Abstract
When modelling unsteady groundwater flow in a large area, the estimation of hydraulic parameters and groundwater levels at node points within a finite element or finite difference mesh is particularly important. Unfortunately, the information on these hydraulic parameters and groundwater levels is usually insufficient. To simulate large-scale unsteady groundwater flow without enough data of groundwater level and hydraulic conductivity, a methodology was proposed in this paper basing on geostatistics, snowmelt and unsteady groundwater flow. Firstly, ROKMT (Residual Ordinary Kriging with Modified Trend) was presented to estimate the spatial distribution of groundwater level, where the spatial trend of the observed groundwater level was modified using DEM (Digital Elevation Model) as an auxiliary information.
Secondly, based on the estimated spatial distribution of groundwater level, the spatial distribution of hydraulic conductivity was identified using the optimization algorithm. Furthermore, the snowmelt in space was estimated by a heat balance model, and the unsteady groundwater flow was simulated using TFDM (Triangular Finite Difference Method).
Finally, as a case study, the unsteady groundwater flow of the Sarobetsu wetland was simulated for the period from January 1998 to December 2000. The results indicate that the proposed methodology can simulate the unsteady groundwater flow in satisfactory accuracy.
