Front Cover
 Tables and figures

Group Title: Bulletin (Tech.)
Title: Assessing phosphorus load reductions under agricultural best management practices
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00026759/00001
 Material Information
Title: Assessing phosphorus load reductions under agricultural best management practices
Series Title: Bulletin (Tech.)
Physical Description: 30 p. : ill. ; 28 cm.
Language: English
Creator: Rice, Ronald W., 1958-
Izuno, Forrest T
University of Florida -- Agricultural Experiment Station
Publisher: University of Florida Agricultural Experiment Station, Institute of Food and Agricultural Sciences
University of Florida, Agricultural Experiment Station, Institute of Food and Agricultural Sciences
Place of Publication: Gainesville FL
Publication Date: 1998
Copyright Date: 1998
Subject: Phosphorus -- Environmental aspects -- Florida -- Everglades   ( lcsh )
Agricultural pollution -- Florida -- Everglades   ( lcsh )
Genre: government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
non-fiction   ( marcgt )
Bibliography: Includes bibliographical references (p. 29-30).
Statement of Responsibility: Ronald W. Rice and Forrest T. Izuno.
General Note: "June 1998."
General Note: Cover title.
 Record Information
Bibliographic ID: UF00026759
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: ltuf - ALY7904
oclc - 39876854
alephbibnum - 002383143
issn - 0096-607x ;

Table of Contents
    Front Cover
        Front Cover 1
        Front Cover 2
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
    Tables and figures
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
Full Text

June 1998 Bulletin (Tech.) 905

Assessing Phosphorus Load
Reductions Under Agricultural
Best Management Practices

Ronald W. Rice and Forrest T. Izuno

Agricultural Experiment Station
Institute of Food and Agricultural Sciences

Ronald W. Rice is Assistant Professor and Crop Nutrition and Water Quality Specialist; Forrest T. Izuno is Professor and Water Management
Specialist; Everglades REC, Belle Glade, FL 33430-8003.

Assessing Phosphorus Load Reductions
Under Agricultural Best Management Practices

Ronald W. Rice and Forrest T. Izuno


The Everglades Forever Act (1994) outlined a comprehensive restoration

program designed to improve water quality and hydroperiod (water quantity and

delivery timing) to the Everglades. Because the Everglades ecosystem evolved as a

low-nutrient (oligotrophic) environment, increased deliveries of nutrient enriched waters

from both agricultural and urban sources is an important water quality concern. The

chemical element phosphorus (P) has attracted the most scrutiny since this nutrient is

most likely to encourage undesirable ecosystem changes in aquatic and wetland

ecosystems (Federico et al., 1981; Davis, 1994).

The Everglades Agricultural Area (EAA) is geographically located south of Lake

Okeechobee and north of the Water Conservation Areas (WCAs). The EAA plays an

important role in the Everglades water supply, either directly through agricultural

drainage runoff, or indirectly by serving as a conduit for large water transfers from Lake

Okeechobee to the WCAs. Drainage waters in the EAA can become nutrient enriched

through crop production inputs, natural mineralization (biological breakdown and

nutrient release) of organic "muck" soils, and use of nutrient enriched irrigation waters



F from Lake Okeechobee. Because water storage options are limited by shallow soils

q o with underlying limestone formations that resist percolation, excess water must typically

"SCIENCE be pumped off-farm into area drainage canals serving the EAA. This canal network
drains a significant portion of EAA runoff into the WCAs through pump stations

managed by the South Florida Water Management District (SFWMD).

Chapter 40E-63 of the Everglades Forever Act placed responsibility with the

SFWMD to design the EAA Regulatory Program (SFWMD, 1992). This program

requires that total P (TP) loads in EAA runoff entering the WCAs must be reduced by at

least 25% relative to the 1979-1988 baseline period of record. This basin-level target is

to be collectively achieved by all growers through the implementation of best

management practices (BMPs) specifically designed to reduce TP discharge from their

properties. Chapter 40E-63 also requires growers to develop their own water quality

monitoring plans consisting of drainage flow measurements and water sample

collection at farm structures discharging into works of the SFWMD (Whalen and

Whalen, 1996).

Basin-level TP drainage is continuously monitored at seven pump stations

operated by the SFWMD. Recognizing that year-to-year differences in rainfall

distributions will influence off-farm drainage requirements and overall basin runoff, the

SFWMD developed a model which adjusts EAA basin TP discharge for annual rainfall

distributions; based on comparisons to the 10-year period of record for similar data

(SFWMD, 1992). For the first annual compliance determination (May 1995 through

April 1996), the EAA basin recorded a 68% TP reduction (SFWMD, 1996).


The EAA Regulatory Program (SFWMD, 1992) is unique in that it assesses

water quality improvements at the basin level rather than at the individual farm level. A

farm-scale computer model (EAAMOD-FARM) currently under development will

eventually provide growers with a sophisticated tool for designing effective P-reduction

BMP systems for their farms (Pickering et al., 1997). At present, growers and water

managers need timely feedback regarding the effectiveness of evolving farm BMP


The objective of this publication is to present several methods that can be used

to quantify TP reductions under BMP implementation at the farm level. Factors that can

complicate or bias the interpretation of farm-level water quality monitoring data will be

identified. The discussion will emphasize that meaningful water quality interpretations

rely on an adequate assessment of rainfall that occurred during the collection of farm

drainage water quality monitoring data.

Water Quality Monitoring Data

In order to monitor farm-level TP discharge, the farm discharge points should be

instrumented to automatically collect water samples during drainage events. Some

provision must also be made to track discharge flow (or volume) during the drainage

event. Details regarding water sampling instrumentation and flow determinations,

which generally involve a pump calibration procedure and the automatic or manual

monitoring of canal depth levels, are beyond the scope of this discussion. Suffice it to


say that it is important to maintain accuracy and consistency across all monitoring and

data collection protocols.

The EAA Regulatory Program (SFWMD, 1992) calls for the collection of

composite water samples. Thus, over the course of a drainage event, the automatic

water sampler is programmed to obtain an aliquot of water at pre-defined time intervals

(say every two hours), with all aliquots combined within a single collection vessel. The

TP concentration of this single composite water sample is considered representative

of the drainage event (or events).

Raw Data

Table 1 includes a listing of raw water quality monitoring data collected over a

1-month period from a single farm discharge structure. Each data entry is

representative of a 24-hr period. During the first four days, there was no pumping

activity (drainage volume=0 gallons) and the TP concentration is simply left blank since

no water sample was collected. The first pumping event, initiated sometime on day 5,

produced a drainage volume of 5,119,240 gallons. Over the course of this event, the

automated water sampling protocol used at this farm site collected a composite sample

with a TP concentration of 0.0882 mg/L.

Table 1 deliberately contains "mixed" units (English and metric). Although

monitoring data are generally collected in English units (gallons), water quality

laboratories often report concentrations in metric units (mg/L). Subsequent use of

these data to generate other calculated data values will invariably require some form of


unit conversion. Details regarding units of measure for English and metric systems and

nutrient load calculation protocols for various water sampling strategies are fully

discussed in Izuno and Rice (1998).

Calculated Data: Load

A "load" is defined as a mass of a chemical or chemical compound that is moved

from one location to another. The P load for any given farm drainage event is simply

the mass of P that is present in the given drainage volume, calculated by the product of

the discharge volume with its associated TP concentration value:

load = volume*concentration*{unit conversion factors}. (1)

For the example data in Table 1, the total P load for day 5 is:

load = (5,119,240 gal)*(0.0882 mg/L)*{(3.785 L/gal)*(0.000001 kg/mg)}
= 1.709 kg,

which can be converted to English units:

load = (1.709 kg)*(2.205 Ibs/kg)
= 3.768 Ibs.

Calculated Data: Unit Area Load (UAL)

Nutrient load calculations are biased by farm size. Direct comparisons of P

loads between farms can be misleading since discharge volume is a factor in the load


calculation and big farms (with larger watersheds) will typically have greater volumes of

water to drain than smaller farms. To avoid this issue, nutrient loads should be

"normalized" for farm watershed area and re-expressed as "unit area loads" (UALs):

UAL = load/(farm area). (2)

The calculation for day 5 UAL (in English units) is:

UAL = (3.768 lbs)/(2250 acres)
= 0.0017 Ibs/acre.

Although Table 1 lists daily UAL values, it is more common to discuss UAL

values representing longer time periods such as a month or year. Keep in mind that

load calculations should be performed on daily data since the calculation (Equation 1)

addresses specific discharge volumes and concentrations that occur on any given day.

Once the daily load values are calculated, the easiest way to find the monthly UAL

value is to calculate all daily loads (Equation 1), total them to find the monthly load

value (177.003 Ibs; Table 1), and then divide this value by the farm area (Equation 2) to

generate the UAL value for the month (0.0787 Ibs/acre). This value can also be found

by summing up all daily UALs over the month, but this option is more labor intensive.

Calculated Data: Cumulative UAL

The last two data columns in Table 1 present UAL and rainfall values as

cumulative summaries. The cumulative UAL for any given day is simply the total UAL


recorded from the first day of monitoring (i=day 1) through the given day of interest

(i=day x):

cumulative UAL (fordyx = E(UAL) for i=day 1 through i=day x. (3)

For the provided example (Table 1), cumulative UAL remains at zero over the first four

days of monitoring when no pumping occurred. On day 5, cumulative UAL (0.0017

Ibs/acre) is exactly equal to the UAL recorded for day 5. Additional pumping activity on

day 6 produced a P load of 35.221 Ibs (Equation 1) which equates to a UAL of 0.0156

Ibs/acre (Equation 2). The cumulative UAL for day 6 is found by applying Equation 3 as


cumulative UAL (foray) = UAL, + UAL2 + UAL3 + UAL4 + UAL, + UAL6
= 0+0+0+0+0.0017+0.0156
= 0.0173 Ibs/acre.

Note that the cumulative UAL value remains unchanged on days when no

pumping has occurred. Thus, cumulative UALs remain at 0.0412 Ibs/acre from day 10

through day 13 (when UAL=0 Ibs/acre). A UAL of 0.0083 Ibs/acre on day 14 results in

the updated cumulative UAL value of 0.0495 Ibs/acre (Table 1). On day 31, the last

day of the monthly monitoring period, the cumulative UAL value (0.0787 Ibs/acre) is

exactly equal to the monthly UAL total recorded for the 31-day period.

Calculated Data: Cumulative Rainfall

Although rainfall is not a direct factor used in load and UAL calculations, an


understanding of rainfall and its influence on farm discharge is critically important in the

interpretation of water quality trends. The utility of a cumulative rainfall database will

become apparent in subsequent discussions. For now, it is only necessary to note that

the cumulative rainfall calculation is similar to the cumulative UAL calculation, except

rainfall values from i=day 1 through i=day x are used:

cumulative rainfall (fordayx) = E(rainfall), for i=day 1 through i=day x. (4)

Why Cumulative?

Once the cumulative UAL and rainfall databases have been calculated, there is

no particular need to focus on any given value for any given day. The important aspect

of this exercise is to generate the entire chronological listing of paired daily cumulative

UAL and rainfall values. These paired data are useful for generating graphical

descriptions of incremental UAL discharge over incremental rainfall. The utility of these

graphs for interpreting water quality monitoring data will be discussed shortly.

Factors that Challenge the Interpretation of Monitoring Data

A number of factors can limit the options available for interpreting water quality

monitoring data. It has already been mentioned that P load data should be adjusted for

farm area since volume discharge requirements can be influenced by watershed size.

For example, over similar 15-month BMP monitoring periods (Table 2), the total P load

discharged from Farm C (768.2 Ibs) was 60% smaller than for Farm B (1947.9 Ibs).


However, after normalizing for farm area, the total UAL discharged from Farm C (1.200

Ibs/acre) was almost three times greater than for Farm B (0.423 Ibs/acre).

Time is also a factor that can bias data interpretations. Calculating the total UAL

discharge for baseline and BMP periods may serve as a convenient data summary, but

these totals should not be compared to one another because the monitoring time

periods differ. For example, the total BMP period UAL (3.251 Ibs/acre) for Farm A is

47% greater than for the baseline period (2.206 Ibs/acre), largely because BMP data

were acquired over an extra 294 days (Table 2).

Normalizing baseline and BMP total UALs by their respective time periods would

minimize the effect of dissimilar monitoring time periods. However, subsequent

comparisons of these UAL-to-time ratios are not instructive since these ratios fail to

address differences in rainfall distributions (and the influence of rainfall on off-farm

drainage requirements) that existed during baseline and BMP data collection.

An alternative to comparing UALs for entire monitoring periods is to assess UAL

discharge trends over time. A natural inclination is to plot monthly UALs over time but

these graphs generally fail to illustrate trend differences between baseline and BMP

data. For example, some of the highest monthly UALs on record for a research farm

site occurred after the implementation of P-reduction BMPs (Fig. la). These

observations highlight the fact that UALs are influenced by rainfall distributions which

vary with season and year (Fig. 1b). In the EAA, water removal through gravity and/or

deep percolation/seepage is negligible due to flat basin topography, shallow soils, and

a relatively impermeable bedrock. Given limited water storage options, farm (and


basin) drainage requirements are closely tied to antecedent and current rainfall


In late-1994, Tropical Storm Gordon caused flooded conditions throughout the

16-county SFWMD regional drainage system which limited the efficient drainage of the

EAA basin. During this time period, many farms were operating under BMP plans but,

none-the-less, recorded elevated UAL discharges during a year that developed into the

fifth wettest on record (SFWMD, 1995).

Clearly, UAL data comparisons between monitoring periods must employ

methods that address rainfall profiles that existed during baseline and BMP data

collection. The first inclination is to normalize UAL data for rainfall. Attempts to achieve

a high degree of resolution by normalizing daily (or weekly) UAL data by daily (or

weekly) rainfall totals are not instructive. A problem arises when there is no rainfall

during these short time periods since the UAL-to-rainfall ratio goes to infinity (i.e.

dividing by zero) which is not interpretable. Ironically, normalizing by longer monthly

increments (Fig. 1 c) fails to illustrate clear trends over time because resolution is lost

regarding daily rainfall distributions and their effect on off-farm pumping requirements.

This point is illustrated herein using data from a research farm site. Note that

1994 rainfall totals were similar in November (8.0 inches) and December (8.1 inches)

(Fig. Ib), but monthly UALs (0.83 and 0.36 Ibs/acre, respectively) were strikingly

different (Fig. 1a). No additional information is gained by reviewing normalized monthly

UAL-to-rainfall ratios (Fig. Ic). Inconsistent trends across monthly data values will arise

due to the effects of dissimilar daily rainfall distributions on pumping activities. In this


example, November rainfall was delivered as a single 3-day tropical storm event and

resulting flooded field conditions demanded unusually high discharges. In December,

rainfall was distributed over three separate events throughout the course of the month.

Under these less extreme rainfall conditions, the grower was able to moderate off-farm

discharge activity.

In summary, a number of constraints limit the options available for interpreting

south Florida water quality monitoring data. The following section highlights water

management trends that have evolved under BMP implementation. Recognizing these

trends is useful in the interpretation of water quality data. This section is followed by a

presentation of three analytical methods that can provide useful assessments of water

quality trends under P-reduction BMP technologies.

Water Management Trends Under BMPs

Many growers have improved on-farm water use efficiency in their efforts to

reduce P (i.e. UAL) discharges. In general, crop production operations under BMP

implementation will allow a more conservative water management response to rainfall.

Popular BMP strategies include redesigning water conveyance networks to enhance

farm drainage uniformity and capacity, hydraulically isolating different crop commodities

within contiguous areas to improve water table management, preferentially retaining

high-nutrient draindown waters on-farm, using field rather than canal water levels to

schedule irrigation and drainage events, maximizing water removal through


evapotranspiration losses, extending fallow flooded field storage periods when

appropriate, and conducting agriculture under higher than traditional water table levels

(Izuno et al., 1995; Bottcher et al., 1995).

Over time, the comprehensive implementation of these strategies (either singly

or in combination) gives growers additional flexibility with respect to water management

decisions. Thus, under BMP operations, improved control over water within the

confines of the farm should lead to reduced off-farm discharge pumping requirements,

particularly in response to minor rainfall events.

Assessing P Reductions Under BMP Implementation

Method 1: Comparing Baseline and BMP UAL-to-Rainfall Ratios

Calculating Meaningful Index Ratios

Given the above trends, an overall summary of water management response

(discharge pumping) to rainfall can be estimated by normalizing monitoring period UAL

totals by their respective rainfall totals (UAL:R). These calculations are performed with

Equations 5 and 6, where i=first day through i=last day of the baseline monitoring

period and j=first day through j=last day of the BMP monitoring period :

baseline UAL:R ratio
= (total baseline period UAL)/(total baseline period rainfall) (5)
= E(daily load)/(farm area)]/_(daily rainfall),, and


BMP UAL:R ratio
= (total BMP period UAL)/(total BMP period rainfall) (6)
= [(daily load)/(farm area)]/E(daily rainfall),.

Using data for Farm B provided in Table 2 (monitoring period total loads and UALs are

given), these equations produce the following UAL:R values:

baseline UAL:R ratio = (1.152 lbs/acre)/(131.0 inches)
= 0.00879 Ibs/acre/inch, and

BMP UAL:R ratio = (0.423 lbs/acre)/(58.5 inches)
= 0.00723 Ibs/acre/inch.

Assessing Ratio Differences

Over time, UAL:R ratios should decline as BMP strategies are phased into crop

production operations. Evidence of a reduction in P discharge can be found when the

UAL:R value for the BMP monitoring period is smaller than for the baseline period. This

comparison can be quantified by finding the relative UAL:R difference, basically

assessing to what extent (as a percentage) the BMP value differs from the baseline


relative UAL:R difference
= [(BMP UAL:R baseline UAL:R)/(baseline UAL:R)]*100. (7)

Using the previously calculated UAL:R values for Farm B, this relative UAL:R difference

is calculated as follows:


relative UAL:R difference = [(0.00723 0.00879)/0.00879]*100
= -17.7%.

Thus, with respect to overall baseline and BMP rainfall, the total UAL discharge

during BMP operations at Farm B was roughly 17.7% less than during baseline

operations. A negative value indicates that BMPs are effecting UAL reductions. Similar

calculations for Farm A indicate a slight (+2.7%), but likely negligible, increase in UALs

under BMPs. The +80.1% relative UAL:R difference for Farm C strongly suggests that

BMP strategies are either inadequate or are improperly managed. This methodology

provides rapid feedback regarding progress under BMPs, and allows the timely revision

to the BMP program.

Method 2: Comparing Baseline and BMP Cumulative Data Distributions

Comparisons between monitoring period UAL:R ratios (discussed above) provide

supporting evidence that P reductions are occurring under BMP implementation.

However, these ratios fail to address the influence of daily rainfall distributions on UAL

discharges. Daily resolution can be achieved by re-expressing UAL and rainfall data as

cumulative data (Table 1, Equations 3 and 4). The purpose is to generate a listing of

paired data that reflect UAL and rainfall totals chronologically over time. Descriptions of

farm P discharge profiles are achieved by plotting cumulative UAL as the dependent

variable (Y-axis) and cumulative rainfall as the independent variable (X-axis). This

description is distinctly advantageous because it assesses water quality trends relative

to rainfall volume rather than calendar time.


Using Regression to Compare Cumulative Data Distributions

For comparative purposes, the cumulative values are first compiled separately

for baseline and BMP data and then plotted together. The comparison is quantified

with linear regression applied separately to the baseline and BMP cumulative

databases. For this analysis, regression is intended to serve a descriptive purpose

rather than a predictive function. Differences between baseline and BMP regression

slopes can be quantified by finding the relative slope difference, which assesses to

what extent (as a percentage) the BMP slope value differs from the baseline value:

relative slope difference
= [(BMP slope baseline slope)/(baseline slope)]*100. (8)

For illustrative purposes, Figure 2 presents schematic diagrams of various

outcomes that can be expected with the linear regression exercise (actual cumulative

data scatter plots are not shown). Evidence of consistent P reductions under BMP

operations is found when the BMP slope is of lower magnitude than the baseline slope.

This regression result indicates that the overall BMP data distribution (cumulative UAL

vs. cumulative rainfall) resides below the baseline distribution, which in turn reflects a

consistent decline in UAL discharge in response to rainfall during BMP operations.

Given this situation, the relative slope difference (Equation 8) will produce a negative

value. Although this scenario is depicted in Fig. 2a, the relative slope difference

(Equation 8) is minimal:


relative slope difference = [(0.0291 0.0300)/0.0300]*100
= -3.0%.

The reverse scenario (declining water quality trends under BMPs, albeit minor) is

presented in Fig. 2b, whereby the BMP slope is +3.1% larger than the baseline slope.

Differences between baseline and BMP UAL discharge trends are more obvious in Fig.

2c, with a relative slope difference (-30.0%) supporting the conclusion that water quality

trends have improved markedly under BMP operations. The reverse scenario (Fig. 2d)

portrays a BMP slope value 42.9% greater than the baseline distribution slope. This

relationship strongly suggests that the BMP program is either not working, is

inadequate in scope, or is improperly implemented.

Application of Statistics to Regression Data

If desired, one can take the regression exercise one step further by assessing

whether or not the difference in baseline and BMP slopes are statistically significant.

The slopes can be compared with a t-test that addresses the entire paired cumulative

data set for both monitoring periods, using the test statistic:

tc = (BMP slope baseline slope)/[(s.e.BMP)2 + (s.e.BL)205, (9)

where s.e.,,P and s.e.BL are the standard errors associated with the regression slope

estimates for the BMP and baseline cumulative distributions, respectively. Given a

"sufficiently large" database, the slopes are significantly different if:


Itl > 1.97 ... significant at the 5% level, and (10)
Itjl > 2.60 ... highly significant at the 1% level, (11)

where Itj refers to the absolute value of to. From a statistical standpoint, "sufficiently

large" refers to a combined cumulative baseline and BMP water quality database

exceeding 242 days (this equates to 240 degrees of freedom for the t-test which,

relative to standard statistical tables for the t distribution, minimizes the critical t-values

to either 1.97 or 2.60, depending on the choice of significance level). In reality, any

meaningful interpretations of water quality data would require monitoring efforts that

greatly exceed 242 days, thus the critical t-values listed in Equations 10 and 11 are


Examples From Farm Research Sites

The interpretative value of cumulative UAL and rainfall distributions will be

highlighted below for four research farm sites that collectively represent a wide range of

cropping systems and BMP implementation strategies. Figure 3a depicts baseline and

BMP cumulative data distributions for Site UF9206A&B, a mixed-crop (sugarcane,

vegetables, rice, and sod) operation. To facilitate comparisons, linear regression lines

for the baseline and BMP data are also provided. Based solely on linear regression

results, one would conclude that BMPs have failed to improve UAL discharge trends

since the BMP slope value is larger than the baseline slope value. This difference is

quantified with Equation 8, using slope values provided in Table 3:


relative slope difference = [(0.04608 0.04302)/0.04302]*100
= +7.1%.

A review of the data scatter plots reveals that new water management practices

during early BMP operations actually supported reduced UAL discharges over the first

46 inches of cumulative rainfall (Fig. 3a). The abrupt UAL increase from 1.21 to 3.68

Ibs/acre occurred during a 2-week pumping period in response to an 8-inch rainfall

event (Tropical Storm Gordon). Placed in perspective; this 3-day storm incurred a

discharge load that exceeded the sum total load incurred over the first 197 days of the

BMP period. As a result, the BMP distribution is displaced above the baseline

distribution (Fig. 3a), masking attempts to reasonably quantify P-reduction trends under

BMP operations.

A visual evaluation of the BMP distribution suggests that, once the farm site

recovered from this single aberrant weather event, BMP discharge trends returned to

those existing prior to the storm. Recognizing that south Florida historically encounters

such conditions, the EAA Regulatory Program (SFWMD, 1992) provides for the

exclusion of extreme rainfall conditions from water quality monitoring databases. Basin-

wide flooding during this November 1994 storm resulted in the temporary abandonment

of water management BMPs in efforts to drain flooded fields during the sugarcane

harvest and winter vegetable production season. In order to address the

disproportionate effects of this single weather event, baseline and BMP distributions

are presented for cumulative databases with November data omitted (Fig. 3b). An

evaluation of this data subset reveals a consistent P-reduction BMP effect of 25.6%


(Equation 8) relative to baseline operations. This UAL reduction under BMP operations

is highly significant, confirmed by using regression data (Table 3) to calculate a t, value

of -46.1 (Equation 9), where ItJ=+46.1 which is greater than 2.60 (Equation 11):

t, = (0.03355 0.04512)/[(1.17x104)2 + (2.22x10-4)20.5
= (-0.01157)/(1.37x10-8 + 4.93x10-8)05
= (-0.01157)/(6.30x10-)0.5
= (-0.01157)/(2.51x104)
= -46.1.

In contrast to mixed-cropping operations, sugarcane monocultures are less

sensitive to immediate water table levels. The ability of sugarcane to withstand short-

term flooding and periodic "wet feet" gives growers additional latitude with respect to

discharge pumping. Although cumulative UALs doubled from 0.18 to 0.36 Ib/acre (data

not shown) during Tropical Storm Gordon, farm operations at Site UF9209A (sugarcane

monoculture) quickly recovered and a retum to BMP strategies ultimately supported a

20.2% UAL reduction (Equation 8; Table 3). A BMP effect is clearly evident for this site

when data encompassing the Tropical Storm Gordon period are omitted from the

analysis (Fig. 4). The baseline and BMP cumulative distributions describe a consistent

divergence and the 30.1% difference between slope magnitudes (Equation 8; Table 3)

reflects a consistent attenuation in UAL discharge during BMP operations.

The linear regression exercise may not be universally appropriate to all cropping

operations. Both the baseline and BMP cumulative distributions for Site UF9201A

(Fig. 5), a vegetable monoculture, describe non-linear UAL discharge profiles. Unlike

other crop production enterprises, pumping activity for vegetable monocultures will


periodically be driven by factors unrelated to rainfall. For example, successive cropping

operations throughout the winter production season necessitate scheduled drainage

events (regardless of rainfall) to accommodate field preparation and crop harvest

machinery. In addition, there is no relationship between off-farm discharge and rainfall

during the summer off-season when fields are deliberately maintained under fallow

flooded conditions.

Intermittent plateaus (no change in cumulative UAL over increasing rainfall)

reflect these deliberate fallow flooded field storage periods when rainfall events have no

bearing on drainage requirements (Fig. 5). However, the clear separation between

baseline and BMP distributions support conclusions favoring P reductions under BMP

operations. In particular, note that the BMP distribution begins to diverge rapidly from

the baseline distribution at 15.7 inches of cumulative rainfall, when baseline UALs

shifted vertically from 4.36 to 5.87 Ibs/acre. This UAL increase under zero rainfall

conditions during baseline operations represents the deliberate discharge of summer

fallow flood waters to accommodate September planting schedules. Under BMP

operations, modified water management strategies and new drainage practices were

implemented to address this traditional "September UAL spike" period. The diverging

distributions (Fig. 5) document declining UAL discharge trends under these BMP


As ar alternative to regression for these type of data, one can take the baseline

and BMP cumulative UAL values recorded at (for this example) every 10-inch rainfall

increment through 70 cumulative inches (the limit of the BMP distribution; Fig. 5). Using


these values, an average cumulative UAL can be calculated for the BMP (4.69 Ibs/acre)

and baseline periods (7.48 Ibs/acre). These average values can subsequently be used

in place of slopes (Equation 8) to calculate an overall 37.3% UAL reduction under BMP


As schematically described in Fig. 2d, comparisons of cumulative data

distributions will also describe declining water quality trends. This scenario is illustrated

for Site UF9204A (Fig. 6) whereby the BMP cumulative distribution remains consistently

above the baseline distribution. This site was under sugarcane production during

baseline operations but five months into BMP operations, half the farm area was

rotated into flooded rice. This major cropping modification occurred without the

concurrent implementation of an adequate water management plan. The 130.6% UAL

increase during the BMP period (Table 3) reflects the inability to properly manage rice

drainage waters at this site. It is important to emphasize that rice production is a

perfectly feasible production practice, given the implementation of appropriately

designed BMPs. The adoption of improved management strategies at a different

sugarcane-rice farm (data not shown) supported a 25.7% UAL reduction during BMP


Method 3: Comparing Rainfall-Adjusted UALs

The EAA Regulatory Program (SFWMD, 1992) requires the EAA basin to

achieve a minimum 25% P load reduction. All growers in the EAA are collectively in

compliance if the basin target is met. The annual compliance evaluation involves a


model that adjusts annual EAA sub-basin P loads for hydrologic variability (SFWMD,

1992). The model also allows for the calculation of estimated farm-level rainfall-

adjusted UALs (AUALs). The model calculates a rainfall-adjusted load based on the

comparison of current "water year" load (and monthly rainfall distribution) to an earlier

10-year (1979-1988) period of record database. By definition, the 1994 water year

(WY94) is the 12-month period beginning May 1, 1993 and ending April 30, 1994.

Required Data for AUA'L Calculations

In order to calculate farm-level AUAL estimates for a given WY, the following

information and data are required:

1. Identify the farm location with respect to EAA sub-basin (S5A, S6, S7, or S8),
2. Identify the water year of interest,
3. Calculate the farm's total UAL for the given WY,
4. Obtain the Thiessen-weighted monthly rainfall values specific to the WY and
EAA sub-basin, and
5. Obtain the "fixed coefficients" (rainfall variation, skewness, and adjustment
factors) specific to the EAA sub-basin that are used in the model equation.

At this writing, the Thiessen-weighted rainfall values and fixed coefficients were

available from the SFWMD website at .

Among several available files, download the archived "EAAB9709.zip" file, "unzip" this

file into its various components, and then open the file titled "EAABASIN.WK4" (a

Windows operating system Lotus 1-2-3 file). The EAABASIN.WK4 file has multiple

pages. For Thiessen-weighted monthly rainfall values, select the page tab titled

"Monthly" to access the data table titled "Basin Compliance Calculations EAA


Regulatory Rule", scroll to the table subheading titled "Monthly Calculated Values:

Basin Average Rainfall (inches)" which includes Thiessen-weighted rainfall values

(beginning October 1978) for EAA sub-basins S5A, S6, S7, and S8. To obtain the fixed

coefficients for each EAA sub-basin, select the page tab titled "Coefficients" to access

the data table titled "Fixed Coefficients for EAA Basin Calculations (From Rule Text)",

and scroll to the table subheading titled "od Mean Rainfall Statistics".

AUAL Calculation

An example of an AUAL calculation is provided below using monthly load data

collected during WY96 from a research farm site located in the S5A sub-basin. These

data are provided in Table 4a along with a listing of S5A sub-basin Thiessen-weighted

monthly rainfall values for WY96. For the readers convenience, the fixed coefficients

for all four sub-basins are provided as well (Table 4b).

In order to calculate the estimated farm AUAL, eight sequential calculations must

be performed (SFWMD, 1992). First, calculate three different factors (ml, m2, and m3)

specific to the Thiessen-weighted sub-basin rainfall values, where ri=individual monthly

rainfall values from i=1 (May) to i=12 (April). Note that m, is simply the average monthly

rainfall for the WY in question. The factor m2 is the average of the squared differences

between individual monthly rainfall values and the average WY rainfall value, while m3

is the average of the cubed deviations:

m, = E(r,)/12, (12)
m2 = ,(r -m)2/12, and (13)


m3 = F(r, -mi)3/12. (14)

Second, calculate the "predictors" (X, C, and S) from the first three rainfall factors. Note

that X is actually the natural log (In) of the total WY rainfall, C is the standard deviation

of the monthly rainfall values divided by the average monthly rainfall value (also known

as the coefficient of variation), and S is a rainfall skewness coefficient:

X = In(12*ml), (15)
C = [(12/11)*m2]S/mi, and (16)
S = [(12/11)*m3]/(m2)1.5 (17)

Third, calculate the adjusted sub-basin rainfall value (Ra) for the WY using the

"predictors" and the fixed coefficients (Cm and S,) specific to sub-basin of interest

(Table 4b), where "exp" signifies the inverse of the natural log function:

Ra = exp[X + 1.053*(C-Cm) 0.1170*(S Sm)]. (18)

Finally, calculate the adjusted unit area load for the WY, using the adjusted sub-basin

rainfall value (Ra) and the fixed rainfall coefficient (Ram) specific to the sub-basin of

interest (Table 4b):

AUAL = UAL*(R./Ra)2868. (19)

Using the data provided in Tables 4a and 4b, these equations produce the following

calculated values to arrive at a WY96 AUAL estimate of 0.5112 Ibs/acre:


m, = (1.33 + 6.56 + ... + 1.73)/12
= 4.4900,

m2 = [(1.33 4.49)2 + (6.56 4.49)2 + ... + (1.73 4.49)]/12
= 14.1591,

m3 = [(1.33 4.49)3 + (6.56 4.49)3 + ... + (1.73 4.49)3]/12
= 39.4395,

X = In(12*4.49)
= 3.9868,

C = [(12/11)*14.1591]05/4.49
= 3.9302/4.49
= 0.8753,

S = [(12/11)*39.4395]/(14.1591)15
= 43.0249/53.2787
= 0.8075,

Ra = exp[3.9868 + 1.053*(0.8753 0.7636) 0.1170*(0.8075 0.9999)]
= exp(4.1269)
= 61.9855, and

AUAL = 0.9319*(50.31/61.9855)2.868
= 0.9319*0.5496
= 0.5122 Ibs/acre.

The above AUAL calculation is specific to WY96 for a farm site located in the

S5A sub-basin. The procedure can be used to calculate AUAL estimates for any farm

and WY period, given the use of monthly Thiessen-weighted rainfall values and farm

UAL data specific to the WY of interest. For any given farm location within a given sub-

basin, the same fixed coefficients (Cm, Sm, and Ram; Table 4b) are used for any WY

period. Keep in mind that baseline and BMP monitoring periods for any given farm site

will not coincide with the WY calendar period. In general, water quality monitoring data


collected from EAA farms during WY94 will predominantly reflect baseline practices,

WY95 data will represent the transition from baseline to BMP operations, and WY96

data will exclusively reflect BMP operating conditions. For any given farm, evidence of

P-reductions under BMPs will be found with lower AUAL values for WY95 and WY96

relative to WY94.

These comparisons can be quantified by finding the relative AUAL difference

across any two WYs of interest (Equation 20), where WY, is the most recent WY and

WYj is the oldest WY in the given comparison:

relative AUAL difference
= [(AUALw, AUALwYj)/(AUALwYj)]*100. (20)

Table 4c provides a summary of UAL and AUAL values for three different WY. The

WY95 to WY94 comparison of AUALs is calculated as follows:

relative AUAL difference
= [(0.9735 -1.9502)/(1.9502)]*100
= -50.1%.

Thus, the 12-month AUAL incurred during the transition from baseline to BMP

operations (WY95) represents a 50.1% reduction in P discharge relative to the AUAL

discharge under baseline operations (WY94). Under full BMP implementation (WY96

AUAL value of 0.5122 Ibs/acre), a reduction of 73.7% was realized (Table 4c).



With less than four years of monitoring data, researchers and water managers

are challenged to assess the effectiveness of farm-level P-reduction BMPs in south

Florida. Straight comparisons between baseline and BMP period UAL data are

inadequate because calendar time frames and rainfall distributions differ for the two

monitoring periods. Meaningful comparisons require some measure of hydrologic

adjustment to UAL data.

Three methods for comparing water quality monitoring data were discussed.

The first method minimizes rainfall differences (which influence drainage pumping)

across monitoring periods by comparing baseline period total UAL to rainfall (UAL:R)

ratios to those for the BMP period. The second method involved the re-expression of

water quality databases into cumulative UAL and cumulative rainfall values. These

cumulative databases are plotted to assess differences in baseline and BMP UAL

discharge trends over incremental rainfall. Linear regression applied to these

distributions allows differences to be quantified through slope comparisons. Finally, the

application of a hydrologic model (developed for P discharge regulatory compliance

assessments) to farm UAL data allows the calculation of rainfall-adjusted UALs

(AUALs) for different water year (WY) periods. Subsequent AUAL comparisons across

different WYs serves to quantify P discharge trends over time.

Although not specifically addressed in this publication, it may be instructive to

briefly summarize water quality trends recorded at 10 EAA research farm sites from


late-1992 through April 1996 (Izuno and Rice, 1997). Using the UAL:R ratio

comparison method, BMP data for six of 10 sites reflected P reductions of 3 to 33%.

Using cumulative databases, the BMP distribution slope magnitudes for six sites were 6

to 35% lower than for baseline, evidence of long-term P load reductions under BMP

strategies. Eight of 10 sites reflected reductions after omitting nonrepresentative UAL

data collected under flooded conditions caused by a 3-day tropical storm. Applying the

hydrologic adjustment model, average AUALs for eight sites declined by 73% over a 3-

year period. Across all three analytical exercises, two sites consistently demonstrated

declining water quality trends as a consequence of large cropping system modifications

in the absence of adequate hydraulic BMP technologies. Despite short baseline

monitoring periods and less than four years of data collected under conditions of highly

variable rainfall, analytical methods discussed herein consistently verify BMP reductions

for a wide range of agricultural cropping systems.



Bottcher, A.B., F.T. Izuno, and E.A. Hanlon. 1995. Procedural guide for the
development of farm-level best management practices for phosphorus control in
the Everglades Agricultural Area. Univ. of Florida Cooperative Extension Service
Circular No. 1177. Gainesville, Fla.

Davis, S.M. 1994. Phosphorus inputs and vegetation sensitivity in the Everglades. In
Everglades: The Ecosystem and Its Restoration, S.M Davis and J.C. Ogden
(eds), ch. 15, 357-378. St. Lucie Press, Delray Beach, Fla.

Everglades Forever Act. 1994. Amendment of the 1991 Marjory Stoneman Douglas
Everglades Protection Act. Chapter 373.4592, Florida Statutes. Tallahassee, Fla.

Federico, A.C., F.E. Davis, K.G. Dickson, and C.R. Kratzer. 1981. Lake Okeechobee
water quality studies and eutrophication assessment. South Florida Water
Management District Technical Publication No. 81-2. West Palm Beach, Fla.

Izuno, F.T., and R.W. Rice. 1998. Calculating nutrient loads. Univ. of Florida
Cooperative Extension Service Circular No. Gainesville, Fla (in review).

Izuno, F.T., and R.W. Rice (eds). 1997. Implementation and verification of BMPs for
reducing P loading in the EAA. Phase V Final Report (vol. 1) submitted to the
EAA Environmental Protection District. Belle Glade, Fla.

Izuno, F.T., A.B. Bottcher, F.J. Coale, C.A. Sanchez, and D.B. Jones. 1995. Agricultural z
BMPs for phosphorus reduction in South Florida. Transactions of the ASAE.

Pickering, N.B., A.B. Bottcher, and J.D. Stuck. 1997. EAAMOD-FARM: Everglades
agricultural area farm-scale hydrologic and phosphorus transport model version
2.0. In F.T. Izuno and R.W. Rice (eds). Implementation and verification of BMPs
for reducing P loading in the EAA. Phase V Annual Report (vol. 2) submitted to
the EAA Environmental Protection District. Belle Glade, Fla.

SFWMD. 1992. Rules of the South Florida Water Management District Everglades
Regulatory Program. Chapter 40E-63, Florida Annotated Code. New: Jan. 22,
1992; Amended: July 7, 1992. South Florida Water Management District, West
Palm Beach, Fla.


SFWMD. 1995. Managing a very wet year. South Florida Water Management District,
West Palm Beach, Fla.

SFWMD. 1996. Rule 40E-63 information update: EAA basin total phosphorus levels
(June 24). South Florida Water Management District, West Palm Beach, Fla.

Whalen, P.J., and B.M. Whalen. 1996. Nonpoint source best management practices
program for the Everglades Agricultural Area. ASAE Paper No. 96-2071. St.
Joseph, Mich.: ASAE.


Table 1. Example of farm-level water quality monitoring data.

Raw data Calculated data

Day Rainfall Drainage Total P P P UALT Cumulative Cumulative
volume concentration load load UAL rainfall

inches gallons mg/L kg Ibs Ibs/acre Ibs/acre inches

1 0 0 0 0 0 0 0
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 2.23 5119240 0.0882 1.709 3.768 0.0017 0.0017 2.23
6 1.32 47847160 0.0882 15.973 35.221 0.0156 0.0173 3.55
7 0.08 43287950 0.0839 13.747 30.311 0.0135 0.0308 3.63
8 0.43 25141490 0.0796 7.573 16.698 0.0074 0.0382 4.06
9 0 10248115 0.0796 3.087 6.806 0.0030 0.0412 4.06
10 0 0 0 0 0 0.0412 4.06
11 0 0 0 0 0 0.0412 4.06
12 0 0 0 0 0 0.0412 4.06
13 1.21 0 0 0 0 0.0412 5.27
14 0.19 22917153 0.0974 8.449 18.629 0.0083 0.0495 5.46
15 1.27 6505400 0.0811 1.997 4.403 0.0020 0.0515 6.73
16 0.15 32179070 0.0648 7.892 17.403 0.0077 0.0592 6.88
17 0.28 20449179 0.0648 5.016 11.059 0.0049 0.0641 7.16
18 0 12774840 0.0613 2.964 6.536 0.0029 0.0670 7.16
19 0 6309890 0.0578 1.380 3.044 0.0014 0.0684 7.16
20 0 0 0 0 0 0.0684 7.16
21 0 0 0 0 0 0.0684 7.16
22 0.29 0 0 0 0 0.0684 7.45
23 0 0 0 0 0 0.0684 7.45
24 0.75 0 0 0 0 0.0684 8.20
25 0 21791085 0.0848 6.994 15.422 0.0069 0.0752 8.20
26 0 0 0 0 0 0.0752 8.20
27 0 0 0 0 0 0.0752 8.20
28 0.44 0 0 0 0 0.0752 8.64
29 0 11845960 0.0779 3.493 7.702 0.0034 0.0787 8.64
30 0 0 0 0 0 0.0787 8.64
31 0 0 0 0 0 0.0787 8.64

Total 8.64 266416532 n.a. 80.274 177.003 0.0787 n.a. n.a.

t UAL = unit area load (assume farm area = 2250 acres).

Table 2. Example of calculating and comparing unit area load (UAL) to rainfall ratios (UAL:R) for baseline and
BMP water quality monitoring data.

Site Farm Monitoring Calendar Time Rain P UAL UAL:R Relative UAL:R
name area period period load ratio differences

acres days inches Ibs Ibs/acre Ibs/acre/inch %

Farm A 1280 baseline 07/23/92 12/31/93 527 85.4 2823.1 2.206 0.02583
BMP 01/01/94- 03/31/96 821 122.6 4161.0 3.251 0.02652 + 2.7

Farm B 4608 baseline 07/24/92 12/31/94 891 131.0 5308.1 1.152 0.00879
BMP 01/01/95 03/31/96 456 58.5 1947.9 0.423 0.00723 17.7

Farm C 640 baseline 07/15/92- 12/31/94 900 159.0 1029.0 1.608 0.01011
BMP 01/01/95 03/31/96 456 65.9 768.2 1.200 0.01821 + 80.1

t A negative "relative UAL:R difference" indicates that the BMP period UAL:R ratio is smaller than the
baseline UAL:R ratio by the given percentage.

Table 3. Linear regression statistics for baseline and BMP cumulative unit area load (UAL) vs. cumulative
rainfall distributions for four research farm sites.

Baseline period BMP period BMP to baseline comparison

Site rt Slope s.e.t r Slope s.e. Itel Relative slope
name differences

Ibs/acre/inch Ibs/acre/inch %

All monitoring data

UF9201A # 0.925 0.15489 1.44x10-3 0.916 0.12421 1.28x103 15.9 19.8
UF9204A 0.978 0.00859 2.74x10-5 0.993 0.01970 6.91x105 149.5 +129.3
UF9206A&B 0.971 0.04302 2.19x104 0.982 0.04608 1.88x104 10.6 + 7.1
UF9209A 0.997 0.01011 1.65x10"5 0.988 0.00807 2.31x105 71.9 20.2

November monitoring data omitted

UF9201A 0.921 0.15228 1.52x103 0.931 0.10835 1.22x10 3 22.5 28.8
UF9204A 0.984 0.00859 2.64x10-5 0.993 0.01981 7.75x105 137.0 + 130.6
UF9206A&B 0.974 0.04512 2.22x104 0.995 0.03355 1.17x104 46.1 25.6
UF9209A 0.996 0.01007 1.90x105 0.995 0.00704 1.39x105 128.7 30.1

t r = correlation coefficient.
$ s.e. = standard error (square root of the variance) of the regression for the slope estimate.
t' = test statistic for comparing baseline and BMP regression slopes, which differ significantly at the 5%
level when It|l > 1.97 and at the 1% level when Itl > 2.60.
A negative "relative slope difference" indicates that the BMP slope value is of lower magnitude (by the
given percentage) than the baseline slope value.
# Strict application of linear relationship estimates are not recommended for UF9201A data (see Fig. 5).

Table 4a. Example data used in rainfall-adjusted unit area load (AUAL)
calculation for the 1996 water year (WY96).

Month Calendar Sub-basin Farm Farm
year rainfall P load UALt

inches Ibs Ibs/acre

May 1995 1.33 0 0
June 1995 6.56 119.60 0.0934
July 1995 6.77 90.50 0.0707
August 1995 10.32 232.37 0.1815
September 1995 4.12 97.94 0.0765
October 1995 12.22 287.03 0.2242
November 1995 1.22 0 0
December 1995 0.81 0 0
January 1996 1.96 193.56 0.1512
February 1996 0.59 26.46 0.0207
March 1996 6.25 91.65 0.0716
April 1996 1.73 53.74 0.0420

1996 WY total 53.88 1192.85 0.9319
Monthly average 4.49 99.40 0.0777

t Thiessen-weighted rainfall values for EAA S5A sub-basin.
t UAL = unit area load (assume farm area = 1280 acres).
Shaded values are used in the WY96 AUAL calculation example.

Table 4b. Fixed coefficients for EAA sub-basins.

Sub-basin C, Sm Ram

S5A 0.7636 0.9999 50.31
56 0.7302 0.7476 49.77
S7 0.7198 0.6112 4627
S8 0.7821 0.8409 45.68

Shaded values are used in the WY96 AUAL calculation example.

Table 4c. Example of AUAL data comparisons for three consecutive WY.

UAL AUAL Relative AUAL difference from:

WY94 WY95 WY96 WY94 WY95 WY96 WY95 to WY94 WY96 to WY94

--bs/acre/year ---- % change -

1.4461 2.1734 0.9319 1.9502 0.9735 0.5122 -50.1 -73.7

0 .8 .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .- . . . . . . . . . . . . . . . . . . . . . . . . . .

o ....----- --6 -- ----





1 2 8 .............................................................- --........................---


4-ll --- - I I I~ ll ---- 1i ---------------
-------- ---- 1 -----I-i1 I-- --L
0.20 ------- -----------------------
2 0.15 -----.--.-------------------------

0.00 -
0.00 - - - ---J -- - - ^ - - - -
1992 1993 1994 1995 1996
Baseline monitoring period BMP monitoring period

Fig. 1. Example of monthly a) unit area load (UAL), b) rainfall, and c) UAL to rainfall (UAL:R) ratio
data for baseline and BMP monitoring periods.

a) b)

o -BL BL

0 BL slope = 0.0300 Ibs/acre/inch BL slope = 0.0291 lbs/acre/inch
BMP slope = 0.0291 Ibs/acrelinch BMP slope = 0.0300 Ibs/acre/inch
Difference = 3.0% Difference = + 3.1%

c) d)



O BL slope = 0.0300 Ibs/acre/inch BL slope = 0.0210 Ibs/acre/inch
BMP slope = 0.0210 Ibs/acre/inch BMP slope = 0.0300 Ibs/acre/inch
Difference = 30.0% Difference = +42.9%

Cumulative Rainfall (inches) Cumulative Rainfall (inches)

Fig. 2. Schematics of linear regression relationships for baseline (BL) and BMP cumulative databases
depicting a) negligible differences, b) negligible differences, c) improved UAL trends under BMPs,
and d) declining UAL trends under BMPs.


46- -------------------------------______------------------------------------------------

5 ------------------ ----- ----------------- -------- ---------------

Cumulative Rainfall (inches)



4 . . .- . -..- .- . . . -. . . .. . . . . ..--.-- - -- - - - - .-- - ---.--- - - - - - -- - - - - - --- -.

^ ^ > ^ ^-----------------------------------
4 --?-------------------------------------------- ------ .-i- -i-

1 ..- ---------- -... ...... .. BL- BL V BMP BMP -

0 20 40 60 80 100 120 140 160
Cumulative Rainfall (inches)

Fig. 3. Site UF9206A&B baseline (BL) and BMP monitoring period cumulative rainfall versus cumulative
unit area P load (UAL) distributions for a) all monitoring data, and b) a data subset omitting the
disproportionate impact of a single tropical storm event.


0 .8 -- --- -- -- -- --- -- -- -- --- -- -- -- --- -- -- -- --- -- -- -- --- -- -- -- --- -- -- -- --- -- -- -- --- -- -- -- --- -- -- -- -


00.2 -- a BL BL v BMP BMP


0 20 40 60 80 100 120 140 160
Cumulative Rainfall (inches)

Fig. 4. Site UF9209A baseline (BL) and BMP monitoring period cumulative rainfall versus cumulative
unit area P load (UAL) distributions for a data subset omitting the disproportionate impact of a
single tropical storm event.


10 --------------------------------------------------------------------------------------------

8 - --- ------------ -

j56 -------------- ------- -- --- ---------------------- -


2 .-- .......---------...-........------- BL -- BL V BMP-- BMP

0 20 40 60 80 100 120 140 160
Cumulative Rainfall (inches)

Fig. 5. Site UF9201A baseline (BL) and BMP monitoring period cumulative rainfall versus cumulative
unit area P load (UAL) distributions for a data subset omitting the disproportionate impact of a
single tropical storm event.


*S 0.6 -------------------= -------------------- ---. -a _------- ----. -------------------. ---------.-. --------- .


r 0.3 -. BL BL BMP BMP -

0 20 40 60 80 100 120 140 160
Cumulative Rainfall (inches)

Fig. 6. Site UF9204A baseline (BL) and BMP monitoring period cumulative rainfall versus cumulative
unit area P load (UAL) distributions for a data subset omitting the disproportionate impact of a
single tropical storm event.

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs