Comparative assessment of empirical random forest family's model in simulating future streamflow in different basin of Sarawak, Malaysia

Abstract

In Sarawak, a region highly vulnerable to climate change, the translation of climate-induced changes in rainfall to river flow is non-linear, presenting a challenge for water resource managers. This research investigates the impact of climate change on hydrological processes in Sarawak, Malaysia, with a specific focus on assessing future spatiotemporal variations in streamflow. The families of Random Forest (RF) empirical models based on data mining techniques were compared and utilized to develop a continuous hydrologic model. Then, by incorporating rainfall and evapotranspiration future projection prepared based on RF past performance statistical downscaling, the top performing RF empirical hydrological model was used for future streamflow projection. The results showed that despite an expected increase in rainfall, the RFR (Random Forest by Randomization) empirical hydrological model demonstrated a potential decrease in river runoff due to heightened evapotranspiration demands associated with rising temperatures. The examination of climate-induced alterations in both rainfall and evapotranspiration patterns revealed a consistent decrease in river discharges during the early to middle period across Sarawak, followed by a shift towards an increasing trend by the end of the 21st century. The central region along the Rajang basin exhibited a prevailing decrease in river discharge, with contrasting patterns in the last part of the century. The northern region displayed diverse trends, with some basins experiencing decreases in river runoff despite augmented rainfall, emphasizing the heterogeneity in response. By employing empirical models, and projecting future scenarios, the study contributes to a better understanding of climate change impacts on hydrology in the region, essential for effective water resource management and environmental conservation.