Water Simulation Model Research Articles

Climate change, water availability and future cereal production in China

Climate scenarios from a regional climate model are used to drive crop and water simulation models underpinned by the IPCC A2 and B2 socio-economic development pathways to explore water availability for agriculture in China in the 2020s and 2040s. Various measures of water availability are examined at river basin and provincial scale in relation to agricultural and non-agricultural water demand and current and planned expansions to the area under irrigation. The objectives are to understand the influences of different drivers on future water availability to support China's food production. Hydrological simulations produce moderate to large increases in total water availability in response to increases in future precipitation. Total water demand increases nationally and in most basins, but with a decreasing share for agriculture due primarily to competition from industrial, domestic and municipal sectors. Crop simulations exhibit moderate to large increases in irrigation water demand which is found to be highly sensitive to the characteristics of daily precipitation in the climate scenarios. The impacts of climate change on water availability for agriculture are small compared to the role of socio-economic development. The study identifies significant spatial differences in impacts at the river basin and provincial level. In broad terms water availability for agriculture declines in southern China and remains stable in northern China. The combined impacts of climate change and socio-economic development produce decreases in future irrigation areas, especially the area of irrigated paddy rice. Overall, the results suggest that there will be insufficient water for agriculture in China in the coming decades, due primarily to increases in water demand for non-agricultural uses, which will have significant implications for adaptation strategies and policies for agricultural production and water management.

Parameterization and Evaluation of the AquaCrop Model for Full and Deficit Irrigated Cotton

Predicting yield is increasingly important to optimize irrigation under limited available water for enhanced sustainability and profitable production. Food and Agriculture Organization (FAO) of the United Nations addresses this need by providing a yield response to water simulation model (AquaCrop) with limited sophistication. In this study, AquaCrop was parameterized and tested for cotton (Gossypium hirsutum L.) under full (100%) and deficit (40, 60, and 80% of full) irrigation regimes in the hot, dry, and windy Mediterranean environment of northern Syria. Model parameterization used the 2006 data and was straightforward within the designed user‐interface, owing to the limited number of key parameters. Accurate simulation of canopy cover was central to sound prediction of evapotranspiration and biomass accumulation. Key user‐input parameters for this purpose were identified as the coefficients defining canopy development and the threshold soil water depletion levels for the water stress indices. The parameterized model was tested using data from the 2004 and 2005 seasons, resulting in accurate prediction of evapotranspiration (<13% error). The predicted yield values were within 10% of measurements, except in the 60 and 80% irrigation regimes in 2004, with errors up to 32%. The model closely predicted the trend in total soil water, but deviation existed for individual soil layers. This study provides first estimate values for cotton parameters useful for future model testing and use. Model parameterization is site‐specific, and thus the applicability of key calibrated parameters must to be tested under different climate, soil, variety, irrigation methods, and field management.

Future cereal production in China: The interaction of climate change, water availability and socio-economic scenarios

Food production in China is a fundamental component of the national economy and driver of agricultural policy. Sustaining and increasing output to meet growing demand faces significant challenges including climate change, increasing population, agricultural land loss and competing demands for water. Recent warming in China is projected to accelerate by climate models with associated changes in precipitation and frequency of extreme events. How changes in cereal production and water availability due to climate change will interact with other socio-economic pressures is poorly understood. By linking crop and water simulation models and two scenarios of climate (derived from the Regional Climate Model PRECIS) and socio-economic change (downscaled from IPCC SRES A2 and B2) we demonstrate that by the 2040s the absolute effects of climate change are relatively modest. The interactive effects of other drivers are negative, leading to decreases in total production of −18% (A2) and −9% (B2). Outcomes are highly dependent on climate scenario, socio-economic development pathway and the effects of CO 2 fertilization on crop yields which may almost totally offset the decreases in production. We find that water availability plays a significant limiting role on future cereal production, due to the combined effects of higher crop water requirements (due to climate change) and increasing demand for non-agricultural use of water (due to socio-economic development). Without adaptation, per capita cereal production falls in all cases, by up to 40% of the current baseline. By simulating the effects of three adaptation scenarios we show that for these future scenarios China is able to maintain per capita cereal production, given reasonable assumptions about policies on land and water management and progress in agricultural technology. Our results are optimistic because PRECIS simulates much wetter conditions than a multi-model average, the CO 2 crop yield response function is highly uncertain and the effects of extreme events on crop growth and water availability are likely to be underestimated.

COSTA a problem solving environment for data assimilation applied for hydrodynamical modelling

A problem solving environment for data assimilation called COSTA is developed at Delft University of Technology. The goal of COSTA is to offer a modular framework where simulation models can be combined with various data assimilation methods. COSTA defines a number of building blocks called components. Examples of components are model, method, stochastic observer and state-vector. New data assimilation systems can be created by combining these components. This paper describes the application of COSTA to the WAQUA/TRIWAQ shallow water simulation model. In the past a model specific RRSQRT Kalman filter has been implemented for WAQUA/TRIWAQ. However, this implementation cannot be used in combination with other models. The WAQUA/TRIWAQ model is changed into a COSTA model component and the original RRSQRT Kalman filter is changed into a generic filter that can be used for other models as well. The new filter is now a part of the COSTA environment. The original filter contained a number of WAQUA/TRIWAQ specific aspects e.g. the drying and flooding of areas in the model. These model specific issues are identified, isolated and moved into the model component. A COSTA based implementation of WAQUA/TRIWAQ with the RRSQRT Kalman filter is realized and compared to the original system in a number of experiments. The experiments show that the COSTA based system produces the correct results and the computational overhead for using COSTA is low. The new RRSQRT Kalman filter is also combined with other COSTA models including the LOTOS-EUROS model for atmospherical transport, chemistry (chemical reaction) and deposition of air pollution on the scale of Europe.

Geological Analysis and Numerical Modeling of Mine Discharges for the Sanshandao Gold Mine: 2. Simulation and Prediction of Mine Discharges

A two-dimensional model is presented as a way to hydrogeologically characterize the controlling factors in the Sanshandao Gold Mine. The finite element method was first applied to simulate the ground water system of the current operation, using leakage data and the calculated recharge. An inverse model was applied to the observed data (e.g., head and discharge) to verify and calibrate the ground water simulation model and to estimate the hydrogeological parameters for the water-bearing zones. Nonlinear mathematic programming was used to solve the inverse model and to estimate the model parameters in 17 districts with different hydrogeological characteristics. The finite element equations were solved by means of a large non-symmetrical sparse equation. The results were in agreement with what is currently observed in the mine. The models and the estimated parameters were then applied to predict the mine water discharge for drifts extending to depths of −330 to −600 m during the next development stage. In order to improve the predicted accuracy of the numerical model, an iterative element mesh was added in the districts near the drainage drifts so that the computed discharges that flowed into the drifts would approach the recharges that flow into the borders of the mine. The model was also used to understand how the mine discharge would be influenced by factors such as unsteady ground water flow and the construction of hydraulic barriers to restrict ground water from entering the pits.

Are water simulation models consistent with steady-state and ultrafast vibrational spectroscopy experiments?

Vibrational spectroscopy can provide important information about structure and dynamics in liquids. In the case of liquid water, this is particularly true for isotopically dilute HOD/D 2O and HOD/H 2O systems. Infrared and Raman line shapes for these systems were measured some time ago. Very recently, ultrafast three-pulse vibrational echo experiments have been performed on these systems, which provide new, exciting, and important dynamical benchmarks for liquid water. There has been tremendous theoretical effort expended on the development of classical simulation models for liquid water. These models have been parameterized from experimental structural and thermodynamic measurements. The goal of this paper is to determine if representative simulation models are consistent with steady-state, and especially with these new ultrafast, experiments. Such a comparison provides information about the accuracy of the dynamics of these simulation models. We perform this comparison using theoretical methods developed in previous papers, and calculate the experimental observables directly, without making the Condon and cumulant approximations, and taking into account molecular rotation, vibrational relaxation, and finite excitation pulses. On the whole, the simulation models do remarkably well; perhaps the best overall agreement with experiment comes from the SPC/E model.

Scanning multichannel microwave radiometer snow water equivalent assimilation

Accurate prediction of snowpack status is important for a range of environmental applications, yet model estimates are typically poor and in situ measurement coverage is inadequate. Moreover, remote sensing estimates are spatially and temporally limited due to complicating effects, including distance to open water, presence of wet snow, and presence of thick snow. However, through assimilation of remote sensing estimates into a land surface model, it is possible to capitalize on the strengths of both approaches. In order to achieve this, reliable estimates of the uncertainty in both remotely sensed and model simulated snow water equivalent (SWE) estimates are critical. For practical application, the remotely sensed SWE retrieval error is prescribed with a spatially constant but monthly varying value, with data omitted for (1) locations closer than 200 km to significant open water, (2) times and locations with model‐predicted presence of liquid water in the snowpack, and (3) model SWE estimates greater than 100 mm. The model error is estimated using standard error propagation with a calibrated spatially and temporally constant model error contribution. A series of tests have been performed to assess the assimilation algorithm performance. Multiyear model simulations with and without remotely sensed SWE assimilation are presented and evaluated with in situ SWE observations. The SWE estimates from assimilation were found to be superior to both the model simulation and remotely sensed estimates alone, except when model SWE estimates rapidly and erroneously crossed the 100‐mm SWE cutoff early in the snow season.

Soil–water and solute movement under precision irrigation: knowledge gaps for managing sustainable root zones

Precision irrigation involves the accurate and precise application of water to meet the specific requirements of individual plants or management units and minimize adverse environmental impact. Under precision irrigation applications, water and associated solute movement will vary spatially within the root zone and excess water application will not necessarily result in deep drainage and leaching of salt below the root zone. This paper estimates that 10% of the irrigated land area (producing as much as 40% of the total annual revenue from irrigated land) could be adversely affected by root zone salinity resulting from the adoption of precision irrigation within Australia. The cost of increases in root zone salinisation due to inappropriate irrigation management in the Murray and Murrumbidgee irrigation areas was estimated at AUD 245 million (in 2000/01) or 13.5% of the revenue from these cropping systems. A review of soil–water and solute movement under precision irrigation systems highlights the gaps in current knowledge including the mismatch between the data required by complex, process-based soil–water or solute simulation models and the data that is easily available from soil survey and routine soil analyses. Other major knowledge gaps identified include: (a) effect of root distribution, surface evaporation and plant transpiration on soil wetted patterns, (b) accuracy and adequacy of using simple mean values of root zone soil salinity levels to estimate the effect of salt on the plant, (c) fate of solutes during a single irrigation and during multiple irrigation cycles, and (d) effect of soil heterogeneity on the distribution of water and solutes in relation to placement of water. Opportunities for research investment are identified across a broad range of areas including: (a) requirements for soil characterisation, (b) irrigation management effects, (c) agronomic responses to variable water and salt distributions in the root zone, (d) potential to scale or evaluate impacts at various scales, (e) requirements for simplified soil–water and solute modelling tools, and (f) the need to build skills and capacity in soil–water and solute modelling.

Impact of spatial data resolution on simulated catchment water balances and model performance of the multi-scale TOPLATS model

Abstract. This paper analyses the effect of spatial input data resolution on the simulated water balances and flow components using the multi-scale hydrological model TOPLATS. A data set of 25m resolution of the central German Dill catchment (693 km2) is used for investigation. After an aggregation of digital elevation model, soil map and land use classification to 50 m, 75 m, 100 m, 150 m, 200 m, 300 m, 500 m, 1000 m and 2000 m, water balances and water flow components are calculated for the entire Dill catchment as well as for 3 subcatchments without any recalibration. The study shows that model performance measures and simulated water balances almost remain constant for most of the aggregation steps for all investigated catchments. Slight differences in the simulated water balances and statistical quality measures occur for single catchments at the resolution of 50 m to 500 m (e.g. 0–3% for annual stream flow), significant differences at the resolution of 1000 m and 2000 m (e.g. 2–12% for annual stream flow). These differences can be explained by the fact that the statistics of certain input data (land use data in particular as well as soil physical characteristics) changes significantly at these spatial resolutions. The impact of smoothing the relief by aggregation occurs continuously but is barely reflected by the simulation results. To study the effect of aggregation of land use data in detail, in addition to current land use the effect of aggregation on the water balance calculations based on three different land use scenarios is investigated. Land use scenarios were available aiming on economic optimisation of agricultural and forestry practices at different field sizes (0.5 ha, 1.5 ha and 5.0 ha). The changes in water balance terms, induced by aggregation of the land use scenarios, are comparable with respect to catchment water balances compared to the current land use. A correlation analysis between statistics of input data and simulated annual water fluxes only in some cases reveals systematically high correlation coefficients for all investigated catchments and data sets (e.g. actual evapotranspiration is correlated to land use, surface runoff generation is correlated to soil properties). Predominantly the correlation between catchment properties (e.g. topographic index, transmissivity, land use) and simulated water flows varies from catchment to catchment. Catchment specific properties determine correlations between properties and fluxes, but do not influence the effect of data aggregation. This study indicates that an aggregation of input data for the calculation of regional water balances using TOPLATS type models leads to significant errors from a resolution exceeding 500 m. Correlating statistics of input data and simulation results indicates that a meaningful aggregation of data should in the first instance aim on preserving the areal fractions of land use classes.

PROPOSAL OF RAINFALL-RUNOFF MODEL FOR FORESTED SUB-BASINS APPLYING TO DISTRIBUTED MODEL

The application of a distributed model for river basins is keen issue. However methodologies for its application have not fully discussed. As for river water simulation models, there are fully developed commercial models with the progress of mathematical techniques for hydraulic calculation. The next topic is a description of rainfall runoff phenomena in sub-basins for river water simulation models. Especially in Japan, the upstream areas of rivers are almost covered with forests. A rainfall-runoff model for forested sub-basins is necessary for continuous simulation including several rainfall events and dry weather periods. In this paper, the authors proposed the pseudo-physical rainfall-runoff model for forested area on the assumption that surface and inter flow are related to void ratio and wetness of soil. They applied this model for a mountainous forested basin as one of distributed models. The simulation results fairly agreed with observed data.

UVQ: A tool for assessing the water and contaminant balance impacts of urban development scenarios

This paper presents the water and contaminant daily simulation model of the total water cycle, called UVQ. The model has been developed to provide a means for rapidly assessing conventional and nonconventional approaches to providing water supply, stormwater and wastewater services to urban allotments, neighbourhoods and study areas. The model is placed the context of other such models developed internationally through a brief literature review. This is followed by a description of the model and output examples, which is used to illustrate the utility of the model. UVQ is an effective preliminary assessment tool for determining the impacts of urban development options on the total water cycle, as well as the performance of a wide range of non-conventional demand and supply side management techniques. It compliments other aspects of an environmental assessment of options, along with more traditional aspects such as infrastructure costing.

Development and Evaluation of the RZWQM‐CROPGRO Hybrid Model for Soybean Production

It is common for agricultural system modelers to enhance their models by learning from other models and incorporating the best state‐the‐science into their models. In this study, the CROPGRO plant growth model of Decision Support System for Agrotechnology Transfer (DSSAT v3.5) was linked to the Root Zone Water Quality Model (RZWQM) to provide RZWQM users an option of using CROPGRO. In the hybrid model, RZWQM supplied CROPGRO with daily soil water and N, soil temperature, and potential evapotranspiration (PET), whereas CROPGRO supplied RZWQM with daily water and N uptake and plant growth variables. The RZWQM‐CROPGRO hybrid model was then evaluated against the original CROPGRO‐soybean model using several data sets from the literature. These data sets represented various drought conditions. Results showed that the RZWQM‐CROPGRO hybrid model simulated higher water stress than the original DSSAT‐CROPGRO model because of higher PET simulated by RZWQM, especially under semiarid climate conditions. Therefore, it was necessary to make some adjustments in the hybrid model under dry and windy conditions, e.g., using a different lower limit of plant available water as DSSAT. The hybrid model with a more detailed soil water balance calculation only affected soil water prediction at the top 60‐cm soil profile where soil water was more dynamic. This study demonstrated a successful linkage between RZWQM and CROPGRO, and the RZWQM‐CROPGRO hybrid model provides users with a tool to conduct detailed simulation of crop production in addition to addressing water quality concerns. This study also demonstrated that, when building models from various sources, compatibility of the interacting modules should be ensured.

Remote sensing of snow thaw at the pan-Arctic scale using the SeaWinds scatterometer

Remotely sensed estimates of snow thaw offer the potential of more complete spatial coverage across remote, undersampled areas such as the terrestrial Arctic drainage basin. We compared the timing of spring thaw determined from approximately 25 km resolution daily radar backscatter data with observed daily river discharge time series and model simulated snow water content data for 52 basins (5000–10,000 km 2) across Canada and Alaska for the spring of 2000. Algorithms for identifying critical thaw transitions were applied to daily backscatter time series from the SeaWinds scatterometer aboard NASA QuikSCAT, the observed discharge data, and model snow water from the pan-Arctic Water Balance Model (PWBM). Radar-derived thaw shows general agreement with discharge increases (Mean Absolute Difference, MAD=21 days, r=0.45), with better agreement (16 days) in basins with moderate-high runoff due to snow thaw. Even better agreement is noted when comparing the scatterometer-derived primary thaw timing with model simulated snow water increase (MAD=14 days, r=0.75). Good correspondence is found across higher latitude basins in western Canada and Alaska, while the largest discrepancies appear at the driest watersheds with lower snow and daily discharge amounts. Extending this analysis to the entire pan-Arctic drainage basin, we compared scatterometer-derived date of the primary (maximum) thaw with the timing of simulated snow water increases from the PWBM. Good agreement is found across much of the pan-Arctic; discrepancies for over half of the analyzed grid cells are less than one week. MADs are 11.7 days for the Arctic basin in Eurasian and 15.1 days across North America. Mean biases are low; 2.1 and −3.1 days for Eurasia and North America, respectively. Stronger backscatter response (high signal-low noise) is noted with higher seasonal snow accumulation, low to moderate tree cover and low topographic complexity. Although our results show inconsistent performance along coastal regions and warmer southerly parts of the study domain, active radar instruments such as SeaWinds offer the potential for monitoring high-latitude snow thaw at spatial scales appropriate for pan-Arctic applications in near real time. Applications include hydrological model verification, analysis of lags between snow thaw and river response, and determination of large-scale snow extent.

Simulating the urban water and contaminant cycle

This paper presents the water and contaminant daily simulation model of the total water cycle, called UVQ. The model has been developed to provide a means for rapidly assessing the impacts of conventional and non-conventional urban water supply, stormwater and wastewater development options on the total water cycle. UVQ provides a valuable insight into the viability of prospective alternative water servicing options in terms of environmental and performance requirements.

Comparative analysis of three methods to generate soil hydraulic functions

An accurate determination of the soil hydraulic functions is crucial for using soil water simulation models. The purpose of this study is to compare different methods to estimate hydraulic properties and their effects on the simulation of soil water movement. Soil water retention and hydraulic conductivity functions were either measured in the laboratory (core method) and in the field (instantaneous profile method and crust method) or estimated from other soil properties (pedo-transfer functions, PTFs). The data were fitted to the van Genuchten equations and the resulting parameter sets were used as input of the mechanistic WAVE model to simulate soil water flow. Calculated and observed water content values were compared to assess the different methods used for determining the soil hydraulic properties. The hydraulic parameter sets generated from the different approaches were found to be significantly different and to have a strong impact on the soil water flow simulation. Soil water retention data obtained from field experiments and laboratory tests differ significantly near saturation but are similar for lower water content values. On the whole, retention curves given by the Rawls and Vereecken PTFs perform best while the other tested PTFs give poorer results. Unsaturated hydraulic functions given by the Saxton, Cosby and Rawls PTFs are relatively close to the measured data in the silt loam layers while the Vereecken PTF leads to good estimations in the pressure head mid-range values, but poor results in the high and low ranges. In the sandy loam layer characterized by low clay and organic matter content, all the tested PTFs perform very poorly. Globally, when considering prediction of both soil hydraulic functions, the Rawls PTFs were found to be the most adequate. The comparison between observed and simulated soil water content values indicates that the best agreement is achieved with hydraulic properties estimated from field experiments. Results are less satisfactory with the parameter sets obtained from the laboratory tests and very poor with the parameter sets given by the Rawls PTFs.

An assessment of the VIC-3L hydrological model for the Yangtze River basin based on remote sensing: a case study of the Baohe River basin

As a first step in our effort to simulate terrestrial hydrological processes for the entire Yangtze River basin, a hydrologically based three-layer variable infiltration capacity (VIC-3L) land surface model is applied to the Baohe River basin, which has a drainage area of 2500 km2. Water fluxes of the Baohe River basin are simulated using the VIC-3L model at a spatial resolution of approximately 4 km. The soil and land cover properties were taken from the resource and environment database (scale 1 : 4 000 000) of China (REDC). The vegetation and land cover data were also derived from the moderate resolution imaging spectroradiometer (MODIS) data over the study area. Measurements from a number of weather stations were used to obtain meteorological forcings for each modeling grid based on a stepwise interpolation approach (SIA). The VIC-3L model was run at a daily time step. Differences in model-simulated water fluxes resulting from using the two different data sources (i.e., MODIS versus REDC) on information of land cover classification and leaf area index (LAI) were compared. Model-simulated daily runoff was routed through the Baohe River basin network and compared with the daily observed streamflow measured at Jiangkou hydrological station at the outlet of the Baohe River basin from 1992 to 2001. Results from the VIC-3L model simulations using REDC data compare well with the observations in general, but the model-simulated streamflows often underestimate the observed peak flows significantly. In comparison, the underestimations of the peak flows are significantly improved when more accurate vegetation information on land cover and LAI from MODIS is used. The REDC land cover was obtained about 20 years ago with only six different types of vegetation that are generally distributed homogeneously throughout the study region. However, the land cover in the Baohe River basin is mainly mixed forest. The unrealistically large land cover area of deciduous broadleaf forest in the REDC data source resulted in an unreasonable increase in evapotranspiration and a decrease in streamflows. This study indicates clearly the important role that remote sensing (e.g., MODIS data) plays in improving model simulations.

Snow process modeling in the North American Land Data Assimilation System (NLDAS): 2. Evaluation of model simulated snow water equivalent

This is the second part of a study on the cold season process modeling in the North American Land Data Assimilation System (NLDAS). The first part concentrates on the assessment of model simulated snow cover extent. In this second part, the focus is on the evaluation of simulated snow water equivalent (SWE) from the four land surface models (Noah, MOSAIC, SAC and VIC) in the NLDAS. Comparisons are made with observational data from the Natural Resources Conservation Service's Snowpack Telemetry (SNOTEL) network for a 3‐year retrospective period at selected sites in the mountainous regions of the western United States. All models show systematic low bias in the maximum annual simulated SWE that is most notable in the Cascade and Sierra Nevada regions where differences can approach 1000 mm. Comparison of NLDAS precipitation forcing with SNOTEL measurements revealed a large bias in the NLDAS annual precipitation which may be lower than the SNOTEL record by up to 2000 mm at certain stations. Experiments with the VIC model indicated that most of the bias in SWE is removed by scaling the precipitation by a regional factor based on the regression of the NLDAS and SNOTEL precipitation. Individual station errors may be reduced further still using precipitation scaled to the local station SNOTEL record. Furthermore, the NLDAS air temperature is shown to be generally colder in winter months and biased warmer in spring and summer when compared to the SNOTEL record, although the level of bias is regionally dependent. Detailed analysis at a selected station indicate that errors in the air temperature forcing may cause the partitioning of precipitation into snowfall and rainfall by the models to be incorrect and thus may explain some of the remaining errors in the simulated SWE.

Flexible TIP4P model for molecular dynamics simulation of liquid water

We have developed a flexible version of the TIP4P model for molecular dynamics simulations. The model does a reasonable job of reproducing the experimental peak frequencies and widths of the bend and stretch peaks in the infrared spectra for both liquid H 2 O and D 2 O .

Water Quality Operation with a Blending Reservoir and Variable Sources

A linear programming approach is used within a simulation model to manage water quality for an urban water supply agency with a blending reservoir and a raw water source of variable quality, the Sacramento–San Joaquin Delta. Schedules of source water quality and availability in addition to the physical storage and conveyance constraints, regulatory restrictions, raw water demand, and urban water management priorities are used as input to the simulation model. Resulting schedules of delivered water quality, reservoir storage, and other relevant data can be used to assist with the evaluation of proposed water project, regulations, and operation strategies that impact water supply performance. This method is applied within the raw water simulation model of the Contra Costa Water District (CCWD), a water district that operates a large lending reservoir with the Sacramento–San Joaquin Delta as its primary water source. A parameter variation study involving salinity levels in the delta quantifies the sensitivity of CCWDs delivered water quality to changes in the source quality.

APPLICATION OF ENHANCED ANNEALING TO GROUND WATER REMEDIATION DESIGN1

ABSTRACT: A methodology for ground water remediation design has been developed that interfaces ground water simulation models with an enhanced annealing optimizer. The ground water flow and transport simulators provide the ability to consider site‐specific contamination and geohydrologic conditions directly in the assessment of alternative remediation system designs. The optimizer facilitates analysis of tradeoffs between technical, environmental, regulatory, and financial risks for alternative design and operation scenarios. A ground water management model using an optimization method referred to as “enhanced annealing” (simulated annealing enhanced to include “directional search” and “memory” mechanisms) has been developed and successfully applied to an actual restoration problem. The demonstration site is the contaminated unconfined aquifer referred to as N‐Springs located at Han‐ford, Washington. Results of the demonstration show the potential for improving groundwater restoration system performance while reducing overall system cost.