The uncertainty of gis-based interpolation methods in constructing shallow groundwater distribution map: A case study at Pleiku city, Gia Lai province

Abstract

Four interpolation methods are applied to interpolate shallow groundwater level in Pleiku city, including inverse distance weighted, tension spline, universal kriging, and ordinary kriging. The cross-validation results record that the ordinary kriging is the best interpolation method which is shown by the lowest RMSE value, the highest R2 value. It is selected to assess the shallow groundwater level in spatial and seasonal change. Based on the groundwater level interpolated by the ordinary kriging, the groundwater level is divided into the northern and southern parts of the study area. The distribution of groundwater level is shallower than that in the southern part wheret groundwater depth is around less than15 m while in the southern part, most of the groundwater level is higher than 18 m. The elevation of groundwater level is found in rainy season; the elevated area accounts for 72.6% of natural area. Additionally, the groundwater level also declines in the rainy season at some region, focusing on Bien Ho lake region and two regions in the southern part of the city whose area accounts for 27.4% of the natural area.