Search

Best utilizing water resources requires coordinating their availability and use in time and space. Required can be: spatially and temporally distributed data; simulators to predict system response to stimuli; procedures for defining management goals, constraints, and scenarios; optimizers to compute optimal management strategies; and appropriate strategy implementation techniques. Here, a strategy is a set of controllable groundwater extraction and injection rates and surface water diversions. Simulation/optimization (S/O) models couple simulators and optimizers to compute optimal strategies for posed management problems. S/O models are becoming more commonly used for policy, planning, system design, and management. For example, water planners and managers sometimes must decide how to control groundwater use to cause a favorable future and avoid serious problems. S/O models can help determine the policies, physical systems, and management strategies that can yield the best consequences. ‘Best’ is defined by the manager/modeler in terms of water availability, sustainability, crop production, economic, social, or environmental criteria, or combinations of those. Addressing multi-objective optimization problems and developing quantified tradeoff curves is simple with a powerful S/O model such as SOMOS. Examples demonstrate data needs and S/O model power for policy and plan development and system design and management.

The Simulation/Optimization MOdeling System (SOMOS) has modules adapted for a range of international water resource management needs. The SOMOA module is best for field groundwater and conjunctive water management situations. It employs several analytical simulation models and several optimization methods to give appropriate optimal guidance for situations having sparse data. SOMOA requires understanding of the limitations of common analytical equations. It does not require experience in finite difference or element numerical modeling. SOMOS also has modules ideal for situations in which there is sufficient data for numerical modeling or aquifer and stream -aquifer systems. Emphasized here are SOMOA and its applications for common field-level water management problems. Sample situations include optimizing: minimizing required surface reservoir size for a period of low surface flows in a stream -aquifer system; conjunctive use for blending irrigation waters of different qualities; and artificial injection.

Making best use of water resources requires coordinating use of groundwater and surface waters--conjunctive use. Conjunctive water management is important on the regional (watershed or aquifer), water district and field scales. Goals might differ for the different administrative levels. A regionally optimal water management strategy is not necessarily optimal for water district managers and individual farmers. Interactions between levels are affected by laws, regulations, traditions and public pressure. To help each organizational level develop water management strategies that best achieve its goals, computer simulation/optimization (S/O) models are useful. Such S/O models couple water flow and transport simulation techniques with formal mathematical optimization algorithms. When properly calibrated, S/O models can be immensely valuable. S/O models have long been used for managing reservoir releases and surface water resources. S/O model use for large-scale groundwater management and contaminant plume management is becoming more common. Advances in computer power and mathematical optimization procedures make it possible to address large problems previously impractical. Likewise, increasing availability and power of PC computers and operating systems make it now practical for water district personnel and field engineers to develop optimal water management strategies. By exercising care appropriate for data and model accuracy, field-level S/O model users can develop better solutions for their management problems. The use of S/O models at all water management levels makes it easier to negotiate compromise solutions when goals differ. S/O models are easily used to develop trade-off curves showing how much of one goal achievement must be forfeited to achieve more of another goal. Communication is enhanced when both negotiating parties have reliable trade-off information. Furthermore, S/O model and trade-off curve use highlights differences in data or understanding of data available to the parties--making it easier to resolve those differences. This paper discusses robust S/O software available and useful for the spectrum of groundwater and surface water managers. This software promotes improved water management and goal achievement. Case studies illustrate benefits of S/O model use.

REMAX is a software package designed to assist water managers in developing optimal ground water or coordinated ground water/surface water strategies for a wide range of management problems. REMAX uses MODFLOW, MT3D, and other simulation models to develop the response matrix or response surface equations employed in the final optimization model. To address nonlinear systems (e.g., unconfined aquifers) accurately, REMAX uses a modified response-matrix method. REMAX several nonlinear functions to represent a wide variety of nonlinear response surfaces. REMAX can be used for deterministic or reliability-based optimization problems. Decision variables are ground-water extraction/injection rates and/or surface water diversion rates. State variables include water flows, stages and concentrations. Available objective fm1ctions and constraints are linear, nonlinear, integer or mixed integer. REMAX has over 80 objective types including linear, quadratic, integer, goal programming, minimax, and maximin functions. Combining these types for multi-objective optimization makes REMAX ideal for cost minimization and/or utility maximization. Available constraints can be used to define relations among decision variables, state variables, or any combination of these variables in time and space. REMAX has been used for designing several ground-water remediation systems for capture and cleanup of ground-water contaminant plumes.

Advances in computer power and mathematical optimization procedures can improve planning and developing sustainable irrigation systems. Simulation and optimization models can help plm1 groundwater and conjunctive use strategies to best achieve management goals while satisfying mm1agement a11d physical constraints. Simulation/optimization models that couple calibrated flow and transport simulation models with optimization algorithms can help design the best water management strategies. Mm1agers can be relatively sure that the groundwater system will respond acceptably when appropriate procedures are employed to develop the water management strategies. Presented case studies illustrate situations in which developed strategies simultaneously address conflicting management goals such as: maximizing sustainable groundwater extraction or conjunctive use versus: minimizing spread or degree of contamination, maintaining adequate artesian flow at springs, or maintaining adequate flow in rivers. Another case illustrates how time-varying future irrigation water needs can be best satisfied by combining groundwater and surface water resources--while assuring adequate saturated aquifer thickness for drought protection and maintaining sufficient river flow for navigation and commerce.

US/REMAX is a computer program designed to assist water managers in developing optimal groundwater and/or surface water strategies for a wide range of management problems. It employs response matrix, regression and other methods adapted for nonlinear systems. US/REMAX performs deterministic or reliability-based, single- or multi-objective optimization. Decision variables are ground-water extraction/injection and/or surface water diversion. State variables include water flows, stages and concentrations. Hard coded objective functions and constraints are linear, nonlinear, integer or mixed integer. Special constraints can be added to address unusual situations.

Concern for groundwater quality and availability is increasing in Pahvant Valley. Ground-water levels are declining due to intense groundwater extraction for irrigation; water having high total dissolved solids concentration is flowing from the southwest toward the pumping sites; and discharge from natural springs in a wildlife refuge is declining. Transient simulation of aquifer response to 20 years of the 1985 pumping rates (beginning with 1985 groundwater level) predicted that spring discharge would decrease by as much as 87% from 1985 rates. Presented are preliminary pumping strategies that maximize sustainable, steady-state groundwater extraction without unacceptably reducing discharge from the springs. A simulation/optimization (s/o) model is used to calculate optimized sustainable groundwater pumping rates that provide prespecified discharge rates from Clear Lake Springs. These optimal pumping strategies are sustainable unless significant changes occur in assumed system recharge and discharge rates. The usefulness of the s/o model for regional management is demonstrated by computing optimal strategies for four scenarios. Scenarios differ in the amount of discharge required from Clear Lake Springs or in the upper limit on pumping at individual locations. The optimal steady-state groundwater withdrawal and spring discharge rates obtained in this study are less than the 1985 rates. To provide sufficient sustainable discharge from Clear Lake Springs, pumping rates should be less than the 1985 rates.

US/REMAX is designed to assist water managers in developing optimal groundwater and/or surface water strategies for a wide range of management problems. US/REMAX uses the response matrix method, which assumes that physical system response to stimuli is linear. However, US/REMAX can also address nonlinear systems via cycling. In one application, a strategy computed using US/REMAX required 40% less pumping than one obtained via a normal simulation model. US/REMAX also easily computes tradeoffs for multiobjective problems.

Simulation/optimization (S/0) models can be used to speed the process of computing desirable groundwater pumping strategies for plume management. They make the process of computing optimal strategies fairly straightforward and can help minimize the labor and cost of groundwater contaminant cleanup.

Long-term water management planning models frequently use large time steps and must employ fairly crude assumptions (such as average climatic conditions, etc.). Managing stream aquifer systems during a dry season requires using finer discretization in time and space. Presented is a computer model, US/REMAX, developed by Utah State University personnel for aiding best management of stream-aquifer systems for both long and short eras. The model computes strategies for optimally allocating surface and ground water resources in time and space. For a water supply problem the model can maximize the sum of delivered surface and ground water. For an environmental protection problem the model can minimize total groundwater pumping (extraction plus injection) needed to capture contaminant plume. The model can provide optimal steady-state or time variant solutions. Weighting coefficients can be used in the objective function: (1) to emphasize substitution of surface water diversion for groundwater pumping or vice-versa, (2) or to achieve linear economic optimization.

We developed a simulation-optimization model which maximizes crop yield while maintaining target salt concentration in the root zone, and/ or preventing salt from leaching to the groundwater. The model performs nonlinear optimization and simulation. Implicit finite difference forms of the nonlinear, transient, unsaturated water flow equation, and the convection-diffusion equation are embedded as constraints. Other constraints include nonlinear functions describing the hydraulic properties of the medium. The objectives of the management model is to develop irrigation strategies which prevent salt leaching to the groundwater when salty irrigation water is used. Five different irrigation strategies are developed for Huntington Research Farm, Utah, and are presented in this paper.

Four alternative simulation/optimization models useful for computing optimal sustained-yield (steady-state) groundwater pumping strategies are compared in terms of formulation, solution procedure, accuracy, and computational efficiency. The different models require different computer processing time and memory. For the aquifer tested system, if more than 10% of the cells have pumping as a decision variable, a fully linearized embedding model will require less computer memory than any other model. All the models address linear and nonlinear steady-state flow in multilayer, unconfined/confined aquifers. They also address several types of nonsmooth external flows. Newly presented are a response matrix model solving external flows described by nonsmooth functions through cycling, and a fully nonlinear embedding model that directly achieves an optimal solution without cycling. Models are tested using a hypothetical three-layer (unconfined/confined) aquifer system (3 layers x 15 rows x 15 columns = 675 cells) . Empirically, globally optimal solutions seem to be obtained. All the models compute the same optimal pumping even if their optimizations are begun using vastly different initial guesses. This addresses a common concern that the solutions to nonlinear problems are not necessarily globally optimal.

Simulation/optimization (S/0) models can be used to speed the process of computing long term groundwater pumping strategies. These combined models can also greatly ease the planning tasks of water management agencies. They make the process of computing optimal perennial yield groundwater management strategies fairly straightforward. S/0 models dramatically improve conjunctive water management and can help minimize the cost of groundwater contaminant clean-up. The differences between S/0 models and the simulation (S) models currently used by over 98 % of practitioners are illustrated with recent applications of S/0 models in regional planning.

A methodology and computer model is developed to determine economically optimum closed subsurface drainage systems in irrigated areas. The mode 1 maximizes net benefits, by comparing profit driven by crop yields to drain system cost and selects an optimum drain layout. The optimization methodology used is the SIMPLEX method (Nelder and Mead, 1965). The SIMPLEX model was linked to the subsurface drainage model DRAINMOD (Skaggs, 1982) and to the surface hydraulic model KINE (Walker and Skogerboe 1987). The selected optimum drainage system maximizes the difference between total revenue, and the total cost of installation, operation and management of a particular drainage system. The optimization sub-program provides a workable and simple procedure for optimizing water management simulation models.

Three well configurations were compared in terms of short-term ground-water contaminant plume containment. These include parallel, octagonal and combination systems. Each system had three extraction wells upgradient and three injection wells downgradient of the contaminant plume. For each system, optimal pumping values and resulting potentiometric surface smoothness were computed for a hypothetical plume. Tested models utilized linear programming optimization and simulation via the response matrix method. The octagonal well configuration required less pumping for a pumping period of 8 days, than did the parallel or combination systems. The octagonal configuration resulted in the smoothest potentiometric surface, in terms of difference in final head at observation wells compared with those at the contaminant source.