An improved streamflow model with climate and land use factors for Hulu Langat Basin

Water is essential for human beings and it is vital in various fields such as agriculture, navigation, energy production, recreation and manufacturing. Rapid urbanization, population growth and economic developments could potentially put stress on the water resources by increasing the water demand. In addition, climate change and land use change could also cause variations in the quantity and quality of water resources. Therefore, assessing the impacts of these changes on water availability is essential and requisite to adapt water resources management and for planning sustainable development strategies especially in a rapid socio-economic development. The aim of this study is to investigate the impact of past and future climate change and land use change on mean monthly and annual streamflows in the Hulu Langat basin, Malaysia utilizing a new generation of physically based hydrological models. The James W. Kirchner (JWK) model is a new physically based model. Although this model does not need any upscaling it is more appropriate for cold and humid areas and it considers the basin as a single storage system. These limitations could have impacts on the applicability of the model. Thus, in the present study, to achieve the objectives, first, the James W. Kirchner (JWK) method was modified and the modified model (MJWK) was then combined with the Soil Conservation Service (SCS) effective rainfall estimation method (MJWK-SCS model) to estimate river flow. An averaging ensemble version of MJWK-SCS model was also proposed (E-MJWK-SCS). Afterwards, the MJWK, MJWK-SCS, EMJWK-SCS, Soil and Water Assessment Tool (SWAT), Artificial Neural Network (ANN), Nonlinear AutoregRessive with eXogenous input (NARX) and waveletNARX models were utilized to predict mean monthly river flow from daily climatic data. The models were calibrated for the period 1985-1988 and the validation was performed for the period 2002-2005. In the calibration phase, the Wavelet-NARX, EMJWK-SCS and SWAT models performed the best with the Nash-Sutcliff Efficiency (NSE) values of 0.85, 0.78 and 0.66, respectively. However, in the validation phase the SWAT and E-MJWK-SCS models performed the best with the NSE values of 0.74 and 0.73, respectively. Since the E-MJWK-SCS and SWAT models performed well in both the calibration and validation phases based on NSE values, they were utilized to assess the climate change and land use change effects on mean monthly and annual streamflows. Prior to applying these models, the uncertainty of their predictions was analyzed utilizing the Sequential Uncertainty Fitting 2 (SUFI2) algorithm. The uncertainty analysis showed that both the models had an acceptable level of uncertainty. However, the EMJWK-SCS model showed lower quantity of uncertainty in prediction with p-factor and r-factor of 0.88 and 0.81 than the SWAT model with p-factor and r-factor of 0.69 and 0.51, respectively. The analysis of the past climate change and land use change impacts on streamflow showed that at annual scale the land use change was more effective than the climate change and it increased mean annual streamflow (11.43% and 5.68% utilizing E-MJWK-SCS and SWAT models, respectively). At monthly scale, both the land use and climate change altered streamflows. The impact of possible future climate change and land use change on mean monthly and annual streamflows was also investigated. Firstly, the climatic variables were estimated under the A1B and A2 climate change scenarios employing the LARS-WG model and the land use map of year 2025 was generated based on the trend of land use changes in the period 1984-2002 utilizing the Land Change Modeler (LCM). Then mean monthly and annual streamflows were forecasted under different combinations of land use and climate change scenarios for the period 2025-2028. At annual scale, a rise in streamflow is expected under the land use change (4.07% and 3.88% utilizing E-MJWK-SCS and SWAT models, respectively) and the combined land use change and climate change scenarios (ranged from 1.81% to 4.54% under various scenarios). The climate changes scenarios represented a decline in mean annual streamflow (ranged from -5.78% to -0.27% for various scenarios). At monthly scale, both increases and decreases in flows were seen under all the scenarios considered (ranged from a decrease of 8.92% to an increase of 11.76% under various scenarios). The findings also showed that the droughts would be possible under the combined climate and land use changes scenarios in the dry seasons. It is concluded that not only both the E-MJWK-SCS and SWAT models are useful tools to simulate mean monthly river flow in the basin but are also suitable for investigating the impacts of climate and land use changes on mean monthly and annual streamflows.

Preview Text FK 2014 150IR.pdf Download (1MB) | Preview

Actions (login required)

View Item