Ecological Thresholds for Groundwater Depth and Ecological Risk Early Warning in the Western Songnen Plain, China

Groundwater plays a key role in maintaining groundwater-dependent ecosystem stability. However, existing ecological thresholds for groundwater depth mainly focus on fractional vegetation coverage (FVC), with insufficient consideration of soil salinity, limiting the accuracy of ecological risk assessments. Focusing on the Songnen Plain, we determined more accurate ecological thresholds and established a groundwater depth-based ecological risk classification standard using groundwater depth, FVC, and soil salinity data from 2001, 2010, 2019, and 2024. A combined Kalman filter and Long Short-Term Memory (LSTM) neural network model was used to predict groundwater depth changes in 2030 based on current conditions and assess ecological risks. Results indicated that both models showed good performance (R2 = 0.95 and 0.79, respectively). From 2001 to 2024, groundwater depth increased, with increasing FVC and decreasing saline soil area. Groundwater depth exhibited significant threshold effects on vegetation growth and soil salinization, with 4.0–6.5 m identified as the optimal range for vegetation growth, while soil salinity accumulated when groundwater depth was <5.5 m. The proposed ecological risk classification revealed significant spatial pattern changes of ecological risk from 2001 to 2030, with medium warning dominating in 2030. The study provides valuable insights for groundwater management, ecological risk early warning, and ecological conservation.