 49,

1992) suggested that infiltration trench performance, at least in the water quality sense,

should be similar to infiltration basins. They can serve areas of 2 to 4 ha (5 to 10 Ac)

(UWRRC, p. 501, 1992).

Sand filters differ from infiltration trenches in that the filter medium is sand, and

some designs do not allow filtered runoff to be infiltrated (Claytor and Schueler, Ch. 1,

pp. 4-6, 1996; CH2MHill, Sec. 8, 1998). In terms of pollutant removal, Claytor and

Schueler (Ch. 4, p. 33, 1996), provides information for various types of sand filters

showing high removals for particulate matter, moderate levels for total phosphorus and

nitrogen but poor removal for nitrate/nitrites.

Experience with infiltration trenches has shown that while sometimes having good

pollutant removal capability, they require frequent maintenance, which is usually not

done (Schueler et al., p 44, 1992; UWRRC, p. 501, 1992). A variety of designs exist

(Claytor and Schueler, Ch. 1, pp. 4-6, 1996), with surface types accompanied by grass

filter strips being the most typical if used in residential areas (UWRRC, p. 501, 1992).

Summary

The conclusion that can be drawn from the discussion above is that typical storm-

water management using conventional end-of-pipe, single practice systems does not

always provide the level of protection of water resources that is desirable-although

some practices are better than others. Particular designs may provide adequate controls in

certain areas of runoff management but not in others. For example, wet ponds provide









good pollutant control but flow control suffers the same drawbacks as dry detention

systems. There is a growing realization that changes are needed in storm-water

management practices. A variety of interesting approaches are available in literature and

are discussed in the next chapter.

General Discussion of Project

The 246.29 ha (608.66 ac) hypothetical property under development was located in

the Camp Blanding Wildlife Management Area that is part of the Camp Blanding Florida

National Guard Training Center. The site is located outside of Alachua County (about 35

miles Northeast of Gainesville) because of the desire to locate a site that was relatively

undisturbed and not in private ownership. Stream stage data were collected on the

upstream side of a culvert crossing an unimproved road. Rainfall data were also collected

with a recording gage located on the site. These data were used to calibrate an existing

conditions hydrologic model that was used to evaluate impacts of development.

The area surrounding the stream was considered a property undergoing residential

development. The hypothetical property was "developed" with a scenario based on

allowable density of housing units. The scenario chosen allowed for minimum 1,857.88

m2 (20,000 SF) lots, the R1-aa zoning classification of the Alachua County, Florida,

Ordinance (ACBCC, 2002). The hypothetical property was then "developed" for design

options:

* Traditional development (full-size lots, pipe/pond storm-water management)
* Cluster development (half-size lots, pipe/pond storm-water management)
* "Partial" Low Impact development (LID storm-water system, full-size lots)
* "Full" Low Impact Development (LID storm-water management, half-size lots)

Low Impact Development (LID) is a design strategy developed by Prince Georges

County, Maryland, which emphasizes protection of vital hydrologic features and replaces









typical storm-sewer system with a distributed, infiltration-based system. The cluster

development option, focused on land preservation only. Each of these alternatives was

analyzed for compliance with hydrologic criteria. The basic criteria for hydrologic

evaluation are discussed in detail in Chapter 2, but generally follow a typical storm-water

management study (i.e., meeting regulatory concerns). In addition, each alternative was

evaluated for how closely it maintained the natural hydrologic response. Each alternative

is compared against the existing conditions of the site. The hydrologic analysis is

discussed in Chapter 4.

Those who are familiar with the economics of land development will note that the

alternatives that reduce lot size will have an economic impact to the developer. Thus, a

second set of analyses was required that dealt specifically with the economics. A market

study was done to measure the impact of reduced lot size and the effect of open space

preservation. In addition, measurement of selected capital improvements was done for

each alternative. The criterion for economic evaluation was that the alternative designs

should have little adverse impact to the developer or potential investors. The methods of

evaluating impact were a comparison of dollar receipts (sales receipts minus construction

costs) and the ratio of receipts to construction costs. Alternatives 2, 3 and 4 are compared

against alternative 1, which represents traditional development. Economics are discussed

in Chapter 5.

Naturally, there was no guarantee that any of these solutions would meet the

economic criterion. The alternatives that were tested essentially represent the corners of a

solution field. Figure 1-1 illustrates this concept. The storm-water management axis

represents the degree that the system tends toward a particular approach, relative to









increasing lot size. This is "measured" by the degree that traditional features of storm-

water management, such as detention ponds or storm inlets, are present. It was expected

that a portion of this solution set would be eliminated by the hydrologic analysis. In

order to determine the optimal solution economically and in terms of storm-water

management system, a decision support system (DSS) was constructed, and is discussed

in Chapter 6.

L II



















Figure 1-1. Conceptual diagram of solution field.

The plans that were developed for this research can be best described as "sketch

plans" that meet land development guidelines (setbacks, lot dimensions, etc.) but are

missing some level of detail that would make them complete. The most visible example

of this is grading. The lots and streets are not graded in detail. They are "rough" graded

to provide runoff flow direction (i.e., to a detention basin) only.

Commercial and industrial development is not addressed in this research. In some

respects, with residential projects, it is easier to conduct an analysis such as this since









development often occurs over a larger site while commercial areas are generally small

parcels.

Structure of This Dissertation

This chapter contains introductory information. The chapters that follow discuss

each of the three components of the project starting with hydrology, followed by

economics and finishing with the decision support system example. There is a certain

degree of overlap among chapters, and certain site design decisions, discussed in chapter

3, were made were based on a principles, for example the economic influence of open

space, that are explained in later chapters, such as Chapter 5 on economics.














CHAPTER 2
DEVELOPMENT ALTERNATIVES

Introduction

This chapter reviews the literature on the four design alternatives that were

introduced in the previous chapter. Land development and storm-water management are

regulated activities. Thus, it is necessary to review the governing regulations. The

applicable design guidelines in the Alachua County Code of Ordinances and SJRWMD

Applicant's Handbook: Regulation of Stormwater Management Systems are reviewed in

this chapter.

Literature on the Development and Storm-Water Management Alternatives

Alternative methods of storm-water management have been proposed to alleviate

the burden placed on water resources by development activities. Proposed alternatives are

not limited to engineered storm-water systems, but also include changes in the way sites

are designed and developed. This section reviews the literature on each of the design

alternatives and presents the rationale for inclusion in this research.

Traditional Development

The traditional development option reflects the minimum standards required by

local ordinance. Local government entities with the power to review and approve new

development require minimum standards of design for both the site plan and storm-water

management system. In the case of site design and storm-water infrastructure, these are

usually physically based standards, such as lot size, lot dimensions, storm sewer pipe

material and allowable lot density. The runoff control portions of the ordinance are









performance based. Typically, these require that a certain discharge may not be exceeded

for one or more theoretical design storm events. It is often the case in the land

development business to design a new development to meet the minimum standards and

nothing more. With this in mind, the rationale for including this approach to land

development in this research is obvious.

The applicable design guidelines for this alternative are the Alachua County Code

of Ordinances (Alachua County Board of County Commissioners [ACBCC], 2002) and

the St. Johns River Water Management District (SJRWMD) Applicant's Handbook:

Regulation of Stormwater Management Systems on the regulation of storm-water

management systems (SJRWMD, 2001). Site design standards are set by Alachua

County, but both entities have concurrent jurisdiction on storm-water management.

Cluster Development

The cluster development concept is an element of most land development

ordinances, including Alachua County. Using the cluster development option allows the

developer flexibility in lot size, while still guaranteeing at least the same number of lots

as would be available if the regular zoning requirements were followed. In return, the

developer provides "open space" (Schueler, pp. 55-56, 1994).

In the context of this research, the cluster development option is significant because

it may be used to lessen the impacts of development on natural hydrologic features,

which is a relatively new application of cluster development (Schueler, p. 55, 1995).

Thus, cluster development design becomes a means of impact minimization rather than

the mitigation approach of traditional storm-water management. The acronym "BMP" is

used, in this research, to refer to structural methods of runoff control. However, use of the

cluster design option for runoff control is itself a BMP albeit a non-structural one.









A related concept: planned development district

Also important to consider in this discussion is the planned unit development

(PUD) also called the planned development district (PDD) in Alachua County (ACBCC,

2002). The Pennsylvania best management practice (BMP) manual states that PUD can

be used to implement the cluster design concept (CH2MHill, Sec. 3, p. 8, 1998). In

Alachua County, these are two design options, although they are quite similar. A detailed

discussion of the County zoning regulations follows later in this chapter; however, to

summarize the PDD design option allows a greater degree of design flexibility in site

planning than the cluster option and is apparently meant specifically for implementing

creative urban planning and encouraging environmental protection (ACBCC, 2002).

Purported benefits

The first benefit of cluster developments identified in the literature is reduction in

impervious area, which in turn provides a reduction in runoff and runoff-borne pollution.

Impervious land uses are mainly dedicated to transportation (Schueler, pp. 19-20, 1995).

With smaller, more compact lots, the size of the interior street system, in the case of

residential projects, is lessened (CH2MHill, Sec. 3, p. 5, 1998). This is the logical

conclusion, since the opposite, large lots increasing the amount of impervious surfaces

for transportation, is frequently true (Schueler, p. 38, 61, 1995). Schueler (p. 61, 1995)

cites a 1989 Maryland Office of Planning study demonstrating that cluster developments

can reduce impervious cover by 10 to 50%, with the greatest benefit coming from

clustering larger lot sizes (reductions in impervious cover decrease as initial lot size

decreases). In addition, cluster developments, with their smaller overall footprint, can

also reduce the amount of uncontrolled areas in a development by concentrating runoff to









one area and providing more options for BMP sites, while reducing the size of the BMPs

(Schueler, pp. 61-66, 1995).

The second benefit of cluster developments is preservation of open space for

natural resource protection. In Pennsylvania, it is recommended that cluster developments

be used to avoid impacting natural drainage features such as ephemeral swales and areas

of high infiltration (CH2MHill, Sec. 3, pp. 4-8, 1998). Cluster developments can also be

used to enhance or ease (because "lost" lots in the buffer area may be transferred to

another portion of the site due to smaller lot size) the creation of buffers around water

resources (Schueler, p. 61, 1995). Logically, these implementations of cluster

developments would require preservation of at least some open space in its natural

condition.

Pollution control from cluster developments may be enhanced because there is less

area generating pollutants. However, the actual composition of the open space is

important. Pollutant reduction, particularly for nutrients, will be lessened if a significant

proportion of the open space is fertilized lawns, such as ball fields or landscaped areas.

The cluster design typically uses conventional BMPs for storm-water management.

Available literature

Other literature has quantified the benefits of cluster development in general or on

issues related to the particular application of cluster developments (i.e., protecting

hydrolgically significant areas). Zheng and Baetz (1999), in a watershed-scale analysis of

development patterns, concluded that development patterns with smaller impacted areas

from combining reduced lots with reduced right-of-way width and mixed dwelling types

with open space result in lower peak flows than more traditional patterns. Zheng and

Baetz (1999) also concluded that this benefit is most visible at the sub-watershed level.









Sloat and Hwang (1989), in their study of detention basin performance, concluded that

preserving a buffer, which can be easier or less contentious through the use of cluster

development, of natural land around a stream reduced peak flows. An application of

some of these principles was the Rock Creek Development in Colorado, a planned

community versus a "true" cluster development (Galuzzi and Pflaum, 1996). Here a

deliberate effort was made to preserve and protect the natural drainage by placing open

space and natural areas along the stream corridors.

Potential negative factors

Cluster-type developments have potential negative factors that can limit their use.

Some of these factors are economic and are discussed Chapter 5. A discussion of

drawbacks that pertain to hydrology is found in Schueler (pp. 67-69, 1995). Although

cluster developments are often not used for water resource protection, Schueler notes that

wording of a particular ordinance can actually decrease usefulness of cluster design. One

concern is the local concept of development density. If the locality has no concept of

unbuildablee" area, due to natural resources etc., impervious area within a cluster

development may actually increase because there are more dwelling units in the cluster

design than there would be in a full-size lot design. This can also be true if "bonus"

provisions exist that allow for more units. Another issue that is discussed is the fact that

local ordinances do not necessarily require that open space to be natural areas, but might

even allow impervious recreation areas to be considered open space (Schueler, pp. 67-69,

1995).

Low Impact Development (LID)

Low impact development is a site planning and design strategy that has recently

been advanced by Prince Georges County, Maryland. It has been presented as an









improvement of the current practice of land development storm-water management. The

goal of LID is a site design that replicates pre-development hydrologic function and

response of the site (Coffman et al., 2000). These goals are accomplished through a mix

of site planning and engineered management practices, which are called integrated

management practices (IMPs) by LID advocates to differentiate them from more

conventional practices (BMPs) (Coffman et al., 2000).

The site planning concepts of LID emphasizes minimization of impacts from

development. Literature advocating LID (Coffman et al., 1998, 2000; Prince Georges

County, Ch. 3, pp. 2-11, 1997, Ch. 2, pp. 2-13, 1999a), recommends minimizing the area

of development to avoid water resources and their associated buffers, and preferably also

include areas of high infiltration potential and natural drainage pathways. Within the

developed area, further lessening of impacts can be accomplished through "site

fingerprinting", which is meant to reduce disturbance and reduce and disconnect

imperviousness on individual lots even preserving woodlands on the lots themselves. The

LID literature cited here does not provide any specific guidance on the extent of land

preservation. Additional site planning goals include reducing impervious area through

alternative road layout, elimination of sidewalks/parking lanes and narrower street

sections. Thus, consideration of the site's hydrologic function and means of maintaining

this function must begin early in the design process.

In addition to site-planning aspects of LID, a fundamental change in the storm-

water management system is proposed. Instead of the conventional end-of-pipe systems,

LID relies on distributed, source controls that are designed to maintain watershed timing

and encourage infiltration. Perhaps the most noticeable practices are use of swales in lieu









of storm sewers and bioretention areas. Bioretention areas are shallow landscaped areas

used to infiltrate runoff on each individual lot (Coffman et al., 1998, 2000; Prince

Georges County, Ch. 3, pp. 11-15, 1997, Ch. 4, pp. 1-25, 1999a).

Purported benefits

The main benefits of LID, the maintenance of natural hydrologic function and

response, have already been discussed. LID is both a prevention and an enhanced

mitigation strategy. This is in contrast to the cluster design concept, which may provide

prevention of some runoff and runoff-borne pollutants, but still relies on the conventional

view of efficient storm-water removal to end-of-pipe management practices. The

prevention aspects of LID are, in most ways, the same as cluster developments, although

nowhere in the LID literature was cluster development explicitly mentioned as a means

of meeting LID site planning goals. On the other hand, LID is also not presented as a

rigid design approach and community, and even site-specific interpretations are

encouraged (Coffman et al., 2000).

A brief definition of disconnection of impervious surfaces, which is not addressed

in discussion of cluster developments, is appropriate. Disconnecting impervious surfaces

means that instead of routing runoff directly to the storm sewer system, runoff from these

areas flows onto pervious lawn areas to allow for infiltration.

The LID storm-water system emphasizes source control and infiltration and even

reuse. Ideally, this approach will replicate the natural hydrology of the site. Andoh and

Declerck (1997) discuss a source control storm-water system that mimics nature and

Kaiser (1997) stresses the need to restructure urban storm-water systems to behave more

like natural systems, including providing opportunity for storm-water infiltration. In the

United States, a similar approach to land development called Conservation Design has









been proposed (Homer, 2000). Heaney et al. (1998) also advocate source control of

runoff and note that if on-site controls can store runoff, most locations could meet their

lawn evapotranspiration (ET) requirements through reuse.

The distributed source controls also provide the advantage that if one unit fails, the

rest are still available, whereas in the conventional system, failure of one management

practice usually means that controls for the entire site are seriously or totally degraded.

Another potential benefit of LID is pollution control. Traditional development, and

even cluster designs, place controls at the end of the storm sewer system. This approach

places a great reliance on a small number of practices. In contrast, LID begins treatment

of runoff at the source and continues treatment through the swale based transport system.

This results in a "treatment train" of management practices that provides the most

effective means of runoff treatment and control (CH2MHill, Sec. 4, p. 2, 1998). Runoff,

however, from the roads will only be treated in the swale system, while runoff from lawn

areas will be treated first in bioretention units and then in the swales if the bioretention

areas should fill. Land protection aspects also assist in pollution control for the same

reason discussed in the cluster development section.

Available literature

Much like the cluster development concept, the benefits of LID implementation are

fairly intuitive to the hydrologist. Prince Georges County, Maryland has some examples

of LID computation in the context of a TR-55 (U.S. Department of Agriculture [USDA],

1986) based analysis (Prince Georges County, Appx. D, 1997). Huhn and Stecker (1997)

discussed the performance of a LID-type approach in Germany, where simulations

showed that a majority of runoff could be infiltrated successfully. This study also

indicated that natural soil conditions impact the effectiveness of such a system. Areas









with better infiltration conditions will have less of a reliance on controlled surface

discharge. A theoretical study by Holman-Dodds et al. (2003) concluded that storm-water

management approaches that encourage infiltration require distributed controls. The most

effective areas for control are in upland areas that tend to have soils with better

infiltration potential and are generally the area where construction is occurring. These

high infiltration areas, however, are also the areas where the greatest impact of

development will be seen. The same study noted that disconnecting impervious surfaces

provided significant benefits over traditional development, but that these benefits

decreased as the rainfall depth increased.

The Jordan Cove Urban Watershed project, currently in progress in Waterford,

Connecticut, will provide an interesting "head-to-head" comparison between traditional

development and LID. Preliminary results are only available for the construction period,

but favor the LID-type design for runoff quantity. The results note pollutant

concentration increases for the LID-type design that were not apparent in the traditional

design but were attributed to specific events (one example was unstabilized swales

resulting in suspended solids output). Mass export of pollutants was higher for the

traditional site; which was attributed to higher runoff volume (Clausen, 2002).

Other publications deal with individual aspects of LID. For example, Braune and

Wood (1999) discuss a matrix developed by others in Colorado that shows disconnecting

impervious surfaces to encourage infiltration as "highly effective". Interestingly, this

matrix also listed some traditional end-of-pipe practices in the same category, while other

LID practices such as swales were somewhat less effective at mitigation.









Other examples in literature deal with performance of individual BMPs

recommended for use in LID. A study of an aquarium parking lot in Tampa, Florida

evaluated the effectiveness of grass swales for runoff quantity and quality improvement.

Areas with swales had less runoff than those without. Pollutant removals varied with

pavement type. Nutrient removal was worse than sediments and metals, particularly for

phosphorus (Rushton, 1999). This variability in pollutant removal was also reflected in

another summary of swale performance in Maryland, Florida and Virginia (USEPA,

2000).

Bioretention is another LID component that has recently been examined. Most of

the focus on bioretention has been on pollutant removal. Hsieh and Davis (2003) reported

that 6 different bioretention facilities showed high removal efficiency for sediments and

lead, variable removal of phosphorus and low removal for nitrate and ammonium. An

EPA literature review (USEPA, 2000) discusses studies on constructed systems in the

field and laboratory systems. All cited studies used synthetic runoff. Performance was

generally high for metals, although one of the studies reported results that were lower for

one of the field systems. Nutrient removal was high for some nutrients like total

phosphorus, but quite low for nitrate. Laboratory studies showed a relationship between

nutrient removal and depth in the bioretention soil matrix.

Potential negative factors

LID, like most development practices, has potential negatives. The LID approach is

generally applicable but dependent on a number of factors (Coffman et al., 1998)

including conflicts with local regulations, property use (land use rights) issues and

development density allowing enough space for LID management practices. It is also

possible that LID will not totally eliminate the need for conventional detention practices









(Coffman et al., 1998). LID also relies on infiltration practices. One problem with

infiltration practices is a high failure rate for a variety of reasons including poor

estimation of infiltration rate and soils compaction (Livingston, 2000). Thus, it would be

wise to consider lessons of the past with regard to infiltration systems, which may not be

reflected in local ordinances or design guidelines. Livingston (2000) discusses of several

practices and the lessons learned. For example, swale infiltration design criteria set by the

SJRWMD were often unattainable due to the space required for swales. The

recommendation was to use check dams or a similar means to provide depression storage

in the swale itself.

Applicable Design Standards

It is important to introduce, briefly, some of the land development requirements

that affect this research before the discussion of site design. The hypothetical residential

development project assumes a property size of 246.29 ha (608.66 ac), that will undergo

residential development. This does not include other areas of the watershed that will drain

through the project site and are included in the hydrology model. There are two

governing regulations: The Alachua County Code of Ordinances (ACBCC, 2002) and the

Applicant's Handbook: Regulation ofStormwater Management Systems from the

SJRWMD (2001), which summarizes rules on storm-water management for the District

and provides design guidance on management practices.

Design standards and metric conversion

Site design was done in metric units. However, the applicable ordinances provide

standards in Imperial units. This required rounding (and therefore slight inaccuracy) of

decimals when dimensions, required by ordinance, were converted. Conversions from

feet to meters, for linear dimensions, were computed with a conversion factor at three









decimal places (i.e., 3.281 feet per meter), and dimensions were rounded to the nearest

0.1 m or, if the conversion was closer, a "common fraction" such as 0.25 or 0.33 (i.e., 50

feet converts to 15.24 m so 15.25 was used instead of 15.2). Most errors are on the order

of inches or fractions of inches, therefore, this error was not an issue for this research or

even site design in general. Area dimensions are rounded to two decimal places.

This conversion convention applies only to the actual layout of the site and is

intended to reflect the fact that most design standards in the County Ordinance do not use

fractions and are frequently rounded to a simple number. For example, the R1-aa zoning

can be considered "half-acre" even though actual minimum lot areas is 20,000 SF instead

of 21,780 SF.

Basic Development Regulations

The traditional development option allowed for a minimum 1,857.88 m2 (20,000

SF) lots and a maximum of 4,046.46 m2 (1 Ac). This is the R1-aa zoning classification for

Alachua County. Allowable density of lots is based on total property area, including

unbuildablee" areas like wetlands, streams and their associated water buffers. The buffer

that is required around water resources varies with the value of the particular feature but

is a minimum of 10.67 m (35 ft). Thus, the allowable number of lots for a given zoning

classification will always be greater than what is practical (ACBCC, 2002).

Minimum lot dimensions appear to be designed to ensure rectangular lots (lots on

curves and cul-de-sacs are exempt from these dimensional requirements). Building

setback lines are also defined for each zoning classification (ACBCC, 2002). These two

standards are important to this research because of the assumption of a single, generic

house footprint that is constant throughout the alternatives. This essentially sets a lower

limit on lot width beyond that required by ordinance.









Alachua County also requires a narrow buffer between subdivisions at the

boundary. Single-family residential districts must have a 4.6 m (15 ft) buffer. No

structures are permitted in the buffer area. There is no requirement that this buffer should

be separate from a platted lot (ACBCC, 2002). Tables 2-1 and 2-2 summarize the

relevant design standards for lots.

Table 2-1. Summary of lot layout standards for R1-aa (ACBCC, 2002).
Area Width Depth Buffer
Minimum 1,857.88 m2 33.5 m 38.1 m 4.6 m
Maximum 4,046.46 m2 None None None

Table 2-2. Yard and setback standards for R1-aa zoning (ACBCC, 2002).
Front Yard Rear Yard Side Yard Side on Street
Minimum 7.6 m 9.1 m 3.8 m 7.6 m
Maximum None None None None

Design standards, listed in Table 2-3, for roads are also of interest. Road lane width

and presence of a turning lane are determined by the level of service as computed by the

average daily traffic (ADT). Road lane widths range from 3.1 to 3.67 m (10 to 12 ft).

Right-of-way (ROW) width is determined by street classification (A through D, with A

being the lowest level of service and Types C and D further divided into C-l and C-2

etc.). Level of service is determined by the ADT. The ROW for a street, of type B or

greater service, with a swale in lieu of storm sewer is wider (ACBCC, 2002).

Table 2-3. Selected road design standards for Alachua County (ACBCC, 2002).
Type A Street Type B Street Type C-1 Street Type C-2 Street
ADT 0 125 126 1,200 1,201 3,200 3,201 7,000
Lane Width 3.1 m 3.1 3.4 ma 3.67 m 3.67 m
ROW Curb 15.25 m 15.25 m 18.3 mb 24.4 m
ROW Swale 15.25 m 18.3 m 24.4 m 30.5 m
a3.4 m lane width applies where ADT is over 400.
bAssumes no on-street parking.









Requirements for the design of on-street parking lanes are also provided, however

on-street parking is rarely seen in developments in Alachua County. Sidewalks are only

required along streets with a high level of service or where certain community features

are present such as schools and parks (ACBCC, 2002).

The street drainage system may either be a closed system with storm sewer and

curb and gutter or a swale section. Swale design has a limiting velocity of 0.9 m/s (3 ft/s)

for the 10-year rain event; otherwise they require a paved invert. Furthermore, swale flow

must not encroach on the road for this design event. Swales are also prohibited where the

groundwater table is within 0.9 m (3 ft) of the ground surface. In general, all storm sewer

pipes must be reinforced concrete under paved roads (asphalt coated corrugated metal is

acceptable elsewhere) and must have a minimum diameter of 38.1 cm (15 in) in a closed

system and 45.7 cm (18 in) where pipe is used in an open swale system. Closed systems

must be designed for the three-year, rain 10-minute event as determined from the local

IDF curve (ACBCC, 2002).

Both Alachua County and SJRWMD govern general storm-water management

requirements. The regulations are, in general, similar. One difference is the required

design storms for peak flow control. SJRWMD requires design for the mean annual (2.5-

year) and 25-year 24-hour storms (SJRWMD, Sec. 9, p. 8, 2001), while Alachua County

states only that peak rate of discharge must not exceed existing conditions for all storms

up to 25-year return interval (ACBCC, 2002). Alachua County does explicitly state that

the volume of any storage system must accommodate the 25-year, 24-hour event.

There is no storm-water quality design event specified in either ordinance. Instead,

water quality storage requirements are defined as a depth of rainfall multiplied by the









contributing area or the impervious area (the depth to be multiplied over the impervious

area is greater than the depth for the whole contributing area). For example wet ponds

must provide water quality storage for the larger of 25.4 mm (1 in) over the entire

contributing area or 63.5 mm (2.5 in) multiplied by the area of impervious cover. An off-

line retention based system with either under drains or infiltration requires half the

treatment storage of a detention system. On-line retention systems, which most end-of-

pipe basins are, require 12.7 mm (0.5 in) more storage than an off-line system over the

entire watershed. Swale systems must infiltrate 80% of the 3-year, 1-hour event. Note

that the swale water quality requirement is inconsistent with the requirement of a paved

invert for velocities above 0.9 m/s (3 ft/s) (this invert requirement is not required in

SJRWMD regulations). Treatment volumes must be recovered in 72 hours (ACBCC,

2002).

Cluster Development Regulations

Development standards for a cluster development are also important for this

research. These apply to lot layout and building setback lines. The minimum lot area for

the zoning district may be reduced by half with dedication of open space that corresponds

to the lot reduction (i.e., 1 m2 reduction in lot area requires 1 m2 dedication of open

space). Natural water resources and their required buffering must be included in the open

space, thus Alachua County indirectly requires a certain portion, depending on the

existing conditions of the site, of the open space to be natural. Recreation areas and

certain storm-water management areas, such as wet ponds, may also be included

(ACBCC, 2002).

Cluster development regulations require that the allowable density, based on

existing zoning, shall not be increased. However, because density calculations can









include unbuildable areas, the actual number of lots for a cluster development may

actually exceed that of traditional development, depending on site conditions (ACBCC,

2002).

Cluster development has different standards regarding lot setbacks. Setback

distance is reduced from standards listed for full-size lots to 4.6 m (15 ft) for front and

rear yards and 1.5 m (5 ft) for side yards, while side yards on streets require 3.1 m (10 ft).

The exception to this rule, according to the wording of the ordinance, occurs when a lot

with reduced area abuts another parcel (in other words, one of the lot lines is also the

subdivision boundary). In this case, building setbacks must be the same as the standard

for the zoning district, though it is not stated if this applies to all setbacks or just the

particular yard that actually abuts the neighboring property. In addition, the lot must also

have a 1.8 m (6 ft) high fence or a 3.1 m (10 ft) wide vegetated buffer with 75% opacity.

Logically, this would replace the 4.6 m (15 ft) buffer, without screening requirements,

which is required under standard zoning. In the previous section, it was noted that the

standard buffer between subdivisions is not mandated as different from platted lots. The

wording of the ordinance would seem to imply that if an off-lot buffer was provided, then

individual lots would not abut the subdivision boundary and cluster development

buffer/setback requirements would not be required (ACBCC, 2002).

Cluster developments have a provision in the County Ordinance that allows for site

density to be increased if a certain percentage of the units in the subdivision are

"affordable" housing (ACBCC, 2002). If this option was applied, site disturbance and

imperviousness may increase above what would normally be the case for a traditional









development. Table 2-4 provides a comparison of the site design standards for standard

R1-aa zoning and a cluster development applied to the R1-aa zoning district.

Table 2-4. Comparison of full-size and cluster lot requirements (ACBCC, 2002).
Standard Cluster
Minimum Lot Area 1,857.88 m2 928.94 m2
Minimum Lot Width 33.5 m 16.75 m
Minimum Lot Depth 38.1 m 19.1 m
Front yard Setback 7.6 m 4.6 m
Rear Yard Setback 9.1 m 4.6 m
Side Yard Setback 3.8 m 1.5 m
Street Side Yard Setback 7.6 m 3.1 m
Buffer/Screening 4.6 m Buffer, no screening Speciala
aRequirement is for screening consisting of 1.8 m high fence/wall or a 3.1 m wide buffer
with vegetation that is 75% opaque.

All storm-water management requirements for cluster developments are the same

for standard developments. As was previously stated, certain storm-water management

features can be counted as open space. The feature should provide some amenity or

aesthetic value (ACBCC, 2002).

Planned Development District (PDD)

The PDD is intended to provide the developer a great degree of flexibility in site

planning for use of the property but also to provide enhanced protection of natural

resources. The developer does not have to follow the minimum design standards that

have been previously discussed. This makes the PDD option even more flexible than the

cluster development option. The design standards of note include avoiding development

within natural areas to the greatest extent possible, ensuring that developed areas are

compact and continuous and maintaining connections between natural areas to avoid

fragmentation (ACBCC, 2002).






32


Summary

The material presented in this chapter presents the concept of the alternative land

development and storm-water management practices that undergo hydrologic and

economic analysis in the following chapters. The governing regulations have also been

presented. With this discussion complete, it is now possible to introduce the actual site

designs in the following chapter. These designs are the "physical" manifestations of the

concepts presented in the preceding discussion.














CHAPTER 3
SITE DEVELOPMENT PLANS

Introduction

This chapter discusses the site design of the four alternatives and the development

of the existing conditions model. Where possible, these alternative land development

plans were designed within governing regulations. However, even in real projects, there

are cases where the design engineer may wish to deviate from the required or generally

accepted design standards. This is usually handled by requesting a variance from the

ordinance. This assumes that the design engineer can provide proper justification for the

design and that public safety is not at risk. The general design discussion for each

alternative describes the various decisions and methods, including clear deviations from

the Alachua County ordinance, which were involved in the site design. Areas where the

County ordinance was vague or subject to interpretation are discussed in a separate

section for each alternative.

Existing Conditions

The existing conditions plan is the basis of all the other plans. Many types of

information were required to conduct even a hypothetical development project. A

detailed site survey was not practical, so existing GIS data were used in lieu of those a

Surveyor would collect. This should not be construed as an endorsement of using GIS

data to replace an actual site survey. The GIS data have errors associated with them and

also may not reflect current site conditions. With the exception of topography, these data

were obtained from the Florida Geographic Data Library (FGDL) (University of Florida









GeoPlan Center, 2000) and are available on CD and World Wide Web download. The

data included shape files for National Wetland Inventory wetlands, 100-year floodplains,

soils and an aerial photograph to determine cover conditions.

Topography was developed from a digital elevation model (DEM) provided by

Camp Blanding. This DEM required re-sampling to a smaller grid size to ensure smooth

contours. The drawback to this approach was that the resulting contours are slightly

inaccurate in areas of steep slopes where they generally underestimate the slope. This is

because the original 30-meter grid interval had insufficient detail to reflect the rapid

change in grade correctly. It was assumed, however, that contours resulting from the

DEM were of sufficient accuracy for a theoretical project.

Another parameter that is important both for development in Florida and for

implementation of LID is the elevation of the water table. Normally this would be

determined through test wells. The only information available was from the Clay County

Soil Survey (Soil Conservation Service [SCS], 1990), which was of questionable

accuracy, and a regression equation developed by for the SJRWMD (Boniol et al., 1993),

which requires user interpolation near surface water features. The water table elevation is

important for two reasons. First, it determines the degree of difficulty in building in a

certain area. While most homes in Florida do not have basements, it may still be

necessary to provide a raised building pad to keep separation between the foundation and

water table. On the other hand, this may be an advisable building technique regardless of

water table depth and is discussed in the various LID manuals (Prince Georges County,

Ch. 3, pp. 15-16, 1997; Ch. 2, pp. 16-18, 1999a). The second reason that water table

needs to be considered is for infiltration management practices. For example,









recommended design parameters for a bioretention area are for a 0.6-1.2 m (2-4 ft) deep

area of prepared soil and a minimum 0.6 m (2 ft) clearance from facility bottom to the

water table (Prince Georges County, Ch. 4, p. 4, 1999a). This research assumed that the

water table was not a limiting factor, although it is acknowledged that certain areas (a

small minority of the "buildable" area) are "danger" zones where infiltration practices

may be unfeasible or extra fill for building pads may be required.

The hypothetical property was considered to be a collection of smaller parcels that

would be grouped together. The total area of the property was 246.29 ha (608.66 ac),

while the existing contributing watershed area (to the reference point located on the

stream where it exits the property to the north-west) was 288.77 ha (713.64 ac). The

allowable density for R1-aa zoning was 1,325 lots. The presence of wetland area, stream

and infrastructure resulted in a density that was much lower. The existing conditions plan

is shown in Figure 3-1 with topography, wood line, soil type, hypothetical property lines,

watershed boundaries, hydrology and existing improvements such as roads.

Traditional Development Plan

Discussion of the traditional design option is divided into three parts. The general

design discussion presents the overall development of the site, including areas where

clear deviations from the governing regulations were made. The next section discusses

the regulations that required interpretation because they were vague or seemed to

contradict other regulations.





































Figure 3-1. Existing conditions plan.

General Design Discussion

The traditional design alternative represents the type of development that would be

seen if the local ordinance was the only guide used. It was assumed that the developer is

taking advantage of two new county roads along the southern and western edges of the

property. The developer is not responsible for construction of these roads or their swale

drainage system except for access improvements (deceleration and acceleration lanes as

required) and the access road running north to south for the eastern most entrance to the

development. It was assumed that the developer had some input on the design of the

swale system on the off-site roads, since these roads would be allowed to drain to the

storm-water management system constructed by the developer on-site.









A prohibition on lots in a water resource buffer area was assumed for this design.

Lots were prohibited in the 100-year floodplain if it was not possible to ensure that the

building pad would be above the floodplain. It was also assumed that streets and storm-

water ponds, and their associated grading, would have minimal impact on the 100-year

flood elevation and, therefore, were permitted in the 100-year floodplain. Lot dimensions,

while acceptable under the local ordinance, are arbitrary. The lot width was set at 37.2 m

(122.1 ft) and the depth at 50.3 m (165.0 ft) for rectangular lots, resulting in lot area of

1,871.16 m2 (20,142.96 SF). Lots on curvilinear street sections need not meet the width

requirements, but must meet the minimum area requirement.

The standard design does, inadvertently, provide a large area of open space. This

consists of the wetland area around the northern stream and a 10.67 m (35 ft) water

resource buffer around the streams and wetlands. Other areas that were unbuildable due

to lack of space for a lot, storm-water management, site buffer etc. were also considered

open space. The site design did not place any other emphasis on protecting natural

resources above that required by ordinance.

The road layout paid no particular attention to topography to reduce grading and

followed a general grid pattern. The road ROW and lane widths are dependent on the

projected average daily trips (ADT). For the most part, ROW width is 15.25 m (50 ft) and

lane width is 3.1 or 3.4 m (10 or 11 ft). Generally a traffic study would be conducted to

determine the ADT, but for the purposes of this study, the ADT was determined from the

number of homes on a particular street using trip data (10 trips per home) from Site

Planningfor Urban Stream Protection (Schueler, p. 133, 1995). When traffic from two

or more streets intersected, their ADT was combined for determination of the service









level of the collector street. Sidewalks are only required on roads with a C or D level of

service. This design assumes however that sidewalks are provided on all internal roads.

This design plan has a generic "house" footprint associated with it. This footprint,

which includes all impervious area including driveways, was used as a guide to lot

dimensions. The footprint was computed at approximately 388 m2 (4,176.8 SF). All

homes are assumed single-story. The homes are only relevant to the hydrologic analysis.

The economic analysis assumes the lots are sold vacant, thus removing the impact of the

home on lot sale price.

The conventional storm-water design option uses curb and gutter, adding 0.6 m to

the width of the road, for collection of runoff and the storm sewer transport system. The

end-of-pipe management practice is a wet detention pond, a rather common practice in

Florida. The eight ponds are situated at natural low points throughout the site and were

designed utilizing the example procedure outlined by the SJRWMD (Ch. 29, pp. 1-6,

2001). Pond computations are based upon the contributing area to that particular pond. It

is clear that many of these ponds require baffles to prevent "short circuiting" of runoff

due to the shape of the ponds and the location of discharge points from the storm sewers.

There are several small, uncontrolled areas of the site that are developed. These

areas are permissible provided the developer provides compensating storm-water

treatment. It was assumed that since county roads are allowed to drain into the

development to the wet ponds, this provided the required compensating treatment. The

storm-water system is not designed hydraulically. The placement interval of the storm-

water inlets is arbitrary. The storm sewer pipe was assumed to be 61.0 cm (24 in)

diameter, reinforced concrete. In reality, pipe width would vary with a 38.1 cm (15 in)









minimum, but this generic diameter was determined to be a reasonable average for the

development. Overall slope of a pipe series was assumed as the elevation of the outlet

subtracted from the elevation of the upstream inlet (minus 1.2 m or 4 ft for inlet depth).

Florida Department of Transportation guidelines for the period of 2000-2001 required a

minimum of 15.25 cm (6 in) of cover between the top of storm sewer and the base of the

road. Thus, the 1.2 m depth in the upstream inlet is sufficient.

Site grading, as was previously mentioned, was rough and for runoff flow purposes

only. In other words, road profiles were not smoothed. However, it was assumed that the

entire lot was cleared and graded (or compacted if no re-grading is required) from

construction vehicle traffic.

Interpretation Issues With The County Ordinance

There were four areas where interpretation of the County ordinance was an issue.

The issue of location of the required buffers at the subdivision boundary was discussed in

the section on the County ordinance. This site design assumed buffers separate from the

platted lots and, thus, considered open space. This approach has been used in

subdivisions in the Gainesville area (for example, Haile Plantation).

Another issue was required maintenance access to a storm-water pond. The final

decision allowed for a 3.67 m (12 ft) access way outside of the platted lots and a strip

around the basin of the same width that has a slope no greater than 8:1. The next issue of

concern dealt with landscaping. This is more important in terms of economics than

hydrology, but it is appropriate to identify it in this section. The landscaping ordinance

was unclear as to whether the 10% landscaping requirement applied to individual lots. It

was assumed that this was the case. Thus, 10% of each lot was assumed to be landscaped.









Swale geometry was another item where deviation from the ordinance was

possible. The SJRWMD allows for swale side slopes to be a maximum of 3:1 with a top

width to depth ratio at least 6:1 (SJRWMD, Ch. 15, p. 2, 2001). Alachua County states

that swale side slopes are "typically" no greater than 4:1 on the side facing the road

(ACBCC, 2002). In this research the SJRWMD geometry standard was applied. The

swale design is parabolic, with a 3.67 m (12 ft) top width and depth of 60.1 cm (2 ft). The

flow depth on which time of concentration calculations are based assumes full flow at

50.9 cm (1.67 ft), resulting in an approximate top width of 3.1 m (10 ft). Swales on the

off-site roads are assumed to have a conservative design requiring a concrete invert.

Swales with this invert were considered to function the same as pipe (tributary

impervious area is directly connected). This contradicts SJRWMD swale design

standards. In this design, however, the swales are not used for runoff treatment or control.

The last item of interest is the access road for the eastern most entrance to the

development. This road is outside the watershed of interest and is only considered in

economic calculation.

Design Plans

The traditional development plan is shown in Figure 3-2, and includes topographic

features for the site including proposed, rough grading for runoff flow and storm-water

management wet ponds. In the pond grading, a contour of a slightly different color may

be observed. This represents the approximate elevation of the permanent pool.





































Figure 3-2. The traditional development plan.

The traditional development plan resulted in 719 lots, in contrast to the allowable

density of 1,325 lots. This represents a 46% reduction and was the result of site

infrastructure (roads, storm-water management etc.) and areas that are unbuildable, such

as the wetland and associated buffer area. These areas are counted in the allowable

density computation.

Figure 3-3 shows the traditional development plan without topographic features or

drainage basins. The lot layout is clearly visible on this plan. While this plan was deemed

consistent with County design regulations, it was accepted that somewhere in the plan the

County engineers might require changes.






































Figure 3-3. The traditional development plan without topography, soils or drainage sub-
areas.

Cluster Development Plan

The second design alternative was cluster development. Much of the design

discussion on the traditional development option is also applicable to the cluster design.

For example, road lane and ROW widths were determined by the same procedure, and

sidewalks are assumed present on all internal roads. This section focuses on differences

between the cluster design and the traditional plan. Also, just like the discussion on the

traditional development, the areas where interpretation of the ordinance was required are

discussed.









General Design Discussion

The most obvious difference between cluster and the traditional options were the

lot dimensions. These were set at a width of 27.9 m (91.5 ft) and 33.5 m (109.9 ft) for

depth. These dimensions result in a standard rectangular lot area of 934.65 m2 (10,061.47

SF). Again, these dimensions are arbitrary but accommodate the generic house footprint.

The footprint, however, is somewhat smaller, 367 m2 (3,950.75 SF), due to lower amount

of pavement required for the driveway.

The water resource buffer was also increased from its minimum size. The Alachua

County ordinance specifies a 22.9 m (75 ft) buffer for Outstanding Florida Waters

(ACBCC, 2002). Based on this regulation and information on forested buffers in two

design manuals (Schueler, p. 86, 1995; CH2MHill, Sec. 8, 1998), a width of 30.5 m (100

ft) was selected even though streams on the site are not specifically listed as Outstanding

Florida Waters.

One potentially important aspect of the road design was changed in the cluster

design option. The road layout for the cluster options mostly dispensed with the grid

layout. Instead, curvilinear loop streets were used, and cul-de-sacs are more frequent;

especially as short branches off loop streets.

The cluster design option also required some thought on appropriation of open

space, which the traditional design did not. Appropriation of open space was based on

several goals. Natural areas, including the wetland and all water buffers, were required to

be considered open space. Beyond these areas, the goal set for this design option was to

preserve areas of natural runoff concentration (i.e., natural swales) to the greatest extent

possible. Generally, this was most successful in the upper areas of the watershed where,

in some cases, virtually the entire time of concentration path was preserved.









The most important goal considered in open space preservation, as far as storm-

water management was concerned, was preservation of upland areas of high infiltration

potential. These areas are roughly defined in Figure 3-4. This description was based on

the Clay County Soil Survey (SCS, 1990) and includes only areas ofNRCS type A soils.

Because of the general nature of soil surveys, boundaries of high infiltration soils could

not be determined with any certainty. Similarly, the GIS data that were used for site

features also contained a certain degree of error. Naturally, this results in errors in Figure

3-4, such as areas of high infiltration within the boundaries of the wetland area identified

by the National Inventory of Wetlands shape file. Still, this figure was useful for planning

purposes. The location targeted to be the center of protection is the ridge between the two

streams.

There are 720 platted lots in this site design. This assumes that the developer will

not take advantage of the fact that the unbuildable areas can be considered open space,

thus allowing an increased number of lots, albeit at reduced size. This assumption,

although potentially positive from a hydrologic perspective, led to another potential

problem. It is common knowledge in the land development business that area often

represents most of the value of a lot and, thus, the sale price. Therefore, reducing lot size

may have a negative economic impact to the developer, and the possibility of the

developer adding more units to the development to compensate for this loss of revenue is

cannot be ignored. On the other hand, proximity to open space may have a positive

impact on lot value. These are discussed in the Chapter 5. For now, it is sufficient to state

only that the cluster site layout is designed to allow for "islands" of open space between

loop streets, thus preventing lots from abutting one another along the rear lot line.





































Figure 3-4. Areas of high infiltration potential.

Distance from the rear lot lines to the subdivision boundary was increased from the

traditional design. The distance used in this option was 4.6 m plus the difference between

lot depth from the traditional and cluster options. These two design features detracted

somewhat from the contiguous area of open space normally resulting from a cluster

development. Effort was made to design the site so islands of open space are as large as

possible to avoid the fragmentation problems discussed in the literature cited on cluster

developments.

The storm-water management system is designed in the same fashion as the

traditional development. Sub-areas that were totally preserved were not included in any

pond computations. There are nine ponds in the cluster design option. Pond computations









for water quality assume a depth of runoff over the entire contributing area. If a portion

of this contributing area was preserved in it's natural state, then the pond would over-

treat runoff

Interpretation Issues With The County Ordinance

Land-use buffers that must exist along subdivision boundaries have an additional

impact on the cluster design option. If a lot of reduced size abuts a neighboring single-

family or agricultural parcel, screening and the more restrictive building setback of the

existing zoning district must be provided. How this requirement relates to the 4.6 m (15

ft) buffer required by the base R1-aa zoning district is unclear. Furthermore, use of the

word "abut" seems to imply that one of the lot lines of the reduced-size lot would be

coincident (the same) as the subdivision boundary. Therefore, it was assumed that

because a buffer was provided outside of the platted lots, the more restrictive building

setbacks and screening would not be required.

Design Plans

Figure 3-5 presents the cluster design option. Figure 3-5 includes topographic

features for the site including proposed, rough grading for runoff flow and storm-water

management wet ponds. In the pond, grading a contour of a slightly different color may

be observed and represents the approximate elevation of the permanent pool.





































Figure 3-5. The cluster development plan.

Figure 3-6 shows the cluster development plan without topographic features or the

drainage basins. Lot layout is clearly visible on this plan. While this plan was deemed

consistent with County design regulations, it was accepted that somewhere in the plan the

County engineers might require changes.






































Figure 3-6. The cluster development plan without topography, soils or drainage sub-
areas.

LID Developments

The final two development design alternatives incorporate concepts of Low Impact

Development (LID). The literature review on LID pointed out that LID seeks to replicate

site hydrology through a two-tiered approach. There is a runoff prevention aspect that

focuses on limiting disturbance on the site being developed, thus leaving a portion of the

site in its natural condition. An infiltration-based, distributed storm-water management

system is used in areas that are disturbed for development. Combination of practices,

such as a primarily LID system that requires one or two ponds to meet local peak flow

regulations is possible where storm-water management requirements cannot be met by









the infiltration-based storm-water management system. This would particularly be true

where the surface area of infiltration practices required is so large that it impacts the

ability to develop the property.

General Design Discussion

The cluster design option is consistent with LID land preservation goals. Thus, it

was decided to use the cluster layout for the LID design and add the infiltration-based

storm-water management system. Examples of LID design included in the Prince

Georges County design manuals show LID site preservation as keeping all development

activity out of high infiltration, hydrologic group A, soils as much as possible while

providing additional preservation on the platted lots that are on the moderate infiltration

soils (types B and C). Development would also take place on type D soils as long as these

are not wetland areas. In Florida, soil infiltration potential is partly determined by the

high water table. This leads to many dual classifications. For example, soil types 19, 29

and 42, on the Clay County Soil Survey (SCS, 1990), are Osier fine sands and are

classified as type D or B, with the B classification applying when the soil is drained.

Furthermore, type B and C soils are less frequent in Florida. They are mainly transitional

areas and most upland areas are high infiltration type A soils. The latter comprise the

majority of the buildable area on the study site. This is both positive and negative. These

areas are ideally suited for infiltration practices but are also likely to see the greatest

hydrologic change from development. Development in these areas cannot be avoided, but

impact should be minimized. It was decided that on-lot preservation of land was not an

ideal solution since there were questions as to whether these areas would truly not be

impacted and of the ability to enforce, either by the County or the Homeowners

Association, the necessary deed restrictions adequately. Thus the decision to forgo any









on-lot preservation and to use the reduced lot size of the cluster development was

justified, even though use of cluster development is not mentioned in LID literature.

The LID storm-water management system was also applied to the traditional

design, although this is not "true" LID because of the lack of land preservation that is

assumed (all lot areas are considered cleared, graded and/or compacted). Thus, this

design is referred too as "partial" LID because only infiltration-based storm-water

management system is present. There are no explicit requirements in any of the LID

literature on how much of the site should be preserved.

Since the site design for the LID options is virtually the same as the traditional and

cluster options, much of the previous discussions remain valid, except for storm-water

management. There are, however, some aspects of the LID designs that need to be

discussed in addition to the design aspects previously discussed on the traditional and

cluster designs.

Sidewalks are generally dropped from the development. For most interior streets,

this still remains consistent with the County codes. However, sidewalks are required to

both sides of type C-l Streets. Therefore, total elimination of sidewalk is not consistent

with the County ordinance on four streets, although it affects only a small proportion of

the total street length. A compromise was selected that eliminated sidewalks from all C-l

streets that crossed no more than 10 lots on a side. This eliminated all sidewalks except

on the street serving the central upland area in the "partial" LID design, where sidewalks

and curbs were used on both sides.

Another area where the LID designs differ from the County ordinance regards road

right-of-way. Use of swales for drainage requires an increase in ROW. For example, a









road requiring 15.25 m (50 ft), which represents all but four interior streets on both the

LID site plans, with conventional drainage would now require an 18.3 m (60 ft) ROW.

The four streets, all entry roads, requiring 18.3 m of ROW would now have a 24.4 m (80

ft) ROW (ACBCC, 2002). The reasons for this additional area are not stated in the

ordinance, but some possibilities are to allow room for sidewalks, and anticipation of

wide swales to meet infiltration requirements or safety.

The 1997 LID design manual for Prince Georges County discusses modification of

their rural road section for use in LID developments. Their standard interior and primary

road have the same ROW requirement as are required in this project. Several sections are

proposed for use in Prince Georges County. ROW widths for these proposed sections are

16.5 m (54 ft) for most internal roads and 18.3 m (60 ft) for higher traffic roads. These

sections use swales for drainage (Prince Georges County, App. D, pp. 15-20, 1997). The

variable lane width allowed in Alachua County, based on level of service, allows for a

fairly narrow road section, similar to the proposed Prince Georges County roads, and was

not altered. The only road design issue was the setting of proposed ROW width and

selecting a swale design. The swale cross-section section selected was the same as the

section described for the off-site County roads (discussed in the traditional development

section). In terms of capacity, this section is probably more than is required for most

interior roads. There are times where flow may be transmitted under a road, and where

this is anticipated, a 60.1 cm (24 in) diameter pipe is assumed. A pipe diameter of 45.7

cm (18 in) is acceptable in an open drainage system, but the 60.1 cm diameter was

selected as a conservative figure. All pipes are reinforced concrete, even under driveway

crossings. The proposed swale section has a top width of 3.67 m (12 ft) and a design full









flow width of 3.1 m (10 ft). The swales on both the interior streets and the off-site roads

were considered completely pervious (no concrete invert). However, the design still did

not strictly follow the SJRWMD infiltration standards because check dams were modeled

to allow for depression storage.

With a 16.5 m (54 ft) ROW, identical to the Prince Georges County standard, the

swale can be accommodated in the ROW of even the smallest interior street. However, it

was decided that the ROW width for most interior streets (type A or B) should remain at

15.25 m (50 ft). Thus, the design maximum flow depth is accommodated in the ROW,

but an additional 60.1 cm (2 ft) of top width is provided on the lot itself. This was

deemed acceptable since an easement might be required on the lot for public utilities.

Type C-l streets will have the same swale section and an 18.3 m (60 ft) ROW and are

similar to the Prince Georges County proposed section for a higher service level LID

street section. A level spreading area or device is intended to be present at the terminal

point of the swales to convert runoff to overland flow.

The final road section provides a travel lane of 3.1-3.4 m (10-11 ft) on most interior

streets with a 1.2-1.5 m (4-5 ft) grass shoulder and 3.1 m (10 ft) of swale at a depth of

60.1 cm (2 ft). On the few higher traffic interior roads at the entrances, the travel lane is

3.67 m (12 ft) and the shoulder is expanded to 2.4 m (8 ft).

Interpretations of The Ordinance

The previous discussions on ROW and swale section are clear deviations from the

County ordinance. The LID storm-water management system design, however, is not

necessarily inconsistent with the storm-water management regulations that would be

enforced on this site. On the other hand, how to apply the current standards defied clear

interpretation. In particular, the issue is computation of water quality treatment volume,









which represents the minimum storage of the system and the starting point for the

hydrologic analysis.

LID assumes a distributed, infiltration-based storm-water management system for

runoff quantity control and quality control. Swale design, at least in theory, in Alachua

County requires infiltration of 80% runoff for a 3-year, 1-hour rain event per both County

ordinance and SJRWMD guidelines. However, Livingston (2000) cited a study that

determined that the swale design guidelines are often not attainable due to space

concerns. In other words, a swale width is required that is so great as to preclude use.

Thus, it was assumed that swales in this design would use some form of check dam or

barrier to encourage ponding for infiltration. The SJRWMD, whose design guidelines

were used for swale cross-section, states that an open conveyance system has no control

structure before discharge to receiving waters (SJRWMD, Ch. 15, p. 1, 2001). Thus, use

of check dams in the swales would mean that the swales are now considered another type

of storm-water management system. They are considered a retention practice. The only

question is whether they are on or off-line detention. They are considered on-line systems

(SJRWMD, Ch. 11, p. 1, 2001). These swales, while providing depression storage, were

not considered bioretention areas.

Bioretention areas appear to also be classified as retention practices (SJRWMD,

Ch. 11, p. 1, 2001). Again, the question arose as to whether they were on or off-line

treatment. Arguably a bioretention unit could be an off-line system since it is not directly

connected to the transport network of the sub-basin and generally runoff would enter as

sheet flow, although SJRWMD may consider them on-line systems.









The storm-water management system for the LID design options was considered a

mix of on-line and off-line retention practices. No clear standard exists for computing

treatment volume for this type of system. Instead of collecting runoff from an entire sub-

area, LID collects runoff from a single lot. This leads to the question as to what is the

appropriate area for the treatment volume calculation. Pond systems are designed to

provide treatment for the entire contributing area, including undisturbed areas, at a single

point. The on-lot nature of LID may make control of undisturbed, preserved areas

impractical because of the space required for the bioretention areas. The method that was

adopted for this research was to compute treatment volume based on the disturbed area.

The actual volume of treatment computed assumed an off-line system and used the 12.7

mm (0.5 in) of runoff over the treated area standard. It was expected that the actual

storage required, due to peak flow control, would be higher, thus providing a level of

water quality treatment exceeding the computed volume, though not necessarily in strict

compliance with SJRWMD rules.

Design Plans

Figures 3-7 and Figure 3-8 show the "partial" LID, based on the 1,871.16 m2 lot,

and the "full" LID design, based on the 934.65 m2 lot, respectively. Topography is shown

on these plans, but any detention basins required due to the results of the hydrologic

analysis are not shown. Also, the bioretention areas are not shown. The pipe used for

driveway crossings are not shown, but any anticipated road crossings are.






































Figure 3-7. "Partial" LID development plan.





































Figure 3-8. "Full" LID development plan.

Figures 3-9 and 3-10 show the LID development plan without topographic features

or the drainage basins. The lot layout is clearly visible on these plans. The two LID plans

represent a new concept in land development. As a result, it cannot be conclusively stated

that these plans would be acceptable for approval by the County Engineer.

Summary

This chapter introduced the actual development plans used for this research. One

aspect that hasn't been discussed is whether these plans truly represent a possible site

plan. Site layout is as much an art as a science. Therefore, it is quite reasonable to expect

that another site planner may have a different interpretation of site plans. Each of the four

development plans represents but one "realization" among an infinite number of










variations. However, they were established in a manor allowing the economic and

hydrologic impacts to be explored in a systematic process.


Figure 3-9. "Partial" LID development plan without topography, soils or drainage sub-
areas.











































Figure 3-10. "Full" LID development plan without topography, soils or drainage sub-
areas.














CHAPTER 4
HYDROLOGIC IMPACTS

Introduction

This chapter covers the hydrologic analysis for four, hypothetical, alternatives for

residential land development and storm-water management. Details of these alternatives

were presented in the last chapter.

Each alternative was compared with the hydrologic response of the existing

conditions of the watershed. Three rainfall scenarios were modeled. The first two were

the 2 and 25-year, 24-hour hypothetical storms that are the design storm events for the

SJRWMD. This research also modeled the proposed alternatives under continuous

simulation based on rainfall collected on the site during the summer of 2001.

Discussion on Modeling the Alternatives

Hydrologic modeling for land development is governed by guidelines set forth by

the regulating agency. Perhaps the most widely used techniques are based on the NRCS

curve number method to determine runoff volume. Hypothetical hydrographs are

generated through a variety of synthetic unit hydrograph methods. A common method for

the complete modeling process is detailed in TR-55 (USDA, 1986).

Two land development alternatives are considered: full-size lots and reduced size

lots typical of those a cluster development. These are paired with two alternatives for

storm-water management: conventional wet ponds and a distributed, infiltration-based

storm-water management system proposed by Prince Georges County, Maryland, in their

LID publications. This second type of system requires a different approach to modeling.









An excellent discussion of the modeling concerns for LID is contained in Wright et al.

(2000). Two interesting points, of several, from this paper are the need for storage based

modeling and continuous models to capture the impact on small, frequent, short duration

storms.

Prince Georges County has presented a NRCS curve number procedure for analysis

of LID. A fundamental assumption of this procedure is that the time of concentration

remains the same as existing conditions. The pre and post-development curve numbers

are used on two sets of nomographs, one for a retention system and one for detention.

The nomographs, which vary by storm depth, provide the required storage distributed

across the watershed (Prince Georges County, App. C, 1999b).

Continuous Simulation Data Collection

The continuous simulation required actual stream flow and rainfall data. Rainfall

data were collected in 10-minute increments from 19 June 2001 until 18 October 2001.

This time corresponds with most of the North-Central Florida rainy season. Stream stage

was also monitored in 10-minute increments. A pressure transducer (In-Situ

TROLL4000) was placed immediately upstream from a large corrugated metal pipe-arch

culvert and data were collected at 10-minute increments. This location became the

reference point for discharges from the site. Cross sections were taken at the level

recorder and at two places downstream of the culvert with a laser level. The stream stage

data at the culvert was translated into stream flow data using the HY8 program. This

program computed stream stage, upstream of a culvert, for a given flow rate. The method

used by this program is an energy balance approach presented in Hydraulic Design of

Highway Culverts (Federal Highway Administration [FHWA], 1985). This approach

allowed a rating curve to be developed. Interpolation between points on this rating curve









resulted in flows for each stage measurement. The rating curve was developed up to a

flow of 2.83 cms (100 cfs). No statistical operations were performed on the rainfall or

stage data.

The collection process for rainfall data resulted in gaps in the data due to periods

where the gage could not be accessed (the site was located on a National Guard post).

The stream stage data did not suffer from data gaps because of higher memory capacity

in the gages. The continuous stream stage record indicated that no rain of consequence

was missed when rain gage memory was full.

Hydrologic Model

The hydrologic model selected for this research was HEC-HMS downloaded from

the Army Corps of Engineers (USACE) Hydrologic Engineering Center. This model

offers the user a variety of choices for precipitation input, runoff volume computation and

transformation. This research used the soil moisture accounting (SMA) module for runoff

volume computation. The soil moisture accounting method is a continuous model. It can

be run as a lumped model or as a distributed, grid-based model. In this research, the

lumped parameter method was utilized mainly because of an error in the distributed

model code that was not fixed until early 2003. The SMA method is well documented in

the Technical Reference Manualfor HEC-HMS (USACE, pp. 44-50, 2000) and a thesis

on the development of the model was also available (Bennett, 1998). Additional updates

are discussed in release notes, although documentation in these cases is often inadequate.

The SMA approach treats the watershed as a series of vertical reservoirs, starting

with canopy storage overflowing to depression storage, soil profile, and two groundwater

layers. For a rain event to produce surface runoff, depression storage must be full. For

water to enter the soil profile through infiltration, canopy storage must first be full. Thus,









a logical progression for generating surface runoff is followed: interception losses take

place in the canopy, then infiltration losses and finally depression storage losses.

Depression storage only fills when the rate of rainfall exceeds the infiltration rate.

Similarly, surface runoff only occurs after depression storage is filled. Groundwater

layers can be used to simulate fast interflow and slow base flow. Groundwater outflow

volume is routed using a linear reservoir technique, while surface volume uses a unit

hydrograph method. In this research, the first groundwater layer is used to represent

interflow, and the second represents the base flow. The base flow portion of the model

may represent a more regional phenomenon. Antecedent moisture conditions at the start

of the simulation can be modeled by defining a percent of volume full (USACE, pp. 44-

50, 2000).

The version of HEC-HMS used in this research contained an important

modification to the SMA that was added after initial release of the program. This update

was the ability to define a percent impervious for the sub-area. Rainfall on this designated

percent impervious is totally converted to surface runoff Thus, if half the sub-area is

impervious and 25.4 mm of rain falls, 12.7 mm of surface runoff is immediately

generated. This modeling behavior is true only for directly connected impervious

surfaces, where the transport system does not allow for significant loss, such as concrete

pipe. For indirectly connected impervious surfaces, such as roofs and roads draining to

swales that allow for infiltration and depression storage loss, this method is not realistic.

One final note on the model is important. At junction points, where stream reaches

or sub-areas are connected, hydrographs are summed. At these points, base flow is not

reported separately. Instead, a composite hydrograph is reported.









Experimental Methods

Design alternatives were modeled under three conditions. In addition to the

continuous simulation, which was run from 19 June until 17 October, the two design

storms required by the SJRWMD for storm-water design were used.

The design storms had depths of 114.3 and 203.2 mm (4.5 and 8.0 in) for the 2 and

25-year, 24-hour, events respectively. These were determined from frequency-depth

maps available in Urban Hydrologyfor Small Watersheds TR-55: 2nd Edition (USDA,

1986). The rainfall distribution for this area of Florida is Type II.

Most storms in the period of rainfall collection were typical convection events.

August was a particularly dry month; however, significant rainfall did occur about mid

month. September was an unusual month that began with a series of particularly intense

convection events with high peak flows. At mid month, a tropical storm hit the area with

sustained rains. The intensity from this event was rather low, but overall precipitation

depth was high.

Existing Conditions Calibration

The first step in the analysis was to calibrate the model against the data collected in

the field. Initial values for the various SMA parameters were estimated using guidelines

provided by Bennett (1998). Even though the SMA model represents a physical process,

the SMA parameters, such as soil storage, might not necessarily reflect actual field

conditions. The exception to this was meteorological parameter ET, which was obtained

from long-term National Oceanographic and Atmospheric Administration (NOAA)

records (NOAA, p. 26, 1982). Initial stream path and watershed boundaries were

generated in HECGeo-HMS.









The field data showed a hydrologic response after most thunderstorms. Initial

calibration attempts indicated that this response could not be adequately modeled as fast

groundwater flow (interflow). On the other hand, it was inconsistent with the character of

the watershed to allow for surface runoff from small storms from upland areas that have

soils with very low runoff potential. Instead, an unconventional sub-area definition was

used. The National Inventory of Wetlands boundary on the smaller (northern tributary)

was selected as a sub-area boundary. Another boundary, roughly following the line of

steeper slopes, was selected on the main channel. Figure 4-1 shows these boundaries.










i' I. ,










Figure 4-1. Existing conditions sub-area division.

These lowland areas were calibrated to allow for surface runoff from most storms,

while the upland areas only produce runoff from the large, tropical event in mid

September 2001. The additional divisions in sub-areas at mid-channel or at the upstream

end of the channel were added to provide load points for wet pond outlets.









Times of concentration (T,) were computed using the procedure and equations

outlined in Urban Hydrologyfor Small Watersheds. TR-55 2nd Edition 5. This method

was selected because it allows for division of the To path into separate elements for sheet

flow, overland flow and channel flow. The stream channel used an estimated cross

section in the To computations. All channel portions of the To are computed assuming a

velocity at full capacity using Manning's equation. This assumption is made because the

flow along a particular channel is not known and, therefore, a detailed hydraulic analysis

cannot be done. The Manning roughness coefficient that is used for the swales assumed

that a level of vegetation was maintained to allow for the appropriate roughness. Off-site

swales along the county roads were modeled with a value of 0.045 for the conventional

storm-water management designs. This value reflects the presence of a concrete invert on

a portion of the off-site swales. All swales were modeled with a value of 0.11 for the LID

storm-water management designs. The roughness coefficient is also assumed to include

the influence of check dams and transitions for small culverts. Roughness used for

uncontrolled analysis, where no depression storage was present, is the same as for

controlled analysis. Thus, the uncontrolled analysis would imply a greater level of

vegetative resistance.

Various synthetic unit hydrograph methods are available in HEC-HMS including

the NRCS dimensionless, Snyder and Clark. Testing of various methods demonstrated

that the Clark Unit Hydrograph method allowed for the best simulation of volume

distribution under the surface runoff hydrographs observed on the site. The Clark method

utilizes a time-area curve: a form of unit hydrograph. A generic formula for the time-area

curve is used in HEC-HMS. The time-area curve is then routed through a linear reservoir









with the storage coefficient R. A linear reservoir relates storage at any time t to the

outflow at time t multiplied by the R coefficient (Viessman and Lewis, pp. 219-221,

1996).

The time-area histogram is dependent on the To (USACE, pp. 60-63, 2000) and will

remain the same if the To value is constant. The R coefficient, however, varies based on

watershed cover and development conditions and can be expressed as a multiple of To

with a range of 8-10 *To for forested areas in the Vancouver, Canada area (Russell et al.,

1979). A value of 8 for upland areas and 5 for lowland areas were the best fits for the

calibration. The ET coefficient was generally held at 0.3; however, it was adjusted as

needed to fine-tune the calibration.

Routing along the existing stream reaches was done using the Muskingum method.

This is a simple form of hydrologic channel routing. Input for this model is a storage time

K and a weighting factor X. The X parameter generally averages about 0.2 and K is

reasonably close to the travel time of the flood wave (Viessman and Lewis, pp. 235-236,

1996). The values of X used were 0.2 except for the channel after the stream confluence

where a value of 0.25 was used. The K values were determined during the calibration

process. Values of 0.25 hours were determined for the main channel segments while a

value of 0.34 was determined for the tributary segment. The routing coefficients are

acceptable for flows within the range observed in the continuous model. Flows that

significantly exceed these, such as the 25-year design storm, may have greater attenuation

effects due to out of bank storage.

Antecedent moisture condition was represented by designating a percent of storage

full at the beginning of the simulation for the soil and groundwater portions of the SMA










storage. These were calibrated for the continuous model. For hypothetical design storms

the calibrated values did not appear to reflect the assumption of "average" conditions

accurately. This could be seen in the interflow groundwater layer where the positive

outflow implied that a rain event response was still underway. This is not particularly

surprising since a storm system had impacted the area before it was instrumented. The

values used for the design storms were determined through observation of the model

storage in late August approximately one week after the single August rain event. Input

data points for the existing conditions model are contained in Appendix A.

The calibrated results are shown in Figure 4-2. The volume error between the

observed and calibrated data is 13.7 mm (0.54 in). The volume for the calibrated model is

below the observed data. The focus of the calibration was on peak flows and volume. The

time of peak flow tended to be earlier than was observed in the field data. This could be

remedied by adjusting the sub-area To and would have required adjusting the results of

the To computations. This was unacceptable since the To results for the predictive models

were computed and consistency between methods was desired.


25 0
2 5 --- -- ---- -- ---- -- -
-0.5

E 2
2-2

S01 25 .E

35
0 4
0 28 56 84 112
Days

Field --Model- Rainfall


Figure 4-2. Comparison of observed flows versus calibrated flows.









Design Alternative Models

Models were constructed for each design alternative for comparison against the

existing conditions model. Several issues that arose during the development of these

models require discussion. Some were unique to either the conventional storm-water

management system or the LID system, while others are universal across all four models.

Naturally, the existing conditions SMA units form the basis of the sub-areas in the

developed conditions. Model input parameters were altered for development, with the

exception of the ET parameters, from the existing conditions parameters.

Universal model assumptions

The Clark unit hydrograph method presented difficulties with the predictive

models. There is no easy means of determining the future storage coefficient, although it

is clear that it will change with development. It was assumed that the future storage

coefficient would reflect the composite nature of the watershed. A weighted average was

used based on natural cover, which used the existing multiple (i.e., 8), and the developed

area, which used a multiple of 1.5 from Russell et al. (1979). The composite multiple was

then multiplied by the To to determine the storage coefficient. The Clark method results

in a somewhat different situation, when modeling LID depression storage required, than

the design procedure used in the Prince Georges County manuals. In the manuals, the

nomographs for determining depression storage are based on the assumption that the sub-

area To remains the same, thus assuming, depending on the synthetic unit hydrograph

method, that the transformation from volume to a hydrograph has no impact on the

problem. To may remain the same in the Clark method, but the storage coefficient does

not, which can result in an increase in peak flow.









Indirectly connected impervious surfaces also presented a significant challenge.

HEC-HMS applies the soil moisture unit, for a particular sub-area, only to the portion of

the sub-area that is "pervious". Rainfall on the percent of the sub-area designated

impervious is considered surface runoff While this is valid for impervious areas that are

directly connected to storm sewers that provide no loss, it is not valid for swales with

check dams for depression storage or for roofs that drain onto lawns. The model assumes

no vertical storage for areas under directly connected impervious area; therefore, the

same assumption must be made for indirectly connected impervious areas.

There were only two potential solutions to this problem. The first option was to

reduce watershed area by the percentage of indirectly connected impervious area. Then,

rainfall depth for each time step would be increased by the same percentage. In other

words, the volume of rainfall was the same but was distributed over a smaller area. This

assumed that rain hitting indirectly connected surfaces such as rooftops was instantly and

evenly redistributed on the pervious portions of the sub-area. For example, if a sub-area

has 10% indirectly connected impervious area, the first proposed solution involves

distributing 110% of the rainfall over 90% of the area.

The first solution was, however, impractical since it required a separate rain gage

file for each sub-area. Another potential solution was to adjust the storage volumes in the

SMA unit proportionally. For example, using the same 10% indirectly connected

impervious area, 100% of the rainfall is distributed on 100% of the area, but with 90% of

vertical storage in each layer of the SMA unit. In either case, the volume of storage and

rainfall is the same. This approach was selected for this research, although there are

drawbacks that are discussed in the limitations section.









The final universal modeling assumption dealt with the impact of compaction. Soil

compaction is assumed on all cleared areas. The assumption is that these areas are

impacted due to grading, clearing activity or both. Schueler (2000) discusses the impact

of compaction in terms of changes in bulk density. In addition, a table is presented

showing the relationship between soil porosity and bulk density. Changes in porosity

imply a change in storage; therefore, this research modeled soil compaction effects by

reducing soil profile storage. Comparing bulk density, from the Clay County Soil Survey

(SCS, 1990), of soils on the site with reported increases in bulk density and changes in

porosity from Schueler (2000), a percent decrease in storage of 25% was determined.

This decrease did not apply to open space areas; thus, the actual reduction in storage is

much less.

An example of such storage reductions is appropriate. Consider an upland area with

an existing soil profile depth of 25 mm and 17% indirectly connected impervious area.

Since impervious area has no storage, all SMA unit storage are reduced by 17%. In the

case of the soil profile in this example, the new depth would be (rounded to the nearest

mm) 21 mm. The soil profile also must be reduced for compaction. The compaction

reduction calculation involved only the pervious portion of the sub-area. In this case 30%

of pervious area is cleared and graded. Therefore, 30% of the site sees an additional 25%,

or 5mm, reduction due to soil compaction, while 70% of the site remains with 21 mm of

storage. On a weighted basis, the final storage depth for the sub-area is 19 mm.

Changes in infiltration/percolation rates have not been discussed. These were not

altered for the developed scenarios. Rate reductions for infiltration were simulated

through storage reduction for compaction. Deep percolation was not impacted.









Conventional storm-water management model notes

The conventional modeling procedure required few assumptions other than the

universal assumptions already noted. Ponds used in the traditional and cluster designs

sometimes captured portions of two existing sub-areas with different parameters (i.e., a

portion of a lowland sub-area and a portion of an upland sub-area). These situations had

upland area overwhelming the lowland portion. Still, storage depths and

infiltration/percolation rates were weighted to account for the composite nature of the

contributing sub-area. Starting percentages of storage full, which represented antecedent

moisture conditions, were not adjusted but used the percentage of the dominant (upland)

existing sub-area. Swales, in the conventional designs, are assumed to have a concrete

invert on a portion of their cross section. Such inverts are generally used to ensure

transmission of flow. Roads draining to these swales are considered directly connected.

The HEC-HMS pond routing techniques had two minor problems. First, no sharp-

crested weir option is available for the outlet structure. The second stage outlet would

normally be one, or more, sharp-crested weirs. In order to accommodate this, the weir

coefficient for the broad-crested weirs was adjusted so that the result of the broad-crested

equation would have an effective coefficient of around 0.4, which is typical of a sharp-

crested weir. Second, the reservoir routines route groundwater through the reservoir,

which is not typically appropriate for storm-water management systems. This was easily

fixed by taking the output hydrograph for base flow and surface runoff from each

contributing sub-area (to a storm-water pond), treating them as separate sources (an

HEC-HMS feature that allows a user defined flow input), and allowing base flow to

bypass the pond and directly discharge to the nearest load point. Outflows from pond

outlets are assumed to translate to the nearest load point instantaneously. This is









reasonable assumption, although a few basins had to be located farther from a load point

than others. Figures 4-3 and 4-4 show the sub-area divisions for the traditional and cluster

designs. Also included are pond locations.


Figure 4-3. Traditional design sub-areas and storm-water system.

Complete tables of input parameters are located in Appendix A. The wet ponds are

treated as directly connected impervious area both in terms of runoff computations and

their impact on SMA units, but are not considered in runoff quality treatment

computations for the pond volume design.






























Figure 4-4. Cluster design sub-areas and storm-water management system.

LID storm-water management model notes

The sub-areas for the LID designs are shown in Figures 4-5 and 4-6. Here, the

storm-water system was designed as a distributed, infiltration-based system relying on

depression storage. Runoff is not deliberately redirected to be concentrated at a particular

pond, thus the sub-area breakdown is much closer to the existing breakdown.

The pond assumptions that applied to the traditional and cluster designs, also

applied to the LID designs. These detention ponds, if required, were conceptual only (not

shown on plans) and were added only if the distributed, infiltration-based storm-water

system could not meet regulatory goals. Ponds were first added to those areas, such as

Area 1, where runoff was already concentrated. If this was insufficient, ponds were then

added to areas where runoff was not concentrated at a single point, such as Area 2 or 3. In

this case, the pond was considered representative of a series of structures.






























Figure 4-5. "Partial" LID design sub-areas.


Figure 4-6. "Full" LID design sub-areas.

Bioretention units on lots and storage from swale blocks and check dams were

considered depression storage. Depression storage was simulated by increasing surface

storage in the appropriate SMA unit above the 2 mm that all lawn and open space areas









were assumed to have naturally. Thus, depression storage is inherently assumed to be

distributed evenly across the sub-area when, in fact, it is not necessarily. The volume of

depression storage was first based on an interpretation of runoff quality regulations. This

required that 12.7 mm (0.5 in) of storage be provided for the disturbed area in each sub-

area. Additional increases in depression storage were allowed for storm-water quantity

management purposes. The physical area required for constructed depression storage

assumed a 152.4 mm (6 in) depth in bioretention areas and swales. Table 4-1 illustrates

the results for adjusting SMA unit depression storage for selected sub-areas in the full

LID design. The portion of the depression storage that is provided by bioretention was

considered uncompacted area. The impact of this assumption on soil storage is small.

Swales, because of their presence in the ROW were considered to be compacted area.

Table 4-1. Adjusted depression storage for selected sub-areas of the "full" LID design.
Sub-area Depth (mm)a Volume (m3) Quality Vol. Area per lot (m2)b
1 7.3 2341.7 1829.3 158.4
2 10.0 4207.0 2839.8 169.4
a Not including the naturally occurring 2 mm.
b Area of depression storage management practices per lot based on 152.4 mm (6 in) cell
depth.

The depression storage parameter is not rounded to the nearest millimeter like the

other SMA parameters. This is mainly for the economic analysis and will be discussed in

the next chapter. Input data tables for the developed design models are located in

Appendix A.

Modeling Results

Tables 4-2 and 4-3 present the modeling results for each design storm. These are

also shown graphically in Figures 4-7 through 4-10 for the design storms only.







76


Table 4-2. Uncontrolled model results for hypothetical design storms.
Alternative 2yr 24hr peak Date and time 25yr 24hr peak Date and time
flow (cms) flow (cms)
Existing 3.344 2 June 0120 9.7969 2 June 0130
Traditional 8.577 2 June 0120 23.304 2 June 0130
Cluster 5.611 2 June 0130 15.426 2 June 0130
"Partial" LID 5.732 2 June 0140 16.076 2 June 0140
"Full" LID 3.991 2 June 0140 11.046 2 June 0140

Table 4-3. Controlled model results for hypothetical design storms.
Alternative 2yr 24hr peak Date and time 25yr 24hr peak Date and time
flow (cms) flow (cms)
Existing 3.344 2 June 0120 9.7969 2 June 0130
Traditional 2.791 2 June 0100 9.815 2 June 0310
Cluster 2.858 2 June 0110 8.090 2 June 0340
"Partial" LID 3.247 2 June 0130 10.633 2 June 0120
"Full" LID 3.269 2 June 0120 9.719 2 June 0140


Figure 4-7. 2-year, 24-hour controlled hydrographs.


-- Risting
-"Full' LID
"Partial" UD
Ouster
-- Traditional


0 60 120 180 240
Time Step (10 minutes)

































Figure 4-8. 2-year, 24-hour uncontrolled hydrographs.


Figure 4-9. 25-year, 24-hour controlled hydrographs


10
g
8
7 Existing
i; 6......... "Full' LID
5 Cluster
S4 Traditional
L 3 "Partial" LID
2
1
0
0 60 120 180 240
Time Step (10 minutes)

























Figure 4-10. 25-year, 24-hour uncontrolled hydrographs

Although the continuous simulation does not lend itself to graphical representation

because of the long time period, hydrographs from these simulations are in Appendix B.

Results of this simulation are presented in Table 4-4 for selected peak events.

Base flow results are shown in Table 4-5 for several time periods during the

simulation. These sample points were times where the flow was relatively steady. This

indicated that only the second groundwater layer was contributing flow.

Watershed Timing

The design storm hydrologic models reveal several points of interest. The timing of

the watershed response, due to the development itself, did not appreciably change

regardless of whether a pipe or swale system is used.

Preservation of entire To paths was quite difficult because the number of paths to

protect increases with the number of sub-areas, which makes it virtually impossible to

prevent development impacts on the To path. Preservation of the initial sheet flow

portions of the To path was possible. Admittedly, many sheet flow areas began in off-site

areas left undeveloped.









Table 4-4. Continuous simulation selected peak flow points and total simulation volume.
Peak Date Peak Date Peak Date Peak Date Total
Alternative (cms) & (cms) & (cms) & (cms) & Vol.
time time time time mmb
1 July 18 2 15
Existing 0.389 2210 0.196 July 1.64 Sept. 1.76 Sept. 524.3
1940 1920 0940
1 July 18 2 15
Traditional 0.41 2210 0.238 July 1.67 Sept. 2.75 Sept. 544.1
1930 1920 1000
1 July 18 2 15
Cluster 0.36 2210 0.2 July 1.67 Sept. 2.54 Sept. 535.5
2030 1930 1000
"Partial" 1 July 18 2 15
LID 0.245 2210 0.18 July 1.65 Sept. 2.95 Sept. 520.0
2000 1930 1000
1 July 18 2 15
"Full" LID 0.325 2210 0.182 July 1.65 Sept. 2.75 Sept. 520.6
1940 1930 0930
aAll alternatives are controlled in this simulation.
bExisting watershed area is 2.88773 km2 and developed watershed is 2.89255 km2.

Table 4-5. Results of the continuous simulation for selected base flow points.
Alternative Flow Date & time Flow Date & time Flow Date & time
(cms) (cms) (cms)
Existing 0.119 6 July 1200 0.115 26 Aug. 0.113 17 Oct.
1200 1200
Traditional 0.095 6 July 1200 0.094 26 Aug. 0.089 17 Oct.
1200 1200
Cluster 0.099 6 July 1200 0.097 26 Aug. 0.093 17 Oct.
1200 1200
"Partial" 0.101 6 July 1200 0.100 26 Aug. 0.095 17 Oct.
LID 1200 1200
"Full" LID 0.103 6 July 1200 0.101 26 Aug. 0.097 17 Oct.
1200 1200

The pond storm-water system had a more pronounced impact on timing than

caused by the development itself for the 25-year storm. The peaks were significantly

delayed, which could create unanticipated flooding problems downstream. This delay

was not present in the LID systems. To times for the LID system were sometimes larger

due to slower transport in the swales and the longer flow path.









Runoff Volume

Uncontrolled runoff volume was increased in each development case, although the

magnitude of this increase may be understated due to limitations in the model. The actual

volume increase is always less than 25.4 mm (1 in) over existing conditions for the

uncontrolled design storms.

Preservation of open space had a positive benefit to storm-water management that

could be seen in comparison of traditional and cluster design uncontrolled results.

Preserving open space, by converting to cluster development from traditional

development patterns, resulted in an uncontrolled volume reduction of 3.8 and 4.2 mm

for the 2 and 25-year events respectively. If the same comparison is made between the

"partial" and "full" LID design, the reductions are 2.3 and 3.0 mm for the 2 and 25-year

events respectively.

The LID designs also showed the benefit of disconnecting impervious surfaces

from the transportation network, at least in a qualitative sense. The off-site swales

allowed for infiltration only in the LID designs. Volume reduction was around 6.0 mm

for a 2-year event and 7.5 mm for a 25-year event when comparing the traditional and

"partial" LID designs. Comparing the cluster and "full" LID designs, reductions of 4.8

and 6.2 mm were observed for the 2 and 25-year events.

Therefore, it would be expected that converting from a traditional development,

with full size lots and conventional storm-water management, to a "full" LID design with

a cluster layout and disconnected impervious surfaces, would result in a volume decrease

of 8.5 mm (0.33 in) for a 2-year rain event. The effect on the 25-year event was observed

as 10.5 mm (0.41 in).









The pond storm-water management system provided little volume control, at least

in the sense that surface outflow volume was decreased. This was expected due to the fact

that HEC-HMS did not model infiltration or ET losses from the ponds. The increased

volume was merely delayed in the slow release, water quality treatment storage of the

pond. The outflow at the end of the simulation from the conventional systems used in

both the traditional land development design and cluster development design are

approximately 0.1 cms (3.5 cfs) greater than existing base flow for both design storms.

The LID systems both returned to sub-surface base flows quite close to the existing

conditions. The volume of water infiltrated by the LID system is dependent on extent of

the area controlled. The LID storm-water management system reduced outflow volume

over the simulation period by 10 mm (0.39 in) over uncontrolled conditions for the

"partial" LID design and by 5 mm (0.2 in) for the "full" LID design.

An interesting observation was made between the uncontrolled "partial" LID

design, where the storm-water transport system was swale-based, and the uncontrolled

cluster design. Volume distributions under the hydrographs were nearly identical. This

indicated that the impact on the Clark storage coefficient from large lots was, possibly,

offset by the swale-based transport, which has a longer T,. The uncontrolled traditional

design had a storage distribution characteristic of many developed watersheds, while the

uncontrolled "full" LID design, combining the grass swale system with open space

preservation, had a distribution closer to that of the existing watershed. The response

from the controlled LID simulations was similar to the existing response.

Peak Flows

Both increased volume discharge and changes in volume distribution increased

uncontrolled peak flows for all development designs. The uncontrolled peak flows for the









cluster and "partial" LID designs are similar because the grass swales may help offset the

volume increase and distribution changes caused by widespread clearing and grading. As

might be expected from the previous discussion, "full" LID results are the closest to the

existing conditions.

The pond system adequately controls peak flows for the 25-year event. The 2-year

event peaks are somewhat lower than the existing 2-year peak. Runoff entering the pond

fills the volume between the circular orifice and second stage weir. This volume is the

water quality treatment storage. This storage volume is released slowly through the

circular orifice to ensure acceptable treatment of runoff The LID designs required more

depression storage in all sub-areas than was calculated for water quality treatment. Table

4-6 shows the final average volumes of management practices required. The depression

storage values are a total site volume and an average per lot of the entire site. Volume and

area of the LID management practices were varied on individual sub-areas. An exact

breakdown on a sub-area basis is contained in Appendix A. The feasibility of

implementing the required area of depression storage, as bioretention, is not addressed in

this research. However, the resulting areas are similar to bioretention unit areas shown in

examples from the Prince Georges County, Maryland design manuals. Therefore, it was

assumed that the area requirements determined in this research are applicable.

Table 4-6. Final volumes of management practices.
Alternative Depression Average per lot Area per lot Total pond
storage (m3) (m3) (m2) volume (m3)a
Traditional 0 0 0 220,171
Cluster 0 0 0 157,122
"Partial" LID 32,059.11 44.59 292.58 29,576
"Full" LID 17,517.82 24.33 159.65 5,497
a Traditional and cluster designs utilize wet ponds with permanent pool while LID
designs utilize dry detention ponds.









The LID designs could not meet peak flow control requirements with the final

depression storage values. The values presented above represent the stopping point at

which dry detention ponds were considered. These ponds are conceptual and not

explicitly designed, although the resulting detention volume is probably quite accurate for

the "full" LID design since it is placed in a sub-area (Area 1) where runoff would

concentrate naturally. The "partial" LID design required ponds in Area 1, Area 6 and

Area 2. Of these three volumes, the one in Area 2 is the most dubious since outflow from

Area 2 is clearly not concentrated. Even with addition of three conventional detention

ponds, the "partial" LID option could not meet peak flow controls for the 25-year design

storm. This underscores the benefit of preserving a portion of the watershed in natural

condition. Certainly, additional ponds or preservation of natural upland area on the lots

would likely improve the performance of the "partial" LID design, but since the ponds

were conceptual, it was decided to end the analysis at this point due to the fact that ponds

were now required in sub-areas that did not have a natural flow concentration point.

Continuous Model

The continuous models were run with all runoff controls implemented. Peak flow

results for the continuous models showed mixed results. Three convection-type storms

are presented first. In these cases, pond systems tended to be very slightly over the

existing peak, while the LID designs were very slightly (with one exception on 1 July for

the "partial" LID design) under the existing peak, although LID performance was close to

the peak for the largest convection event on 2 September. The LID designs were under

the existing peak for the first two events because the depression storage controls in the

lowland sub-areas, which are the main runoff producers for these events, reduce surface

runoff. The fact that pond system peak flows were slightly higher than most existing









peaks was possibly due to increased surface runoff from development impacts in the

lowland sub-areas and pond outflow.

Two storm events where all designs failed to meet peak flow controls were the

convection event on 2 September and the tropical storm on 14-15 September. Each

design was about 1.0 cms (35.31 cfs) over existing conditions for the tropical event, but

only slightly so for the convection storm. A sustained period of flow observed after the

large convection on event 2 September indicated that the interflow (1st groundwater

layer) was full. Inspection of the model results showed that just prior to the 14-15

September event, the interflow layer had a higher percent of storage full than existing

conditions. The reason for this might be attributed to runoff from indirectly connected

impervious surfaces being routed over the pervious areas resulting in higher antecedent

moisture conditions. As would be expected, based on the previous discussion, open space

preservation in the uplands reduces the impact. However, it is also clear that LID designs

are less capable of dealing with this storm.

Overall volume discharged through the simulation period increased 10-20 mm

(0.39-0.79 in) for the pond system while the LID options were slightly lower, about 4.0

mm, than the existing conditions. Each design alternative resulted in slightly lower flows

at the three base flow periods observed due to a smaller area contributing groundwater

flow (the volume of groundwater flow is routed through a linear reservoir to determine

flow at any given point in time). Therefore, the conclusion was that any increase in total

volume is due to increased surface runoff or interflow. The magnitude of base flow

difference is not great; however, the cumulative impact would result in a lower base flow

outflow volume.









Structural Controls

Table 4-6 clearly shows the impact of preserving natural open space on structural

storm-water management. The reduced lot designs all required less structural storm-water

management. If the storm-water management system failed in the future, the impact on

the receiving waters may be less for the reduced lot designs due to less developed area.

Model Limitations

This model and procedure had certain limitations. The procedure used for

indirectly connected impervious surfaces presents the biggest concern. The ability to

identify a percent impervious was added after the Technical Reference Manual for HEC-

HMS was written and is not fully documented. The same option was implemented for

other loss methods, such as the curve number method, from the first release. From this

documentation, it was determined that this option only represents directly connected

impervious surfaces and that the SMA unit would not apply in this portion of the

watershed. This was confirmed by running HEC-HMS on two test sub-areas with

identical SMA parameters and one sub-area with 10% imperviousness defined. One sub-

area (with the 10% imperviousness) is 1.0 km2 and the other is 0.9 km2. The groundwater

time-series, soil moisture time-series and groundwater outflow volume are identical. This

is logical since the SMA model represents vertical storage. Horizontal storage is modeled

through the linear reservoir groundwater routing routine. Therefore, indirectly connected

surfaces must also have zero vertical storage. The procedure used in this research to

model indirectly connected impervious surfaces may introduce error to the computations.

The SMA percolation equations are dependent on the ratio of storage filled in the upper

contributing subsurface layer to the lower receiving layer. Percolation between one layer

and the next lowest is greatest when the upper layer is full and the lower layer is empty.









Storages fill faster no matter which modeling "trick", previously described, is used.

However, simple recreation of the spreadsheet computations presented by Bennett (1998)

for each option yielded slightly different results. A HEC-HMS test run also yielded

slightly different results; however, the overall impact was very slight: a difference of 17

m3 for a 12.7 mm storm on 1.0 km2 sub-area (0.9 km2 for the increased intensity sub-

area).

Regardless of the method of modeling indirectly connected, impervious surfaces,

another source of error will occur. The model assumes that the runoff from these surfaces

is instantly and uniformly distributed over the rest of the sub-area. This is, of course, not

generally the case. The degree of error that is added is not known, but is common

throughout each design simulation. Any hydrologic model that does not explicitly have

routines for indirectly connected impervious surfaces will present this problem.

A similar concern exists for depression storage. Surface runoff can only occur after

all depression storage is filled. Depression storage was assumed to be uniformly

distributed across the sub-area, which is not true. Furthermore, runoff from road surfaces

would not normally flow into bioretention units. Therefore, it is possible that runoff from

the road will exceed the depression storage provided in the swales and flow to the basin

outlet before the lawn and roof areas are contributing runoff to basin outlet. The model

does not consider this distinction.

Finally, there is the matter of the Clark storage coefficient. This unit hydrograph

method is not widely used because of difficulties in determining this coefficient. The

calibrated coefficients are fine, but the method of computing the future coefficient, by

weighting the proportion of the sub-area in natural condition versus developed condition









to compute the multiplication factor used with the To to determine the Clark storage

coefficient, cannot be validated. The assumption is that the greater the proportion of

natural open space area, the closer the coefficient will be to the existing one.

A distributed parameter model could be useful for modeling residential storm-water

management systems. This would be particularly true for the distributed, infiltration-

based LID storm-water management system. The SMA module has a distributed

parameter option. However, this option is based on a grid fixed to a geographic map

projection. The grid lines can not be adjusted to coincide with lots, which would be the

best way to model LID systems.

Hydrologic Summary and Conclusions

Comparison of the four design alternatives in this hydrologic study led to several

conclusions. Application of the infiltration-based, distributed storm-water management

system appears to result in a developed watershed response that is closest to natural

conditions, particularly if it is accompanied by a program of land preservation around

stream corridors and upland high infiltration areas. Land preservation also decreased

reliance on storm-water management control practices. This, at least in a qualitative

sense, reinforces the LID design philosophy presented by Prince Georges County,

Maryland. Hydrologic performance could certainly be improved (and possibly eliminate

the conceptual ponds in Sub-area 2) in the "partial" LID design if a portion of the lots

were preserved in their natural condition. However, this approach would seem to violate

the LID concept of minimizing development in high infiltration soils because large areas

of uplands would still be impacted while the "full" LID impacted a smaller area.

LID storm-water systems may perform worse than a conventional pond system

when antecedent moisture conditions are above average. The land preservation aspect of