Vertical Soil Moisture Heterogeneity Analysis and Model Reconstruction in Moist Zones of Aquaculture Pond Dikes Based on Multi-depth Soil Probe Observations

Due to frequent irrigation and drainage, the vertical distribution of soil moisture in pond dikes plays an important role in water retention and pollutant transport. This study selected five pond dikes from a rice–crayfish co-culture area in Xinghua, Jiangsu Province as sampling sites. Frequency Domain Reflectometry (FDR) sensors were installed at depths of 5, 15, 30 and 50 cm to continuously monitor soil moisture content and groundwater level over two aquaculture cycles (approximately 300 days). A layered Van Genuchten–Mualem model was used to construct the soil hydraulic functions. The results showed that during high water-level periods, the average difference in soil moisture content between surface and deep layers across the five sites reached 0.17 m³/m³. This led to an average error of 25% in the estimated hydraulic conductivity. After correction, the root mean square error (RMSE) of the model decreased from 0.031 to 0.014, indicating a clear improvement in model accuracy. This study proposed a moisture transport calculation method based on vertical gradient constraints, providing parameter support for seepage management and soil function evaluation in moist aquaculture zones.