Parameter Optimization of Conceptual Tank Model for Groundwater Level Prediction

Abstract

Groundwater is regarded as one of the critical factors that can affect slope stability. Thus, groundwater levels may render useful information regarding the stability conditions of a slope. This preliminary study focused on developing a simple and quick analytical tool to evaluate the groundwater levels due to rainfall for slope stability assessment. To achieve this objective, a well-established rainfall-runoff model known as tank model was adopted in this study. An instrumented soil slope located in Malaysia was used as the case study to investigate the effectiveness of the proposed approach. Rainfall and groundwater levels data for a period of 8 months were used to calibrate the tank model unknown parameters representing runoff, infiltration, groundwater flow and head. The tank model was able to produce a satisfactory root mean square error (RMSE) of 0.185 for the computed groundwater levels compared to the observed groundwater levels. To produce a more accurate prediction, it is recommended to utilize the multi tank models that are position at crest, middle and toe of the slope. An accurate groundwater levels prediction will contribute to a reliable slope stability analysis which is valuable for the landslide early warning system applications.