|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 MAXIMIZING SUSTAINABLE REGIONAL PRODUCTION: BALANCING WATER, ENERGY, AND CARBON By DAVID AARON PFAHLER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENT S FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013
2 2013 David Aaron Pfahler
3 To my wife, Alicia, and my two girls, Anna and Ellie
4 ACKNOWLEDGMENTS I wish to thank m y advisor, Dr. Mark Brown, for his constant encouragement and guidance in this effort, as well as his great patience throughout the process. I also wish to thank my committee members, Dr. Matt Cohen, Dr. Joseph Delfino, and Dr. Alan Hodges for their contri butions and insights to the development of this dissertation This work would not have been possible without the funding received through the NSF AMW3 IGERT program. I would like to thank Dr. Brown for giving me the opportunity to be a part of this prog ram. I would also like to thank my colleagues and friends in the IGERT cohorts for their friend ship and encouragement along the way. Finally, I would like to thank my wife, who has been so s upportive and patient in this endeavor. Without her constant sacr ifices this dissertation c ould never have been written
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ ........ 10 LIST OF ABBREVIATIONS ................................ ................................ ........................... 11 ABSTRACT ................................ ................................ ................................ ................... 13 CHAPTER 1 MODELING REGIONAL SUSTAINABILITY ................................ ........................... 15 Introduction ................................ ................................ ................................ ............. 15 Regional Sustainability ................................ ................................ ..................... 15 Regional Scale ................................ ................................ ................................ 16 Measuring Sustainability ................................ ................................ ................... 17 Emergy S ynthesis ................................ ................................ ...................... 18 Ecological Footprint Analysis ................................ ................................ ..... 18 Multi Criteria Decision A nalysis ................................ ................................ .. 19 Maximum Sustainable Production ................................ ................................ .......... 20 D efining and Valuing Production ................................ ................................ ...... 21 Defining System Boundaries ................................ ................................ ............ 21 Limiting Factors ................................ ................................ ................................ 22 Study Area ................................ ................................ ................................ .............. 24 Conclusion ................................ ................................ ................................ .............. 25 2 DEVELOPMENT OF A REGIONAL EIO LCA MODEL ................................ ........... 27 Introduction ................................ ................................ ................................ ............. 27 Methods ................................ ................................ ................................ .................. 27 Economic Input Output Life Cycle Assessment ................................ ................ 27 Economic Model ................................ ................................ ............................... 32 Resource Consumption Data ................................ ................................ ........... 32 Development of water intensity vectors ................................ ..................... 33 Development of energy intensity vectors ................................ ................... 35 Development of a GHG emission intensity vector ................................ ...... 38 Development of a value added intensity vector ................................ ......... 39 Model Construction ................................ ................................ .......................... 40 Emergy Evaluation of Regiona l Industries ................................ ........................ 40 Results ................................ ................................ ................................ .................... 41 Water Intensities ................................ ................................ ............................... 41 Energy Intensiti es ................................ ................................ ............................. 42
6 GHG Emission and Value Added Intensities ................................ .................... 42 Regional EIO LCA Models ................................ ................................ ............... 43 Emergy Evaluation ................................ ................................ ........................... 43 Discussion ................................ ................................ ................................ .............. 44 Regional Resource Intensity Vectors ................................ ............................... 44 Direct and Indirect Resource Use and Emergy Evaluation ............................... 45 Environmental Impacts ................................ ................................ ..................... 48 3 DEVELOPMENT OF A LAN D USE OPTIMIZATION MODEL ................................ 56 Introduction ................................ ................................ ................................ ............. 56 Methods ................................ ................................ ................................ .................. 56 Water and Carbon Balance Models ................................ ................................ .. 56 Linear Programming Optimization Models ................................ ........................ 57 Land Use ................................ ................................ ................................ .......... 58 Water Balance Model ................................ ................................ ....................... 58 Environmental water flows ................................ ................................ ......... 60 Economic water flows ................................ ................................ ................ 61 Economic Model ................................ ................................ ............................... 62 Economic Linkages ................................ ................................ .......................... 62 Combined Economic/Ecological Land Use Model ................................ ............ 64 Linear Optimization Model ................................ ................................ ................ 65 Goal function ................................ ................................ .............................. 66 Variables ................................ ................................ ................................ .... 67 Constraints ................................ ................................ ................................ 67 Results ................................ ................................ ................................ .................... 71 Land Use ................................ ................................ ................................ .......... 71 Water Balance Model ................................ ................................ ....................... 71 Land Use Models ................................ ................................ ............................. 72 Optimization ................................ ................................ ................................ ..... 73 Sensitivity Analysis ................................ ................................ ........................... 74 Shadow Pricing of Ecosystem Services ................................ ........................... 75 Discussion ................................ ................................ ................................ .............. 76 Linearizing Land Uses ................................ ................................ ...................... 76 Accounting for Ecosystem Services ................................ ................................ 78 Choosing Goal Functions ................................ ................................ ................. 79 Static Economic Structure ................................ ................................ ................ 81 4 ENERGY AND CARBON BALANCE IN A REGIONAL MODEL ............................. 94 Introduction ................................ ................................ ................................ ............. 94 Methods ................................ ................................ ................................ .................. 95 Energy Use of Regional Land Uses ................................ ................................ 95 GHG Emissions of Regional Land Uses ................................ ........................... 96 Definition of New Land Uses ................................ ................................ ............ 97 Definition of Energy and Emissions Constraints: ................................ .............. 99 Results ................................ ................................ ................................ .................. 100
7 Land Use Models ................................ ................................ ........................... 100 Initial Optimization ................................ ................................ .......................... 102 Greenhouse Gas Constraint ................................ ................................ ........... 103 Fossil Fuel Constraint ................................ ................................ ..................... 104 Optimization with Combined Constraints ................................ ........................ 105 Comparison of Constraints ................................ ................................ ............. 106 Discussion ................................ ................................ ................................ ............ 106 Impact of Sustainability Constraints ................................ ............................... 107 Impact on Population ................................ ................................ ...................... 108 Impact of Renewable Energy Land Uses ................................ ....................... 109 Regional Implications ................................ ................................ ..................... 110 5 EVALUATING THE MODEL ................................ ................................ ................. 122 Contributions of the Model ................................ ................................ .................... 122 Limitations of the Model ................................ ................................ ........................ 126 Conclusion ................................ ................................ ................................ ............ 128 ABSTRACTS A RE GIONAL WATER ALLOCATION ................................ ................................ ...... 129 B REGIONAL ENERGY ALLOCATION ................................ ................................ ... 137 LIST OF REFERENCES ................................ ................................ ............................. 150 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 156
8 LIST OF TABLES Table page 2 1 Directly allocated water intensities for Florida ................................ .................... 49 2 2 Directly allocated water intensities for the Peace River region ........................... 50 2 3 Comparison of electric power generation energy intensities .............................. 51 3 1 Peace River land use areas ................................ ................................ ................ 82 3 2 Water outflows from environmental inputs ................................ .......................... 83 3 3 Water inflows from economic use ................................ ................................ ....... 84 3 4 Water outflows from economic use ................................ ................................ ..... 85 3 5 Land use inputs to the Peac e River optimization model ................................ ..... 90 4 1 Carbon sequestration estimates of Peace River region land uses ................... 113 4 2 Energy and GHG emission characterization of existing land uses ................... 114 4 3 Characterization of alternative energy land uses ................................ .............. 11 5 4 4 Impact of GHG emission constraint on output and shadow price ..................... 116 4 5 Impact of fossil fuel constraint on regional output ................................ ............. 118 A 1 Florida water use adapted from Marella ................................ ........................... 131 A 2 ................................ ...... 132 A 3 Average water consumption by sector ................................ .............................. 134 A 4 Peace River region water withdraws ................................ ................................ 135 A 5 Peace River region water use correction factors ................................ .............. 136 A 6 Peace River region wastewater disposition ................................ ...................... 136 B 1 Florida energy use 2002 adopted from SEDS ................................ .................. 143 B 2 Comparison of SEDS and FOKS energy values ................................ ............... 144 B 3 Adjustment of Florida gasoline use ................................ ................................ ... 145 B 4 Average highway fuel use adopted from TEDB ................................ ................ 146
9 B 5 Modification of highway fuel use with VIUS data ................................ .............. 147 B 6 Florida truck fuel use from 2002 VIUS ................................ .............................. 148 B 7 ................................ .............. 149
10 LIST OF FIGURES Figure page 1 1 Peace River wate rshed and region ................................ ................................ ..... 26 2 1 C omparison of regional direct and indirect water use ................................ ......... 52 2 2 Comparison of regional direct and indirect energy use ................................ ....... 53 2 3 Emergy v alues for Peace River regional industries ................................ ............ 54 2 4 Composition of emergy values for Peace River regional industries .................... 55 3 1 Water i nflows for regional land uses ................................ ................................ ... 86 3 2 Water outflows for regional land uses ................................ ................................ 87 3 3 Percent change in monetary output when account ing for economic linkages .... 88 3 4 Percent change in resource intensities due to accounting for economic linkages ................................ ................................ ................................ .............. 89 3 5 Change in land use distribution for regional optimization ................................ ... 91 3 6 Sensitivity analysis of the groundwater recharge constraint ............................... 92 3 7 Total regional production and shadow prices for ecosystem services ................ 93 4 1 Sensitivity analysis of optimized land use distribution under GHG constraints 117 4 2 Sensitivity analysis of optimized land use distribution under fossil fuel constraint ................................ ................................ ................................ .......... 119 4 3 Comparison of initial, optimized, and fully constrained land use ....................... 120 4 4 Impact of sustainability constraints on regional output ................................ ..... 121
11 LIST OF ABBREVIATIONS B EA United States Bureau of Economic Analysis CO 2 Carbon dioxide CH 4 M etha ne D OE Department of Energy E FA Ecological Footprint Analysis E IA Energy Information Agency E IO LCA Economic Input Output Life Cycle Analysis E PA Environmental Protection Agency E T E vapotranspiration F DEP Florida Department of Environmental Protection F LU CS Florida Land Use Characterization System F OKS Fuel Oil and Kerosene Sales G DP Gross Domestic Product G HG Greenhouse Gas I MPLAN IMpact analysis for PLANing I O Input Output I PCC International Panel on Climate Change I SO International Standards Organizatio n L CA Life Cycle Assessment L CI Life Cycle Inventory M CDA Multi Criteria Decision Analysis M EA Millennial Ecosystem Assessment M TCO 2 E Metric Tons of Carbon Dioxide Equivalents N AICS North American Industry Classification System
12 N ASS National Agricultural S tatistics Survey N PP Net Primary Production N 2 O D initrogen oxide P RCIS Peace River Cumulative Impact Study S EDS State Energy Data System S ES Social Ecological System S FWMD South Florida Water Management District S IT State Inventory Tool S JRWMD Saint John S WFWMD Southwest Florida Water Management District S WUCA Southwest Water Use Caution Area T EDB Transportation Energy Data Book U EV Unit Emergy Value U N United Nations U SGS United States Geological Service V IUS Vehicle Inv entory and Use Survey
13 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 MAXIMIZING SUSTAINABLE REGIONAL PRODUCTION: B ALANCING WATER, ENERGY, AND CARBON By David Aaron Pfahler May 2013 Chair: Mark Brown Major: Environmental Engineering Sciences Regional sustainability is a question of increasing i mportance for regional planners with water and energy use and the produc tion of greenhouse gas es (GHG) emerging as chief constraints on achieving sustainability in the long term. Modeling the sustainability of regions is difficult as one must predict the impacts of cascading interactions within complex social ecological system s. In this dissertation a combination of Economic Input O utput Life C ycle Assessment and linear optimization methods was explored as a way to model regional sustainability. The Peace River region of central Florida was used as a case study for the model. T he goal function of the optimization wa s to maximize regional economic output measure in both monetary and emergy terms. The model optimized regional production under water, energy, and GHG emission constraints by changing t he area devoted to individual la nd use s Each of the constraints was tested separately and in combination. The results of the model showed that changing the mix of land uses could potential ly provide a 1.8% increase in economic output, and a 6.2% increase in the supported population whi le maintaining groundwater and flood storage constraints In addition, it was also shown that by including renewable energy land uses in the regional
14 analysis a 20% reduction in greenhouse gas emission and a 20% reduction in fossil fuel could be achieved and still provide a 1.3% increase in economic output and a 2.5% increase in regional pop ulation. This result suggests that increased sustainability for the region is attainable, and highlights solar photovoltaics and bioethanol from water efficient sorghu m as key technologies for the region to pursue to achieve these goals. This study make s several steps toward a better integration of regional sustainability modeling. The accounting of direct and indirect changes in resource consu mption within regional lan d uses, the linearization of land uses in the optimization model, and the calculation of shadow prices for ecosystem services for the region are novel capabilities that can be used to give decision makers increased insight into how development decisions im pact the regional system as a whole.
15 CHAPTER 1 MODELING REGIONAL SUSTAINABILITY Introduction Regional Sustainability In 1987 the UN Brundtland Commission published its report Our Common Future 1 in which it stated that global societies should strive toward sustainable deve development that meets the needs of the present Translating global initiatives to local levels remains a key issue for achieving sustainable patterns of humanity and environment since wit hout sustainable action at all scales, the overarching global goal is not possible. Of particular importance is sustainable action at intermediate (regional) scales, since personal and national sustainable action seems more easily accomplished. At the regi onal level, managers and policy makers are assigned the task of trying to sustain prosperity and at the same time limit interacting parts, limited data availability and hi gh uncertainty of regional systems The question of how to develop sustainable human societies may be one of the most pressing issues facing humanity today. The human population has grown exponenti ally over the last century 2 as has human consumption of natural resources 3 resulting in the fact that today, h umans have become one of the greatest forces drivi ng change in the biosphere 4 Several research groups have develope d sustainability indicators which reveal that many nations of the world have exceeded sustainable levels of consumption of the eart 5 At the same time, environmental indicators reveal th g rapidly
16 degraded. 6 As the earth nears limits to human carrying capacity, answering questions of efficient resource distribution, sustainable scale, and just d istribution take on new urgency. 7 While sustainability questions are being actively addressed with research and indicator development at the global scale 8 few tools exist at the local and regional scale. At these smaller scales, different pictures of sustain ability can emerge since regions vary dramatically in their available ecological resources, population level s and consumption patterns. Some regions may already be over their local carrying capacity, while other regions may still have capacity for growth. If regions are to negotiate a sustainable pattern in the coming years they will need appropriate tools to identify environmental constraints, and reveal which development options increase long term productivity, and which will decrease long term productiv ity. Regional Scale Methods for measuring sustainability developed at the global and national scale have been applied at regional level as well. 9 But while these methods can provide a measure of the s ustainability of the region, they do not answer some of the critical questions managers and policy makers need to understand about the path a region should follow to become more sustainable. Me thods are needed at the regional level that can compare various development options, and that can suggest a development path that leads to a productive yet sustainable system. Regional managers need to be able to understand the tradeoffs the region faces i n making sustainable development decisions. Most current sustainability analysis methods are far more descriptive than they are prescriptive.
17 Measuring Sustainability To determine if a region is sustainable, data is needed on the flow s of different kinds o f resources, and the environmental impacts of those flows. Environmental acco unting methods have been developed to try to account for all the flows within complex socio ecological systems. Life Cycle Assessment methods have been used to track both resource flows and their associated impacts within systems. 10 Often they are used to provide sustainability comparisons between competing development options. A combination of Life Cycle Analysis with Economic Input Output modeling (EIO LCA) has resulted in models of national resource use and th eir associated environmental impacts. 11 EIO LCA models have been developed at the national and state 12 and the multi state level 13 but they have not been developed below the state level due to difficulties in collecting all the required data. While these models are very successful at tabulating resource flows and pollution impacts generated within an economy, they do not integrate all the environmental, economic, an d social criteria into a sustainability model or metric. While t hey are able to account for interact ions between economic industries, environmental impacts that are produced There are two broad approaches used to try to integrate all the environmental economic, and social aspects to determine the sustainability of socio economic systems The first approach is to try to put all resources o n a commensurate scale. Emergy S ynthesis and Ecological F ootprint Analysis (EFA) integrate the various aspects of sustainability by expressing production and absorption capacity in common terms. A second approach to integration is to develop a non commensurate set of sustainability c riteria, and the n assign a weight to each criterion in a combined
18 index. Sustainability analysis using Multi Crite ria Decision Analysis (MCDA) models relies on this second approach Emergy S ynthe sis Emergy accounting evaluates all resource flows on the basis of t he equiv alent energy of solar energy which was required to generate the resource flow 14 This embodied energy basis allows emergy to account for both ecological and economic aspects of systems in commensurate terms. Emergy synthesis is a valuab le tool in that it can be applied at any scale. It has been applied to measuring and comparing the sustainability of systems at national 15,16 state 17,18 and regional watershed 19,20 levels An Emergy Sustainability Index (ESI) has been defined as an integrated metric to evaluate the sustainability of systems. 21 The index accounts for the level of system pr oduction that is attained by importing emergy sources as well as the associated load pla ced on the environment. The ESI accounts for both the level of production of the system and the environmental impacts in a si ngle aggregated metric. T he ESI can give va luable insight into how far away from lo ng term sustainability a system is currently. However, the synthesis process does not provide tradeoffs between different organizations of the region In addition, emergy synthesis relies on highly aggregated regiona l flows, and has a unique algebra that requires careful analysis to avoid dou ble counting. 22 The more regional flows that are included in the regional model, the more difficult it can be to track emergy flows without double counting. Ecological F ootprint Analysis Ecological F ootprint Analysis was ini tially introduced in the 1990s as a way to cal culate the carrying capacity of social ecological systems 23 EFA selects land area as the integrating factor in its sustainability analysis. It calculates t he land area required to
19 provide the resource flows required for the population of a region, as well as provide all the land area requi red to dissipate wastes. EFA compares the total land area required per capita to support a region to the global average of productive land per capita Systems that are above the global carrying capacity are considered beyond sustainable limits. This methodology puts s ustai nability into terms of the amount of land area that is being used to support the system which is easily understandable for most people T he method has been applied widely to socio ecological systems at vario us scales, including global 24 national 25,26 and regional 9,27 scales. The EFA approach integrates ecological limitations i nto its evaluation, but it faces the same shortcomings as emergy synthesis in that it does not give regional managers insight into potential development paths. It helps the manager know where the system is relative to a sustainable level, but does not help regional managers know how to more efficiently allocate reso urces Multi Criteria Decision A nalysis The second approach to address ing the problem of dispa rate sustaina bility metrics is to combine multiple metrics using a weighting scale. This approach can be used with MCDA to perform sustainability analysis 28 MCDA is able to combine dispa rate measures of economic, environmental, and social sustainability goals as cri teria in an optimization model. The inclusion of an optimization model in MCDA allows for analysis of different resource allocations and development scenarios making it a much more forward looking approach than other sustainability analysis methods Sever al MCDA models have been built to address sustainability issues at the regional scale 29 31
20 However, in order to integrate the various sustainability criteria, MCDA relies on a priori weighting of the criteria to generate a value function that is then maximized in the optimization 32 E ither regional stakeholders or modeling experts must make the determination of which sustainability metrics are most important befor e the model is run, or what minimum levels they must achieve within each criteria What is then actually maximized in the model is the value function that was d efined by the researchers by weighting the various sustainability measu res. This ranking of which goals are most important defines what the end goal of the system should be, and the optimization finds a solution that is closest to that goal. The drawback of this approach is that not all of the sustainability measures are necessarily simultaneously satisfied. The weighting of the value function determines which sustainabil ity measures will be met, and which will not. Maximum Sustainable Production Regional modeling methods are needed that are simple enough for regional managers to use and adapt to their particular regio nal level. The models should address gaps in data acco unt for the interdependencies of social, economic, and environmental factors, and be able to model the trade offs between these factors. A novel combination of modeling techniques is introduced in this dissertation in order to address shortcomings of previ ous sustainability modeling. One of the first questions to be answered in order to construct such a modeling tool is the goal for the system. While sustainability is a desired outcome for the system, sustainability on its own is not a metric for human welf are. Minimizing the human population might achieve system sustainability, but if high value is placed on human society, this is a less desirable solution than other options. The overall goal for this modeli ng effort is defined as maximizing the long term p roduc tion of the system.
21 Maximum sustainable production, then, requires that the level of production can be maintained for future generations and should be considered as an average production level over seve ral decades to allow for long term variation in e nvironmental variables and natural disturbance cycles. Defining and Valuing P roduction All social ecological systems provide multiple products as output. Not all products are equally important or desirable, and so some method of valuing system products is needed. In this dissertation, t wo methods are used to evaluate regional production The first uses monetary value as measured in the price paid for all goods and services produced A common measure of a social production is the Gros s Do mestic Product (GDP) and it is the primary valuing tool used today in social systems A second valuation measure is the emerg y evaluation method 14 Emergy is a donor based value system proposed by H.T. Odum to capture both economic an d environmental value s It is a quality corrected measure of the available energy that was required to make something, expres sed in common units of solar emJ oules System em power is a measure of the total emergy flowing through a system per unit time, and it can be used as a meas ure of total system production. Defining S ystem B oundaries As in any system analysis, the boundaries of the regional system must be defined. There exist many criteria that could be used to choose a regional system boundary. In this dissertation research a clo se match between physical and social system boundaries was needed. The surface watershed was chosen as the boundary for the physical system The boundary for the social system was chosen to match the watershed boundary as closely as possible using political county boundaries that
22 approximate d the watershed area. This wa s a convenient boundary because both the physical and economic data required for a regional system model are often reported at the county level. Limiting Factors N o model can completely capture all the interactions of a complex system. A good model economizes by selecting the most important factors, and may ignore secondary factors. For a model of system production, factors that are likely to be limiting are conside red most important. Three critical resource categories that require trade offs between economic and ecological production are chosen for this model: water, energy, and land area. In addition, the regulation of carbon as a greenhouse gas is modeled as a glo bally important ecosystem service that may be regulated in the future and comprise a part of system constraints. Energy is a fundamental input into all production processes. For change to happen in a system, energy must be supplied. Energy flows within th e environmental system are derived from the energy that drives the biosphere. Energy can also be obtained from outside the system in exchange for goods and services produced within the system. Current global energy supplies and economics allow this to happ en at a net energy gain for many regional systems, and this strategy is heavily employed to raise a mix of a region, and the future availability of energy sources are i mportant considerations in determining a maximum sustainable production of a region. Water is a critical material component of almost all production processes. In many regions of the world, water is already a limiting factor in social ec ological systems 33 Water, however, is rarely imported from outside a watershed because of
23 high transport cost and high political barriers. Reg ions often instead draw on available water stored in underground aquifers. This water is recharged slowly, an d pumping it faster than recharge results in dropping water levels. Water has the potential to become even more constraining for regions in the future as thes e storages of groundwater are extracted faster than they are replenished. Water in the system is c onsidered to be consumed once it leaves the boundaries of the system, either by evapotranspiration, river flow, groundwater flow, or embodied in physical products that are exported. Land area can also be a limiting resource in a region. Regions need to und erstand and manage the tradeoffs between regional growth and consequent land use changes. Land use change alters the level and mix of ecosystem services that are provided by a landscape. Ultimately long term sustainability will require that regions provide production limiting ecological resources and services from within instead of relying mainly on outside sources or energy intensive replacements driven by non renewable resources. Therefore, achieving a maximum sustainable level of regional domestic produc tion will require finding a balance between the production of the of ecosystem services that come from that capital. Greenhouse gas emission and carbon sequestration are also considered within the model. The greenhouse gases considered are carbon dioxide (CO 2 ), methane (CH 4 ), and dinitrogen oxide (N 2 O). Carbon is not considered a limiting materiel in regional production, but rather its storage represents an environment al service that is provided by the region to the global system. The storage of carbon helps to control greenhouse gas concentrations at a global scale. The impacts of this sequestration are
24 not confined to the region in question, but impact the wider biosp here and have far reaching consequences for climate regulation. Allocation of globally important ecosystem services to regions is a matter of current debate, and includes questions of just distribution of resource consumption. The model deals with allocati on by allowing for external policy directives to set allowable emission levels. The research in this dissertation sought find and maintain the maximum sustainable regional product by developing an input/output model of a regional social ecological system t ha t allows the production of that region to be explored in terms of its li miting factors. The model identified the maximum sustainable regional product based on an extrapolation of the current system organization and then was used to predict changes to t hat maximum level due to chan Energy and water provision were selected as the most important limiting factors of production within the model. Greenho use gas (GHG) emissions were also considered as a system waste product that ha s impacts beyond the scale of the region itself. These three factors were used to develop sustainability constraints for the region. Study A rea This study was focused on the Peace River watershed in southwestern Florida. This watershed has several charact eristics that make it an interesting case study. The Peace River watershed lies within the Southwest Water Use Caution Area (SWUCA). SWUCA is a 5,100 square mile, eight county area in southwest Florida where depressed aquifer levels have caused saltwater t o intrude into the aquifer along the coast, contributed to reduced flows in the upper Peace River, and lowered lake levels in portions of Polk and H ighlands counties 34 Development within the region has tapped available water resources to the point where restrictions have been issued on water
25 supply and wat er quality has been adversely affected. The question of the sustainability of current water use is actively being addres sed in the region 35 The Peace River also h as significant areas of phosphate mining within its boundaries. By state law, phosphate lands mined since 1975 must be reclaimed after the mining and dewatering process is complete. T his represents over 10% of the land area that coul d shift its land use in the future. Developing this reclaimed land with appropriate land uses could potentially improve the long term sustainability of the system. Conclusion This chapter has laid out the purpose of the regional modeling effort, given an overview of the applicab le modeling methods, and discussed some of the assumptions that must be made in the creation of a regional sustainability model. Chapter 2 develops a regional Economic Input Output Life Cycle Analysis (EIO LCA) model that predicts changes to energy and wat er consumption and GHG emission due to changes in economic activity within the region. Chapter 3 develops a regional land use optimization model that incorporates the EIO LCA data developed in Chapter 2. The optimization model is used to maximize the regio constraints. Chapter 4 applies the resulting optimization model while considering future maximum sustainable production i s evaluated. The goal of this effort is to make steps toward a better integration of ecological and economic modeling, with the hope that it will give decision makers increased insight into how development decisions impact the regional system as a whole.
26 Figure 1 1 Peace River watershed and region Available from http://www.swfwmd.state.fl.us/waterman/lakehancock/img/peace river.jpg
27 CHAPTER 2 DEVELOPMENT OF A REGIONAL EIO LCA MODEL Introduction This chapter outlines the construction of a regional E IO LCA model for both the state of Florida and for the Peace River region within Florida. The model is built using a combination of a commercially available regional economic input output model and publicly available data on regional water, energy, and gre enhouse gas emissions. The result of the Florida and Peace River regional model s is a list of resource intensities for all the model industries in the regions. The regional EIO LCA model is used to calculate direct and indirect input requirements for all t he industries in the model. This input information is then used to calculate an emergy signature for each industry within the model, accounting for both direct and indirect resource consumption. Methods Economic Input Output Life Cycle Assessment Environm ental accounting techniques have been developed to track flows of resources through economic network interactions. Life Cycle Assessment (LCA) is concerned with quantifying both resource inputs and environmental impacts of waste products for the entire lif e cycle of a product, from cradle to grave. 10 A Life Cycle Inventory (LCI) traces material and energy flows used as inputs in a production chain. Two LCA methods have been developed and widely applied. Process based LCA is a bottom up method of accounting. It relies on an in depth knowle dge of the inputs used in a production process, and then links these inputs back to their own production processes to form an entire production chain. It relies on large databases, and requires a boundary to be set on how far back the supply chain will be followed. Economic Input
28 Output Life Cycle Assessment (EIO LCA) is a top down asses sment method. 36 It uses an economic input output (IO) model to capture information about all the purchases r entire economy, it accounts more completely for resource flows. However, it suffers from accuracy in that it must aggregate industry production functions and use averag ed resource flows. EIO LCA models have previously been developed at the national and also at the state level. 12 However, this type of model has yet to be applied at the watershed scale. EIO LCA models are based on economic input output models which were originally developed by Wassily Leontief in the 1930 s. An IO model separat es the economy into a set of industry sectors by aggregating businesses with similar end products. A transaction matrix is developed that uses these sectors as row and column headings. Financial transactions are represented as row industries selling produc ts to column industries. Reading across a row gives the sales of go ods and services from one industry to all other industrie s in the economy, while reading down a column gives all t he purchases made by that industry in order to produce its product. Several columns are added beyond the transaction matrix to represent final demand, the goods and services that are sold to end consumers who are not producing additional goods and services. Final demand categories include the end consumption of the population, go vernment organizations, and exports. Several rows are added below the transaction matrix to represent value added transactions which include the purchasing of labor, payment of re nts and taxes, and profits. The transaction matrix, Z captures the flows of money paid for goods and services throughout the economy for the year of the data.
29 Leontief recognized that if the production relationships in the matrix were assumed to remain linear, a predictive model could be developed that would calculate the change in total economic activity that would occur for any given change in the final demand for a product. The first step in creating the model is to define the technical coefficient matrix, or the A matrix. The A matrix is defined by starting with the transactio n matrix Z and dividing each column in the transaction matrix by its respective column total as shown in equation 2 1 (2 1) In this equation a ij is the technical coefficient, z ij is the transaction, and x j is the column total. Reading down a column of the A matrix gives the production function for a particular column industry, showing how much it must purchase from each industry in the matrix the A matrix reveals the sales output that row industry must provide to each industry in the matrix If there is an increase in final demand for a product in the economy, the in dustries that make that product must produce that additional final demand and their suppliers must produce more of the goods and services required to supply the industry producing final demand. C hanges to the final demand of the whole economy can be repres ented as a vector, f The additional production required to meet the new final demand is then ( I x f ) where I is the identity matrix. The additional production of the direct suppliers needed to supply the final demand producing industries can be represent ed as ( A x f ), and t he total change in economic activity, x is then x = ( I + A ) f This equation captures the change in output due to the addition al final demand and the additional production of its first level suppliers, also called its direct inputs. How ever,
30 there are also indirect inputs to consider. The suppliers of the suppliers must also provide more output, and so the effect of the new final demand ripples through the network of the economy causing additional production in each successive level of s uppliers. The total direct and indire ct requirements are given by equation 2 2. (2 2) In matrix algebra, this expansion series is equal to the quantity ( I A ) 1 which is termed the Leontief matrix. The change i n output is expr essed by equation 2 3. (2 3) Input output modeling is based on several simplifying assumptions that are important to understand, be cause they reveal the approximate nature of the model: production functions are linear; this is to say if any additional output is required, all inputs will increase in linear proportion. This is a good approximation over a limited range of change, but large changes in production may introduce different efficiencies of scale. No supply constraints: The model must assume that an industry has unlimited access to raw materials, and that these materials are available at close to the same price s output is limited only by the demand for its products. Homogeneous sector output : The model assumes that the proportions of commodities that the model industries produce r emain the same, regardless of proportionately increasing the output of all its other products. Industry technology assumption : This assumption comes into play when data is collected on an industry by commodity basis and then converted to industry by industry matrices. The model assumes that an industry uses the same technology to produce all its products. In other words, an industry has a primary or main product and all other product s are byproducts of the primary product. The economic IO model is based on monetary data, and these monetary flows are related through price to the flow of physical goods in the economy. Several research
31 groups have developed methods of including physical flows in IO models. The EIO LCA model developed by Hendrickson et al. 37 includes physic al flows by supplementing the I O matrix with a n external set of resource intensity vectors. Resource intensities are defined by dividing the independently reported resource use of an industry by the monetary output of that industry, and so they have units of resource/$ output. T his operation is performed for each industry in the mo del r esulting in the resour ce intensity vector, r shown in equation 2 4. (2 4) The symbol r i is used to denote the resource use in sector i and x i is the total dollar outpu t for sector i A vector of the total resource use, b can be obtained by multiplying the total economic output by the resource in tensities as shown in equation 2 5. (2 5) R is a matrix with the elements of the vector r along the diagonal and zeros elsewhere, and x is the vector of relative change in total output based on an incremental change in final demand. This methodology can be used for both input resources such as water and energy use, or for waste products such as GHG emissions. The EIO LCA framework can be used to track resource use in the economy, and can be used to predict the change in reso urce requirements as the demand changes. In addition to the assumptions of the ec onomic input output model, the EIO LCA model assumes that the resource/dollar output ratio remains constant as production scales up and down, and that no resource substitution is carried out. These simplifying assumptions allow an estimate of the total res ource requirements associated with
32 Economic Model The regional economic model wa s created using a commercial economic impact ( IMPLAN ) soft ware. IMPLAN uses as its base the national benchmark input output (IO) models created every five years by the U.S. Bureau of Economic Analysis (BEA). In order to create regional economic models, IMPLAN uses the national technical coefficients matrix and co mbines that matrix with regional industry output totals and regional value added data, which include labor costs and taxes. This modeling method proportional to the purch ases of the corresponding national level industry. Since production functions are rarely available for regional industries, this appr oximation is necessary to complete the model, and is generally considered to be acceptable because the technology to produc e a given set of goods or services is likely to be fairly industry purchases. The IMPLAN transaction matrix and its associated databases are used as the economic model for the regional EIO was used which divides the economy into 509 input output industries. Two models were created within the software, a n economic model of Flor ida, and a model of the four county area that encompasses the Peace River watershed, and includes Polk, Hardee, De Soto, and Charlotte counties Resource Consumption Data Resource consumption data for the regional economy wa s obtained from reports by stat e a nd federal agencies. This data wa s reported at various levels of aggregation,
33 so a mapping of categories was created between the 509 industries of the IMPLAN model and the reported r esource consumption categories so that resource use could be allocated across all industries. Development of water intensity vectors Water intensities were previously developed for 428 industries at the national level for the EIO LCA database 38 The authors of this study pointed out that these values are national averages, and should not be used for regional analysis because of the potential for large differences in water requirements between different regions. Using regional water use data, water intensity vector s were developed for both Florida and for the Peace River region. The 2000 USGS report on state wide water use 39 was used as the data source for the Florida regional mo del. Total water use by state has been estimated by the USGS every five years from available data sources, which ter management districts, water use permits, and water utility reports. The USGS data wa s reported in more highly aggregated categories than the regional economic data, and a methodology was developed to allocate it to the individual industry level The al location method proceeded in several steps. First, water use was assigned directly to any industry that wa s reported at the same level of aggregation as the economic industry. For the state of Florida, direct assignment of water use was made for electric p ower generation, 14 agricultural industries, 4 mining industries, 1 recreational industry, and also for residential use. The remaining water use was assigned to the aggregated sectors reported by the regional data sources, and then allocated within those s ectors to individual industries using the same allocation proportions observed in the national EIO LCA model. The calculation s employed in the allocation method are explained in greater detail in Appendix A.
34 A water use intensity vector was also developed for the Peace River region. Water use data for 2002 was obtained from the Southwest Florida Wate r Management Estimated Water Use Report 40 The water management district reports on water use only within its boundaries, while portions of Polk and Charlotte county lie outside the SWFWMD Correction factors were developed in order to scale water use to represent the full county area These factors were created by comparing year 2005 county level data obtained from the USGS Florida report 41 with year 2005 dat a from the SWFWMD 42 Data for Polk and Charlotte coun ties for 2002 was then scaled by these correction factors to estimate total water use in the counties. Within the Peace River region, self supplied agricultural, mining, citrus processing, phosphate mining, power gen eration and residential water use were matched directly with their corresponding industries. Publicly supplied water and the remaining self supplied water were allocated among the remaining industries using the same methodology developed for the Florida water use intensity vector. Water use for each industry was then divided by the economic output of each industry to create the water use intensity vector. I n addition to total water use intensity, inte nsity vectors were developed for both groundwater and surface water use. Since not all the water that is withdrawn from the environment is considered to be consumed within the economic act ivity to which it is allocated, a consumed water intensity vector was also developed. Water that evaporates during use or is embodied in products in the course of economic use is considered to be consumed as it is diverted from any further surface and groundwater interactions. water is returned to the regional system as wastewater
35 flows. Florida state averages of water consumption for various econo mic activities were used to calculate consumed water intensities for the region 39 Development of e nergy intensity vectors National level energy intensity vector s were previously developed by the Carnegie Mellon Un iversity Green Institute 12 These energy intensity estimates were made b y allocating data on U.S. fuel use to 428 industry categories. Directly reported data was used wherever possible, and economic allocation was used when further allocation was required in acco rdance with ISO 6000 standards. For this effort, a n energy intens ity vector was developed for the state of Florida. total energy u se in the power, industrial, commercial, transportation, and residential categories 43 Available public data sets were then used to allocate this energy use among the 509 industry categories of the regional economic model. Energy use was directly allocated for all industries in which data was available for the allocation These industries include d the electric power generation, transportation, agric ulture, mining, and residential sectors. Whenever data wa s reported at a more aggregated level than the individual industries within the economic model, it wa s first allocated to the lowest aggregated sector possible. Energy use within these aggregated sec tors wa s then allocated to individual industries in the same relative proportion as the allocation of the national energy intensities within that same sector Finally, the energy input for each industry was divided by the economic output of the electric po wer industry to give the en ergy intensity Energy consumed for electric power g eneration wa s the largest energy use categor y 43
36 Power Annual State Data Tables 44 gas, petroleum fuels, biomass and waste, nuclear, and other alternate energy sourc es. Private and government power generation and distribution industries were combined to form a single electric power industry for the model Transportation wa s the second largest category of energy use, accounting for was obtained from the EIA SEDS database 43 In the SEDS database, all fuels used throughout the state for transpo rt and mobility applications were included within the transportation category, including fuel used by non transport industries to pr ovid e their own transportation, as well as fuel used for residential transport. The national econom ic model, on the other hand, had ten transportation industries that include d only the industries that provide transport as a paid service to other industries In order to allocate transportation fuels to the economic transport industries, a methodology was developed using three different fuel use data sets. First, fuel use was allocated based on fuel type 45 data set Allocations were made from this data set to the air, rail, and water transport industries. Next, highway fuels (TEDB) 46 47 The TEDB was used to allocate fuel use by vehicle type, and the VIUS was then used to allocate truck vehicle types to specific industries. Energy intensities were developed for the truck transport, and ground passenger transport industries from this data. In addition, residential transporta tion fuels were allocated to the residential sector. All remaining
37 transport fuel use was allocated to the commercial and industrial sectors. The details of these alloca tions are included in Appendix B Energy intensity vectors for Florida agricultural in dustries were developed using the 2002 Agricultural Census 48 The census reported energy expenditures for each state, using the North American Industry Classification System (NAICS), which was then mapped to the IMPLAN industry categories. In the 2002 census, all fue l expenditures were reported in one aggregated category. However, in the 1997 Agricultural Census 49 fuel costs were reported in separate categories for natural gas, gasoline, diesel, and liquefied petroleum gas. The 1997 data was used to estimate 2002 fuel consumption for each fuel type by assuming that the percentage of expenditures on each fuel remained constant across both years. Fuel consumption for 2002 was then calculated by multiplying the resulting 2002 fuel expenditures by 2002 fuel prices 50 for industrial uses in Florida for 2002 Residential energy consumption wa s reported directly in the EIA SEDS database for fuels delivered directly to households 43 The transportation fuels that remained after industry fuel use allocation were also allocated to the residential sector. In contrast with other sectors, the energy intensity for the res idential sector was calculated in units of assumed to scale linearly with population. Energy intensities for Florida mining operations were developed using USGS data on the physical amount of min era ls mined in Florida 51 and data on the national average mining fuel use per unit of mined mater iel as reported by NAICS category for several mining industries in the year 1997 52 These energy intensities per ton of mined
38 material were assumed to remain constant over time and were used to estimate the energy use for three major Florida mining industries: phosphate, limeston e and sand, and mineral mining. All remaining fuel use was allocated either to the industrial or the commercial sector. The industrial sector included agriculture and mining services, forestry, utilities other than electricity generation, construction, an d manufacturing. The commercial sector included wholesale, retail, professional services, and government industries. No data was available at the state level to develop energy intensity measures for the se use was then allocated in proportion to the allocation of the national energy intensities within t hese sectors. as there were no data sets available for energy allocation at t he county level w ithin Florida. The only exception was within the electric power generation industry. Data on power grid database 53 Data from the year 2000 was used as no data was available for 2002. The commercial generation facilities that lie within the Peace River region w ere selected from the database Industries that self supply electricity were excluded from the power sector calculations, as they were considered to consume all the electricity they generate marketed electric power supply. The Florida Development of a GHG emission intensity vector GHG emission intensities were developed fo Inventory Tool (SIT) 54 The SIT was developed to assist s tates in performing a
39 standardized reporting of greenhouse gas emissions based on the International Panel 55 The tool consists of a series of Excel spreadsheets that accept data inputs on state fuel use and industrial producti on activities. The tool calculate s the emis sions of differe nt greenhouse gases, including carbon dioxide (CO 2 ), methane (CH 4 ), and nitrogen dioxide (N 2 0) emissions which were the greenhouse gases considered for the regional model. The SIT provides default data by state for energy consumption, emission activities, and GHG estimated emissions were tabulated using the default GHG emission rates specified for the state of Florida for the year 2002. from the regional EIO LCA model was multiplied by the default e mission rate to arrive at the overall emissions for the industry. T he emissions were then divid ed by the economic output of each industry to calculate the emission intensity vector Four emission intensity vectors were developed, one for carbon dioxide (CO 2 ), methane (CH 4 ), and nitrogen dioxide (N 2 0) emissions, and one for total GHG emissions. All emissions intensities were reported in units of metric tons of CO 2 equivalents (MTCO 2 e). Only the emissions associated with fuel use were included in the regional EIO LCA model, land use emissions were accounted for separately within the optimizat ion model developed in Chapter 3 Development of a value added intensity v ector A value added intensity vector was developed for Florida and the Peace River region s Value added measure s the value of indus try output paid directly to human input s in the production process Value added categories in cluded in the model were labor payments, profit for owners, an d rent income from the IMPLAN database 56 This total value added for each industry was divided by the total economic output of the
40 industry to give a value added intensity in units of dollars of value added per dollar of economic output. This vector is later used in the calculation of indirect labor inputs in each industry. Model C onstruction The economic data for each region was combined with the corresponding set of regional resource intensity vectors to create the input tables for the regional EIO LCA model. The model was constructed as a macro running in an Excel spreadsheet The with in Matlab and then imported into Excel. The regional EIO LC A model was used to calculate the direct and indirect resource us e for a one million dollar change in production for each of the industries in the model. Emergy Evaluation of Regional Industries Emergy intensities were calculated for each individual indust ry in the region using the output of the regional EIO LCA model. The output of the EIO LCA model included both the direct and indirect resource consumption for water, energy, and labor inputs. In the emergy calculation, e ach resource component was multipli ed by its respective unit ground and surface water were obtained from a recent evaluatio 57 s for fossil fuels 58 E lectricity was not included as a direct energy input, but was included as an indirect energy input based on the purchases each industry made from the electric power generati on industry. In order to complete the emergy valuation, the emergy of labor within the region had to be calculated. This was done by summing the emergy of the direct water and fuel inputs used to support the residential population. In this case, e lectricit y was included as
41 a direct energy input as no indirect inputs were considered for the residential sector The residential emergy was then divided by the total value added provided by the regional labor force. The result was an average labor UEV in units o f sej/$M. The eme rgy of labor for each individual industry wa s then calculated by multiplying this average labor UEV by the total value added for that industry. The emergy of i ndirect value added calculated in the EIO LCA model was used to represent the em ergy of service contributions to each industry as it capture d the payments made to labor in the purchases from other industries. Results Water Intensities The w ater intensities for Florida that were calculated by direct allocation to an industry are repo r ted in Tab le 2 1 intensities with the national water use intensities Florida agriculture is shown to have significantly higher water intensities than the national average in vegetable, fruit, greenhouse, su mining sector is more water intensive, likely due to the need to extract groundwater to facilit ate surface mining national average as well Finally, the recreation industry is also much more intensive than the national average due to golf course irrigation Table 2 2 shows the directly alloca ted water i ntensities for the Peace River r and fruit production is lower than the national average. Mining water intensities remain
42 higher than the n ational average, a s does the region recreation water intensity. Water intensity for the power generation sector is shown to be lower than the national average. Energy Intensities Electric power generation is the largest energy use in Florida and the Peace River region, an d this category makes the larg est impact in differentiating national and regional energy intensities. T he energy intensities for electric power generation were directly allocated for both Florida and the Peace River region, and are reported in Table 2 3. F uel intensities were calculated for coal, natural gas, petroleum, waste and biomass, and alternate fuel sources, which included nuclear, hydro, wind, and solar electricity These intensities are compared to the national aver age intensity, showing significa nt differences in fuel sources between regions is almost entirely nuclear power, with a very small percentage from hydroelectricity. Florida has a much higher use of petroleum and nuclear power than the national average, w hile its coal and natural gas use are lower than average A biomass fuel intensity wa s not reported in the national energy intensity vector, and so no comparison could be made in this category. The Peace River region has much higher than average use of nat ural gas and petroleum, with coal use being much lower than average The Peace River did not have any alternative energy sources generated within the region for the year of the model. GHG Emission and Value Added Intensities The calculated GHG intensity ve ctor for Florida was also used for the Peace River region. energy intensities, the emission intensities are the same for the two regions. The only exception is in the power generation ind ustry, where data was available to differentiate
43 The value added vectors for F lorida and the Peace River region have differences due to the difference in labor costs within these two regions. Regional EIO L CA M odels T he regional EIO LCA models were used to calculate the dir ect and indirect resour ce use of each industry in the respective regions H ow much of a prod footprint is masked by indirect effects occur ing outside the producing industry can be sho wn by c omparing the magnitude of direct and indirect inputs for each industry F igure 2 1 and F igure 2 2 show the normalized cumulative distribution of water use and energy use in the region respectively The se distribution s are plotted against the log of the ratio of direct to indirect resource requirement. For log values above zero, the direct resource consumption represents more than half of the total resource requirement for that product ; while for log values below zero, the indirect resource values co nstitute a greater portion of the resource requirement Both Florida and the Peace River show similar patterns within their resource requirements, in that the highest resource consuming industries also have the highest direct to indirect resource ratios. F or both water and energy consumption, the industries that have higher indirect consumption than direct consumption make up less than 10% of the total direct consumption in the region. However, the vast majority of industries f all into the negative log quad rant where indirect fuel use i s greater than direct fuel use. Emergy Evalua tion The emergy signature of Florida and the Peace River wa s calculated using the results of the regional EIO LCA model. Figure 2 3 shows the emergy values for the Peace River regio nal industries compared on a log scale There are 5 orders of
44 magni tude between the lowest and highest industry emergy values The i ndustries with the hig hest emergy inputs include the electric power generation and the transportation industries. Figure 2.4 shows the breakdown of inputs into each emergy value on a normalized scale. The blank spaces in the graph are industries that are not present within the Peace River region. The e mergy contribution s of direct w ater inputs are visible primarily in the agric ultural industries, while they make small contributions elsewhere. Indirect water use is a very small factor in the total emergy value. Direct energy inputs dominate the electric power generation and the transportation industries. However, m ost industries are dominated by indirect energy inputs due to purchased electricity Commercial industries can be seen to have higher labor and service components than most other industries. Discussion Regional Resource Intensity Vectors The goal of the regional EIO LCA model is to relate regional economic production to environmental resource use. The res ource intensity vectors that were developed for the r egional model were critical to defining these relationship s The more accurately the regional resource use can be spe cified, the more the mod el will represent the regional resource use as being different from the national model. Large differences between average intensity values for the nation, the state, and the r egion are demonstrated in Table 2 1, Table 2 2, and Table 2 3. The methodology developed in this dissertation of allocating regional data first and then filling in missing values with national and state level averages is a valuable tool for regional modelers with data access limitations. T he down casting of the n ational and state models allows for
45 regional modelers to make reasonable estimates for resource intensities in the industries where no regional data is available. The national input output model is updated every five years, and efforts are continuing to d evelop national resource intensities. The potential also exists for a regional modeler working in a particular region to develop time series of resource intensities for the region as new data comes available This could provide additional information in tw o ways, both capturing technological change in resource eff icien cy over time, as well as providing estimates of the uncertainties associated wi th the different resource intensities of the region Direct and Indirect Resource Use and Emergy Evaluation A ma jor benefit of the regional EIO LCA model is its capability to estimate both the direct and indirect resource consumption for each industry in the model. Because it takes the entire regional economy into account, it can make comprehensive estimates of reso urce requirements without suffering from cutoff errors. The regio nal modeler can use this data on direct and indirect resource use to construct regional emergy models with unprecedented detail. Normal ly emergy evaluations of regional systems must rely on j ust a few internal compartments with highl y aggregated resource flows Figure 2 3 shows the emergy value of each industry included in the Peace River model, totaling several hundred industries. An emergy analysis with this level of detail has never been co nducted for a region. The potential exists to use this method to compare emergy valuations across multiple regions to develop robust ranges for the emergy values of a wide range products and services. The emergy evaluations constructed with data from the m odel have several important properties that should be pointed out that make them different than other
46 emergy evaluations. The first property that is important to recognize is that the construction of the model treats all industry outputs as splits that fol low monetary flows. This avoids the issue of trying to track co products within the economic matrix in order to avoid double counting A potential draw back to this approach is its reliance on prices to determine splits Prices can shift relative to each ot her over time, which will change the percentage of splits in the model. In addition, waste flows that have no economic value, but may serve as inputs to other processes will not be accounted for in the model. A second property of the regional EIO LCA data is that it does not include resource use in products that are imported from outside the defined region. This is because th e economic matrix only includes regional transactions. Inputs that are purchased from o utside the region could be estimated using IMPL AN regional purchase coefficients Th e resource requirement of the imported products could then be estimated by assuming that imports have the national average resource requirement s, using the national EIO LCA model for the calculation The emergy values that are calculate d in the regional model represent a n average value for the production, and not a value for a specific final product. A ny given industry includes all the output products that industry produces, in the proportions dictated by th e economic input output model. Only in the case where byproducts are a small component of total production c ould direct compariso ns be made between industry emergy values and product emergy values However, s ince the reported emergy unit values are in sej/ $ M deriving a n average product emergy from the industry emergy would just require price information for the product in the year of the model. The regional EIO LCA model is built using producer prices, so to convert to a
47 retail price the transportation, wh olesale, and retail margins must be added to the producer price. Estimates of these margins are available as part of the IMPLAN database. This model then has the potential to provide regionalized emergy values for a vast array of products that currently ha ve no emergy valuations published. The method employed in this chapter for estimating the emergy of industry outputs accounts only for water, e nergy, and labor inputs. T hese are likely to be the largest contributors to the emergy signature, but they are n ot necessarily the only contributors. Other material inputs such as fibers and minerals have been excluded. The difficulty with including other materials is that in dependent data sources are incomplete or non existent at the regional level, and often at th e national level as well. Zhang 59 made estimates of material inputs in the national economy, but due to data limitations, the model had to assume that material flow through the economy was directly proportional to dollar flow. This simplifying assumption could be used to give rough estimates of material inputs for the industries in the model, but represents a different methodology than using resource intensities based on repo rted resource use While m aterial inputs may be significant factors in the emergy signature of primary manufacturing and mining industries, it is not clear that they will be limiting factors in future production. Many materials either have substitutes, or can be recycled if appropriate processes are implemented. Energy and water have no substitutes in production processes. It is interesting to note that e ven though water as a material is required in large amounts by many indus tries, its cont ribution to total emergy is shown to be fairly small in this model It makes a significant contribution only
48 in the agricultural industries. Other materials are likely to account for even less of the total emergy of industries. Environmental Impacts The re gional EIO LCA model provides insight into the effects of an inter connected ec onomy on direct and indirect resource consumption, and how that resource consumption will change with changes in economic activity To fully cap ture the impacts of economic chan ges on the sustainability of a region a consideration of the environmental impacts due to these changes is also needed, including the impacts on ecosystem services. In Chapter 3 the regional EIO LCA model is combined with a land use model in order to capt ure environment al impacts from regional production
49 Table 2 1. D irectly allocated water intensities for Florida Total Water withdrawn 39 Economic Output 56 Florida Water Intensity National Water Intensity 60 Percent Difference Industry (gals) ($M) (gal/$M) (gal/$M) 1 Oilseed farming 3.91E+08 55.55 7.03E+06 8.78E+06 19.9% 2 Grain farming 2.57E+10 79.00 4.62E+08 1.19E+09 61.2% 3 Vegetable and melon farming 1.58E+11 1455.22 2.85E+09 2.36E+08 1106.2% 4 Tree nut farming 3.39E+08 57.47 6.11E+06 4. 63E+08 98.7% 5 Fruit farming 6.80E+11 1858.92 1.22E+10 4.50E+08 2621.8% 6 Greenhouse and nursery production 1.49E+11 1684.65 2.69E+09 5.21E+07 5057.0% 7 Tobacco farming 1.97E+09 24.80 3.56E+07 1.92E+07 85.2% 8 Cotton farming 3.30E+09 27.94 5.95E+07 1. 24E+09 95.2% 9 Sugarcane and sugar beet farming 3.13E+11 439.88 5.63E+09 7.55E+08 645.8% 10 All other crop farming 8.50E+10 40.24 1.53E+09 3.85E+07 3874.8% 11 Cattle ranching and farming 5.63E+09 743.43 7.57E+06 2.13E+07 64.4% 23 Gold, silver, & othe r metal ore mining 2.56E+09 71.41 3.58E+07 4.95E+07 27.7% 24 Stone mining and quarrying 2.44E+10 281.91 8.65E+07 4.91E+07 76.1% 25 Sand, gravel, clay, and refractory mining 1.32E+10 128.60 1.03E+08 5.84E+07 76.1% 26 Other nonmetallic mineral mining 2.7 4E+10 848.58 3.23E+07 1.20E+06 2588.3% 30 Electric Power 4.60E+12 9603.28 4.79E+08 2.54E+08 88.7% 478 Other amusement and recreation industries 1.08E+11 7200.73 1.50E+07 2.29E+05 6474.5%
50 Table 2 2. D irectly allocated water int ensities for the Peace River region Total Water Withdrawn 40 Economic Output 56 Peace Water Intensity National Water Intensity 60 Percent Difference Industry (gals) ($M) (gal/$M) (gal/$M) 3 Vegetable and melon farming 6.74E+09 41.50 1.62E+08 2.36E+08 31.1% 5 Fruit farming 1.10E+11 485.16 2.27E+08 4.50E+08 49.5% 6 Greenhouse and nursery production 4.97E+09 59.11 8.41E+07 5.21E+07 61.5% 10 All other crop farming 6.30E+08 3.74 1.68E+08 3.85E+07 337.3% 11 Cattle ranching and farming 4.74E+09 55.87 8.49E+07 2.13E+07 298.9% 12 Poultry and egg production 2.99E+08 11.34 2.64E+07 3.87E+07 31.7% 13 Animal production, except cattle and poultry 6.26E+08 20.29 3.08E+07 1.59E+07 94.1% 24 Stone mining and quarrying 2.15E+07 17.79 1.21E+06 4.91E+07 97.5% 25 Sand, gravel, clay, and refractory mining 1.30E+09 36.50 3.57E+07 5.84E+07 3 8.8% 26 Other nonmetallic mineral mining 8.64E+09 377.42 2.29E+07 1.20E+06 1807.5% 30 Power generation and supply 3.77E+09 460.86 8.18E+06 2.54E+08 96.8% 32 Water, sewage and other systems 4.69E+09 10.23 4.59E+08 0.00E+00 N/A 60 Frozen food manufactu ring 9.01E+07 552.23 1.63E+05 2.37E+05 31.2% 61 Fruit and vegetable canning and drying 9.10E+08 153.94 5.91E+06 8.59E+06 31.2% 157 Phosphatic fertilizer manufacturing 7.26E+09 1102.63 6.59E+06 8.44E+06 22.0% 478 Other amusement and recreation industr ies 2.67E+09 254.77 1.05E+07 2.29E+05 4481.7%
51 Table 2 3. Comparison of electric power generation energy intensities National Florida Peace River region Fuel Type Energy Intensity 46 Fuel Use 47 Output 41 Energy Intens ity % Diff Fuel Use 48 Output 41 Energy Intensity % Diff (TJ/$M) (TJ) ($M) (TJ/$M) (TJ) ($M) (TJ/$M) Coal 78.3 726684 14541 59.32 24% 41840 461 90.79 16% NG 22.9 564636 14541 46.09 101% 97389 461 211.32 823% Petro 3.8 3583700 14541 29. 29 671% 4077 461 8.85 133% Biomass 0 47475 14541 3.88 N/A 1766 461 3.83 N/A Alternate 2.5 410579 14541 30.47 1114% 0 461 0 N/A
52 Figure 2 1 Compar ison of regional direct and i ndirect water u se for each industry in the regions. The largest w ater users in both regions have high direct to indirect water use ratios.
53 Figure 2 2. Comparison of r egional direct and indirect energy u se for two different regions. The highest energy consumers in both regions have the highest direct to indirect en ergy use ratios.
54 Figure 2 3. E mergy values for Peace River regional indust ry output showing a range of five orders of magnitude in emergy values.
55 Figure 2 4. Composition of emergy values for Peace River regional industries showing distinct signat ures for agricultural, power, manufacturing, transportation, and commercial industries
56 CHAPTER 3 DEVELOPMENT OF A LAND USE OPTIMIZATION MODEL Introduction This chapter outlines the development of a linear programming optimization model for the Peace River region. The goal of the optimization is to maximize the sustainable production of the region. Land use was chosen as the variable that would be manipulated within the optimization model in order to achieve maximum sustainable production. This required def ining 19 different land uses within the Peace River region for which the environmental and economic resource flows were modeled. A set of constraints were then defined for the optimization in order to bound the possible mi x of land uses within geographic, economic, and sustainable limits Within this chapter, the sustainability constraints employed were based on regional water use and included groundwater recharge and flood water storage. The sustainability constraints incorporate d information on economic w ater consumption obtained from the regional EIO LCA model developed in the previous chapter. The results of the optimization for the Peace River region are reported and the impacts of the selection of sustainability constraints are explored. Methods Water and Carbon Balance Models The regional EIO LCA model predicts resource flows within the economic system, but it does not capture all the resource flows in the environment. To predict changes to these flows, an environmental model is constructed of the Peac e River region using simple water and carbon mass balance models for each land use within the Peace River model Since water flows are sourced from the environment, and waste
57 water flows are returned to the environment, this type of model helps account for environmental water flows due to economic activity. Linear Programming O ptimization M odel s Optimization models are used widely to determine how to allocate limited resources to maximum effect. Optimization using linear programming was developed for use in military logistics planning during WWII, and revealed pu blically in 1947 The main feature of the optimization model is a goal or objective function that is to be either maximized or minimized. Along with the objective function, a set of constraints are d eveloped that set the limits of the available resources that are required by the objective. In linear programming, these constraints are expressed as linear equations and are written in terms of a limiting variable. Combining all the constraint equations d efines the solution space. Search algorithms are used to efficiently explore the solution space and find the maximum or minimum value for the objective that still meets all the constraints. Linear programming models have been used in a wide range of appli cations, including land use planning 61 wat ershed management 62 and ecosystem serv ice analysis. 63 In this modeling effort, the objective function is set to maximize system producti on, as defined earlier. The constraints are designed to maintain the long term sustainability of the region. The constraints may be defined as being physically based, or they can be defined by the social side of the system as laws or policies that apply to the specific region. This optimization model assumes both linear objective functions and linear constraints. The resulting model scales all land use impacts linearly with area. While this is an approximation, it is a defensible assumption for most natura l, agricultural,
58 logging, mining, and even residential land uses as long as the interactions of the land uses are appropriately accounted for This assumption likely does not hold for commercial and industrial land areas, where production can be increased independent of land area. The nonlinearity of these land uses was addressed in the optimization model by forcing their area to remain constant. The construction of the regional optimization model required defining major land uses within the region, and ass igning average economic and environmental flows of water, energy, and GHG emissions to these land uses. Publically available data sources were used to define both the regional land uses and resource flows within those land uses Data sources were selected to correspond as closely as possible to the year 2002 so that they matched the y ear of the regional EIO LCA An exact match was not always possible, and in these cases data as close as possible to the year 2002 was selected. Land Use Land use data for the Peace River region was compiled from the three water management districts that contain porti ons of the four counties within their boundaries: the Southwest, South, and St. Johns River Water Management Districts. A single data set was not available for any one year, and so data from 2004 2006 were combined to yield a complete land use data set for the four county region This land use data 64 65 66 was reported according to the Florida Land Use Characterization System (FLUCS). FLUCS land uses were aggregated to form 19 economic and environmental land uses within the regional model. Water Balance Model A water balance model was developed for the Peace Ri ver region in order to allocate water flows to specific land uses. The water balance model incorporated both
59 economic and environmental water flows. D ata for environmental flows was obtained from the Peace River Cumulative Impact Study (PRCIS) 35 cond ucted jointly by the Southwest Florida Water Management District and the Florida Department of Environmental Protection The PRCIS report created a w ater bal ance model for the sub basins of the Peace River watershed for several different time periods in order to D ata for the years 1997 1999 from the PRCIS report was used to define average environm ental flow condit ions The first step in the water balance modeling process was to align economic and environmental boundaries. The surface watershed boundary of the Peace River is almost completely contained within the political boundaries of four counti es: Polk, Hardee, DeSoto, and Charlotte County. However, the political boundary of these four counties encompasses more area than the watershed itself. In the optimization model, the watershed boundary wa s redefined to coincide with t he political boundary. S urf ace runoff and stream base flow contribution values for the land uses within the watershed area were applied across the entire four county area. This modification resulted in a significantly greater volume of stream flow in the water balance model tha n in the actual physical system, but it preserved the relative magnitudes of environmental flows within each of the land uses and still allowed the calculation of a groundwater recharge balance across the region The alternate approach of limiting the econ omic data to the area within the natural watershed boundary was deemed not to be feas ible because the economic data wa s reported only at the county level
60 Environmental water flows Rainfall wa s the major environmental input flow. The annual rainfall variat ion for the region is large, varying from 30 up to 70 inches a yea r. The PRCIS 35 reported an average rainfall value of 50 inches per year from 1997 1999. T his rain fall value wa s used as the input value f or rainfall in all land uses Evapotranspiration (ET) wa s the largest water outflow in the system. ET was not measured directly in the PRCIS study, but was estimated based on a regional land use analysis. In the stu dy a reference evaporation for the region was calculated using a modified Penman Montief equation and regional data on solar irradiance and temperatures. Regionally specific crop coefficients were then used to estimate an actual ET for each land use. C rop coefficients used in the study for Peace River land uses were obtained from literature values that were appropriate for central and southern F lorida. The crop coefficients from the PRCIS report were ado pted for the water balance model The PRCIS study al so reported runoff coefficients for land uses in the region for two hydrologic soil types and for dry and wet rainfall seasons. In the regional water balance model, the mid points of these two reported ranges for two di fferent soil types were combined in a weighted average to develop a single runoff coefficient fo r each land use. The weighted average adopted the PRCIS study assumption that 38% percent of the rain falls in the dry season, and 62% falls during the wet season. The runoff coefficients for the r the dry season and having runoff only during the wet season. The estimated ET and runoff for each land use were subtracted from rainfall to give a value for groundwater recharge. The surficial, in termediate, and Floridan aquifers
61 were all considered a single groundwater unit for the water balance model. Data was not available to differentiate between these different underground reservoirs. Within the PRCIS study, a percentage of the groundwater re charge was allocated to base flow, based on hydrograph separations for the Peace River. The regional water balance model maintained this percentage of base flow contribution to stream flow as a constant for all land uses. Economic water f lows Modification of the environmental water flows by economic use was accounted for in characterizing t for economic use were obtained from USGS reports 39 Florida stat e averages were used to determine the amount of this water that is consumed as additional ET or is embodied in a final product, and the amount returned to the wate rshed in wastewater flows. These water extractions and return flows were included in the wate r balance model of each land use In the water balance model, wastewater applie d at the surface of the ground wa s considered to enter the vadose zone and recharge the surf icial aquifer. Wastewater that wa s injected underground into the aquifer without in te racting with the vadose zone wa s also considered to be a contribution to ground w ater recharge. Wastewater that wa s discharged di rectly to surface water bodies wa s considered to contribute directly to stream flow. For each land use, wastewater return flows were assigned to the land use that initially extracted the water regardless of where the return flows are re applied. Following this convention is necessary to allow net groundwater flows for each land use to be calculated in the model. The total groundwa ter extracted ( represented as negative recharge values) for each land use type was added together with the groundwater
62 recharge values (represented as positive recharge values) to give a net groundwater recharge for each land use. The sum of groundw ater re charge over all the land area in the model represents the net groundwater rec harge for the entire watershed. Economic Model The economic model of the Peace River region relies on year 2002 economic data from IMPLAN 56 for the four county region. The 509 industries in the regional EIO LCA model developed in chapter 2 were allocated to 12 different economic land uses defined for the region Individual land uses were defined for industries that were land use intensive, and f or industries where production wa s proportional to land area. Commercial and industrial land uses that were not land intensive were grouped together. Resource consumption due to economic activity in the ind ustry was allocated to each land use Economic Linkages The area of each land use is the variable that is being manipulated in the optimization model. However, the initial definition of a land use only accounts for its d irect resource consumption. Increasing or decreasing any economic land use also causes resource consumption to occur outside that land use, within the economic sectors that provide the required production inputs. This resource consumption occurs outside th production processes. To account for these interactions in the model, three ki nds of economic linkages are considered: backward, forward, and induced l inkages Backward linkages are the purchases that each industry makes from other regional industries in order to make their product. The indirect resource consumption from backward linkages was calculated in the regio nal EIO LCA model developed in C hapter 2 using a Leontief
63 predictive model. However, increasing an economically productive land area not only creates additional resource demand within the region, it also creates an additional supply of economic goods that can feed additional production processes in the region. Forward linka ges consider the additional output produc ts that could be produced within the re gion based on an increase in the supply of inputs These outputs can be either directly exported, or they could be further processed within the local economy to produce higher value added goods. The processing of these goods within the region will generate further economic activity and require additional resource consumption. The Ghosh model makes the assumption that the percentage of regional products processed within the regi on into higher value adde d goods remains constant for an y increase in regional supply. The model then uses the current economic structure to calculate how much more resource use the processing of this additional supply will consume in the local economy. As a linear model, it assumes that there will be a market for these additional goods, and that their creation will not change the prices re ceived for these goods. The Ghosh model is created from the economic transaction matrix by dividing each row by the tot al output of the row as shown in equation 3 1. (3 1) The predictive model is then created as shown in equation 3 2 where x is the total output vector, the quantity (I GT) 1 is the Ghosh matrix, and v is the vector of value a dded. (3 2)
64 As with the Leontief predictive model, the resource intensity matrix can be multiplied by the change in total output from the Ghosh model to give the direct and indirect resource consumption due to changes on the supply side as shown in equation 3 3. (3 3) The final economic linkage to be considered was the induced linkage. This linkage considers the resource consumption imp act caused by an increase in labor payments to the population living in the region. When a land use increases, thereby increasing economic activity, it also increases income to the workers that are providing labor to that economic activity. A portion of th is income is usually spent within the region, creating an additional demand for regional products and services. IMPLAN software provides a set of induced impact multipliers that are multiplied by the change in final demand from the Leontief model to determ ine changes due to increased spending by the population. The induced impacts calculated in IMPLAN are combined with the backward and forward linkages to give a more complete estimate of regional impacts based on changing economic output. C ombined Economic /E cological Land Use M odel To complete the modeling of the individual land uses, the indirect resource consumption that occurred in other land uses due to backward, forward, and induced linkages was subtracted from those land uses and re allocated to the d riving land use. This re allocation was done for each land use in the model that had an economic component, resulting in both additions and subtractions for each land use. Each resource category was redefined in this way, acco rding to equation 3 4 (3 4 )
65 In this equation, R i stands for the resource use of land use i and category r D i,r stands for the direct resource use of that land use, L i,r, Out G i,r, Out and I i,r, Out stand for the indirect resource use from the Leontief, Ghosh, and induced linkages respectively that occur outside land use i but are driven by its production. These indirect resources are added together for land uses 1 through n L i,r,In G i,r,In and I i,r,In stand for the indirect resource use from land uses 1 through n that were i n side land us e i and are subtracted out of it T his total resource flow R i ,r was then d ivided by the total area of the land use, A i to yield area based resource intensities for each resource category for each land use (3 5 ) These area base d resource intensities were the inputs for the optimization model used to characterize the change in re source consumption as land areas were changed in the model Linear Optimization Model The optimization model for the Peace River Region w as constructed in Microsoft Excel using the solver add in feature, which is a commercial optimization package employing the Simplex linear programming algorithm. Building the model required the definition of a regional goal function and development of cons traints for the model. A 50 year time hori zon was selected for the model in order to include activities such as phosphate mining, which is a major part of the current economic structure. Based on curre nt mining rates, this activity wa s projected to remain in the region over the time horizon of the model.
66 Goal f unction As stated in Chapter 1, the objective function of the optimization model was to maximize regional production. Two measures of production were defined: total monetary throughput and regional em ergy throughput. Total monetary throughput was defined as the sum of the production of each land use as measured in dollars. It should be noted that this is not the same measure as gross domestic product, which d oes not include intra industry flows in its calculation, but instead includes only final consumption. The total throughput includes intra industry flows. Total emergy throughput was defined by adding the largest environmental emergy flow plus the direct and indirect emergy flows in the economy for e ach land use. In the case of emergy as well, intra industry flows are included in the total throughput calculation. For all land uses, the evapotranspiration of water was the largest environmental emergy flow. Economic emergy flow for each land use was cal culated using the following categories of resources use: direct and indirect water input, direct and indirect energy input, and direct and indirect labor input. Indirect labor input represents the category usually referred to as goods and services in an em ergy valuation. It represents the labor input embodied in inputs of goods and services purchased from outside the land use. Th e land use emergy valuation did not include material use other than wa ter. Since the emergy flow wa s calculated using the regional model, it included only indirect flows from purch ases within the region, and did not include an emergy evaluation of purchases from outside the region other than the primary energy sources. This conv ention limited the objective function to a maximization plus the primary fuel inputs
67 Variables The manipulated variables in the linear programming model were the land areas allocated to different land uses within the region. To construct the model, resou rce flows were defined in terms of flow per unit of land area. The assumption of the model is that these flows will vary in direct proportion with land use. As each land use area was changed, the economic and environmental flows associated with that land u se change d linearly Constraints Several types of constraints were defined for the optimization model. Physical constraints dealt with the physical characteristics of the region. T he primary physical constraint wa s the total area available in the system. After optimizatio n, the sum of all land uses had to be equal to the total land area. In addition to the total area, several other land areas were held constant in the model. The area of water was set as constant in the model, including river, lake, and est uary area, as these land types are not readily converted to other land uses. In addition, wetland area was constrained by the maximum extent of wetland area in the system before wide spread development, defined as the total wetland area from the 194 0 land use analysis 35 No minimum wetland area was defined in the model. Water resource constraint: Sustainability constraints limit resource use to levels that can be sustained indefinitely within the system. Groundwater sustainability limits were defined by limiting groundwater consumption so that the currently modeled amount of groundwater recharge in the system wa s maintained into the future Groundwater consumption was defined as groundwater that was withdrawn and subsequently either evaporated, or was transported out of the region though means
68 other than groundwater flow. The intent of the constraint was to maintain average aquifer levels in the system. A flood con trol constraint was also developed for the r egion. This constraint considered the storage of runoff from a large storm event in the system. The 25 year, 6 hour return storm for the region was used to define the rainfall event 67 with the assumption tha t upland land areas would have 2 inches of initial abstraction, and the remaining 4 inches of water would become runoff. All developed upland areas and forested upland areas were assumed to have this average runoff. Water bodies including lakes and streams were assumed to have enough freeboard to hold the 6 inches of rainfall, plus an additional foot of rain. Wetlands were assumed to hold the 6 inches of rain, plus an additional 6 inches of rainfall for an average of a foot of rain. Unimproved pasture land was assumed to hold its own runoff in low lying areas essentially causing no net runoff of water. The flood control constraint was defined so that storm water storage in the region must be equal to the projected runoff. Spatial characteristics of the lands cape were not considered in the runoff model, so the location of the flo od storage was not evaluated. Several economic constraints were included to govern the economic interactions within the system. The first economic constraint dealt with the population supported in the region. According to IMPLAN data, 56 40% of the 2002 population was employed within the region A constraint was defin ed so that the working population of the region must be large enough to provide all but 1% of total employment requirements. A separate constraint was defined so that unemployment could not exceed 5% of the working population. This co nstraint was implement ed by bounding the total number of
69 workers available from the residential land use with the total jobs available from the economic land uses. The effect wa s that the residential area of the region was constrained to grow only large enough to provide the la bor workforce required to meet of the workforce, 60%, remained constant in the model. Electricity generation constraint : The Peace River region produced electricity fro m multiple power generation plants in 2002. Based on calculations from the regional EIO LCA model of average commercial, industrial, and residential electricity demands, approximately 25%, or 10,000 TJ of the electricity produced in the region in 2002 was electricity production was constrained so as not to exceed the amount of electricity exported in 2002 any additional electricity production must be consumed within the region Since there are large urban markets bordering the Peace Ri ver region, the potential may exist to export more electricity, but additional export was not evaluated in the model. Commercial and industrial a rea c onstraint: The commercial and industrial land ar eas were held constant in the model. The production within these land uses does not necessarily vary linearly with land area. Production can be increased and decreased in these land uses without increasing land area by changing other input factors, such as labor hours, or capital investments. Increases that result ed from changes in the other land uses were previously separated out of these two land uses. The remaining production, then, is not tied strongly to the production occurring across the lands cape of
70 the region The model assumes that the level of production can be maintained within the current area footprint. Mining industry constraint: In the model the mining industry is constrained by defining an amount of land area that can be mined each year. Cur rent mining in the region consumes about 2000 ha of new land area per year. 68 T he t otal annual mined area was defined as the actively mined land plus all land still in various stages of reclamation. Mined land area was modeled as requiring an average of 12.5 years to be reclaim ed t o an alternate use. 68 At t he end of this time, the land area is considered available for other commercial use. The total mined area including active mines and land being reclaimed was was set as 31,000 ha per year and this amount of land use was he ld constant in the model. At this rate of mining, phosphate deposits are projected to remain in the region beyond the 50 year time horizon of this model allowing this industry to be included in the land use analysis Agricultural industry constraints : Constraints were also defined for th ree of the agricultural industries. These constraints prevented industries that have limited export markets from increasing beyond a reasonable growth estimate The industries constrained included vegetable production, greenhouse production, and other agri culture. These are comparatively small land uses in the region, but they have high profitability per land area Vegetable production was constrained to ha ve a maximum 50% growth in area. G ree nhouse production and other agriculture production were considere d to have more limited markets for their products and were limited to only 10% growth in area.
71 Results Land Use Table 3 1 reports t otal land areas for each land use category in the model Natural areas still comprise 39.8% of the total land area, while ind ustrial, commercial, and residential together make up only 6.9% of land area. The majority of the land area is under mining and agricultural development, with unimproved range land making up the largest single land use at 27% and citrus and mining contribu ting 10.4% and 8.4% of the land area. The current region has 3.9% of land area classified as open or disturbed that is not under any economic land use. Water Balance M odel The water budgets developed for regional land uses result in each land use having a characteristic signature of water inflow and outflows. The env ironmental inflow in the model wa s rainfall, and the characteristic outflows for rainfall into ET, runoff, and recharge are given in Table 3 2. Economic inflows come from extraction of ground an d surface waters. The percentage of economic inflows from different sources are given in Table 3 3, while the disposition of those inflows into outflows are given in Table 3.4.When combined, these environmental and economic flows define the water budget fo r each land use. Figu re 3 1 shows the contribution of extracted surface and ground water to the total input flow for each land use in the region. Land uses that represent the most significant changes in input flows per hectare include vegetable farming wit h a 69% increase, industrial land use with a 59% increase, and power generation with a 46% increase. Figure 3 2 shows the allocation of output flows between ET, runoff, and groundwater recharge for all land uses. The figure includes the contributions to th ese flows from extracted ground and surface waters used in economic production
72 processes. Significant redistributions of water to ET occur in both industrial and agricultural industries. While the power industry has large extrac tions of water for cooling, estimates for Florida are that only 9% of that water evaporates, 39 and the rest is returned to surface storage where it recharges groundwater. Residential and industrial wastewater flows make up the major economic contributions to surface flows. Land Use Models Modifying the land use models to account for indirect resource consumption resulted in significant changes to the flows within each land use. Figure 3 3 shows the percent change for the monetary value of pro duction after indirect economic linkages had been accounted for In this figure, the original dir ect value of production is shown as 100% of the initial value Net additions or subtractions of backward, forward, and induc ed linkages are shown in relative s cale to th e original direct requirement. The c ommerc ial and industrial land uses were held c onstant in the model, so they did not drive any increases in other land uses in the model. Therefore, they only display economic activity that wa s removed and added to the other land uses. Mining, citrus, vegetable, nursery, other crops, and cattle land uses all have net positive additions from their economic linkages The power sector and logging sector were only allocat ed the backward and induced linkage impacts. F orward linkage impacts were not allo cated to these two industries since additional production within these two industries was not considered to drive any additional production of goods and services within the region. Accounting for indirect economic linkag es created between 50% and 150% increases in output for these land uses. For most of the land uses, the induced impact created the largest change.
73 The indirect resource consumption for each land use was cal culated by multiplying the changes in indirect out puts by the matching resource intensity This was done for applicable backward, forward and induced linkages for each land use. Figure 3 4 show s the percent change in resource consumption or economic impact for each land use as a result of accounting for the indir ect contributions to that land use The changes are compared in groundwater use, job provision, GHG emission, as well as emergy flow and economic output. Comparing across land us es, a wide range of changes due to indirect impacts is observed, vary ing from single digit decreases t o an almost 900% increase in GHG emissions within the logging land use. The economic and environmental characteristics of each regional land use were combined and then divided by th e area of land use. This created the comb ined economic and environmental land use model that served as in input into the optimization model. Table 3 5 gives the results for the two objective functions, monetary output and emergy flow, and the variables for each land use in the model. O ptimization The optimization of regional production resulted in land use shifts as shown in Figure 3 5. The results represent a tradeoff between economic production which requires groundwater extraction, land uses that provide groundwater recharge, and land uses tha t provide floodwater storage. Agriculture industries with high production value per land area increase to their internally constrained limits. The largest land use shift is a tenfold increase in irrigated cattle ranching, while citrus and un irrigated catt le ranching experience 35% and 43% declines respectively. All upland forest and logged forest is converted into other land uses, while total wetland area actually increases in value. The model increases total wetland area by 46%, while converting all wetla nd
74 area to wetland forest. These land use changes result in an additional $465M output, a 1.8% increase in economic output for the region. In addition, the region now supports a 6% larger population than before the optimization. The initial optimization wa s also run with the objective of maximizing emergy throughput instead of monetary t hroughput. Only 4 areas of land use differed in the emergy optimized scenario from the monetary optimized scenario, and all of these varied by less than 1.5%. The increase i n emergy flow was 3.9% overall, as compared with the increase in monetary flow of only1.8%. Sensitivity Analysis The sensitivity of optimized production to the groundwater sustainability constraint was evaluated by varying the volume of groundwater that m ust b e recharged into the system. This constraint was varied over a range from 74% to +57% from the target constraint. Beyond this range, the optimization had no feasible solution, meaning all of the constraints could not be satisfied simultaneously. Indu strial and commercial production not associated with agricultural production would have to be allowed to decrease in the region to meet any further constraints. Figure 3 6 shows the resulting shifts in land use as the groundwater constraint is varied. Requ iring less recharge resulted in increased water available for economic production. With more irrigation water available, the citrus industry expanded. Requiring more groundwater recharge resulted in land use shifts away from citrus into cattle ranch ing, fi rst on improved pasture which is partial ly irrigated, and then to range land which has no irrigation Higher profitability agriculture such as vegetable and nursery production remained at maximum limits until the most severe groundwater recharge restrictio ns were implemented. At the highest recharge restrictions almost all irrigated
75 agricultural production is eliminated to provide the required water flow for commercial and industrial production. Shadow Pricing of Ecosystem Services A result of setting up the regional optimization with sustainability measures as constraints is that shadow prices can be calculated for the ecosystem services that are constraining production. A shadow pr ice measures the extra value that c ould be added to the goal function as a result of increasin g a constraining resource by a single unit of that resource a shadow price for each binding resource constraint. Only binding constraints have shadow prices, because increasing a constraint that is not limiting production do es not result in any additional production capability In the case of the groundwater sustainability constraint, the constraint is how much water must be recharged to maintain aquifer leve ls. The shadow price reveals the value an additional unit of groundwater recharge has in increasing the goal function of the region. This value can be used to represent the value of the environmental service of groundwater recharge being provided by each l and use to the region as a whole. It is important to note that the shadow price does not remain constant over the entire range of production. As the groundwater constraint is increased, the price per unit of groundwater chang es at certain inflection points By varying the value of the groundwater constraint used in the optimization, the range of the shadow price for the constraint can be explored, in effect giving a price curve for the constraining factor. Figure 3 7 shows the price curves generated in the current model for both groundwater recharge and storm water storage. When the requirement for groundwater recharge is low, high water intensity activities like citrus production are selected in the
76 model. Citrus fields have high er runoff coefficients relat ive to the land uses they replace and so more storm water storage is required in the region. The shadow price of storm water storage is highest when the constraint for groundwater recharge is lowest. As the requirement for groundwater recharge is increase d in the region, high water intensity land uses like citrus are replaced with lower water intensity land uses such as cattle production on pasture and rangeland. These land uses provide more storm water storage than citrus, and the need for storm water sto rage area in the region falls as does its shadow price. Eventually, storm water storage is no longer limiting, and the shadow price falls to zero. Groundwater recharge demonstrates a different pattern. Initially, the shadow price for groundwater recharge i s very low. As the recharge requirement increases, so does the shadow price for the ecosystem service. Discussion Linearizing Land Uses Linearizing the land uses in the model is a unique aspect of this model. The land uses initially are only defined by the direct impacts that take place in the land area itself. Because the resource consumption of the land uses in this model was determined using the regional EIO LCA model, the indirect resource consumption could be calculated for each land use. For the each type of economic linkage this indirect resource consumption is subtracted from the upstream land uses, and added instead to the land use that generates the demand for it. This accounting for indirect resource consumption within the driving land use captu res resource demand that would have been missed by a conventional optimization model Because the indirect resource consumption is subtracted from the land use it occurs in, the model avoids double counting these resources in the optimization. What would n ormally be perceived as nonlinearity in
77 resource demand is now a ccounted for in a linear manner, allowing more accurate modeling of the region. Accounting for backward linkages, and for induced impacts is standard practice within economic input output mode ls and social accounting models. However, accounting for the forward linkages as in this application is not standard practice. Figure 3 3 shows that the forward linkage component is in many cases larger than the backward linkage component in the model. The meaning of this component should be considered carefully. In order to calculate this linkage the model must assume that the region will process new outputs from the land use into the same mix of products as are already produced. For example, if 30% of ora nge production is made into orange juice in the region, the forward linkage says that 30% of any increased production will also be made into orange juice. In reality, this depends both on markets for the product and capacity within the region. However, it can be taken as a reasonable estimate of the upper limit of what additional production could be expected to be captured within the region, and the resources required to do so. Models that include forward linkages then should be viewed as estimating the upp er limit of economic development. Forward linkages were not used for all land uses, as s ome forward linkages in the model may not drive any additional production. For example, a dditional ele ctric power generation is not likely to drive the consumption of t hat power in producing new products within the region unless new industry moves into the region. An additional consideration of forward linkages is how well the industry is represented by the national economic model. The forestry industry in the Peace Riv er does not supply the same mix of products that are representative of the national forestry industry as a whole. Tree
78 harvest in the Peace River is not likely to become lumber for houses or pulp for papermills and so it will not drive the production of t hese products, even though the model based on the national economy may specify t hat. This limitation on using forward linkages could be remedied in part by developing regionally specific product allocations instead of relying on national averages. While th is is impractical for all the industries in the model, if this were done for the largest industries in the region, it c ould increase the accuracy of the model results. Accounting for Ecosystem S ervices A key concern for sustainability models has been accou nting for necessary ecosystem services. Ecos ystem services are included in this model through the development of sustainability constraints. Groundwater recharge and floodwater storage are both ecos ystem services that are consumed by the production of some land uses, and produced by other land uses in the model. The ability of the model to consider which services are needed and to what level they need to be provided relative to regional production is vital information for regional managers co ncerned with pr oviding these services in the future. The ability of the model to provide shadow prices for limiting services is powerful feature. Valuing ecosystem services has proven a difficult task in environmental science, as they are produce by the environment free of charge. This model provides a shadow price for any ecosystem services that are limiting in the system. The shadow price tells the regional manager how much more value of production could be generated if one additional unit of the ecosystem service were available. The shadow price is not an absolute value of the service, as it does not relay information on what it takes to provide that service. However, it is valuable information to the regional manager who
79 must make cost benefit calculations on providin g services in the region. Shadow prices can be used to design subsidies and incentives for land owners to provide additional units of ecosystem service within the region. They can also be used to justify the cost of infrastructure investments designed to p rovide ecosystem services in areas where they are needed Choosing Goal Functions Two goal functions were considered for this effort, maximizing emergy throughput and maximizing monetary throughput. The choice of the g oal function for the region was not f ound to have a significant impact on th e outcome of the optimization. Maximization objectives based on m onetary output and emergy flow gave very similar optimization results. At first this result may appear surprising given that monetary valuation assigns zero value to land uses that have no economic output while emergy valuation assigns the value of the ET flow of these land uses. However, comparing the value per area of the monetary and emergy flows shows that they follow the same order for the economic land uses in the region. Industrial and commercial land uses have the highest intensities, and environmental land uses have the lowest. The implication is that land uses will be substituted for each other in the same order regardless of whether the monetar y flow or the emergy flow metric is optimized. However, because the relative magnitudes of the values for each land use are different between monetary and emergy e valua tion, inflection points are likely to be different when using different goal functions. The similar order of land uses for monetary and emergy values may not hold true for all regions, or for the future. The order in the Peace River region is determined primarily through the energy of fossil fuels that are consumed in each region. I
80 Environme ntal land uses have bo t h the lowest monetary and emergy values. These land uses are selected in the model based primarily on the fact that they provide more of limiting ecosystem services such as flood control and groundwater recharge per area than other l and uses Environmental land uses that provide the highest lev els of limit ing services a re then selected. This helps to explain the reason the model selects wetland forest over upland forest, because the water storage and water recharge per area are higher for the wetland forest land use Since the model is seeking only to maximize output, it does not valu e diversity of landscapes, instead i t selects the land use that will give it the highest amount of the constrained variable per area. Valuing the land use s with energy flow raises the question as to the interpretation of an emergy shadow price for a land use. Following the logic of the monetary example would suggest that the emergy shadow price represents the additional amount o f emergy that could flow thro ugh the compartments of the regional system by providing one more unit of the binding constraint. However, it should be recog nized that this is summing all the emergy flow through each compartment in the model, and so emergy is being double counted within this value The same is true for the monetary output of the region, where total money flowing through each compartment is what is being optimized and not GDP The implication is that the greatest value to the region is in having maximally connected compa rtments (or land uses in this case), as this maximizes throughput. Comparing the composition of the emergy flows for each land use shows that energy use dominates the emergy valuation for each economic land use. This underscores the reliance even a highly agricultural region has on imported energy
81 sources. These energy sources are a primary driving force for the production of the region. This is reflected both in the monetary and the emergy valuation of the output. Both measures seem to work as goal functi ons, at least in the short term. In the long term, prices for energy sources may shift dramatically in the future based on factors of supply and demand outside of the region itself. However, emergy values should maintain more stability over the long term b ecause the energy required to make products is i ndependent of supply and demand Emergy valuation may prove to be a better measure for planning efforts with long time horizons and uncertain future energy costs Static Economic Structure A major limitation of the current model is that it is predicting a future in which the underlying economic structure of the region remains the same and only existing land uses change linearly. New industries a re not evaluated for their impact Land use then is the limitin g factor that is being addressed most strongly in the model Chapter 4 begins to explore how the optimization model can be used to address changes in economic structure by introduc ing new land uses to the region
82 Table 3 1. Peace River land use areas La nd Use Area 64 66 (ha) % Total Residential 48181 4.8% Commerci al 20946 2.1% Industrial 3113 0.3% Mining 88714 8.8% Power 995 0.1% Citrus 109433 10.8% Vegetable 2930 0.3% Nursery 5162 0.5% All Other crops 9607 0.9% Cattle Pasture 27813 2.7% Cattle Range 285285 28.2% Logged forest 15043 1.5% upland forest 117805 11.6% Wetland forest 110761 11 .0% Wetland 69583 6.9% Lakes 40118 4.0% Stream 3970 0.4% Salt marsh 10308 1.0% Open Land 41726 4.1%
83 Table 3 2. Water outflows from environmental inputs Land Use ET 35 Run Off 35 Recharge Residential 57.30% 25.60% 17.10% Commercia l 49.10% 37.80% 13.20% Industrial 40.90% 55.00% 4.10% Mining 61.40% 34.50% 4.10% Powe r 49.10% 34.50% 16.40% Citrus 73.60% 26.20% 0.20% Vegetable 69.50% 30.30% 0.20% Nursery 69.50% 30.30% 0.20% All other crops 69.50% 30.30% 0.20% Cattle pasture 65.50% 21.40% 13.10% Cattle range 61.40% 21.40% 17.20% Logged forest 65.50% 18.10% 16.50% Upland forest 65.50% 18.10% 16.50% Wetland forest 69.50% 10.00% 20.50% Wetland 73.60% 10.00% 16.40% Lakes 85.90% 0.00% 14.10% Stream 85.90% 14.10% 0.00% Salt marsh 73.60% 26.40% 0.00% Open land 61.40% 25.60% 13.10%
84 Table 3 3. Water inflows from e conomic use Land Use Supply 40 Source 40 Self Public GW SW Residential 20.60% 79.40% 91.10% 8.90% Commerci al 90.80% 9.20% 83.10% 16.90% Industrial 34.30% 65.70% 95.90% 4.10% Mining 100.00% 98.00% 2.00% Power 100.00% 85.00% 15.00% Citrus 100.00% 94.40% 5.60% Vegetable 100.00% 94.40% 5.60% Nursery 100.00% 94.40% 5.60% All o ther crops 100.00% 94.40% 5.60% Cattle pasture 100.00% 94.40% 5.60%
85 Table 3 4. Water outflows from economic use Land Use ET 39 Wastewater 40 Self Supplied Public Supplied Vadose Surface Water Injected Residential 39.00% 60.00% 61.00% 31.00% 8.00% Commercial 80.00% 59.00% 61.00% 29.00% 10.00% Industrial 19.00% 38.00% 62.00% 34.00% 4.00% Mining 30.00% 100.00% Power 9.00% 100.00% Citrus 72.00% 100.00% Vegetable 72.00% 100.00% Nursery 72.00% 100.00% All Other crops 72.00% 100.00% Cattle Pasture 72.00% 100.00%
86 Figure 3 1 Water inflows for regional land uses showing the contribution of economic groundwater and surface water flows to the water intensity of land uses in the Peace River region.
87 Figure 3 2 Water outflows for regional land u ses showing the distribution of total flow into ET, runoff, and groundwater recharge. Each outflow is further divided to show the inflow source for that portion of the flow, whether that be rainfall or economic water in flows.
88 Figure 3 3. Percent c hange in monetary outp ut when accounting for economic linkages within the region. Direct output measures the total economic output of the land use itself, while the three additional contribu tions represent economic activity occurring the oth er land uses but required by the direct production of the primary land use
89 Figure 3 4 Percent c hange in resource intensities of land uses due to accounting for economic linkages T he changes are significant within all resource categ ories Small initial values, such as logging, tend to have the largest percent increases.
90 Table 3 5. Land use inputs to the Peace River optimization model Land Use Output Emergy Jobs Water Balance Flood Balance ($M/ha) (seJ/ha) (jobs/ha) (m^3/ha) (m^3 /ha) Residential 0.0000 2.94E+17 6.898 1151 1016 Commer c i al 0.7621 6.10E+17 11.475 667 1016 Industrial 2.3154 3.01E+18 18.221 9410 1016 Mining 0.0252 9.14E+16 0.200 284 1016 Power 0.8410 3.89E+19 6.163 210 1016 Citrus 0.0097 1.95E+16 0.151 2 467 1016 Veggie 0.0336 6.00E+16 0.352 5619 1016 Nursery 0.0263 4.97E+16 0.410 2551 1016 Crops 0.0078 1.55E+16 0.169 498 1016 Cattle pasture 0.0025 5.40E+15 0.034 1228 1016 Cattle range 0.0001 4.63E+14 0.002 2184 0 Logged forest 0.0003 6.53E+1 4 0.006 2089 0 Upland forest 0.0000 1.62E+14 0.000 2091 0 Wet forest 0.0000 1.72E+14 0.000 2598 1524 Wetland 0.0000 1.82E+14 0.000 2078 1524 Water 0.0000 2.13E+14 0.000 1790 3048 Stream 0.0000 2.13E+14 0.000 0 3048 Salt marsh 0.0000 1.82E+14 0.000 0 0 Estuary 0.0000 2.13 E+14 0.000 0 0 Open 0.0000 1.52E+14 0.000 1658 1016
91 Figure 3 5 Change in land use distribution for regional optimization showing a shift away from citrus and in to more cattle production.
92 Figure 3 6 Sensi tivity analy sis of the groundwater r echarge c onstraint showing how increasing the required groundwater recharge results in shifts into first pastured cattle, and the range cattle, while decreasing the required groundwater results in a shift into citrus.
93 Figure 3 7 Total regional production and shadow prices for groundwater recharge and floodwater storage
94 CHAPTER 4 ENERGY AND CARBON BALANCE IN A REGIONAL MODEL Introduction This chapter outlines the addition of energy and GHG emission consideration into the optimization model for the Peace River region. This required the definition of the GHG emission characterist ics of the land uses in the model. GHG emissions and as previously defined in the regio nal EIO LCA mod el were land uses. In addition, new land uses that produce ren ewable energy for the region were defin ed based on information on recently constructed and planned alternative energ y projects in the region. Sustainabil ity constra ints dealing with GHG emission s and fossil fuel use are incorporated into the optimization model, and the maximum sustainable production is recalculated with these new constraints and land uses At the conclusion of the chapter, the results of t he optimization model are repor ted and the impacts of the new sustainability constraints and the new renewable energy land uses are evaluated. Energy is a key limiting resource that is required for regional production. Regional production utilizes both ren ewable energy sources from the sun and rain, and non non renewable energy is imported, and is purchased with m oney earned through exports. A s to imported energy is the ratio of money received for its exports to money spent on importing energy. I n the future, increased competition for energy sources, combined with decreasing supplies could substantially decrease this ratio and increase the cost s of imported energy. In the future, e nergy
95 security could become a significant challenge to systems seeking to maintain sustainable levels of produ ction. While a region may achieve higher productivity using imported energy sources while energy pric es are low, long term sustainability will be higher if a region can increase its reliance on internally generated energy sources. Energy and GHG emission are closely linked, and should be con sidered in concert with one another This is because f ossil fuels the m ajor energy source in developed regions, emit greenhouse gases as they are combusted. While a greenhouse gas balance is not required to maintain sustainable production at the regional scale, a global balance of greenhouse gas emissions is important to clim atic stability, which has the potential adversely impact the region. In addition, n ational level emission policies may eventually be translated to the regional level for implementati on producing a limiting constraint on emissions. One strategy for reducin g GHG emissions is to shift to renewable en ergy sources that have lower GHG emissions. Part of t he motivation for this chapter is to begin building analysis tools that will help region s evaluate and plan a transition to increased reliance on internal energ y sources Methods Energy U se of Regional Land Uses Each land use in the Peace Region was characterized in terms of its f ossil energy consumption The ener gy use of each land use was calculated using the regional EIO LC A model developed in Chapter 2 The GRID database 69 provided electricity production data for the Peace River region. Since data for 2 002 was not available; data from 2000 Industries that generate electricit y for their own internal use were not included in the fuel use of the power generation industry. The electricity generated by these indu stries was considered
96 to be consumed within the industry it self, and was not counted as electricity entering the regional power grid Residential electricity and fuel use for the Peace River region was estimated using the per capita energy use of Florida m ultiplied by the population of the region GHG E mission s of Regional Land Uses To calculate the net GHG emission for each land use in the model, the env iro nmental carbon sequestration of each land use was estimated and added to any economic emission s to o btain a n et emission value for that land use Economic emissions for each land use were calculated using the regio nal EIO LCA model from Chapter 2 Indirect emissions that occur red in other land uses but were driven by the economic activity in another land use were added to the driving land use and subtracted from the emitting land use, using the methodology described in Chapter 3 for economic linkages. An average environmental GHG sequestration was estimated for each land use as well Average n et primary p roduction (NPP), the difference between total production wn consumption, was used as an upper bound on how much carbon could be stored annually in a particular land use. For upland land uses, each land use was assigned an average carbon seq uestration value based on a characteristi c NPP for that land use. 70 NPP was then converted to metric tons of carbon dioxide equivalents (MTCO 2 e/ha/yr) to give an equivalent GHG sequestration rate Wet land uses, including wetlands, lakes, and streams w ere assigned a zero value of GHG sequestration Wet land uses emit CH 4 and N 2 O as part of the ir biogeochemical cycling, and although these emissio ns are small with respect to carbon f lows, their higher warming potentials give them disproportionate impact As a result, GHG emission and sequestration effects can
97 cancel out in wet environments The literature reports a wide range of values for wet land uses from net emission to net sequestration based on varying climate and hydrology conditions 71 Based on the unresolved uncertainty associated with freshwater wetland hydrology the assumption was made that on average the wide variations in emission and sequestration cancel out f or wet land areas. Definition of New Land Uses For a region to increase i ts energy sustainability, it needs to increase the use of renewable sources of energy from within the region. There are relatively few options for renewable energy within the Peace R iver region Wind and geothermal energy reso urces are limited with in the region and h ydropower is limited because of low topo graphic relief Solar insolation is high, however, and represents a significant energy source. B iomass production is also high in the region due to high insolation, plentiful rainfall and high average annual temperatures that allow for extended growing seasons. These two resources currently represent the best options for renewable energy production in the region. Incorporating biom ass energy into the regional model requires an accounting of the carbon contained in the biomass fuel sources. Fuels derived from regionally harvested biomass re emit their stored carbon into the atmosphere as they are consumed. The sequestration of carbon was assigned to the land use where the net primary production occurred. Re emission of carbon was assigned to the land use where the emission occur r ed, which was not always the land us e from which the stored carbon originated. This allocation convention a llowed renewable energy sources to be substitute d within the optimization model for non renewable energy sources while still accurately tracking total GHG emissions for the region
98 T o evaluate the potential impact of implementing renewable technologies on a regional scal e, each proposed renewable energy technology had to be transformed into a new regional land use. Economic and environmental characteristics at th e landscape scale were collected using available data on existing or proposed installations. A p recently constructed DeSoto Solar Energy Com plex located near the city of Arcadia within the Peace River watershed 72 A bioetha nol production land use was modeled after a proposed bioethanol plant to be built in Hi ghlands County, just south of the Peace River region. 73 This plant is to be constructed by US Envirofuel and plans to use a combination of sugar cane and sorghum as bioethanol feed stocks. The third alterna tive energy source was a biomass fueled electric power plant. Such a plant is currently under construction by US EcoGen within Polk county in the Peace River region 74 This plant plans to use wood from eucalyptus trees grown in plantations on reclaimed mining land to produce electricity for wholesale within the region Data to characterize the resource requirements and land use impacts of these three land uses was collected from publically available sources 72 74 The environmental water requirements for each new land use were also defined. The ph otovoltaic land use was not considered to have any continuing economic input requirements once the plant was constructed. Environmental water flows for the land use were estimated using the same values as were used for the open land use. Water requirements for growing sugarcane and sorghum in southern Florida were obtained from Evans and Cohen 75 Sugar cane was reported to have 1100 mm of ET per year, and have an average irrigation rate of 725 mm/yr. Sweet sorghum is a much more water
99 efficient crop, reported to have 650 mm of ET during its 120 day growing season. If the land area is left fallow between harvests an annual ET for the land use of 825 mm/yr of ET is estimated to be required with an average irrigation rate of 250 mm/yr Water requirements for the eucalyptus plantation were assu med to be the same as for a Florida pine plantation gi ving an annual ET of 889 mm year. Definition of Energy and Emissions Constraints: Two additional constraint s were added to the model, a GHG emission constraint and a renewa ble energy constraint. Total GHG emissions are a concern at the global scale, and so any local constraints are likely to be based on implementation of policies at either the global or national level. An emissions constraint was selected to meet the proposed Kyoto protocol global polic y objectives, which called for a 9% reduction from 1990 emission le vels within the U.S. 76 Ca lculating the total reduction for the region required an estimate of both the 2002 and 1990 emissions for the state o f Florida. These were calculated m 1990 to 2002 was found to be an 11% increase Adding this to the proposed 9% reduction gave a total reduction of 20% required from 2002 emission levels in order to meet the Kyoto standard. In order to transfer this standard to the regional level, t he ass umption was made that each region within Florida would be responsible to meet the same percentage reduction as required for the state. The initial conditions for the Peace River regional optimization mode l were used to calculate the GHG emiss ions accounting for both environmental and economic emissions. T he goal of a 20% reduction was applied to this value to set the emissions constraint
100 A fossil fuel constraint was also developed for the region. Energy use within the model was separated int o two categories: energy use requiring direct consumption of fossil fuels, which included the production of electricity, and energy use from the consumption of electricity. For both of these categories, n et energy demand from a land use was assigned positi ve values, while net energy production was assigned negative values. This convention allowed any net renewable fuel production to off set fossil fuel demand, and net renewable electricity production to offset electricity demand S ince renewable energy produ ction could be used to replace fossil fuel use the constraint had the effe ct of limiting fo ssil fuel but not necessarily reducing the total energy use with in the region Total energy use could however decrease if required in order to meet the constraints and maximize An additional constraint was employed for solar electricity production. Since large scale electrical energy storage is not yet commercially viable solar electricity is effectively constrained to provi de only peaking power during the day. A constraint was employed in the opt imization to limit solar electricity to account for only 10% of the total electri city production for the region. This preve nted the model from assuming base electricity needs could b e met with solar energy. R esults Land Use M odels The GHG emission and energy use characteristics for the original land uses in the model were defined. The average carbon sequestration rates for the environmental part of the land uses are given in Table 4 1. Forest and Logged forest had the highest GHG sequestration rates followed by improved pasture. Citrus and nursery land uses had intermediate sequestration values, followed by range land, open land, and mined
101 land, the majority of which is in various st ages of reclamation. Residential, commercial, industrial, and power generation land uses had decreasing amounts of sequestration, due to the increasing intensities of these land uses. Finally, vegetable and other crops were considered not to have sequestra tion because these crops are removed from the land use annually. Wet land uses all had zero sequestration values The environmental GHG sequestration values were combined with the economic GHG emissions for each land use from the EIO LCA model. Table 4 2 gives the list of combined energy and GHG characteristics per hectare for the original land uses in the model Of all the land uses that had an economic component, only citrus, cattle, and logging retain ed net sequestration characteristics after the eco nom ic emissions were accounted for. Upland forest wa s the only environmental land use that had net sequestration values. The new land uses were also characterized on a per hectare basis so they could be added to the model. Table 4 3 gives the list of metrics and associated values calculat ed for the new land uses Comparing these on a per hectare basis, s olar electricity has the high est production value and the highest groundwater recharge. After construction is complete solar power generation has minimal wa ter or fuel requirements, and no carbon emissions from electricity production. Some fossil fuels will be required for mowing during summer months but these requirements are small and they were c onsidered negligible within the model. Ethanol production i s the only alternative energy land use that has an output that can substi tute directly for fossil fue ls in transportation activities It is also the only renewable energy land use that has net ca rbon sequestration, due to the carbon
102 retained in the ethanol it produces as an end product While this carbon will be re released as it is consumed in other industries, those emissions will be attribu ted to the consuming industry. Sugarcane bioe thanol production is a high net consumer of groundwater due to sugarcan e irrigation. The sorghum only bioethanol land use increased the la nd area needed to produce the same amount of energy but dramatically de creases groundwater use requirements The biomass fired electricity land use emits the carbon stored in the trees as it burns them to produce electricity Since both the sequestration and emission of this carbon are occurring in the same land use, no net GHG emission is created. However, the harvesting of trees requires fossil fuel use as does the initial start up of the power plant, so the land use acquires a net GHG emission. Its GHG emission rate however, is significantly less than the emission of the power generation land use that is based on burning fossil fuels allowing it to dec rease net regional GHG emissions if it is substituted for fossil fuel based electricity generation. The groundwater recharge of biomass fired electricity is higher than that of bioethanol, as there is no irrigation of the tree crop. Initial Optimization Th e optimization model was updated with the new ly defined land uses. In the initial model run, regional output was optimized without adding GHG emission and fossil fuel constraints. The resulting land use distribution included the maximum allowable solar ele ctricity generation in the region. Only the solar energy land use was selected in the unconstrained model run out of the new land uses. The inclusion of solar power
103 resulted in a $14M increase in regiona l production over an optimized region with out any new alternative energy land uses Greenhouse Gas C onstraint A s et of model runs were performed to evaluate the impact of incorporating a greenhouse gas constraint on regional production The initi al constraint required a 20% GHG reduction consistent with th e Kyoto protocol Under this constraint, l ogg ing land use increased from 0 to 254,615 ha, while electric power generation decreased by 2.5 %. Regional economic production dropped by $77 M, a 0.28% drop from the ma ximum production without the GHG constra int For comparison, the 20% GHG reduction constraint was also evaluated without the addition of renewable energy land uses, and a $170M drop in regional product ion was incurred The addition of the solar land use then, resulted in an increase of $93 M in re gional output when under the 20% GHG emission reduction constraint. A sensitivity analysis was conducted varying the emission constraint between a 20% and 100% reduction. The results of these model runs are shown in Table 4 4 The regional output de crease d by $77M for a 20% reduction and by $1,463 M for a 100% reduction in GHG emissions. A regional shadow price was c alculated per metric ton of carbon dioxide equivalents, and was found to increase from $90 to $413 per MTCO 2 e as the GHG reduction requirement increased Figure 4 1 shows th e changes in land use required to meet the GHG emission constraint for each model run As the constraint increased land use shifted into logging and away from citrus, pastured cattle, and power production. This allowed the la ndscape to sequester more carbon in the products of the logged forest. This pattern held up to the 70% reduction constraint when sorghum based bioethanol began to be substituted for logging land use. The
104 optimization was able to achieve a 100% reduction i n GHG emission, meaning that GHG emissions were completely balanced with sequestration. However, a t this most string ent constraint condition, $1.46 billion had been lost from regional economic output, and the regional population h ad decreased by 14%. Fossi l Fuel Constraint A second set of model runs evaluated the fossil fuel constraint independent of the GHG constraint. The initial constraint was based on a proposed fossil fuel reduction of 20%. Under this scenario, the region added 132,752 ha of sorghum ba sed bio fuel land use. Total regional production decreased by $125M, which was a 0.47% loss from the maximum production without the constraint. In comparison, meeting this cons traint without any offset from renewable energy sources resulted in a $315M loss to regional production. Incorporating the 20% fossil fuel offset constrain t had the additional impact of reducing GHG emissions by 16%. A sensitivity analysis was conducted varying the fossil fuel constrain t between a 20% and 70% offset of non renewable fossil fuel. The results of these m odel runs are shown in Table 4 5 At the 70% offset constraint, the loss in production was $1,433 M, a 5.35% drop from the maximum pro duction without this constraint. In this scenario almost all agricultural and forest la nd had been converted to energy production. The 80% cons traint had no solution for the optimization as not enough land area that was not already under constraint was availab le to meet the additional energy production requirement required Figure 4 2 show s the land use shifts required to meet the increasing fossil fuel constraint in the region. Pasture land for cattle production steadily gives way to sorghum production to meet the energy requirements of the region. Because of water limitations,
105 sugarcane b ioethanol is never selected for inclusion in the region As pastureland decreases and sorghum bioethanol increases, water runoff in the model declines, and the need for wet forest which is providing runoff storage declines. It is subsequently converted to rangeland pasture I n stark contrast to the previous model, all upland forest is removed from the model, because it does not add to the energy supply, and its ability to sequester carbon is not valued without a GHG emission constraint. Optimization with Co mbined Constraints A final optimization was performed with both the 20% GHG reduction constraint and the 20% renewable fossil fuel constraint operating simultaneously. The resulting l and use distribution is given in Figure 4 3 and compared with both the in itial land distribution and the optimized distribution with no energy or GHG constraints In the fully constrained model, s olar energy and sorghum based biofuels are both selected for incorporation i nto the region Land use shifts away from citrus and catt le, to accommodate the significant increase in sor ghum based bioethanol. Logged forest also increases due to it s ability to sequester carbon. Total f orested land area in the region is dramatically increased, as all freshwater wetland was converted to fores ted wetland to maximize water recharge. The scenario that was optimized without GHG emission or fossil fuel constraints resulted in a 1. 83% increase in economic output, while the constrained scenario resulted in a slightly lower increase of 1.34%. The loss due to increased sustainability constraints was only .5% of total output, having a value of $128M. This develop scenario still represents a $352M gain from the initial economic output.
106 Comparison of Constraints Figure 4 4 shows a comparison of three susta inability constraints, groundwater recharge, fossil fuel reduction, and GHG emission, varied independently of each other The impact of each constraint is shown in terms of its effect on regional production. The graph reveal s tha t in terms of percent chang e g roundwater recharge is the most limiting factor in this region The loss in production is steep er with percent changes of groundwater recharge than percent changes in GHG emissions or fossil fuel reduction T he change in production is fairly linear for the percent changes of groundwater rec harge. This is not the case for changes in ener gy and GHG constraints, which have fairly low impacts on regional production for small percent changes, but have large impacts as their percent changes increase. Discussi on This chapter considers a wider picture of sustainability for the Peace River region, adding energy and emission constraints to the water constraints evaluated in Chapter 3. One of the questions that this model is designed to answer is to find where the region s current production is in comparison to its maximum sustainable production level. there is still room for development in the region. Optimization u nder water constraint s alone resulted in a 1.8% increase in economic output. This increase comes largely from re allocating the open land in the region, including un reclaimed mining lands, to productive land uses. Including the GHG and fossi l fuel constraints lowered this inc rease in economic output to 1.34% This suggests that proposed 20% GHG emission and fossil fuel sustainability constraints could be m et while still allow ing room for economic growth in the region
107 It should be noted that the land use shifts required to ach ieve this growth are substantial The optimized solution meets t he constraints through a combination of shifting land uses to renewable energy production, redu cing electricity production beyond what the system itself requires and reducing the supported po pulation. The primary reason for the minimal economic impact is that initially, the monetary output for the new renewable energy land uses are only marginally lower than the moneta ry output for the land uses they are replacing. However, a s sustainability r equ irements are increase d the gap between the value of production of the replaced land use and the renewable energy land use widens and regional losses in production are incurred more rapidly Impact of Sustainability Constraints A consideration of Figu re 4.4 shows that for the initial conditions of the model, economic produ ction is most sensitive to the groundwater constraint. The slope of the g roundwater constraint is initially higher than the slope of the GHG or the fossil fuel constraint. This sugges ts that t he actual groundwater recharge required for maintaining aquifer levels will be an important question to answer for any regional planning organization. The optimization model does not answer that question. For this model, the groundwater constraint was chose to maintain the level of recharge that was already occurring in the region for the year of the model, but this may or may not be sufficient to maintain aquifer levels over the region. As percent changes increase for the constraints the slope of the fossil fuel and GHG emissions constraints become steeper. Eventually, the slope of the fossil fuel constraint becomes steeper than that of the water constraint, and fossil fuel use has the potential to become the limiting production factor. However, t he solution space of the
108 model closes before the crossover point with the water constraint is reached. This lower limit to the solution space is caused by holding constant the commercial and industrial production that is not related to the other land uses. This requires that a minimum amount of water and energy resources be made avail able for this production. Allowing shifts in the area of these land uses would extend the solution space of the model and crossover points could be observed This boundary on t he solution space also prevents a scenario with 100% of energy coming from inside the region from being evaluated. Implementing the renewable energy constraint has the additional effect that it results in a decrease in GHG emissions. However, the resultin g land use distribution has significant differences. Implementing GHG reductions alone resulted in increases in logged forest area because of the carbon sequestration that forestry land uses provide. If the region has a high value for natural landscapes, o r desires to maintain higher storages of both energy and fiber in the region, then the GHG emission constraint helps to value those landscape functions. The renewable energy constraint by itself values producing energy on the landscape at the maximum rate. It selects sorghum biomass because of the high energy yield, the low water requirement, and the higher price that it commands per hectare than other renewable energy sources. Impact on Population A n important result of the model is that t he populat ion sup ported in the region declines with increased sustainability cons traints. While the optimized region under only the groundwater constraint will support a 6 .3 % increase i n population, adding the 20% GHG and fossil fuel constraints results in only a 2.5% incr ease in supported population In the 70% renewable energy constraint, renewable energy land uses consume 52% of the variable land use in order to support the residential population and industrial and
109 commercial production. While the monetary loss is only e stimated to be 5% of total economic production in this scenario, the residential population loss exceeds 15%. In this case the model is trading addition population for additional production. The model is operating under the assumption tha t the average resi dential resource consumption is always maintained, and so therefore population must decrease constraints. In reality there are other options available to a region to meet these constraints Efficiency increases in the residential reso urce consumpt i on including water, energy, and land requirements could be employed to continue support for a higher population. Allowing d ecreases in average stan dard of living would also allow the region to support a higher population in the region Ultim ately, the regional stakeholders must agree to the goal the region wants to achieve, whether that is maximizing the supported population, or maximizing the standard of living of the final population, or a balance in between. The valuable aspect of the opti mization model is that it allows for an estimate of the extent to which efficiency measures would need to be employed to maintain the desired regional population at the maximum sustainable production level Impact of Renewable Energy Land Uses Th e introduction of the renewable energy land uses gives the region additional options to meet both GHG emission and fossil fuel reduction constraints. These land uses generate both jobs and economic output within the region. However, depending on the land u se, they can also be more resource intensive than the land uses they replace The optimization model provides a way to evaluate the mix of renewable energy land uses that provides the best value to the region.
110 Including renewable energy land uses also resu lts in a decrease in GHG emissions. However, the resulting land use distribution has significant differences. Implementing GHG reductions resulted in increases in logged forest area because of the carbon sequestration that forestry land uses provide. If th e region has a high value on natural landscapes, or desires to maintain higher storages of both energy and fiber in the region, then the GHG emission constraint helps to value those landscape functions. The renewable energy constraint by itself values prod ucing energy on the landscape at the maximum rate. It selects sorghum biomass because of the high energy yield, the low water requirement, and t he higher price that transportation fuels can generate Groundwater recharge continues to be a limiting factor i n providing renewable energy in the region. However, it should be recognized that this is due in part to technological limitations in solar energy. If solar energy could be used as base power, and not just as peaking power, then it becomes the best option for the region in terms of resource requirements. However, even if the energy storage issue were to be solved, capital requirements to implement solar remain a significant barrier. Regional Implications The results of the current optimization model have so bering implications for regional growth of the Peace River system When under even partial sustainability constraints, t he production capacity of the landscape is limited in its ability to provide the additional production needed to drive economic or popul ation growth Instead, population decrease is suggested by the model to maximize regional output. A lternative energy land uses are limite d in their ability to create economic growth. Solar energy and low water requirement bioethanol production show promise in helping meet initia l renewable energy and GHG reduction goals. However, given the current renewable
111 energy options, the land, water, and capital requirement s are too large to make even an agricultural region like the Peace River fully energy ind ependen t at current resource consumption rate s Even with renewable energy land uses, the Peace River region has limited capacity for sustainable growth. The implications of the model should be considered in light of predictive capability. The model o nly extrapolates the current economic structure of the region. Significant economic restructuring can occur in the commer cial and industrial sectors that were held constant in the model. These changes could be a future driver of economic and population gro wth. However, it should also be considered that any additional commercial and industrial production will require additional resource consumptio n. Future resource intensities may not remain constant into the future. Energy and water efficiency can be increa sed in many production land uses to alleviate some of the resource constraints encountered in the model The ability of the model to forecast which resources may be most limiting as the region develops, and a shadow price for those resources, can be helpfu l in directing resource efficiency investments. Finally t he technology of renewable energy production can improve over time Current technologica l barriers can be overcome. The optimization model can provide insight into the characteristics that renewable technologies and land uses will need to have in order to compete for resources with in a region. This information could help plan both technology investments and implementation. The regional optimization model is a useful tool to envision the outcome of sy stem if
112 forecast limitations to future production can be used to help transition smoothly to a maximum sustainable production.
113 Table 4 1. Carbon sequestration estimate s of Peace Riv er region land uses Land Use NPP Carbon Stored Equivalents of CO2 Total GHG sequestered (g C/m2/yr) (MT C/ha/yr) (MT CO 2 /ha ) (MT CO 2 /ha ) Residential 172 0.86 3.16 3.16 Commerci al 27 0.13 0.49 0.49 Industrial 0 0.00 0.00 0.00 Mining 300 1.50 5.50 5.5 0 Power 27 0.13 0.49 0.49 Citrus 325 1.63 5.96 5.96 Vegetable 325 1.63 5.96 0.00 Nursery 325 1.63 5.96 5.96 All Other crops 325 1.63 5.96 0.00 Cattle Pasture 400 2.00 7.33 7.33 Cattle Range 300 1.50 5.50 5.50 Logged forest 600 3.00 11.00 11.00 Upl and forest 600 3.00 11.00 11.00 Wetland forest 800 4.00 14.67 0.00 Wetland 1000 5.00 18.33 0.00 Lakes 200 1.00 3.67 0.00 Stream 200 1.00 3.67 0.00 Salt marsh 1000 5.00 18.33 0.00 Open Land 300 1.50 5.50 5.50
114 Table 4 2. Energy and GHG emission ch aracterization of existing land uses Land use Carbon Balance Electricity Balance Fossil Energy Balance (MT CO2eq/ha) (TJ/ha) (TJ/ha) Residential 53.89 0.3432 0.8054 Commercial 39.51 0.3756 0.5469 Industrial 350.77 1.3759 7.2171 Mining 11.20 0.2 082 0.2673 Power 8425.33 46.0169 135.7594 Citrus 3.29 0.0045 0.0412 Veggie 8.13 0.0137 0.1270 Nursery 0.82 0.0121 0.1058 Crops 2.72 0.0057 0.0421 Cattle pasture 6.32 0.0020 0.0156 Cattle range 5.44 0.0001 0.0009 Logged forest 10.93 0.0001 0.00 13 Upland forest 11.00 0.0000 0.0000 Wet forest 0.00 0.0000 0.0000 Wetland 0.00 0.0000 0.0000 Water 0.00 0.0000 0.0000 Stream 0.00 0.0000 0.0000 Salt marsh 0.00 0.0000 0.0000 Estuary 0.00 0.0000 0.0000 Open 0.00 0.0000 0.0000
115 Table 4 3 Chara ct erization of alternative energy land u se s Land Use Economic Output Jobs Water Balance Flood Balance Carbon Balance Electricity Balance Fossil Energy Balance ($M/ha) (jobs/ha) (m^3/ha) (m^3/ha) (MT CO 2 e/ha) (TJ/ha) (TJ/ha) Biomass Electric 0.0012 0.003 2 300 1016 0.083 0.063 0.000 Solar 0.0314 0.0211 3307 1016 0.000 1.592 0.000 Sugarcan e Bioethanol 0.0052 0.0042 1357 1016 6.125 0.044 0.087 Sorghum Bioethanol 0.0049 0.0040 11 1016 4.755 0.042 0.071
116 Table 4 4 Impact of GHG emission con straint on output and shadow price Output Percent change Shadow price Percent change ($M) ($/MTCO 2 e) No Constraint 26755 20% 26678 0.28% 90 0.0% 30% 26595 0.59% 90 0.0% 40% 26510 0.92% 90 0.0% 50% 26423 1.24% 98 9.6% 60% 26329 1.5 9% 98 9.6% 70% 26233 1.95% 121 34.9% 80% 26027 2.72% 381 325.7% 90% 25662 4.08% 381 325.7% 100% 25292 5.47% 416 364.0%
117 Figure 4 1 Sensitivity analysis of o ptimized land u se distribution under GHG c onstraints showing increases in plantation forest and then sorghum bioethanol as the GHG constraint increases.
118 Table 4 5 Impact of fossil fuel constraint on regional output Constraint Output Percent Change Shadow Price ( $M ) ( $/TJ ) No constraint 26755 20% 26630 0.47% 3523 30% 26551 0.76% 6102 40% 26415 1.27% 6102 50% 26233 1.95% 9586 60% 25835 3.44% 23017 70% 25322 5.35% 23017 80% No solution
119 Figure 4 2 Sensitivity a nalysis of o ptimized land use distribution under fossil fuel c onstraint showing a loss of citru s and pastured cattle and an increase in sorghum bioethanol to meet fossil fuel reductions.
120 Figure 4 3. Comparison of initial, o ptimized and fully constrained land u se
121 Figure 4 4 Impact of sustainability constraints on regional o utput sh owing the region is initially most sensitive to percent changes in groundwater recharge constraints. GHG emission and fossil fuel constraints initially display a smaller sensitivity, but increase in sensitivity as constraints are increased.
122 CHAPTER 5 EVALUATING THE MODEL Contributions of the Model The combination of EIO LCA with a land use optimization model provides a novel appr oach to regional sustainability modeling that addresses several shortcomings of other available methods The contributions of this wor k include a method for building a regional EIO LCA model, and a method of constructing a linear programming model that both captures sustainability constraints and accounts for internal interactions between the model components The regional EIO LCA model is attractive for use in regional planning in that it can be implemented quickly with modest effort The use of commercially available economic modeling software combined with publically available data can dramatically lower the tim e and cost required to develop the regional model without sacrificing too much regional specificity using a top down allocation scheme. The resource flows that comprise the largest portions of the total flow are speci fically allocated, while state and national averages are used to estimate the remaining flows. The use of the commercial ly available economic input output model provides standardized economic categories so that results are easily comparable with other reg ions and regions can also be scaled easily to larger geographic areas The commercial software also provides for readily available annual updates of the economic model. Public resource consumption data varies in collection timefr ames, but is generally upd ated every four to five years. This allows for the possibility of updating the model at regular intervals. A time series of regional EIO LCA models could give insight
123 into how the economic structure is changing over time, and the rate at which resource eff iciency may be increasing in certain industries. By analyzing these trends over time, projections can be made as to what future resource intensities of certain industri es may become, allowing for increasingly accurate optimization results. Construction of a time series of models can also provide an estimate of the uncertainties involved in the calculated resource intensities, allowing for uncertainty analysis to be conducted within the model. The way the linear optimization model is conceived helps to resol ve several challenges inherent in other modeling techniques. In order to move away from having multiple decision criteria to evaluate, a single objective function is selected. That objective is for the region to ma ximize its productivity to be as producti ve as it can be within the limits of the system. While regional stakeholders will need to agree to this objective, it has potential to be a common starting point of agreement even between those who disagree over the meaning of sustainable development. Sele cting a single objective function has the advantage that it removes sustainability as the end goal of the system, and instead sustainability consider ations become the constraints in the model. This gives more flexibility in how the sustainability constrain ts are handled. Sustainability constraints for a region can be levied from outside the system as national and state policy, or they can be selected from inside the system by stakeholders that want certain aspects of their system to be sustained into the fu ture. A valuable aspect of this model is that the required sustainability constraints themselves do not necessarily have to be known with certainty before model implementation T he sustainability constraint can be added to the model and the impact of diffe rent levels of
124 constraint can be evaluated in a sensitivity analysis In multi criteria decision models, sustainability considerations are made into systems goals, and these goals must be weighted relative to the other goals prior to running the model. The current construct allows for the sustainability constraints to operate concurrently in the model without an externally imposed hierarchy. There remain theoretical considerations as to how the productivity of the region should be measured. As was discusse d in chapter 3, both monetary and emergy valuations were considered and no difference in outcome was noted The use of monetary measures makes the model translate more easily into terms that regional planners will be familiar with. Emergy should be furthe r explored as a stable measure of output over the long term The use of total throughput as a production measure includes the flows that travel through different land uses multiple times. It would be possible to build the model and use value added or GDP a s a measure instead and avoid double counting. The decision to use to tal throughput instead of GDP was based on a desire to capture the internal cycling within the region. Regions with higher levels of internal cycling will be more resilient in the face of external disturbances. This represents a theoretical question that deserves additional attention, and could be an avenue for future work. A significant contribution to sustainability science is the method of constructing land uses so that they account for the indirect effects they have within th e region. Accounting for the indirect resource use helps to evaluate all the tradeoffs that must be made in the sys t em when a major resource becomes limiting. The use of both backward and forward linkages in the mod el helps to maximize the benefit the region could see
125 from increases in land uses. The optimized region then represents an upper bound on higher value added products. There are many opportunities to expand and improve the capabilities of the model. The current model operates with just 16 economic land uses when including the renewable energy land uses and only 4 variable environmental land uses A more detailed model with additional land uses would increase the number of variable economic interactions considered in the model. Commercial and industrial sectors which are currently highly aggregated co uld be separated into additional land uses within the model. The reside ntial sector could also be separated into low, medium, and high i ntensity housing. The model, however, is fundamentally limited to land uses that have resource requirements that change linearly with land area and so many industries will need to remain agg regated Once the regional model is developed, it remains flexible in its ability to accommodate new information. It can continue to be updated with new land uses and land use characteristics as this information comes available. The wide scale implementat ion of best management practices may significantly shift the resource consumption of certain land uses in the future Within the model, any land use can be divided into a land use utilizing old practices, and a land use utilizing new management practices t o evaluate the impact of best management practice within t he region. In the same way, renewable energ y land uses can be updated as technology improves efficiency and yield.
126 Additional work can also be done to make the model better reflect the region in qu estion. IMPLAN models can be modified with regional data to better tailor the interactions of the region. More specific data on regional industries will allow mo dels to be refined to more closely match the economic characteristics of a region. T he resource consumption of major regional industries can also be further tailored with the use of detailed studies of these industries, or through the use of estimates of similar industries using process based LCA. This data can be incorporated in the place of nation al average estimates to increase the accuracy of the regional model. A final area for extension of the model is the array of sustainability constraints that are considered. The current implantat ion of the regional model was limited in scope to water, energ y, and GHG emissions. In a highly agricultural region, soil conservation and nutrient availability can be important limiting factors to production as well. A more in depth analysis would consider carbon and nutrient storage in the In additi on, the eutrophication of streams and lakes in th e region is of concern and could also be addressed with information on nutrient use and emission from each land use. Data is also available on regulated air emissions from certain industries in the region th at could be considered. Limitations of the Model The regional model has several important limitations and the interpretation of the results should incorporate an understanding of these limits The current model has no spatial analysis. Only total area and average land use characteristics are considered, and the impacts of spatial relationships are neglected. While the simplification of regional interactions is necessary to build a useful model, lack of a spatial component ignores many important questions. For example, the location of groundwater recharge
127 has a major impact on its ability to maintain sustainable aquifer levels within the region. Groundwater recharge needs to be spread appropriately across the region to maintain aquifer levels. While it was c onsidered to be beyond the scope of this modeling effort, spatial modeling is necessary to fully understand the spatial constraints of the system. An additional drawback of the model is that does not include any analysis of the impact of loss of diversity within the region. Two kinds of diversity should be considered. Biodiversity within the region is not valued as an ecosystem service by the current set of constraints. While it could be assumed that natural areas in the region will be appropriately managed for biodiversity, there is not yet any mechanism for ensuring that an appropriate mix of habitats is provided. The result is that entire land uses can be eliminated in the optimization, i.e. upland forest, or herbaceous wetland. One way to address this sh ortcoming is to add constraints that provide minimum levels of certain habitat types in the region that help ensure the maintenance of diverse plant and animal life. A second kind of diversity to consider is the diversity of economic activity. A system tha t devotes all its land use to one type of activity is more likely to suffer from rapid changes in market conditions, or natural disasters and climatic variation. The inclusion of a metric to weigh the stability of the optimized system would help regional p lanners decide how much optimization they want to pursue at the expense of system stability. Finally, not all factors have been included in the evaluation of land uses. Capital costs of develo ping the land use are important but not fully included in the cu rrent model. Although regional productivity may be increased by a particular land use, high capital cost s or low return rates on that capital investment could stand in the way of their d evelopment, particularly if the land use is dependent on external sour ces of capital. In
128 the current regional model, the photovoltaic solar land use appears to be an obvious choice for development, and it competes economically with current land uses. This is because the high capital cost of installing solar was not evaluated by the model. Capital costs should be better integrated into future implementations of the model. Conclusion The regional EIO LCA optimization model can be a useful tool to evaluate future provides value in that it i ncorporates both the environmental and economic spheres, and explicitly considers both limiting resource consumption and sustainability constraints. The model is simple in its form, and does not explicitly account for certain nonlinear interactions within regions. However, because the model is constructed so that it bounds the possible solution space, it is still possible to explore the limits of available development options. The model provides a much needed link between the complex environmental interacti ons, and the complex economic interactions within a region. Finally, the model provides a means to estimate the maximum sustainable production level for a region. While acknowledging that much information is still uncertain about the future of the system it still give s regional planners a picture of what a future system would look like based on an extrapolation of the current organization. This allows important questions to be explored in terms of the goal s of system development and resource constraints t hat will be encountered along the way.
129 APPENDIX A REGIONAL WATER ALLOCATION This appendix outlines the steps that were taken to develop the regional water intensities for Florida and the Peace River region. The first step was to collect data from gover nment reports on total water use for each region. Water use for the state of Florida is reported every 5 years by the USGS. The closest data set was the year 2000 water use report. The total consumption data is reported in Table A 1. M atching Florida econo mic data for the year 2000 was 2000 data wa s reported in an older industry categorization scheme and a bridge table was used to match industries between the year 2000 and the year 2002 categories The second ste p was to d evelop water intensities for individual industries USGS estimates of water use had sufficient category resolution that water use could be matched directly with 13 agricultural industries, 4 mining industries, and the e lectric power i ndustry. Tab le A 2 gives the USGS data for agricultural industries. The rest of the industrial and commercial water use was highly aggregated and could not be directly allocated to corresponding IO industries. Instead, c orrection fa ctors were developed to adjust natio nal water intensities into regional water intensities. The correction factors ensured tha t the water use of each economic sector summed to the US GS reported value for that sector. The procedure was to multiply the regional economic output by the national w ater intensity for the corresponding industry. The resulting water use was summed for all industries in an economic sector. The USGS reported water use for that sector was then divided by the calculated water use from the national water intensities.
130 (A 1) This percent difference in totals was used as a correction factor for all the industries within that economic sector. (A 2) This allocation method wa s carried out separat ely for self supply water and public supply water sources as the economic aggregation was reported different ly for these two sources. The allocation of the self supplied water subcategories of food production, pulp/paper, chemicals, and other manufacturin g help ed to differentiate some of the high water use industries in Florida from national averages Self supplied water for the residential sector was estimated in the USGS report by assuming that residential per capita water use was the same as publicly s upplied per capita water use. The per capita water use is then multiplied by the population of Florida that is not served by a public water supply syste m to calculate total water use. Consumed water intensities were calculated by multiplying the water inte nsity by Florida average consumption perce ntages as reported in Table A 3 A water intensity vector was als o developed for the Peace River region. Water use is reported individually for each county in the Peace River region by the SWFWMD. The repo rted data is given in Table A 4 Correction factors for Polk and Charlotte County were developed since they have areas outside the SWFWMD that needed to be accounted for. The list of correction factors is given in Table A 5 Wastewater disposition was included in t he Peace River water budget model, and the values for the Peace river region are given in Table A 6.
131 Table A 1. Florida water use adapted from Marell a 39 Sector Subsector Industrial Electric Power Commercial Resi dential (gals) (gals) (gals) (gals) Public Supply 3.56E+10 3.11E+11 5.43E+11 Industrial Self supply Agriculture 1.43E+12 Mining 6.80E+10 Chemical industry 4.12E+10 Paper industry 5.56E+10 Food manufacturing 1.24E+1 0 Other manufacturing 8.24E+09 Electric Power Self Supply 4.60E+12 Commer c i al Self supply Golf courses 1.08E+11 Other recreation 4.21E+10 Other commercial 2.06E+10 Residential Self Supply 7.25E+10 Total 1.65E+12 4.60E +12 4.82E+11 5.43E+11
132 Table A 2 Fl 39 Ground Water Surface Water Reclaimed Water Total Water Use Water withdrawn Industry Crop Type (gals) (gals) (gals) ( gals) (gals) Oilseed farming 3.61E+08 2.92E+07 3.91E+08 3.91E+08 Soybeans 3.61E+08 2.92E+07 3.91E+08 3.91E+08 Grain farming 1.28E+10 1.29E+10 2.30E+08 2.59E+10 2.57E+10 Field corn 1.24E+10 6.83E+08 2.30E+08 1.33E+10 1.30E+10 Rice 0.00E+00 1.2 1E+10 1.21E+10 1.21E+10 Sorghum 3.32E+08 4.75E+07 3.80E+08 3.80E+08 Wheat 1.06E+08 1.10E+07 1.17E+08 1.17E+08 Vegetable and melon farming 1.12E+11 4.57E+10 1.20E+09 1.59E+11 1.58E+11 Vegetables 1.03E+11 4.33E+10 1.20E+09 1.48E+11 1.46E+11 Potat oes 9.34E+09 2.41E+09 1.17E+10 1.17E+10 Tree nut farming 3.03E+08 3.65E+07 3.39E+08 3.39E+08 Pecans 3.03E+08 3.65E+07 3.39E+08 3.39E+08 Fruit farming 3.98E+11 2.82E+11 6.94E+09 6.87E+11 6.80E+11 Blueberries 4.38E+08 1.10E+07 4.49E+08 4.49E+08 Citrus 3.85E+11 2.81E+11 6.94E+09 6.73E+11 6.66E+11 Grapes 1.17E+08 1.83E+07 1.35E+08 1.35E+08 Peaches 3.65E+06 3.65E+06 3.65E+06 Strawberries 4.78E+09 1.83E+08 4.96E+09 4.96E+09 Miscellaneous 6.75E+09 1.22E+09 7.97E+09 7.97E+09 Non specifi c fruit 5.66E+08 4.38E+07 6.10E+08 6.10E+08 Greenhouse and nursery production 1.03E+11 4.59E+10 1.63E+09 1.51E+11 1.49E+11 Field grown 1.99E+10 4.16E+09 1.09E+09 2.51E+10 2.41E+10 Greenhouse grown 5.11E+08 2.88E+08 7.99E+08 7.99E+08 Container grow n 3.81E+10 9.63E+09 5.40E+08 4.82E+10 4.77E+10 Sod 4.49E+10 3.18E+10 7.67E+10 7.67E+10
133 Table A 2. continued Ground Water Surface Water Reclaimed Water Total Water Use Water withdrawn Industry Crop Type (gals) (gals) (gals) (gals) (gals) T obacco farming 1.94E+09 3.65E+07 1.97E+09 1.97E+09 Tobacco 1.94E+09 3.65E+07 1.97E+09 1.97E+09 Cotton farming 3.01E+09 2.88E+08 3.30E+09 3.30E+09 Cotton 3.01E+09 2.88E+08 3.30E+09 3.30E+09 Sugarcane and sugar beet farming 1.13E+10 3.01E+11 3.1 3E+11 3.13E+11 Sugarcane 1.13E+10 3.01E+11 3.13E+11 3.13E+11 All other crop farming 6.85E+10 1.65E+10 2.28E+10 1.08E+11 8.50E+10 Peanuts 7.34E+09 4.60E+08 7.80E+09 7.80E+09 Pasture hay 5.88E+10 1.53E+10 5.37E+08 7.46E+10 7.41E+10 Other (grasse s) 1.59E+10 1.60E+10 0.00E+00 Miscellaneous 2.40E+09 7.34E+08 6.37E+09 9.50E+09 3.13E+09 Cattle ranching and farming 1.13E+10 5.51E+08 1.19E+10 1.19E+10 Livestock 1.13E+10 5.51E+08 1.19E+10 1.19E+10 Animal production, except cattle and poultry 2. 85E+09 7.67E+07 2.93E+09 2.93E+09 Fish farming 2.85E+09 7.67E+07 2.93E+09 2.93E+09 State Totals 7.26E+11 7.06E+11 3.28E+10 1.46E+12 1.43E+12
134 Table A 3. Average water co nsumption by sector 39 Category % Co nsumption Public Supply 38.7% Residential Self Supply 38.7% Industrial Commer c i al Self Supply 19.1% Agricultural Self Supply 69.5% Recreational Irrigation 80.0% Power Generation 9.0%
135 Table A 4. Peace River region water withdraws 40 Charlotte De Soto Hardee Polk Total Sector Subsect or (gals) (gals) (gals) (gals) (gals) Public Supply Residential 3.63E+09 3.11E+08 3.76E+08 1.76E+10 2.19E+10 Commercial 1.01E+09 1.65E+08 1.36E+08 4.53E+09 5.85E+09 Industrial 9.39E+07 2.92E+07 1.65E+07 5.83E+08 7.23E+08 Recreation 5.6 4E+07 0.00E+00 0.00E+00 2.67E+08 3.23E+08 Public Use 7.81E+08 8.27E+07 8.68E+07 3.74E+09 4.69E+09 Agriculture Self Supply Citrus 1.70E+10 2.53E+10 1.83E+10 4.96E+10 1.10E+11 Other Crops 1.87E+08 9.53E+07 2.05E+08 1.43E+08 6.30E+08 Nursery 8.7 5E+08 1.34E+09 1.16E+09 1.60E+09 4.97E+09 Vegetable 3.22E+09 9.32E+08 1.92E+09 6.75E+08 6.74E+09 Pasture 6.33E+08 1.19E+09 1.16E+09 9.51E+08 3.93E+09 Livestock 9.51E+06 4.02E+08 3.47E+08 9.79E+08 1.74E+09 Mining Self Supply Limestone 0.00E+00 0.00E+00 0.00E+00 2.15E+07 2.15E+07 Peat 0.00E+00 0.00E+00 0.00E+00 2.56E+06 2.56E+06 Phosphate 0.00E+00 0.00E+00 2.90E+08 8.35E+09 8.64E+09 Sand & Shell 2.12E+08 6.24E+07 0.00E+00 1.03E+09 1.30E+09 Other 6.28E+07 5.48E+06 0.00E+00 4.00E+08 4.68E+ 08 Industrial and Commercial Self Supply Manufacturing 2.17E+08 0.00E+00 3.65E+05 8.35E+09 8.57E+09 Food Processing 0.00E+00 1.50E+07 3.83E+07 9.47E+08 1.00E+09 Commercial 0.00E+00 0.00E+00 0.00E+00 1.86E+06 1.86E+06 Power 0.00E+00 9.49E+06 7 .12E+07 3.69E+09 3.77E+09 Other Industry 6.59E+06 0.00E+00 0.00E+00 4.37E+08 4.43E+08 Recreational Self Supply Parks 1.82E+08 3.29E+06 1.57E+07 1.86E+09 2.06E+09 Golf 6.94E+08 5.29E+07 8.83E+07 1.83E+09 2.67E+09 Other Recreation 2.56E+06 0.00 E+00 1.83E+06 6.19E+08 6.23E+08 Residential Self Supply Residential 1.38E+09 7.17E+08 3.74E+08 5.16E+09 7.63E+09 Total Water Use 3.03E+10 3.07E+10 2.46E+10 1.13E+11 1.99E+11
136 Table A 5. Peace River region water use correction factors Agricult ure Industry Domestic Recreation Public County Self Supply Self Supply Self Supply Self Supply Supply Charlotte 4.34 18.04 1.24 1.00 1.00 Polk 1.23 1.01 2.55 1.09 1.04 Table A 6. Peace River region wastewater disposition 39 Charlotte De Soto Hardee Polk Total (gals) (gals) (gals) (gals) (gals) Surface Applied 1.75E+09 8.98E+07 1.32E+08 6.32E+09 8.30E+09 Deep Injection 1.18E+09 0.00E+00 0.00E+00 0.00E+00 1.18E+09 Stream 0.00E+00 2.14E+08 2.73E+08 3 .63E+09 4.11E+09 Total Wastewater Treated 2.93E+09 3.03E+08 4.05E+08 9.95E+09 1.36E+10 Pe r cent Disposition Surface Applied 59.8% 29.6% 32.5% 63.6% 61.1% Deep Injection 40.2% 0.0% 0.0% 0.0% 8.7% Stream 0.0% 70.4% 67.5% 36.4% 30.3%
137 APPENDI X B REGIONAL ENERGY ALLOCATION This appendix outlines the steps that were taken to develop the regional energy intensities for Florida and the Peace River region. The first step was to collect data from government reports on total energy use for each regio n. Total energy use for Florida is given in Table B 1 adapted from the SEDS database. This data was used to set the total energy use for the state. The next step in developing the energy intensity vector was to disaggregate energy use by fuel type. Fuel oi ls and kerosene purchases are reported by state for (FOKS) report. Fuel totals in this data set were not the same as values reported in the SEDS database. The reported val ues were corrected so that they added to SEDS totals This was done A comparison of SEDS and FOKS values and th e correction factors used is given in Table B 2. Aft er adjustment to SEDS values, fuel use wa s allocated to specific transportation industries. Jet fuel and aviation gasoline from the SEDS report were l fuels reported in the FOKS report and off highway fuels were removed from the transportation sector and allocated to the industrial sector. Military fuel use was removed from the transportation sect or and assigned to the commercial sector as part of government fuel use. Vessel bunkering fuels were allocated between two
138 industrial sector. Since no state level data existed to allocate these fuels, several assumptions were made in their allocation. equipped to use this type of fuel. Commer cial fishing fuel use was calculated by economic output. The resulting commercial fishing industry fuel use was subtracted from the vessel bunkering distillate total, and th e remaining distillate fuel was allocated to water transport. The highway fuels reported in the FOKS report required additional data for further allocation. Federal Highway Administration (FHWA) data was used to allocate highway related fuel use. The FHWA reports state estimates of highway use of gasoline in Table MF 21 77 and off roa d gasoline use in Table MF 24 78 The FHWA data includes all gasoline sales, incl uding gasoline additives such as ethanol, and aviation gasoline. Transport ethanol and aviation gasoline use from the SEDS table were added together along with commercial, industrial, and transportation gasoline use from the SEDS table in order to arrive a t a total energy value of 982,015 TJ for all gasoline consumed in as shown in Table A 3. Off road fuel use wa s estimated by the FHWA using empirically based models for agricultural, construction, and industrial/commer cial vehicles. For the regional model, t hese fuels were removed from the transportation industries, and allocated to their respective industrial, commercial, and resi dential categories. Data from the Transportation Energy Data Book (TEDB) was used to allocate fuel use between
139 veh icle types. The TEDB reports highway energy use by mode of transport and fuel type for the entire U.S. Florida does not report energy use by v ehicle type, and so national percentages of energy use by vehicle type were used for an initial allocation of and medium and heavy trucks. Table A 4 gives the data from the TEDB and the share of energy use that each fuel represents. The 2002 Vehicle Inventory and Use Survey 47 was used to adjust national highway energy use to better reflect Florida energy use. The VIUS is a national survey instrument used to determine national truck inventories and uses. S urvey data is reported by state, and was analyzed to estimate truck fuel use by industry category and fuel type. Estimates of fuel use derived from the survey data are included in Table B 5 These estimates are then used to tailor the TEDB allocation of tr uck and automo bile use to Florida conditions as shown in Table B 6. The Florida VIUS fuel use estimates that light trucks account for only 37% of gasoline use, as opposed to national averages of 40%. The gasoline powered heavy trucks are estimated to accou averages of 3.5%. The resulting 6% difference in total gasoline use is assumed to be accounted for by increased automobile use within Florida from 56% to 62%. For diesel fuel, light trucks jump fro m a national average of 5% to over 10% of total diesel fuel. However, heavy trucks account for only 38% of diesel fuel use, as opposed to the national average of 90%. It is unreasonable to assume that diesel automobiles account for this 42% difference in f uel use. The more likely explanation is that the heavy trucks that consume the missing fuel are long haul vehicles traveling between states and are
140 not registered in the state of Florida. These vehicles would not be reported within hese heavy trucks would not be included in the economic total diesel fuel use and not allocated to any other sectors. This assumption can be S/D ratio for truck transport, which estimates that 83% of truck transport is supplied by local businesses. The national energy intensity vector assumes all automobile, motorcycle, and light truck fuel use is for personal transport, and removes their fuel use entirely from the sector. Light trucks, however, are often used in the self provided transport of industries. The state level VIUS data can be used to separate transport industry provided and self provided transport within both the light and heavy d ut y truck categories. Table B 7 shows the fuel use associated with VIUS industrial categories. The agricultural, construction, mining, and manufacturing categories are removed from the transportation sector and assigned to the industrial sector. All service truck fuel use is removed and assigned to the commercial sector. Personal transportation fuel use is removed and gned to the truck transport industry. The remaining automobile and motorcycle use is allocated to the residential categor y, a s this is the predominant use and no other data sets are available to allocate this category. Buses are used primarily within two allocates the fuel use based on the percentage of economic output of these two
141 industries. In Florida, over 80% of the rev from water transport, and only 20% comes from land transport (U.S. Economic Census, 2002) The water transport portion of the sightseeing industry is considered to come entirely from gasoline powered watercraft Marine gasoline fuel use is allocated to sightseeing by the ratio of commercial to privately registered boats in Florida This allocation assumes that privately owned and commercial boats have similar usage patterns and fuel consumption. Once commercial marine gasoline use is calculated, the remainder of gasoline marine fuels are removed from t he transport sector and allocated to residential recreational use. The assumptions involved are necessarily crude, and introduce large uncertainty in the water tran sport industries. Better methods are sought for their allocation. Pipeline transport represents natural gas, crude oil, and water transportation. In Florida, all pipeline transport is natural gas transportation. The EIA reports the natural gas consumed in pipeline transport ation, and this value wa s allocated to pipelines. The transportation electr icity use reported for Florida wa s also allocated entirely to pipeline use, as no electric rail existed in Florida in 2002. The final three transport industries h ave no state level data available. The national energy intensities are used for these industries. Their petroleum fuel use is assumed to be truck transport, while natural gas and electric usage is assumed to be for facilities and are subtracted from the co mmercial use categories. This method gives Florida industries within a sector the same relative resource use per dollar as the national model within a defined sector. However, it allows changes
142 in the percentage of resource use that each sector represents sumption is unique to the state.
143 Table B 1. Florida energy use 2002 adopted from SEDS 43 Residential Commercial Industrial a Transportation Power Totals Fuel ( TJ ) ( TJ ) ( TJ ) ( TJ ) ( TJ ) ( TJ ) Coal 30 223 32289 0 725989 758532 Natural Gas 16515 60924 90308 12647 564069 744463 Distillate Fuel 574 15765 43443 224765 22704 307251 Residual Fuel 0 470 10531 69164 285684 365850 Motor Gasoline 0 2178 13460 1016749 0 1032386 Kerosene 379 93 10 0 0 482 LPG 7638 10125 2570 651 0 20984 Aviation Gas 0 0 0 2618 0 2618 Jet Fuel 0 0 0 161567 0 161567 Other Petroluem b 0 0 13240 4862 0 18102 Petroleum Coke 0 0 0 0 50006 50006 Biomass 5095 1370 98022 39 47430 151956 Geothermal 2108 632 0 0 0 2740 So lar 29407 0 0 0 0 29407 Nuclear 0 0 0 0 370903 370903 Hydroelectric 0 0 0 0 2003 2003 Purchased Electricity 389031 299441 68194 211 0 756877 Totals 450778 391223 372067 1493272 2068787 4776126 a The industrial values report estimates of fuels with feed stock use removed b Other Petroleum includes asphalt and road oil, lubricants, pentanes, napthas, and waxes
14 4 Table B 2. Comparison of SEDS and FOKS energy values Distillate Fuel Residual Fuel SEDS 43 FOKS 45 Difference Adjusted SEDS 43 FOKS 45 Difference Sector Subsector (TJ) (TJ) % (TJ) (TJ) (TJ) % Residential 574 591 2.9% Comm ercial 15765 16215 2.9% 470 505 7.3% Industrial 43443 45256 4.2% 10531 10949 4.0% Oil 32 31 Farm 19535 18714 Off Highway 18237 17471 Other 325 312 Transport 224765 231180 2.78% 69164 74215 7.3% Railroad 11413 11096 Vessel Bunker 22411 21790 74215 On Highway 196510 191057 Military 846 822 Power 22704 18320 19.3% 285684 285924 0.1% Total 307251 311562 365850 371593
145 Table B 3. Adjustment of Florida gasoline use Sectors Subsectors Subsectors EIA SEDS a Table MF 21 77 Table MF 24 78 % Total Adjusted Value (TJ) (TJ) (TJ) (TJ) Highway Use 1010967 96.51% 998971 Private & Commercial 997152 95.20% 985320 Public Federal Civilian 1185 0.11% 1171 Public State & Local 12629 1.21% 12479 Non Highway Use Priv ate & Commercial 35900 3.43% 35474 Agriculture 3874 10.79% 3828 Aviation 3295 9.18% 3256 Commercial & Industrial 5266 14.67% 5204 Construction 4473 12.46% 4420 Marine 17394 48.45% 17188 Miscellaneous 1597 4.45% 1578 Public State & Local 606 0.06% 598 Total 1035043 1047472 1035043 a EIA SEDS total includes aviation gasoline and ethanol
146 Table B 4 Average highway fuel use adopted from TEDB 46 Gasoline Diesel LPG NG Total Gasoline Diesel LPG TBTU TBTU TBTU TBTU TBTU % % % Light vehicles 15871.1 310.6 10 16191.7 96.5% 6.3% 37.2% Automobiles 9273.9 52 .0 0 9325.9 56.4% 1.1% 0.0% Light trucks 6573.3 258.6 10 6841.9 40.0% 5.3% 37.2% Motorcycles 23.9 23.9 0.1% Buses 6.7 171.7 0.2 11.6 191.1 0.0% 3.5% 0.7% Transit 0.2 77.5 0.2 11.6 90.4 0.0% 1.6% 0.7% Intercity 29.2 29.2 0.6% School 6.5 65 .0 71.5 0.0% 1.3% Medium/heavy trucks 569.7 4440.4 16.7 5 026.8 3.5% 90.2% 62.1% HIGHWAY TOTAL 16447.5 4922.7 26.9 11.6 21409.6 100.0% 100.0% 100.0%
147 Table B 5. Modification of highway fuel use with VIUS data TEDB US Avg Gasoline FL VIUS Adjusted Gasoline FL VIUS Adjusted Gasoline TEDB US Avg Di esel FL VIUS Adjusted Diesel FL VIUS Adjusted Diesel % % TJ % % TJ Total Highway Fuel 998971 191057 Automobiles 56.38% 59.88% 598184 1.06% 1.06% 2018 Light trucks 39.97% 39.28% 392351 5.25% 10.98% 20979.32 Motorcycles 0.15% 0.15% 1452 Buses 0.04% 0.05% 456 3.49% 4.14% 7901 Transit 0.001% 0.006% 61 1.57% 2.22% 4245 Intercity 0.59% School 0.04% 395 1.32% Med/heavy trucks 3.46% 0.65% 6528 90.20% 40.06% 76538.66 Total Accounted 100.00% 100.00% 100.00% 56.23% Unaccounted 0 43.77% 83620
148 Table B 6. Florida truck fuel use from 2002 VIUS 47 Gasoline Diesel LPG Total Petro Sector Subsector TJ TJ TJ TJ Truck Transportation 39793 31028 109 70930 For hire transportation/warehousing 11511 14626 40 Vehicle rental ser vices 11430 9347 38 Not reported 16852 7055 31 Industrial 50854 27804 32 78689 Agriculture, forestry, fishing 4692 4130 10 Mining 31 1113 0 Utilities 4374 1858 8 Construction 34464 17100 14 Manufacturing 7293 3602 0 Commercial 36021 2 4035 77 60133 Wholesale trade 5319 7104 26 Retail trade 8561 5985 21 Information services 61 105 0 Administrative/support services 4529 5925 17 Recreation services 45 77 0 Hospitality services 2954 3226 8 Other services 14552 1614 5 R esidential 272211 14651 0 286863 Personal transportation 272211 14651 0 TOTAL 398879 97518 218 496615
149 Table B 7 ndustries LPG Gasoline Aviation Gas Jet Fuel Diesel Residual Total Petroleum NG Transport Industry TJ TJ TJ TJ TJ TJ TJ TJ Air transportation 638 2618 161567 164823 Rail transportation 11096 11096 Water transportation 20029 69164 89193 Truck transportation 404 36372 29694 66470 1144 Transit and ground passenger t ransport 5 456 6768 7229 10 Pipeline transportation 11493 Scenic and sightseeing transportation 4297 1133 5430 Postal service 2557 998 3555 Couriers and messengers 24033 9378 33411 Warehousing and storage 863 337 1200 Totals 409 69217 2618 161567 79432 69164 382406 12647
150 LIST OF REFERENCES (1) United Nations Report of the World Commission on Environment and ; 1987. (2) United Nations, Department of Ec onomic and Social Affairs, World Populati on Prospects: The 2010 Revision http://esa.un.org/unpd/wpp/Documentation/ publications.htm. (Accessed September, 2012). (3) Ewing, B.; Moore, D.; Goldfinger, S.; Ousler, A.; Reed, A.; Wackernagel, M The Ecological Footprint Atlas 2010 ; http://www.footprintnetwork.org/en/index.php/ GFN/page/ecological_footprint_atlas_2010. (Accessed May, 2012). (4) Zalasiewicz, J.; Williams, M.; Steffen, W.; Crutzen, P. The New World of the Anthropocene. Environment al Science and Technology 2010 44 2228 2231. (5) World Wildlife Fund Global Living Planet Report 2012 ; http://wwf.panda.org/ about_our_earth/all_publications/living_planet_report/2012_lpr/. (Accessed Jan 2013). (6) United Nations Millenium Ecosystem As sessment Ecosystems and Human Well being: Synthesis ; Washington, DC, 2005. (7) Daly, H. E. Allocation, distribution, and scale: towards an economics that is efficient, just, and sustainable. Ecological Economics 1992 6 185 193. (8) Wackernagel, M. Ecol for planning toward sustainability PhD Dissertation. School of Community and Regional Planning. The University of British Columbia. Vancouver, Canada.1994. (9) Graymore, M.; Sipe, N.; Rickson, R. Regional sustainability: How useful are current tools of sustainability assessment at the regional scale? Ecological Economics 2008 67 362 372. (10) ISO 14040 International Standard Environmental management Life cycle assessment Principles and fra mework ; International Organization for Standardization: Geneva, Switzerland, 2006. (11) Hawkins, T.; Hendrickson, C.; Higgins, C.; Matthews, H. S.; Suh, S. A Mixed Unit Input Output Model for Environmental Life Cycle Assessment and Material Flow Analysis. Environmental Science & Technology 2007 41 1024 1031. (12) Carnegie Mellon University Green Design Institute Economic Input Output Life Cycle Assessment (EIO LCA) software; http://eiolca.net/ (accessed June 2012).
151 (13) Cicas, G.; Hendrickson, C. T.; H orvath, A.; Matthews, H. S. A Regional Version of a US Economic Input Output Life Cycle Assessment Model. International Journal 2007 12 365 372. (14) Odum, H. T. Environmental accounting: Emergy and environmental decision making ; John Wiley and Sons, Inc.: New York, 1996. (15) Ulgiati, S.; Odum, H. T.; Bastianoni, S. Emergy use environmental loading and sustainability An emergy analysis of Italy. Ecological Model l ing 1994 73 215 268. (16) Brown, M. T.; Cohen, M. J.; Sweeney, S. Predicting nationa l sustainability: The convergence of energetic, economic and environmental realities. Ecological Modelling 2009 220 3424 3438. (17) Campbell, D. E.; Ohrt, A. Environmental Accounting Using Emergy: Evaluation of Minnesota EPA/600/R 09/002 January 2009 ; http://www.epa.gov/nheerl/ download_files/ publications/ MNEmergyEvalfinal2009_1_16.pdf (Accessed May 2011). (18) Campbell, D.; Meisch, M.; Demoss, T.; Pomponio, J.; Bradley, M. Keeping the books for environmental systems: An emergy analysis of West Virg inia. Environmental Monitoring and Assessment. 2004 94 217 230. (19) Agostinho, F.; Ambrsio, L. A.; Ortega, E. Assessment of a large watershed in Brazil using Emergy Evaluation and Geographical Information System. Ecological Modelling 2010 221 1209 1 220. (20) Campbell, D. E.; Garmestani, A. S. An energy systems view of sustainability: emergy evaluation of the San Luis Basin, Colorado. Journal of environmental management 2012 95 72 97. (21) Brown, M., Ulgiati, S. Emergy based indices and ratios to evaluate sustainability: monitoring economies and technology toward environmentally sound innovation. Ecological Engineering 1997 9 51 69. (22) Rugani, B.; Benetto, E. Improvements to Emergy evaluations by using Life Cycle Assessment. Environmental Scie nce & T echnology 2012 46 4701 12. (23) Wackernagel, M. Ecological Footprint and Appropriated Carrying Capacity: A Tool for Planning Toward Sustainability, 1994. PhD Dissertation. School of Community and Regional Planning. The University of British Colum bia. Vancouver, Canada.1994. (24) WWF Living Planet Report 2008 ; 2008; http://wwf.panda.org/ about_our_earth/ all_public ations/living_planet_report/2008_lpr/ (accessed Feb 2011).
152 (25) Medved, S. Present and future ecological footprint of Slovenia The inf luence of energy demand scenarios. Ecological Modelling 2006 192 25 36. (26) Monfreda, C.; Wackernagel, M.; Deumling, D. Establishing national natural capital accounts based on detailed Ecological Footprint and biological capacity assessments. Land Use Policy 2004 21 231 246. (27) Zhao, S.; Hong, H.; Zhang, L. Linking the concept of ecological footprint and valuation of ecosystem services: a case study of economic growth and natural carrying capacity. International Journal of Sustainable Development & World Ecology 2008 15 448 456. (28) Van Pelt, M. J. F. Ecologically sustainable development and project appraisal in developing countries. Ecological Economics 1993 7 19 42. (29) Lpez ridaura, S.; Masera, O.; Astier, M. Evaluating the sustainabilit y of complex socio environmental systems the MESMIS framework. Ecological Indicators 2002 2 135 148. (30) Graymore, M. L. M.; Wallis, A. M.; Richards, A. J. An Index of Regional Sustainability: A GIS based multiple criteria analysis decision support s ystem for progressing sustainability. Ecological Complexity 2009 6 453 462. (31) Cao, K.; Huang, B.; Wang, S.; Lin, H. Sustainable land use optimization using Boundary based Fast Genetic Algorithm. Computers, Environment and Urban Systems 2012 36 257 269. (32) Rowley, H. V; Peters, G. M.; Lundie, S.; Moore, S. J. Aggregating sustainability indicators: beyond the weighted sum. Journal of Environmental M anagement 2012 111 24 33. (33) Gleick, P. H.; Palaniappan, M. Peak Water: Conceptual and Practical Limits to Freshwater Withdrawal and Use. Proceedings of the National Academy of Sciences of the United States of America 2010 107 11155 11162. (34) Southwest Florida Water Management District Southern Water Use Caution Area Recovery Strategy: Final Rep ort ; 2006. (35) Southwest Florida Water Management District; Florida Department of Environmental Protection Peace River Cumulative Impact Study ; 2007. (36) Hendrickson, C. T.; Lave, L. B.; Matthews, H. S. Environmental life cycle assessment of goods and services: an input output approach ; RFF Press: Washington, DC, 2006.
153 (37) Hendrickson, C.; Horvath, A.; Joshi, S.; Lave, L. Economic input output models for environmental life cycle assessment. Environmental Science and Technology 1998 32 184A 191A. (38 ) Blackhurst, B. M.; Hendrickson, C.; Vidal, J. S. i Direct and Indirect Water Withdrawals for U.S. Industrial Sectors. Environmental Science & Technology 2010 44 2126 2130. (39) Marella, R. L. Water Withdrawals, Use, Discharge, and Trends in Florida 2000: U.S. Geological Survey Scientific Investigations Report 2004 5151 ; 2004. (40) Southwest Florida Water Management District 2002 Estimated Water Use in the Southwest Florida Water Management District ; 2004. (41) Marella, R. L. Water Withdrawals, Use and Trends in Florida, 2005: Scientific I nvestigations Report 2009. (42) Southwest Florida Water Management District 2005 Estimated Water Use Report Southwest Florida Water Management District ; 2007. (43) US Energy Information Administration (EIA) Stat e Energy Data System, 1960 2009, Florida ; 2009. (44) US Energy Information Administration (EIA) Electric Power Annual State Data Tables, 2010 ; 2010. (45) US Energy Information Administration (EIA) Fuel Oil and Kerosene Sales, 2002 ; 2002. (46) Davis, S. C.; Diegel, S. W. ; 2004; Vol. 6973. (47) US Census Bureau Vehicle Inventory and Use Survey: Florida, 2002 ; 2004. (48) National Agricultural Statistics Service Census of Agriculture, State of Florida, 2002 ; 2004; Vol. 1. (49) N ational Agricultural Statistics Service Census of Agriculture, Florida, 1997 ; 1999. (50) US Energy Information Administration (EIA) Energy Price and Expenditure Estimates, Selected Years, 1970 2009, Florida ; 2009. (51) United States Geological Survey Flo rida Mineral Yearbook, 2002 ; 2003. (52) US Census Bureau Economic Census, Mining Sector, Fuels and Electric Energy, 1997 ; 2001.
154 (53) US Environmental Protection Agency eGrid2010 v1.1 Database ; 2010. (54) US Environmental Protection Agency State Inventor y and Projection Tool ; 2012. (55) International Panel on Climate Change 2006 IPCC Guidelines for National Greenhouse Gas Inventories ; 2006. (56) MIG Inc. IMPLAN software 2002 (57) Buenfil, A. Emergy Evaluation of Water, PhD Dissertation. University of Florida, Gainesville, Florida. 2001. (58) Brown, M. T.; Protano, G.; Ulgiati, S. Assessing geobiosphere work of generating global reserves of coal, crude oil, and natural gas. Ecological Modelling 2011 222 879 887. (59) Zhang, Y.; Baral, A.; Bakshi, B Accounting for Ecosystem Services in Life Cycle Assessment, Part II: Toward an Ecologically Based LCA. Environmental Science and Technology 2010 44 2624 2631. (60) Blackhurst, M.; Hendrickson, C.; Vidal, J. S. I. Direct and Indirect Water With drawa ls for US Industrial Sectors Supplemental Material. Environmental Science & Technology 2010 44 1 12. (61) Bouman, B. Quantifying economic and biophysical sustainability trade offs in land use exploration at the regional level: a case study for the North ern Atlantic Zone of Costa Rica. Ecological Modelling 1998 114 95 109. (62) Zoltay, V. I.; Asce, A. M.; Vogel, R. M.; Asce, M.; Kirshen, P. H.; Westphal, K. S. Applied to the Ipswich River Basin. Journal of Water Resources Planning and Management 2010 566 575. (63) Nelson, E.; Mendoza, G.; Regetz, J.; Polasky, S.; Tallis, H.; Cameron, Dr.; Chan, K. M.; Daily, G. C.; Goldstein, J.; Kareiva, P. M.; Lonsdorf, E.; Naidoo, R.; Ricketts, T H.; Shaw, M. R. Modeling multiple ecosystem services, biodiversity conservation, commodity production, and tradeoffs at landscape scales. Frontiers in Ecology and the Environment 2009 7 4 11. (64) Southwest Florida Water Management District Southwest Florida Water Management District Land Use and Cover 2006 District Wide ; 2006. (65) South Florida Water Management District South Florida Water Management District Land Use and Cover 2004 05 ; 2005.
155 (66) Saint Johns River Water Management District St. J ohns River Water Managment District Land Use and Cover 2004 ; 2004. (67) Hershfiled, D. Rainfall Frequency Atlas of the United States for Durations from 30 Minutes to 24 Hours and Return Periods from I to 100 Years ; 1963. (68) Florida Industrial and Phosp hate Research Institute Reclamation ; http://www.fipr.state.fl.us/research area reclamation.htm. (accessed June 2012) (69) US Environmental Protection Agency eGrid2002 Database ; 2002. (70) Primary Productivity of the Biosphere ; Lieth, H.; Whittaker, R. H. Eds.; Springer Verlag: New York, 1975. (71) Kayranli, B.; Scholz, M.; Mustafa, A.; Hedmark, Carbon Storage and Fluxes within Freshwater Wetlands: a Critical Review. Wetlands 2009 30 111 124. (72) Florida Power and Light DeSoto Next Generation Sola r Energy Center FAQs http://www.fpl.com/environment/sol ar/desoto_faq.shtml#4 (accessed September 2012). (73) Krohn, B. Highlands EnviroFuels, LLC Sugar Biomass to Advanced Biofuel Commercial Production. In Florida Energy Summit ; 2012. (74) Matt Soberg U.S. EcoGen plans 60 MW Florida biomass power plant http://biomassmagazine.com/articles/5828/u s ecogen plans 60 mw florida bioma ss power plant (accessed June 2012). (75) Evans, J. M.; Cohen, M. J. Regional water resource implications of bioethanol produc tion in the Southeastern United States. Global Change Biology 2009 15 2261 2273. (76) United Nations Framework Convention on Climate Change (UNFCCC) Kyoto Protocol ; 2011. (77) Federal Highway Administration Highway Statistics, Motor Fuel, 2002, Table MF21 ; 2003. (78) Federal Highway Administration Highway Statistics, Motor Fuel, 2002, Table MF24 ; 2003.
156 BIOGRAPHICAL SKETCH David Pfahler completed a Bachelor of Arts in chemistry from Cedarville University and entered the U was in the Air Force Research Laboratory in Kirkland AFB, New Mexico. He helped develop chemical lasers during his three years there. In his second assignment, he completed a Mas ter of Science in chemistry from the University of Florida. David then worked in the Air Force Research Laboratory at Wright Patterson AFB, Ohio developing energy systems for Air Force applications. In June 2007 David left the Air Force and was selected a s an associate for the NSF Integrative Graduate Education and Research Traineeship (IGERT) program in a daptive management of water wetlands and wa tersheds.