WaterSubject

In arid regions, human activities like agriculture and industry often require large ground water extractions. Under these circumstances, appropriate ground water management policies are essential for preventing aquifer overdraft, and thereby protecting critical ecologic and economic objectives. Identification of such policies requires accurate simulation capability of the ground water system in response to hydrological, meteorological, and human factors. In this research, artificial neural networks (ANNs) were developed and applied to investigate the effects of these factors on ground water levels in the Minqin oasis, located in the lower reach of Shiyang River Basin, in Northwest China. Using data spanning 1980 through 1997, two ANNs were developed to model and simulate dynamic ground water levels for the two subregions of Xinhe and Xiqu. The ANN models achieved high predictive accuracy, validating to 0.37 m or less mean absolute error. Sensitivity analyses were conducted with the models demonstrating that agricultural ground water extraction for irrigation is the predominant factor responsible for declining ground water levels exacerbated by a reduction in regional surface water inflows. ANN simulations indicate that it is necessary to reduce the size of the irrigation area to mitigate ground water level declines in the oasis. Unlike previous research, this study demonstrates that ANN modeling can capture important temporally and spatially distributed human factors like agricultural practices and water extraction patterns on a regional basin (or subbasin) scale, providing both high-accuracy prediction capability and enhanced understanding of the critical factors influencing regional ground water conditions.In China, a huge country with a population of more than 1.3 billion, half of its land is situated within arid or semiarid regions of which 26.6% has an average precipitation of less than 200 mm/year. Ground water plays an important role in the economic development and ecological balance in these arid and semiarid areas, particularly in Northwest China. Over the past several decades, human activities such as ground water extraction for irrigation have resulted in aquifer overdraft in these areas, disrupting the natural equilibrium of these systems. Excessive ground water level declines have produced serious ecological problems such as land desertification and soil salinization, displacing inhabitants from their ancestral homeland.

In China, there is 150 × 10³ km² of desert area, which is increasing at an annual rate of approximately 2000 to 3000 km²/year. The saline areas produced by irrigation with the highly mineralized deep ground water encompass roughly 2 million hectares, which occupy approximately one-third of the country’s total saline Because of these severe consequences and recognizing China’s growing reliance on increasingly scarce ground water resources, it has become extremely important to accurately simulate and predict potential ground water level changes in these regions so that appropriate water resources management and environmental protection policies can be developed and implemented.

A number of previous researchers focused on the impact of human activities on ground water systems in arid and semiarid areas, leading them to conclude that overexploitation of these systems has produced excessive ground water level declines However, these studies have analyzed only the relationships between ground water levels and human activities on a qualitative level. Others have used advanced numerical models to simulate and quantify the impact of human activities (e.g., ground water extraction) on ground water conditions simulation models had been used successfully for simulating and predicting ground water levels for many years, extending back into the 1960s. As noted by “The power of these models is they can capture high spatial and temporal variability of aquifer properties and conditions inherent to natural hydrogeologic systems. However, this capability renders numerical models data intensive, and to achieve acceptable simulation and prediction performance, the properties and conditions of the ground water system must be accurately represented within the model’s space and time domains. The unavoidable discrepancies between the model and the real world system inevitably produce simulation and prediction error.” Because the properties and conditions of the ground water system can never be ascertained with absolute accuracy, empirical models may provide an appropriate alternative method and can provide useful results without costly calibration time.

The artificial neural network (ANN) methodology is an alternative modeling and simulation tool, especially for dynamic nonlinear systems. One of the most important features of ANN models is their ability to adapt to recurrent changes and detect patterns in a complex natural system. As discussed, unlike traditional physical-based numerical models, ANNs often do not require explicit characterization and quantification of physical properties and conditions and are not based upon simplifying mathematical and physical assumptions (e.g., porous media). Rather, ANNs learn the system behavior of interest from representative data that often consist of easily measurable variables.

The advantages and disadvantages of ANNs over conventional simulation methods have been discussed in detail by. In hydrology, ANNs have been largely applied to the rainfall-runoff modeling, precipitation forecasting, and water quality modeling . ANNs have also been applied successfully to ground water level prediction under variable weather conditions and under pumping conditions without explicitly accounting for this variable developed ANN models that accurately predicted transient ground water levels in response to variable weather and pumping conditions and extended this work to water quality for an upconing problem in a coastal aquifer. Some ANN ground water prediction models have been used for ground water management, where the models are combined with formal optimization methodology. This body of research collectively demonstrates that ANN models may serve as efficient and accurate models for simulating ground water systems and can be used for developing effective management and protection strategies.

In this study, ANNs were developed to predict average ground water levels in a semiarid region using monthly stress periods, with predictor input variables addressing meteorological, hydrological, population, and agricultural ground water extractions. Therefore, unlike previous studies by others like this article demonstrates that important and numerous ground water extractions that are temporally and spatially distributed over a large regional-scale system can be accounted for by an ANN, with their corresponding effect on the system accurately predicted by the model. The ANN models were used to perform valuable sensitivity analyses, identifying the relative importance of different factors on the regional ground water system. In addition, the models were used to perform extended simulations over hypothetical 1-year periods, using different sets of input values, to assess the impacts of agricultural activities on the ground water system. This modeling simulation and analysis helped quantify average expected ground water level responses to different levels of agricultural activities, which can be used to help develop appropriate long-term strategies to promote the long-termed sustainability of the resource and the surrounding environment.

This research demonstrates that ANN modeling can capture important temporally and spatially distributed human factors like agricultural practices and water extraction patterns on a regional basin (or subbasin) scale, and achieve high predictive accuracy, as well as improving understanding of complex ground water systems. The modeling results indicate that human activities and surface water inflows are the most important factors affecting monthly ground water level changes in the Minqin oasis. As the simulation exercises demonstrate, which is supported by data, the ground water usage has been not sustainable, and if human stresses remain unchecked, ground water levels will continue to decline. The simulation results indicate that reducing irrigation areas and increasing surface water inflows are critical measures for reducing ground water level declines within the Minqin oasis.

For regional-scale ground water basins, water scientists and decision makers need to understand the effect of hydrologic, meteorologic, and human activities on ground water conditions. The sensitivity analysis and simulation capability afforded by ANN models, as shown in this study, can be an extremely effective and efficient tool for ground water analysis and management, and for the Minqin oasis, helped achieve the following objectives: (1) gaining a better understanding of the system by semiquantifying the relationships between human activities and environmental conditions on ground water levels; (2) identifying the appropriate levels of agricultural activities and surface water reservoir inflows (i.e., upstream diversions) for maintaining ground water levels; (3) revising data collection strategies for improving models and increasing confidence in simulation projections.

The modeling results and analysis will help decision makers understand the influence of human actions on the ground water system, promoting its sustainable use and thereby preserving the long-term economic viability of the region. At the same time, additional work is required. The ANN models developed in this study have limited ability to reveal differences in ground water responses over space in response to variable agricultural practices and environmental conditions. In future work, we will develop ANN models to predict ground water levels at multiple locations to delineate spatial variations of ground water responses across the regional-scale basins, as well as perform multiobjective optimization.