Research on water supply and agricultural water use forecasting in arid regions: a case study of Xinjiang

Water resources are the lifelines of the agricultural development in Xinjiang. Currently, the problem of water shortage for agriculture in this region is becoming increasingly severe. Against this backdrop, predicting the changing trends of water supply and agricultural water use in Xinjiang and analyzing the supply and use relationship between them are of great practical significance for ensuring the sustainable development of regional agriculture. Firstly, we conducted an in-depth analysis of the water supply and agricultural water use patterns in Xinjiang over the past two decades. Secondly, we evaluated and compared several mainstream water resource prediction models, ultimately developing a novel GM(1,1)-NN essemble model. Validation results demonstrated that this model exhibits superior accuracy in forecasting water supply and agricultural water use compared to other existing models. Finally, we utilized the newly developed GM(1,1)-NN essemble model to predict short-term water supply and agricultural water use trends in Xinjiang. Based on these findings, we proposed recommendations for water resource conservation from both technological and regional planting perspectives. The key results are as follows: (1) There are significant regional disparities in water resources in Xinjiang, primarily attributed to uneven precipitation distribution and imbalanced economic development. (2) The GM(1,1)-NN essemble model demonstrates high short-term predictive accuracy for both water supply and agricultural water use in Xinjiang. (3) According to our GM(1,1)-NN essemble model’s projections, both water supply and agricultural water use in Xinjiang are expected to exhibit a downward trend in the coming years. The reduction in agricultural water use will help allocate more water resources to non-agricultural sectors. (4) Despite these improvements, the contradiction between water shortage and the high proportion of agricultural water use (approaching to 88%) remains unresolved. Therefore, it is recommended to reduce agricultural water use through the widespread adoption of water-saving facilities and the optimization of crop planting structures across different regions.