Assessment of resources and ecological risks induced by groundwater utilization in the unconfined aquifer in the western Jilin Province: A case study in the Taoer River catchment

Abstract

The western part of Jilin Province is one of the major agriculture zones in China. It is located in a semi-arid zone with limited and unreliable water resources, and also with serious soil salinization. It is of critical importance to manage the groundwater levels in the unconfined aquifer, considering both the risks of water resources and the ecological problems related to the shallow buried depth of groundwater levels. However, there is still lack of quantitative risk analyses on the current strategies of water use in this area. In this study the recurrent neural network is used to predict the rainfall and river flux from 2019 to 2023, which are then used as the input variables in the stochastic groundwater flow models to predict the spatial distribution of groundwater levels in the unconfined aquifer. The groundwater levels under six scenarios, including the present-day recharge and discharge, drought, chaos extraction, managed aquifer recharge, drip irrigation and paddy farming, are calculated. Following a risk assessment, the drip irrigation with a net extraction rate ranging from 2.0×108 to 3.0×108 m3/a is considered as the best strategy for groundwater resources utilization, which can effectively prevent the water resources in the unconfined aquifer from being over-exploited (with the buried depth of greater than 12 m), and also maintain the depth of groundwater table of greater than 1m to reduce the risk of soil salinization. Meanwhile, chaos extraction inducing the water resources depletion is the major factor, and both the managed aquifer recharge and paddy farming are helpful in water resources conservation, but may worse the soil salinization. The methodology employed in this study can be widely used in other arid and semi-arid areas for groundwater resources management.

Keywords

• groundwater resource management
• risk assessment
• soil salinization
• stochastic modeling
• recurrent neural networks
• numerical simulation