Integrating economic and groundwater models for groundwater-management can help improve understanding of trade-offs involved between conflicting socioeconomic and biophysical objectives. However, there is significant uncertainty in most strategic decision-making situations, including in the models constructed to represent them. If not addressed, this uncertainty may be used to challenge the legitimacy of the models and decisions made using them. In this context, a preliminary uncertainty analysis was conducted of a dynamic coupled economic-groundwater model aimed at assessing groundwater extraction rules. The analysis demonstrates how a variety of uncertainties in such a model can be addressed. A number of methods are used including propagation of scenarios and bounds on parameters, multiple models, block bootstrap time-series sampling and robust linear regression for model calibration. These methods are described within the context of a theoretical uncertainty management framework, using a set of fundamental uncertainty management tasks and an uncertainty typology.

A DRASTIC-model method based on a geographic information system (GIS) was used to study groundwater vulnerability in Egirdir Lake basin (Isparta, Turkey), an alluvial area that has suffered agricultural pollution. ‘Lineament’ and ‘land use’ were added to the DRASTIC parameters, and an analytic hierarchy process (AHP) method determined the rating coefficients of each parameter. The effect of lineament and land-use parameters on the resulting vulnerability maps was determined with a single-parameter sensitivity analysis. Of the DRASTIC parameters, land use affects the aquifer vulnerability map most and lineament affects it least, after topography. A simple linear regression analysis assessed the statistical relation between groundwater nitrate concentration and the aquifer vulnerability areas; the highest R ² value was obtained with the modified-DRASTIC-AHP method. The DRASTIC vulnerability map shows that only the shoreline of Egirdir Lake and the alluvium units have high contamination potential. In this respect, the modified DRASTIC vulnerability map is quite similar. According to the modified-DRASTIC-AHP method, the lakeshore areas of Senirkent-Uluborlu and Hoyran plains, and all of the Yalvaç-Gelendost plain, have high contamination potential. Analyses confirm that groundwater nitrate content is high in these areas. By comparison, the modified-DRASTIC-AHP method has provided more valid results.

Management of groundwater resources can be enhanced by using numerical models to improve development strategies. However, the lack of basic data often limits the implementation of these models. The Kharga Oasis in the western desert of Egypt is an arid area that mainly depends on groundwater from the Nubian Sandstone Aquifer System (NSAS), for which the hydrogeological data needed for groundwater simulation are lacking, thereby introducing a problem for model calibration and validation. The Grey Model (GM) was adopted to analyze groundwater flow. This model combines a finite element method (FEM) with a linear regression model to try to obtain the best-fit piezometric-level trends compared to observations. The GM simulation results clearly show that the future water table in the northeastern part of the study area will face a severe drawdown compared with that in the southwestern part and that the hydraulic head difference between these parts will reach 140 m by 2060. Given the uncertainty and limitation of available data, the GM produced more realistic results compared with those obtained from a FEM alone. The GM could be applied to other cases with similar data limitations.

Several groundwater-level forecasting studies have shown that data-driven models are simpler, faster to develop, and provide more accurate and precise results than physical or numerical-based models. Five data-driven models were examined for the forecasting of groundwater levels as a result of recharge via tailings from an abandoned mine in Quebec, Canada, for lead times of 1 day, 1 week and 1 month. The five models are: a multiple linear regression (MLR); an artificial neural network (ANN); two models that are based on de-noising the model predictors using the wavelet-transform (W-MLR, W-ANN); and a W-ensemble ANN (W-ENN) model. The tailing recharge, total precipitation, and mean air temperature were used as predictors. The ANN models performed better than the MLR models, and both MLR and ANN models performed significantly better after de-noising the predictors using wavelet-transforms. Overall, the W-ENN model performed best for each of the three lead times. These results highlight the ability of wavelet-transforms to decompose non-stationary data into discrete wavelet-components, highlighting cyclic patterns and trends in the time-series at varying temporal scales, rendering the data readily usable in forecasting. The good performance of the W-ENN model highlights the usefulness of ensemble modeling, which ensures model robustness along with improved reliability by reducing variance.

Hydrologic budgets to determine groundwater availability are important tools for water-resource managers. One challenging component for developing hydrologic budgets is quantifying water use through time because historical and site-specific water-use data can be sparse or poorly documented. This research developed a groundwater-use record for the Ozark Plateaus aquifer system (central USA) from 1900 to 2010 that related county-level aggregated water-use data to site-specific well locations and aquifer units. A simple population-based linear model, constrained to 0 million liters per day in 1900, provided the best means to extrapolate groundwater-withdrawal rates pre-1950s when there was a paucity of water-use data. To disaggregate county-level data to individual wells across a regional aquifer system, a programmatic hierarchical process was developed, based on the level of confidence that a well pumped groundwater for a specific use during a specific year. Statistical models tested on a subset of the best-available site-specific water-use data provided a mechanism to bracket historic groundwater use, such that groundwater-withdrawal rates ranged, on average, plus or minus 38% from modeled values. Groundwater withdrawn for public supply and domestic use accounted for between 48 and 74% of total groundwater use since 1901, highlighting that groundwater provides an important drinking-water resource. The compilation, analysis, and spatial and temporal extrapolation of water-use data remain a challenging task for water scientists, but is of paramount importance to better quantify groundwater use and availability.