Quantifying uncertainties of multi-model climate change scenarios on the water crisis in Malaysia

Malaysia has a relatively limited capacity to deal with the effects of climate change while being one of the most vulnerable nations to its effects. As a developing country, the lack of a consistent temporal and spatial data source has always been an issue, and the region is also considered data-scarce. This study's primary goal is to evaluate the effects of climate change on Malaysia's water resources, particularly the Selangor River Basin (SRB). Instead of using a single source input dataset, cross-combined datasets from multiple sources were used in order to optimise the hydrological model. Five input variables, including precipitation, temperature, solar radiation, relative humidity, and wind speed, were used to define seven scenarios using single and cross- combined method. To improve the hydrological model multi-site calibration method is employed. Finally, climate change prediction data from several Global Climate Models (GCMs) is utilised to assess the effects of climate change on SRB water supplies. The CFSR and CMADS global reanalysis datasets show a highly significant relationship on precipitation, with an r-value of 0.81 for both datasets. However, for temperature data, CMADS surpasses CFSR on maximum and minimum temperatures, with 0.6 and 0.7, respectively. In the SWAT model, most of the scenarios achieved a 'good' performance range on the calibration and validation processes. However, SWAT model with CFSR as input data achieved an 'unsatisfactory' range with R2 of 0.35, NSE of 0.16, Pbias of 0.00, KGE of 0.50, and RSR of 0.92. For a cross-combined approach, the result shows the combination of the observed and CMADS datasets performed better than the combination of the observed and CFSR datasets. The sequential technique outperformed the simultaneous and basin-by-basin techniques by achieving 'satisfactory' range at all outlets. The SRB's assessment of climate change predicted an increase in precipitation and temperature from 2030 to 2050. Climate data from 'ensemble average' realisation predicted SRB would receive a huge amount of precipitation in November and April every year, and high temperatures from February to June. Additionally, a few sub-basins are expected to have water availability greater than 5 m3/s for three consecutive years.