Impact of dam and land use change on the flow characteristics of Selangor River basin

Abstract

Rainfall-runoff response is the inter-play between rainfall, and the properties of a catchment. The aim of this study is to evaluate the influence of dam and land use change on the hydrological properties specifically on runoff, stormflow volume, time to peak and flow duration curve. This study chose Selangor River Basin (SRB) in view of its importance in providing about 40% of the water supply to Selangor and Klang Valley. There are two dams in the SRB, namely Sungai Selangor Dam and Sungai Tinggi Dam. In this study, SRB is divided into two catchments, namely CA (1529 km2) and CB (220 km2) of which CB is a sub-basin of the CA. Fifty two percent of the land area in CA is covered by forest, 11.7% is urban area and the rests are oil palm and agricultural. 94% of CB is covered by forest and forms the entire basin of the Sungai Selangor Dam. Prior to the dam construction in the year of 2007, the monthly runoff coefficient for CA ranges from 0.19 to 0.48 while for CB ranges from 0.59 to 0.79. This data was observed and collected in ten years period of time. The monthly runoff in CB shows a larger variation after the completion of dam with runoff coefficient between 0.28 and 0.89. The baseflow and stormflow before the dam construction made up about 51.5% and 48.5% of the total flow, respectively. The effect of dam operation was evident from the shift in the flow duration curve with reduction in the proportion of extreme high flow especially for CB and increased in the proportion of extreme low flow for CA. However, both low flow and stormflow in CA recorded a decrease of 70% and 58%, respectively. This must be due to heavy abstractions of river water for the treatment plants. The HEC-HMS model was calibrated and validated for three rainfall categories: light, medium and heavy. The findings showed that the future changes in land use are expected to increase the peak flow, and stormflow volume simultaneously shorten the time to peak. It is also found that the model parameters for years 2006 and 2016 were different and attributed to changes in the land use. The simulated peak flow in 2016 for 100-year ARI storm is three-fold higher than the base line condition in year 2006. In conclusion, the higher the imperviousness area the higher the peak flow and stormflow volume but lower the time to peak.