当前位置:首页 >> 电力/水利 >>

Scenario development for water resources planning and watershed management Methodology


Environmental Modelling & Software 26 (2011) 873e885

Contents lists available at ScienceDirect

Environmental Modelling & Software
journal homepage: www.elsevier.com/locate/envsoft

Scenario development for water resources planning and watershed management: Methodology and semi-arid region case study
Mohammed I. Mahmoud*, Hoshin V. Gupta, Seshadri Rajagopal
Department of Hydrology and Water Resources, The University of Arizona, 1133 E James Rogers Way, Tucson, AZ 85721, United States

a r t i c l e i n f o
Article history: Received 28 January 2010 Received in revised form 18 January 2011 Accepted 1 February 2011 Available online 26 February 2011 Keywords: Scenarios Water resources management Scenario development Scenario planning Water resources planning

a b s t r a c t
Utilizing the scenario development framework from Mahmoud et al. (2009), a set of scenarios were developed for and applied in the Verde River Watershed in Arizona, USA. Through a scenario de?nition exercise, three dimensions of future change with respective axis extremes were identi?ed: climate change (periodic droughts vs. sustained drought), demographics (water-conservative population vs. water-consumptive population), and the economy (booming economy vs. poor economy). In addition to the various combinations of dimension extremes, each scenario was given a unique event or theme that was characteristic of the combination of dimension extremes it possessed. The scenarios were then ?eshed out into narrative forms that expanded on the details of each scenario’s internal temporal evolution. The scenarios were analyzed by a water supply and demand model that was speci?cally constructed for their simulation. Following the analysis of scenario results, assessment narratives were provided to outline the impact of each scenario on the Verde River Watershed and management operations in that basin. ? 2011 Elsevier Ltd. All rights reserved.

1. Introduction Climate change and population growth are exerting ever-increasing pressure on water resources in arid and semi-arid regions, forcing water managers and stakeholders to make critical decisions regarding water resources under varying degrees of uncertainty. With increased water scarcity placing stresses on existing water systems, all of the factors that contribute to changes in water use and consumption must be identi?ed and examined. Furthermore, with water management issues becoming progressively more focused on sustainability, the impacts of these factors on water supply and water demand have become even more pertinent. The adverse effects of climate change and population growth are particularly severe in the American Southwest; especially in the state of Arizona. The population of Arizona grew by 20.2% from 2000 to 2006, compared with a 6.4% growth rate for the entire United States (U.S. Census Bureau, 2008). During the next 10e20 years, the expected effects of climate change on Arizona include the enhanced possibility of extended droughts and diminished rates of precipitation (CLIMAS, 2008). These effects will translate into reduced regional water supply and increased levels of water demand.

* Corresponding author. Tel.: ?1 520 248 1914. E-mail address: mahmoud@email.arizona.edu (M.I. Mahmoud). 1364-8152/$ e see front matter ? 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.envsoft.2011.02.003

Scenario analysis has become recognized as a useful tool for dealing with such issues, and for generating solutions to contemporary water resource problems (Tarboton, 1995; Zacharias et al., 2005; Pallottino et al., 2005). Scenario analysis is the process of evaluating alternative future states of a system by examining several plausible pathways towards them. These pathways to the future; which comprise of various descriptors of change within the respective system, are referred to as scenarios. Analysis of these scenarios provides a better understanding of how and why an alternative future may evolve, and most importantly, the necessary steps that affected decision-makers and stakeholders are required to take in order to adapt to such a future. Although the use of scenario analysis for future planning is a relatively new concept, and scenarios are not commonly used to generate future outlooks, their value in producing better management decisions and strategies is now recognized (Maack, 2001; Schwartz, 2000). Because scenarios explore plausible descriptions of trajectories of change leading from the present into a number of possible alternative futures (various views of how the future may unfold), and by considering the many interrelated dynamic elements of the system (Schwartz, 1991; Van der Heijden, 1996), such studies can help to greatly reduce the risks inherent in decision making. Several contemporary and ongoing environmental scenario analysis studies exist e primarily focused on climate change. The carbon dioxide emission climate scenarios produced by the Intergovernmental Panel on Climate Change (IPCC) have elicited a large

874

M.I. Mahmoud et al. / Environmental Modelling & Software 26 (2011) 873e885

number of follow-up scenario analysis studies by the IPCC and other researchers (e.g. Girod et al., 2009; Beyene et al., 2010; Manning et al., 2010). Many of these studies focus on different aspects of the global climate-based scenario projections; such as their viability and validity, their potential impacts and consequences, and suggested improvements and enhancements associated with their projection. On a more localized scale are scenario studies conducted by the UK Climate Impacts Programme (UKCIP). The emphases of the UKCIP scenario studies are the consequences of and adaptations to climate change within the United Kingdom (UKCIP, 2010). Examples of these different studies include regional effects within speci?c areas of the country, adaptation strategies for different sectors such as transportation and business, and overall response preparedness of local entities to the changing climate. Furthermore, numerous small and medium-scale scenario analysis studies can be found in the literature, covering various issues of environmental signi?cance besides climate change. This paper discusses the application of scenario development to water management issues associated with the Verde River Watershed in Northern Arizona. Three of the eight scenarios developed in this study were analyzed with respect to future impacts on the Verde River Watershed, utilizing a simple supply and demand model of the watershed. Based on an analysis of the scenario simulations, we discuss the implications to water resources management in the Verde River Watershed. 2. Background The Verde River Watershed is a signi?cant source of water supply for the state of Arizona. Surface water stream?ow from the river is used to meet the water demands of the greater Phoenix Area (the capital of Arizona and its largest city). The watershed consists of three segments; the Upper, Middle, and Lower Verde River Watersheds (see Fig. 1). The Upper and Middle Verde Watersheds overlay the Big Chino and Little Chino aquifers; two of the main sources of groundwater supply in the watershed. Groundwater discharges from the Big Chino and Little Chino aquifers contribute signi?cant volumes of ?ow to the Upper Verde River (Wirt, 2005). Other major cities in the watershed include Prescott, Sedona, and Camp Verde; where the middle watershed borders with the lower watershed. The scenic Lower Verde Watershed stretches from Camp Verde to the Horseshoe dam reservoir. A primary water management concern in the Verde River Watershed is the heavy pumping of groundwater taking place in the upper portion of the watershed from the Big Chino and Little Chino aquifers. As groundwater pumping increases in the Upper Verde Watershed, the volume of stream?ow in that section of the river will continue to shrink. Increased urbanization, population growth, and the availability of land to cater to socio-economic expansion will only serve to amplify existing water demands. A proposed project to pump groundwater from the Big Chino aquifer and transport it to the city of Prescott and its surrounding area can only worsen this problem (Wolfe and Meyer, 2006). Another pertinent issue of concern is the unregulated diversion of Verde River water (Sonoran Institute, 2007). As the largest consumptive users of water in the Middle Verde Watershed, irrigation companies with water rights are diverting water from the Verde River without any accountability for the volume of diverted water. This activity contributes to reduced Verde River ?ow during the summer months e at a time when base?ow is already at its lowest point. 3. Scenario development In light of these growing concerns in the basin, we adopted a scenario planning approach to evaluate the effects of these issues

in the Verde River Watershed. As a representative stakeholder, and due to its vested interests in the basin, the Salt River Project (SRP) was consulted to assist in the development of future watershed management scenarios. The SRP is a private corporation that provides water and electricity to customers within Central Arizona, generating, transmitting, and distributing electric power to 920,000 homes and businesses. As the largest water supplier for the greater Phoenix area (a service area that covers over 375 mi2), the SRP delivers approximately 1 million acre-feet of water to agricultural, municipal, and irrigational water users (SRP, 2005, 2006, 2007). Consequently, the SRP is the primary electricity and water provider for the city of Phoenix (SRP, 2008). The SRP manages several dams and corresponding reservoirs in the watershed: Horseshoe Dam, Bartlett Dam, and Granite Reef Diversion Dam (see Fig. 1). As such, it depends strongly on surface water stream?ow to meet the demands of its shareholders and the greater Phoenix Area. To develop a set of watershed management scenarios for the Verde River Watershed, we followed the scenario development framework presented in a series of papers by Wagener et al. (2006), Liu et al. (2008), and Mahmoud et al. (2009). The framework organizes the scenario development process into ?ve key phases: scenario de?nition, scenario construction, scenario analysis, scenario assessment, and risk management. The scenario de?nition phase is used to identify key management questions pertinent to the future development and use of water resources in a given basin. Scenario dimensions of interest based on these questions are incorporated into the ?nal product of the phase: a set of scenario de?nition narratives describing the projected futures in terms of different dimensions of change. In the scenario construction phase, the model necessary to simulate alternative futures based on the de?ned scenarios is implemented. The different relationships, equations, and assumptions behind the model are speci?ed, along with the sources needed to create the required scenario data for the model. In the scenario analysis phase, the scenario simulation results are analyzed to determine dominant system behavior based on trends, patterns, and regime shifts. The results of the scenario analysis phase are then used in the scenario assessment phase to draw conclusions via a set of assessment narratives, to examine and explain the consequences of each alternative future. Finally, in the risk management phase, management strategies are suggested; as derived from the assessment narratives of the scenario assessment phase.

4. Scenario de?nition Discussions with the SRP helped to reveal three primary scenario dimensions of interest with respect to future change for water resources management, these being: 1) climate change, 2) demographics, and 3) the economy. These themes emerged from the discussion of several management questions raised in meetings with the SRP, including:  What are the impacts of economic and climate change on the Verde River Watershed?  How could runoff be affected under various scenarios?  How can mitigation strategies be developed against droughts and ?oods?  From a management perspective, what are the effects of land use on the Verde watershed? Potential scenario development questions raised in regards to the climate change dimension included:

M.I. Mahmoud et al. / Environmental Modelling & Software 26 (2011) 873e885

875

Fig. 1. Verde river watershed.

 How will climate change alter existing vegetation types and precipitation?  How will changes in precipitation impact the timing of ?oods, occurrence of snow vs. rain, and the amounts of sediment yield affecting reservoir storage?  What will be the cumulative impacts on the magnitude and frequency of stream?ow? Further discussion on the dimensions of demographics and the economy yielded the following necessary inclusions:  Shifts in demographics should account for population growth, development patterns, varied lifestyles, and conservation  Scenarios should explore the effects of population migration; i.e. residents moving in/out of the watershed  The outlook of funding and monetary resources should be considered Finally, land use and land cover changes were linked to the three dimensions by establishing that land use changes were connected with demographics, and land cover changes were connected with climate. To allow a suf?cient length of time for the evolution of future possibilities, a 50-year time horizon was selected. Having identi?ed the three primary scenario dimensions, we considered two possible extremes for each dimension: periodic droughts vs. sustained drought for climate change, water-conservative population vs. water-consumptive population for demographics, and booming economy vs. poor economy for the economy. By taking all possible combinations of the two extremes on each dimension, a total of eight possible scenarios were generated (see Fig. 2). Next, we proceeded to draft a set of scenario de?nition narratives derivable from the de?ned scenarios. Depending on the scenario, different logical evolutions of change were described. To

further reinforce the relevance of these narratives within the application area of the Verde River Watershed, an event that relates to a critical management issue in the watershed was prescribed to each scenario de?nition narrative. Each scenario de?nition narrative was assigned a different event based on the compatibility of the event to the actual attributes of that scenario. These events gave each scenario a distinct identity. The ?nal result was eight scenario narratives; of which a summary of their highlights is listed below: 1. Water Rights Settlement (Periodic Droughts/Water-conservative Population/Booming Economy) e implements additional water allocations to meet the water demands of all outstanding Native American water rights in the basin. 2. Restrictive Water Use (Periodic Droughts/Water-conservative Population/Poor Economy) e enforces groundwater use restrictions such that water shortages occur and actual water

Fig. 2. Scenario dimensions.

876

M.I. Mahmoud et al. / Environmental Modelling & Software 26 (2011) 873e885

3.

4.

5.

6.

7.

8.

demand has to be scaled down to the restricted available supply. Urban Sprawl (Periodic Droughts/Water-consumptive Population/Booming Economy) e explores rampant urban growth and its effects on water demand and supply. Flash Floods (Periodic Droughts/Water-consumptive Population/Poor Economy) e includes ?ash-?ooding events during summer months of extremely wet periods. Rainwater Harvesting and Surface Storage (Sustained Drought/ Water-conservative Population/Booming Economy) e enhances water supplies by applying rainwater harvesting practices virtually basin-wide, and shifts the storage of water aboveground instead of using it to recharge groundwater aquifers. Environmental Awareness (Sustained Drought/Water-conservative Population/Poor Economy) e prohibits the use of surface water to protect riparian areas and natural habitats. Business-oriented Growth (Sustained Drought/Water-consumptive Population/Booming Economy) e examines the effects of growth induced by pro?t-seeking motives; i.e. expansions of water uses in the industrial and agricultural sectors. Forest Fires (Sustained Drought/Water-consumptive Population/Poor Economy) e adds forest ?res to the scenario during summer months that exceed a speci?ed threshold temperature.

Table 1 Summary of model inputs, parameters, and variables. Inputs Qualitative Employment sectors Gross national product Numerical Industrial water use Gross income per capita Temperature Precipitation Land use cover Vegetation change Irrigation crops Irrigation ef?ciency Ef?uent recharge Population Age distribution People per household Employment level Demand allocation Vegetation transpiration Soil permeability Crop irrigation requirements Income tax bracket Runoff Snow melt Snow water equivalent Evaporation Evapotranspiration In?ltration Seepage Sediment yield Total water demand Residential water demand Agricultural water use Soil moisture Disposable income per capita Groundwater storage Parameters Calculated variables

model output variables. Calculated variables are model outputs that help to shape and describe each respective scenario’s resulting alternative future. The interaction between all these components is illustrated in Fig. 3. 5.1. Qualitative model inputs

5. Scenario construction To begin the process of simulating these future scenarios, a conceptual model was developed. The model implemented a monthly time-step of progression to capture seasonality effects of various hydrologic variables, and only accounted for the Upper and Middle segments of the Verde Watershed. The justi?cation behind this decision was to maintain model simplicity by avoiding reservoir regulation effects; since that information wasn’t available and due to the fact that the lower Verde Watershed contained the SRP reservoirs of the Horseshoe and Bartlett dams. It was also assumed that any evaporative losses or water diversions between the southern border of the Middle Verde Watershed and Horseshoe Dam reservoir were negligible due to the scenic area designation of that section of the river. Additionally, to simplify the simulation process, the model treated the portion of the Verde River Watershed under analysis as a lumped region where the output of the watershed region serves as input to the SRP reservoirs. The ?nal piece to the conceptual model required converting the 50-year planning time horizon into concrete temporal years for the model. The 50-year time horizon was therefore assigned in the model as the years 2011e2060. When the conceptual framework of the model was complete, the associated equations, relationships, and assumptions used to describe the system were scripted into a MATLAB code as a computational model. Table 1 presents a summary of the model’s inputs, parameters, and calculated variables. Qualitative inputs are employed to distinguish propagation trends that can in?uence and steer the evolution of speci?c numerical inputs. Numerical inputs into the model represent the driving forces of the eight drafted scenarios. Each input undergoes a modi?cation of values per scenario that is indicative of the changes prescribed for that input in each scenario de?nition narrative. Data for the inputs are derived from a number of sources that are considered reference data upon which the input’s scenario projection can be based on. Parameters are model inputs that do not vary with time and are considered numerical constants with respect to temporal simulations. Model parameters are used in conjunction with numerical model inputs to assist in calculating 5.1.1. Employment sectors This qualitative input is utilized to propel numerical values for the industrial water use input. The model operates under the assumption that trend changes in employment sectors will affect the projection of industrial water use through each scenario’s time period. 5.1.2. Gross national product The Gross National Product (GNP) is used to drive the numerical input of gross income per capita. The model assumes that trend changes in the GNP will affect the projection of gross income per capita through each scenario’s time period. 5.2. Numerical model inputs 5.2.1. Industrial water use (acre-ft) Industrial water use scenario projections are partly guided by qualitative changes in the study area’s anticipated outlook of active

Fig. 3. Computational model components.

M.I. Mahmoud et al. / Environmental Modelling & Software 26 (2011) 873e885

877

employment sectors. Also, industrial water use in this study lumps commercial and other uncategorized water uses and explicitly accounts for Indian water rights and water use. Reference data for industrial water use was supplied by the Arizona Department of Water Resources (ADWR) Verde River Watershed Study (ADWR, 2000). 5.2.2. Gross income per capita ($) Gross income per capita is the average untaxed amount of income each person residing in the study region receives per year. Projections of this input were in?uenced by the qualitative input of the GNP. Reference data is adopted from Yavapai County data in the Arizona Statistical Abstract (Arizona Statistical Abstract, 2003). 5.2.3. Temperature ( C) Temperature data generated for each scenario was projected from historical data using information on previously recorded minimum, maximum, and average monthly temperature for each month. Reference data for temperature projections was provided by Maurer et al. (2002); a hydrologically based dataset of land surface ?uxes and states for the contiguous United States. This dataset has been commonly manipulated to force land surface hydrology models such as the Variable In?ltration Capacity Model (VIC). Depending on whether a scenario dictated a normal/wet climate, periodic droughts, or a sustained drought during its timeframe, temperature was randomly generated according to the scenario’s climate conditions. 5.2.4. Precipitation (ft) Precipitation data for each scenario was also projected from historical data provided by the same dataset (Maurer et al., 2002) using information on previously recorded minimum, maximum, and average monthly precipitation for each month. Total precipitation was randomly generated for each month according to each scenario’s climate conditions. 5.2.5. Land use cover and vegetation change (acres) Land use and vegetation cover changes were derived from reference data found in the Reconnaissance Watershed Analysis on the Upper and Middle Verde Watershed report (Barnett and Hawkins, 2002) that was administered by the ADWR. For the purposes of the model, land use and vegetation change were combined as a single cover with each cover type assigned an initial amount of acres that changes per each scenario’s description. Therefore, changes in one cover type can signi?cantly in?uence the other and thus, are directly related. The land use component of the combined coverage consists of agricultural acreage and urban development. The vegetation component of the combined coverage consists of grass-land, desert shrubegrass, pinyonejuniper, ponderosa pine, Arizona chaparral, water, and riparian areas. Water and riparian areas were combined as a single unit and although changes in river discharge volumes can affect the size of water and riparian coverage, the cover area of the combined components did not change. This is because expansions/reductions in rivers and water-bodies in the model occur within this combined area, and consequential changes to riparian areas do not exceed the initial combined coverage acres. 5.2.6. Irrigation crops (acres) Crop irrigation acres refer to the types of irrigated crops that constitute the land use of agricultural acreage and their respective acres in production. Crop acres include those belonging to alfalfa, corn, pasture, turf/landscaping, vegetables, orchards, and nursery trees. Changes in the amount of acres for any of these irrigation crop-types result in a change in the total number of

agricultural acres. Reference data for crop irrigation acres was supplied by the ADWR Verde River Watershed Study (ADWR, 2000). 5.2.7. Irrigation ef?ciency (%) Irrigation ef?ciency is a measure of how much water used to irrigate agricultural crops is actually retained by the plants in comparison to how much excess water is wasted in irrigating agricultural crops. Reference data for irrigation ef?ciency is derived from the ADWR Verde River Watershed Study (ADWR, 2000). 5.2.8. Ef?uent recharge (acre-ft) Ef?uent recharge in the model is produced from septic systems and waste water treatment, and accordingly has some connection to industrial water use. The ADWR Verde River Watershed Study (ADWR, 2000) is the source of ef?uent recharge reference data. 5.2.9. Population Population rates increase monthly in the model and are based on reference data from the ADWR Verde River Watershed Study (ADWR, 2000). 5.2.10. Age distribution (%) The age distribution input variable for the model distinguishes age groups that categorize the watershed’s population. Age distribution was estimated from Yavapai County reference data found in the Arizona Statistical Abstract (Arizona Statistical Abstract, 2003). 5.2.11. People per household The people per household input variable aggregates the average number of persons living within each home in the analyzed portion of the Verde Watershed. Reference data for this variable was provided by the ADWR Verde River Watershed Study (ADWR, 2000). 5.2.12. Employment level (%) The employment level describes the percentage of the population that is actually employed; i.e. the inverse of the unemployment rate. Employment level reference data is derived from unemployment rate data for Yavapai County in the Arizona Statistical Abstract (Arizona Statistical Abstract, 2003). 5.2.13. Demand allocation (%) The input of demand allocation for each scenario denotes the percentage of the total water demand calculated by the model that is met by stream?ow diversions and groundwater pumping respectively. Therefore the model requires an input of demand supply source for both surface water and groundwater. The percentage values for demand allocation are purely based on the prescribed condition described in each scenario. 5.3. Model parameters 5.3.1. Vegetation transpiration (ft) Vegetation transpiration accounts for standard transpiration rates of particular vegetation cover types: ponderosa pine, Arizona chaparral, and pinyonejuniper. These average values for transpiration were obtained from Black (1996). 5.3.2. Soil permeability (ft) The average permeability of the soils in the Upper and Middle Verde Watersheds was obtained from permeability values of soiltypes in the watershed; which included frigid subhumid, mesic

878

M.I. Mahmoud et al. / Environmental Modelling & Software 26 (2011) 873e885

subhumid, mesic semi-arid, and thermic semi-arid soils. The reference data for this information came from the ADWR Verde River Watershed Study (ADWR, 2000). 5.3.3. Crop irrigation requirements (acre-ft/acre) The irrigation requirement for each agricultural crop is the amount of volume in acre-feet required to irrigate an acre of the crop type. Irrigation requirements differed for each crop type and neglected the effects of leaching, conveyance losses, and other water needs. Average irrigation requirements were estimated from reference data provided by the ADWR Verde River Watershed Study (ADWR, 2000). 5.3.4. Income tax bracket (%) The tax bracket for gross income per capita corresponds to a progressive increase in tax percentage applied to higher levels of personal income. The actual tax applied is a function of the income range for each person in the watershed. Data for the tax bracket was pooled together from various Arizona tax sources. 5.4. Calculated model variables Calculated variables fall under two main sections in the model: 1) the water balance, and 2) the demand function. The water balance side of the model incorporates hydrological variables of the system through a set of equations that aim to determine the amount of runoff leaving the lumped regions of the Upper and Middle Verde Watersheds. Some simpli?cations to these equations have been applied to reduce model complexity and to allow for potential enhancements to the model. The basic water balance equation used in the model is:

Vegetation transpiration rates from the model parameter input is weighted according to the available number of acres for each vegetation type that has its own designated transpiration rate; i.e. ponderosa pine, Arizona chaparral, and pinyonejuniper. Calculated evaporation is weighted to all other land cover types and is estimated using Hargreaves temperature-based evaporation (Hargreaves, 1975; Shuttleworth, 1993). One of the main reasons that Hargreavesbased evaporation was utilized in this model to estimate evaporation was its exclusion of energy balance terms. This simpli?ed the model simulation process by reducing the number of terms necessary to calculate evaporation. Additionally, it allowed for the unique conservation and representation of climate exclusively through two input variables: temperature and precipitation. The water balance equation is applicable for temperatures above 0  C. For freezing temperatures of 0  C and less, the water balance relationship becomes:

Runoff ? 0 Infiltration ? 0 Snow Pack ? Snow Water Equivalent ? Precipitation ? Evapotranspiration
Since the soil is frozen in these conditions, any precipitation contributes to the snow pack and there is no soil in?ltration or surface runoff. Additionally, the source of water for evapotranspiration comes from the snow pack. Sediment yield is the amount of sediment traveling past a certain point in a river or basin in a set period of time. Sediment yield is dependent on stream?ow, ?ow velocity, and river conditions. To approximate sediment yield in the model, the Dency and Bolten (1976) general watershed empirical calculations were adopted since they estimate sediment yield exclusively as a function of watershed area and runoff. The demand function portion of the model formulates relationships between socio-economic inputs and hydrological variables to approximate total water demand. Most of the relationships are linked by subjective socio-economic assumptions that otherwise would have been statistically dif?cult to numerate. The primary demand function equation is:

Runoff ? Snow Melt ? Precipitation ? Evapotranspiration ? Infiltration ? Seepage
where runoff, snow melt, precipitation, evaporation, in?ltration, and seepage are in units of feet. Runoff and in?ltration dynamics are governed by Hortonian overland ?ow; i.e. water available from precipitation and seepage that is not evaporated enters the soil, and water that is unable to in?ltrate due to permeability limits becomes runoff. In?ltration is also limited to the number of non-urban acres in the watershed; i.e. water that could potentially promote to in?ltration in urban portions of the watershed study area does not actually in?ltrate the soil surface but instead directly contributes to surface runoff. The water that does in?ltrate into the soil then contributes to existing soil moisture conditions. Precipitation is determined from the generated scenario input, seepage is calculated in the demand function portion of the model, and in?ltration is dependent on the permeability of the soil and its capacity to store water. Evaporation and snow melt are based on the following equations and relations:

Total Water Demand ? Residential Water Use ? Agricultural Water Use ? Industrial Water Use
where industrial water use is provided through scenario inputs, agricultural water use is driven by watering requirements for irrigation crops, and residential water use is determined from various socio-economic variables. The units for the demand function and subsequent water demand and water use equations are acre-ft. Agricultural water use is primarily dominated by crop water demand:

Snow Melt ? Stored Snow Pack ? Snow Water Equivalent
The snow melt component in the model is basically stored snow from prior monthly time steps. The occurrence of snow melt depends on the availability of a stored snow pack resulting from the combined effect of precipitation and temperatures at or below 0  C in immediately previous months. Evapotranspiration in the model is a weighted calculation based on land cover types:

Agricultural Water Use ? Total Crop Water Demand ? Soil Moisture
The amount of moisture already in the soil partially lessens irrigation requirements. Individual crop water demand is calculated by multiplying the number of acres of each crop type; i.e. alfalfa, corn, pasture, turf/landscaping, vegetables, orchards, and nursery trees, by the corresponding irrigation requirement of each crop. Total crop water demand is the sum of each crop’s water demand:

Evapotranspiration ? Vegetation Transpiration ? Calculated Evaporation

M.I. Mahmoud et al. / Environmental Modelling & Software 26 (2011) 873e885

879

Crop Water Demand ? Crop Acres*Crop Irrigation Requirements
Seepage is estimated by determining the quantity of water that is not absorbed by the crops during irrigation via the ef?ciency of irrigation practices and tools:

Groundwater Storage ? Stored Groundwater ? Effluent Recharge ? Groundwater Pumping River Diversions ? function?Demand Allocation? Groundwater Pumping ? function?Demand Allocation?
Groundwater storage is in?uenced by ef?uent recharge and groundwater pumping. Initial groundwater storage is derived from approximations of available groundwater to a depth of 1200 feet from the Little Chino, Williamson, and Big Chino Valley aquifers (ADWR, 2000). Both groundwater pumping and river diversions are functions of the demand allocation input. The amount of pumping and diversions that take place is directly related to the percentage of total demand allocated by each scenario to groundwater and surface water respectively. With the computational model in working order, scenario datasets were generated for all scenarios conforming to the restrictions and guidelines placed in each scenario de?nition narrative. Depending on the scenario, some amendments to the original model code were necessary to investigate special conditions prescribed by the scenario. 6. Scenario analysis For further analysis and assessment of the scenarios, this paper will only focus on three of the eight scenarios: Water Rights Settlements (Scenario 1), Flash Floods (Scenario 4), and Businessoriented Growth (Scenario 7). Prior to implementing the scenarios to the computational model, some special model and scenario data modi?cations had to be included. 6.1. Scenario analysis modi?cations 6.1.1. Water right settlements (periodic droughts/waterconservative population/booming economy) The key alteration in the model with respect to scenario 1 involves the allocation of demand. Since this scenario’s theme of water rights settlement focuses on the allotment of Indian water rights under industrial water use, the required water quota for this water right settlement is set at 50,000 acre-ft. The goal of the scenario is to be able to provide 50,000 acre-ft by the end of the 50-year timeline. This volume of water can then be carried forward to meet any required water rights for these outstanding settlements. To achieve this water quota, demand allocation throughout the scenario is split between industrial water uses; which includes settled Indian water rights, and non-industrial uses. With respect to water supply sources under this scenario, groundwater pumping is the source of supply for non-industrial water use and stream?ow diversions are the main supply provider of industrial water demand. However, during months when diversions cannot fully meet industrial water demand, stored groundwater is expected to make up the demand difference. 6.1.2. Flash ?oods (periodic droughts/water-consumptive population/poor economy) The inclusion of ?ash ?oods in scenario 4 consists of adjusting in?ltration volumes during the summer months of wet years. Only 25% of the water that can potentially in?ltrate into the soil actually contributes towards in?ltration. Additionally, demand allocation is

Seepage ? ?1 ? Irrigation Efficiency?*Agricultural Water Use
Residential water use is derived from assumptions on relationships between residential water demand and socio-economic conditions:

Residential Water Use ? function?Disposable Income; People per Household;Age Distribution?
Three different socio-economic relationships for residential water demand have been explicitly modeled: residential water demand that is dependent on disposable income per capita, residential water demand governed by the number of people per household, and residential water demand based on the population’s age distribution. Disposable income per capita utilizes information from the gross income per capita and employment level inputs along with the income tax bracket parameter:

Disposable Income ? GrossIncome*Employment Level*?1?Tax?
The relationship between disposable income and residential water demand is assumed to increase towards a maximum residential water demand value. The more disposable income that is available, the more likely people will be able to spend it on residential water to meet their needs. However there comes a point where all residential water needs are met, including wasteful water use. Beyond that point adding more disposable income does not increase the amount of residential water demanded. The underlying logic behind the relationship between the number of people per household and residential water demand is that as the number of persons living in a household increases, the amount of residential water demand will be less. This is attributed to the fact that multi-person households have common water uses that are not exclusive to each person; e.g. lawn watering, dishwashing, laundry, etc. Therefore with more people in a home, less water is required to meet these common demands. However, having less persons per household increases residential water demand because much of those common water uses are no longer distributed over more residents. The relationship between age distribution and residential water demand assumes that each of the age distribution groups has corresponding values of residential water demand. These values are based on socially in?uenced perceptions of water consumption habits associated with age. Using the age distribution input, a weighted value of residential water demand as a function of age distribution was calculated with the percent of population in each age group and each age group’s corresponding representative residential water demand. When all three sources of residential water demand were calculated, the mean of all three residential water demand functions was appointed as the representative value of residential water use in the model. Finally, the volume of stream?ow leaving the Upper and Middle Verde Watersheds and the volume of groundwater stored in the watershed was found using the following equations:

Streamflow ? Runoff ? Diversions

880

M.I. Mahmoud et al. / Environmental Modelling & Software 26 (2011) 873e885

not constant throughout the entire scenario period; it shifts according to climate conditions. During drought years groundwater is the primary source of supply and during wet years stream?ow is the main source of supply. When the climate represents normal conditions, demand is split evenly between the two supply sources. 6.1.3. Business-oriented growth (sustained drought/waterconsumptive population/booming economy) In the business-oriented growth of scenario 7, expansion and development undergo a dramatic shift between two portions of the scenario period. During the ?rst 30 years, the large entire-scenariospanning increases in population, employment levels, urbanization, and gross incomes are accompanied by aggressive development of agricultural lands and enhancement of industrial water use. These changes during the ?rst 30 years of the scenario are intended to illustrate the effects of business-oriented growth by boosting agricultural and industrial water demand. The augmentation of industrial water use also has the added bene?t of producing higher ef?uent recharge volumes for the whole projection period. However, in the last 20 years of the time horizon, industrial water use and agricultural acreage suffer a slight reduction in volume and acreage respectively. This represents the shift in agricultural and industrial sector priorities in pursuing only the most pro?table crops and industries. 6.2. Scenario analysis results Following the inclusion of the above adjustments to the model for each scenario, the scenario analysis phase then proceeded by generating simulations for each scenario using the computational model and each scenario’s respective dataset. When the model simulations were complete, an analysis was conducted for all the scenarios to examine the effects of each scenario’s conditions on the Upper and Middle Verde Watershed’s water supply. 6.2.1. Water right settlements (periodic droughts/waterconservative population/booming economy) Since the emphasis of this scenario is on the expansion of industrial water demand to accommodate the new water rights emerging from the settlement of outstanding Native American water rights, the analysis for scenario 1 examined the evolution of demand during the scenario projection period. The effect of this scenario on total demand is gauged by contrasting the demand volumes of the three different demand types: agricultural, industrial, and residential. Comparatively, agricultural water use represents the largest consumptive use in the watershed. Demand levels for industrial and residential water uses pale in comparison to those of agricultural water demand; in some cases by several orders of magnitude as other scenario simulations con?rm. However, in scenario 1, the expansive growth of industrial water use is also accompanied by a reduction in total agricultural water use (see Fig. 4a). This creates an interesting effect in total water demand that causes it to level off at around 33,600 acre-ft; as visualized in Fig. 4b. The decrease in total water demand is not especially smooth and demonstrates some seasonality. This is because industrial water use increases on an annual basis (at the beginning of each year) due to new water allocations resulting from water rights settlements, while other water uses (i.e. agricultural and residential) continue to change on a monthly basis. If projected trends of diminishing agricultural water demand were to continue beyond the 50-year time horizon, industrial water use may indeed become the dominant water use in the watershed, as is intended by the scenario’s theme. However, groundwater storage in this alternative future is dangerously low, and virtually over-drafted at a storage volume of approximately 43,500 acre-ft (see Fig. 4c).

6.2.2. Flash ?oods (periodic droughts/water-consumptive population/poor economy) The addition of ?ash ?oods in scenario 4 creates an excessive amount of surface runoff during the summer months of wet years. Since a variable climate (that ranges from periodic droughts to wet periods) drives this scenario, ?ash-?ooding events produce outcomes that initiate an amendment to demand allocation strategies. In response to these ?ooding conditions, demand allocation in this scenario is not uniform. When the possibility of ?ash ?oods occurring is strong; i.e. in wetter-than-usual years, all of the watershed’s demand is extracted exclusively from surface water supplies. In drought periods groundwater becomes the sole source of water supply. Demand allocation is split evenly between surface and groundwater during normal climate episodes. Demand allocations for surface and groundwater are presented in Fig. 5a and b respectively. Both ?gures demonstrate that, in some months, stream?ow diversions are not capable of providing their assigned allocation during intervals of normal climate. As expected, during those months, groundwater pumping makes up the needed difference. What is the collective impact of these changes in demand sources on groundwater storage? To properly answer that question, recharge must be considered. Ef?uent recharge in this scenario drops to 50% of its initial volume. Consequently, except for short periods where groundwater demand is zero, groundwater storage maintains an overwhelmingly negative trend; plummeting to a volume equivalent to 28% of its initial volume (see Fig. 5c). These results suggest that varying demand allocations is a good strategy but that overdraft of aquifers may be inevitable due to rising levels of demand. 6.2.3. Business-oriented growth (sustained drought/waterconsumptive population/booming economy) Business-oriented growth in scenario 7 is characterized by marked increases in industrial water demand and agricultural water demand. These boosts to demand can in turn impact longterm groundwater stores as subsurface pumping constitutes 50% of demand allocation. Fig. 6 tracks the changes to the different categories of demand and their combined effect on groundwater. All the different types of water demand classi?ed in the model are illustrated in Fig. 6a. Consistent with historical trends in Arizona, agricultural water use is the largest consumptive use. Augmentation of industrial water use in this scenario pales in comparison to that of agricultural water use and, to a lesser extent, residential water use. Residential water demand follows suit in this scenario with respect to rising volumes of industrial and agricultural water demand due to socio-economic conditions boosting residential water use; i.e. more income, higher levels of employment, and fewer people per household. After 30 years into the scenario, agricultural water use begins to decline as marketing policies dictate the cultivation of more pro?table and water-reliant crops. This drop signi?es the reduction of agricultural acreage to a more manageable size. Unfortunately, this slowing down of agricultural growth does little to prevent the inevitable overdraft of groundwater supplies after 32 years of future projection, as seen in Fig. 6b. The resulting overdraft in this scenario is large, due to uncontrolled rampant demand growth motivated by market forces with no consideration to the condition of local water resources. The combined volume of water needed to meet demand during the last scenario segment of overdraft is a staggering 1419.7 million acre-ft. The alarming management question to answer is: where will this extra volume of water come from? The only option includes supplementary sources outside the current modeled system; otherwise a water shortage crisis of epic proportions is to be expected.

M.I. Mahmoud et al. / Environmental Modelling & Software 26 (2011) 873e885

881

Fig. 4. Scenario 1: demand types and groundwater. a) Scenario 1: agricultural water use vs. industrial and residential water uses. b) Scenario 1: total water demand. c) Scenario 1: groundwater storage.

7. Scenario assessment Scenario assessment involves taking the results of each simulated scenario and presenting them in a short narrative form that communicates the impacts of each proposed scenario over the projected 50-year time horizon as an alternative future. In addition to assessing the analysis results in a basic narrative form, an implication theme is adopted for each scenario assessment narrative based on the suitability and appropriateness of theme with respect to narrative content.

7.1. Water right settlements: evolutionary change/good news and bad news (periodic droughts/water-conservative population/ booming economy) Evolutionary change in this scenario is characterized by the adjustment of typical demand magnitudes; as classi?ed by source, and its cumulative effect on total water demand. The scenario’s theme of growing industrial demand resulting from the additional allotment of settled Indian water rights under the designation of

Fig. 5. Scenario 4: demand allocation and groundwater. a) Scenario 4: surface water demand. b) Scenario 4: groundwater demand. c) Scenario 4: groundwater storage.

882

M.I. Mahmoud et al. / Environmental Modelling & Software 26 (2011) 873e885

Fig. 6. Scenario 7: Water demand and groundwater. a) Scenario 7: water demand by type. b) Scenario 7: groundwater storage.

industrial water use is accompanied by the decline of agricultural water use; traditionally the largest source of water consumption in Arizona. The combination of these growing and declining forces causes total water demand to reach a near-stationary level of demand in the last half of the scenario timeframe. This positive result however is overshadowed by the fact that this alternative future has a mere 42,616 acre-ft of groundwater storage to sustain its subsequent water demands. This state of water resources is achieved due to the demand allocation scheme of this scenario: abundant stream?ow supplies are constrained to industrial water uses while the limited groundwater supplies take on all other sources of water demand. Evolutionary change:  The transition of dominant water demand source from the agricultural sector to the industrial sector. Good news:  Total water demand reaches a lower near-constant demand level. Bad news:  Groundwater storage is utilized to the point where it is virtually over-drafted.

from drought-stricken stream?ow. An equal distribution of total water demand is enforced throughout regular climate years. However, the smaller number of wet years as compared to other climate-type years still renders groundwater the principal source for providing water throughout the entire scenario period. This is evidenced by the 72% drop in aquifer water storage. Cycles:  Demand allocation undergoes cycles of shifting supply sources as determined by climate conditions. Crisis:  Excess surface water volumes during ?ash ?ood events contribute to greater magnitudes of potentially transportable sediment yields. Response:  Changing demand allocations as a function of climate attempts to curb the volume of surplus discharge volumes during wet summer months and the correspondingly higher concentrations of sediment yield.

7.3. Business-oriented growth: perpetual transition (sustained drought/water-consumptive population/booming economy) Business-oriented growth initiates boosts in water demand volumes for the agricultural and industrial sectors. Rampant economical development that follows this trend also enhances residential water demand due to higher population growth and favorable socio-economic conditions; e.g. less unemployment and better income wages. Consequently, overdrafts occur within the 32nd year of scenario projection (2042) and demand volumes thereafter are alarmingly high. The uncertainty in this scenario is highlighted by a critical management question: Will water demand during the end-years of the scenario be sated by external (out-ofbasin) water supply sources or will a catastrophic state of water shortage ensue?

7.2. Flash ?oods: cycles/crisis and response (periodic droughts/ water-consumptive population/poor economy) The recurrence of ?ash ?oods in the wetter summer months of this scenario produces unfavorable conditions with regards to excess ?ow volumes and large sediment yield concentrations. This prompts the establishment of a management strategy that alters demand allocation according to source (surface water vs. groundwater) depending on prevalent climate conditions. Allocating demand to surface water supplies during wet intervals can assist in lessening the potential load of transported sediments, and utilizing stored groundwater in drought periods can alleviate the burden

M.I. Mahmoud et al. / Environmental Modelling & Software 26 (2011) 873e885

883

Perpetual transition:  The uncertainty in the outlook of this alternative future beyond the modeled time horizon is characterized by the reality that anything can happen; although not necessarily for the better. 8. Risk management Although risk management is traditionally viewed as a stakeholder activity wherein decision-makers develop and apply management strategies and operations/policies while giving consideration to scenario analysis results and subsequent scenario assessments, this section will focus on some suggestions that can aid in mitigating risk with respect to the water supply generated from the Verde River Watershed. Based on the scenario analysis and scenario assessment phases, the risk imposed by upstream water users in the Verde River Watershed is composed of three components: A) Groundwater Pumping, B) Stream?ow Diversions, and C) Sediment Yield. As the following paragraphs convey, the management repercussions of each of the three components are strongly correlated. 8.1. Groundwater pumping Continued excessive pumping of groundwater in the Upper Verde Watershed could potentially overdraft the aquifers in that segment of the watershed. This could indirectly affect the water supply coming from the Verde River Basin, since recent studies (Wirt, 2005) have shown that base?ow from the Big and Little Chino aquifers signi?cantly enhances stream?ow volumes in that section of the Verde River. Although some groundwater pumping from those aquifers may not cripple the contribution of subsurface discharges, the complete overdraft of the aquifers surely will. Therefore it is in the SRP’s best interests to work closely with water management entities in the Upper Verde Watershed to identify other sources of supply that can offset the overwhelming dependency of the Upper Verde Watershed’s population on a supply source that is well on its way to complete depletion. One possible solution, that can partially mitigate this problem, is to have municipalities in the Upper Verde Watershed enter a set of “surplus” water contracts that basically allow them to purchase water, from the SRP, that can be diverted from the Verde River during periods when the watershed’s reservoirs are nearing full capacity and excess river ?ows into them are expected. Upper Verde Watershed municipalities can then use the purchased water to either meet their local water demands or recharge the nonsustainable aquifers. It would be more advantageous if the latter choice was followed, as this delays the inevitable overdraft of the aquifers and maintains some level of discharge from the Big Chino and Little Chino aquifers into the Upper Verde River. 8.2. Stream?ow diversions Stream?ow diversions upstream of the reservoirs on the Verde River can either be a bene?t or a detriment to the river’s water supplies depending on the prevailing climate. Climate conditions greatly dictate the impact that river diversions have on the river’s reservoir supplies. During drought periods, whether extensive or short-term, the water supply stored behind the Verde River Watershed dams (mainly Horseshoe and Bartlett), is highly sensitive to any diversions taking place upstream. However, in years where there is abundant stream?ow contributing to the reservoirs, stream?ow diversions may curb the frequency of operating outlet works and spillways to pass ?ood waters and lessen the volume of water to be discharged from the dams at safer capacities. Out of these two possibilities, the greater threat is that of pervasive

drought conditions that can potentially jeopardize the amount of stored water in the reservoirs through higher evaporation rates and lower volumes of in?ow into the reservoirs. Therefore, the SRP must ?nd ways to regulate stream?ow diversions in the Upper portions of the Verde River Watershed during drier seasons. Similar to the groundwater pumping issue, the SRP can explore other sources of water for stream?ow-dependant upstream users during drought episodes or can enter into agreements with upstream entities to limit their diversions; with added incentives for their compliance. An example of such an incentive can entail the SRP providing excess water to those upstream entities during wetter periods at costs lower than other alternative water suppliers. 8.3. Sediment yield Sediment yield transported from the Upper and Middle Verde Watersheds into the reservoir system in the Lower Verde Watershed is principally a function of stream?ow, as well as other river characteristics; e.g. stream velocity, river cross-section characteristics, etc. This is because sediment is eroded and transported downstream primarily by stream?ow and surface runoff. However, it should also be noted that the potential amount of sediment that can be transported by a river depends on the actual amount of sediment yield available. So although equations may be used to approximate sediment yield; e.g. Dency and Bolten (1976), predicting sediment yield is dif?cult because most empirical equations do not consider initial sediment concentration. Furthermore, even if equations incorporate sediment concentrations, obtaining their physical measurements is dif?cult. Sediment yield is an important factor because the accumulation of sediment yield in the reservoirs diminishes the dam structures’ storage capacity. Additionally, sediment yield that enters water storage facilities, tends to remain in that location for long time periods. Accumulated sediment yield in dams does not tend to ?ush out of dam reservoirs unless outlet works are activated to pass waters of ?ood-level proportions (White, 2005). The mitigation of this predicament is connected to the issue of stream?ow diversions. Suf?cient water must ?ow into the reservoir system to meet demand requirements, yet excessive volumes can worsen the abundance of sediment yield in the reservoirs. Since the release of excess waters from dam spillways at ?ood water levels is not common, mitigation strategies must target reservoir in?ows. Therefore, the minimum amount of water ?ow necessary to ?ll the reservoirs and/or meet downstream demands should be pursued to restrict sediment yield concentrations in the reservoirs. However, this recommendation is provisional, until a method can be devised to measure the actual amount of sediment concentration that is available for transportation by the Verde River and the amount of sediment yield currently residing within the river’s reservoirs. 9. Discussion Out of the eight original scenarios, the three scenarios examined in this case study represented a range of combinations for the extremes of the three selected scenario dimensions (climate change, demographics, and the economy). Prior to ?eshing out the scenarios through scenario de?nition narratives, it was expected that the best possible combination of scenario dimension extremes that would yield the most favorable alternative future included periodic droughts, a water-conservative population, and a booming economy. The best possible alternative future was analyzed in this study through the scenario results of the Water Rights Settlements scenario (Scenario 1). The premise of the scenario was that waterrelated conditions were favorable enough that existing and future water right settlements could be resolved. The upside of this

884

M.I. Mahmoud et al. / Environmental Modelling & Software 26 (2011) 873e885

scenario was that water demand reached a lower near-constant level, but unfortunately not soon enough, as impending groundwater overdraft loomed in the very near post-scenario future. This scenario simulation indicated that even if scenario conditions seem optimal; the resulting alternative future may not be as optimal, thus validating the need for scenario analysis and scenario assessment beyond the scenario de?nition phase to con?rm or disprove the initial premise of a scenario. Alternatively, the worst possible future was expected to comprise of a sustained drought, a water-consumptive population, and a poor economy. Although the scenario represented by these conditions (Scenario 8 e Forest Fires) was not analyzed in this paper, another scenario with two of the three same scenario dimension extremes was examined. The Business-oriented Growth scenario (Scenario 7) included a sustained drought, a waterconsumptive population, and a booming economy. The only difference between the dimension extremes of these two scenarios was the state of the economy; which proved to be the driving force in the Business-oriented Growth scenario. As the analysis and assessment of the Business-oriented Growth scenario indicated, rampant business growth and spending continued in spite of the fact that the state of water resources was very poor e a sustained drought coupled with a water-consumptive population. This set of circumstances led to a great state of uncertainty as to how the correspondingly high demand for water would be met. Although the economy was in bad shape in the Forest Fires scenario, it was not the dominant theme in the scenario as evident by the scenario’s name. In fact, the state of water resources was much better in this scenario than in the Business-oriented Growth scenario; since the presence of wild forest ?res temporarily increased surface water supplies. Overall, with respect to water resources, the Businessoriented Growth scenario proved to have the least favorable alternative future out of all the scenarios. The results of this study point to a very important message regarding the status of water supply and demand in the future. For water resources to become sustainable either available water supply needs to be enhanced or water demand needs to be reduced. However, analysis of the eight Verde River Basin scenarios suggests that augmentation of water supplies is not a feasible long-term solution because supplementary supplies need to be obtained from elsewhere or water shortages will ensue. Depending on out-of-basin resources is not a permanent solution to the problem because other supply sources will eventually fall under the risk of depletion. Therefore the only way to truly maintain the longevity of local water resources is to change habits of water consumption and the behavioural patterns underlying them. This is not an easy objective to achieve, yet it is one that water managers and policy-makers must confront before the issue of water sustainability becomes even more acute. 10. Conclusions Implementation of the scenario development framework towards critical management concerns in the Verde River Watershed proved to be a successful endeavour with respect to establishing stakeholder engagement. The methodical approach of the scenario development phases provided a systematic process for achieving the goals of the study. As the representative stakeholder component in the development process, the SRP was an excellent source of information, guidance, and input. Additionally, their participation ensured that the study was very relevant to pertinent issues and hence validated the pursuit of such an approach. One of the challenging aspects of this study’s scenario development activity pertains to the inherent subjectivity required by the scenario development process. In this case, there was some dif?culty in constructing the conceptual model underlying the

simulation/computational model. The conceptual model had to connect all model components in a plausible manner in order to represent a cohesive system, which would eventually take form as a set of equations and assumptions that constitute the computational model. The subjectivity required in making assumptions that link together some of the key variables is not an easy task (e.g. socio-economic relationships), since the assumptions need to be logical and consistent. Finally, scenario development is not a “one-and-done” enterprise; it is a reiterative process that requires the revaluation of the created scenarios. Following that logic, an adaptive management policy pertaining to the eight Verde Watershed scenarios should be considered by integrating aspects of regular monitoring of indicators and sequential post-audits. Acknowledgments Support for this research was provided by the Salt River Project (SRP). Additionally, the authors would like to acknowledge the SRP hydrology group for their consultation in this work. References
ADWR, April 2000. Verde River Watershed Study. Arizona Department of Water Resources. Arizona Statistical Abstract, 2003. Data Handbook, 2003. Economic and Business Research Program, Eller College of Business and Public Administration, The University of Arizona. Northland Press. Barnett, L.O., Hawkins, R.H., June 30, 2002. Reconnaissance Watershed Analysis on the Upper and Middle Verde Watershed. Report. Beyene, T., Lettenmaier, D.P., Kabat, P., 2010. Hydrologic impacts of climate change on the Nile River Basin: implications of the 2007 IPCC scenarios. Climatic Change 100 (3), 433e461. Black, P.E., 1996. Watershed Hydrology. CRC Press. CLIMAS, 2008. Climate Assessment for the Southwest. http://www.climas.arizona.edu/. Dency, Bolten, S., 1976. Ecological Issues in Floodplains and Riparian Corridors. White Paper. Center for Streamside Studies, University of Washington. Girod, B., Wiek, A., Mieg, H., Hulme, M., 2009. The evolution of the IPCC’s emissions scenarios. Environmental Science and Policy 12 (2), 103e118. Hargreaves, G.H., 1975. Moisture availability and crop production. Transactions of the American Society of Agricultural Engineering 18 (5), 980e984. Liu, Y., Mahmoud, M., Hartmann, H., Stewart, S., Wagener, T., Semmens, D., Stewart, R., Gupta, H., Dominguez, D., Hulse, D., Letcher, R., Rashleigh, B., Smith, C., Street, R., Ticehurst, J., Twery, M., van Delden, H., White, D., 2008. In: Jakeman, A., Voinov, A., Rizzoli, A.E., Chen, S. (Eds.), Environmental Modelling and Software and Decision Support e Developments in Integrated Environmental Assessment (DIEA)Formal scenario development for environmental impact assessment studies, vol. 3. Elsevier, pp. 145e162. Maack, J., 2001. Scenario analysis: a tool for task managers. In: Social Analysis: Selected Tools and Techniques. Social Development Paper No. 36. World Bank Available from the Social Development Department, The World Bank, Washington, D.C. Mahmoud, M., Liu, Y., Hartmann, H., Stewart, S., Wagener, T., Semmens, D., Stewart, R., Gupta, H., Dominguez, D., Dominguez, F., Hulse, H., Letcher, R., Rashleigh, B., Smith, C., Street, R., Ticehurst, J., Twery, M., Delden, H., van Waldick, R., White, D., Winter, L., 2009. A formal framework for scenario development to support environmental decision making. Environmental Modelling & Software 24, 798e808. Manning, M.R., Edmonds, J., Emori, S., Grubler, A., Hibbard, K., Joos, F., Kainuma, M., Keeling, R.F., Kram, T., Manning, A.C., Meinshausen, M., Moss, R., Nakicenovic, N., Riahi, K., Rose, S.K., Smith, S., Swart, R., Van Vuuren, D.P., 2010. Misrepresentation of the IPCC CO2 emission scenarios. Nature Geoscience 3 (6), 376e377. Maurer, E.P., Wood, A.W., Adam, J.C., Lettenmaier, D.P., Nijssen, B., 2002. A long-term hydrologically-based data set of land surface ?uxes and states for the conterminous United States. Journal of Climate 15 (22), 3237e3251. Pallottino, S., Sechi, G.M., Zuddas, P., 2005. A DSS for water resources management under uncertainty by scenario analysis. Environmental Modelling & Software 20 (8), 1031e1042. Schwartz, P., 1991. The Art of the Long View: Planning for the Future in an Uncertain World. Doubleday, New York. Schwartz, P., 2000. The of?cial future, self-delusion and the value of scenarios. Financial Times Mastering Risk, Part Two, May 2. Shuttleworth, W.J., 1993. Evaporation. In: Maidment, D.R. (Ed.), Handbook of Hydrology. McGraw-Hill (Chapter 4), various pagings. Sonoran Institute, 2007. Sustainable Water Management: Guidelines for Meeting the Needs of People and Nature in the Arid West. White Pine Inc., Ann Arbor, Michigan. http://sonoran.org/index.php?option?com_docman&task?doc_view& gid?131&Itemid?5.

M.I. Mahmoud et al. / Environmental Modelling & Software 26 (2011) 873e885 SRP, 2005. Salt River Project 2005 Annual Report. Salt River Project. SRP, 2006. 2006 Salt River Project Annual Report. Salt River Project. SRP, 2007. Salt River Project 2007 Annual Report. Salt River Project. SRP, 2008. Salt River Project. www.srpnet.com. Tarboton, D.G., 1995. Hydrologic scenarios for severe sustained drought in the southwestern United States. Water Resources Bulletin 31, 803e813. UKCIP, 2010. United Kingdom Climate Impacts Programme. http://www.ukcip.org.uk/. U.S. Census Bureau, 2008. United States Census Bureau. http://www.census.gov/. Van der Heijden, K., 1996. Scenarios: The Art of Strategic Conversation. John Wiley & Sons, New York. Wagener, T., Liu, Y., Stewart, S., Hartmann, H., Mahmoud, M., July 2006. Imagine e scenario development for environmental impact assessment studies. CD ROM. Internet:. In: Voinov, A., Jakeman, A., Rizzoli, A. (Eds.), Proceedings of the iEMSs Third Biennial Meeting: “Summit on Environmental Modelling and Software”.

885

International Environmental Modelling and Software Society, Burlington, USA http://www.iemss.org/iemss2006/sessions/all.html. White, S., 2005. Sediment yield prediction and modeling. Invited Commentary. Hydrological Processes 19, 3053e3057. Wirt, L., 2005. Sources of base ?ow in the Upper Verde River. In: Wirt, L., DeWitt, E., Langenheim (Eds.), Geologic Framework of Aquifer Units and Ground-water Flowpaths, pp. F2eF34. Verde River Headwaters, North-Central Arizona, U.S. Geological Survey Open-?le Report 2004-1411. Wolfe, E.W., Meyer, W., 2006. Water-resource Issues in the Upper Verde Watershed. http://www.upperverdewaterissues.org/full%2520reports/Upper%2520Verde% 2520water-resource%2520issues%2520WAC%2520data.pdf, 5 p. Report. Zacharias, I., Dimitriou, E., Koussouris, Th, 2005. Integrated water management scenarios for wetland protection: application in Trichonis Lake. Environmental Modelling & Software 20, 177e185.


相关文章:
地理专业名词中英文对照.txt
Waterlogged lands 流域管理 Watershed management 水边开发 Waterside development ...resources 环境规划 Environmental planning 发展合作 Development cooperation 发展...
swat校准技巧.pdf
Srinivasan and L. M. Hauck. 2001. Validation ...Scenario 1 Sediment 0.60 0.50 Observed Simulated...for water quality (AI1 in.wwq) Srin@brc.tamu...
雨洪最佳管理措施决策支持系统_BMPDSS_介绍.pdf
scenario for stormwater ment,modeling and layout ...for the design, assessment and planning of ...watershed water quality management[R]. Springer ...
环境中英文.doc
Waterlogged lands 流域管理 Watershed management 水边开发 Waterside development ...resources 环境规划 Environmental planning 发展合作 Development cooperation 发展...
ArcGis软件生成的河网径流节点标识提取_图文.pdf
et al.GIS and scenario analysis aid to water pollution control planning of...[13] Maidment D R.Arc hydro GIS for Water Resources[M],ESRI press,...
海河流域二元水循环模型开发及其应用_图文.doc
for scenario analysis of water resources planning and management in the HRB....Development of WEP model and its application to an urban watershed [ J ]...
环境英文对照.doc
Waterlogged lands 流域管理 Watershed management 水边开发 Waterside development ...resources 环境规划 Environmental planning 发展合作 Development cooperation 发展...
SJC_ Water_Resources_Management_Planning Update_Mok....ppt
SJC_ Water_Resources_Management_Planning Update_...Watershed Water Resource Division MORE WATER Project...Scenario development f... 13页 2下载券 ...
北大考研-环境科学与工程学院研究生导师简介-刘永.pdf
EstuariesandCoasts,WaterResearch,WaterResourcesresearch...basedonscenarioa nalysesforwatershedmanagementunder...sinurbanfringe.LandscapeandUrbanPlanning,82(4):233...
硕士答辩--SWAT模型_图文.ppt
for water resources in rural basins 1995 Arnold,...for the Upper North Bosque River Watershed [J]....Planning on Crop Yield, Nitrate Leaching and ...
...Approach to Simulate Urban Water Resources Susta....pdf
stormwater management and water supply for ...water resources, support improved watershed planning...Scenario development f... 13页 2下载券 低影响...
...of Water Pollution Control Measures for Lake Bas....pdf
Wuhan Institute of Water Resources Planning and ...for sustainable policy makings in the watershed. ...marginal cost analysis or scenario analysis are tradition...
地下水模型Feflow介绍_图文.ppt
5.3 WASY Institute for Water Resources Planning and Systems Research Ltd. ...scenario flow model 地下水补给 地下水抽水井 GW-recharge (Source 10-4 m/...
Watershed Conf First Announcement and Call for Papers.pdf
But management for sustained use of water and ...resources requires an integrated watershed-based ...(CVR), ITALY Coastal Research and Planning ...
Human Resource Planning.doc
Planning - An Analytical Study Abstract: The ...Management in the Post Globalization Scenario. Since...diverse resources for its growth and development. ...
水利专业英语-Jiaocongcong.doc
Chapter 2 Planning for Water Resources Development ...(2)An overall water-management plan,developed ...(6)watershed:流域 (4)peak discharges:洪峰流量 (...
梁子湖流域水环境功能区划及水质现状分析_顾自强.pdf
Reasonable planning, development and utilization of ...watershed, according to the water environment ...cance for water resources protection. Keywords: ...
paper holland.pdf
for the development of management scenario and the simulation of the effect of management practices on water and sediment yielding in Gharasu watershed (...
分散式污水处理系统流程,设计,管理,使用_图文.pdf
Planning & design Construction O&M ?...Watershed Management, Stormwater Permits, and TMDLs...bay-wide for subdivisions, commercial development ...
Organisation Development_图文.ppt
(MHCM Programme) BH 6043 Human Capital Planning ...development, improvement, and reinforcement of the ...Scenario planning for uncertain future & preparing ...
更多相关标签: