Modelling streamflow response under different climate change scenarios in Luvuvhu River Catchment, South Africa

Abstract

Climatic change vulnerability assessments needs the evaluation of climatic threats to water systems as well as the design of effective strategies. The factors and repercussions of climate change are multifaceted, encompassing several ironic elements, particularly the reality that low-income nations are the most vulnerable to its impacts while contributing the least to the human activities that exacerbate climate change. This research examines climate change scenarios utilizing four potential future scenarios, referred to as Shared Socio-economic Pathways (SSPs), encompass SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5 across twelve general circulation models (GCMs) from the CMIP6 framework. Streamflow responses were evaluated for two distinct timeframes: near future (2015 - 2050) and far future (2051 - 2100), with historical data analysed for a period of 1980 to 2014. Climate stress tests were conducted, and an analysis was carried out to model the streamflow response in the Luvuvhu River Catchment (LRC) to changing climate conditions using Soil Water Assessment Tool (SWAT). Readily accessible spatial and temporal data were utilized and the model was calibrated with recorded monthly streamflow data. The performance of the model was assessed through computation of Nash-Sutcliffe Efficiency (NSE), Percent Bias (PBIAS), and the coefficient of determination (R²). The calibration findings for monthly simulation for the (1991-1999) period fell within acceptable model performance limits, with an NSE of 0.57, PBIAS of -0.35, and R2 of 0.74. Validation findings for monthly simulations from 2001 to 2009 were mostly acceptable ranges for model performance with NSE (0.49), PBIAS (-0.50) and R2 (0.52). The observed and simulated streamflow showed strong agreement, as confirmed by both graphical analysis and quantitative statistics. Projected annual precipitation and temperature changes showed that the LRC climate will be warmer in most scenarios than wetter as most of the streamflow respond fell under high risks of about 30% decrease in precipitation. As a result, the study findings highlights the need for well-informed management and adaptation strategies by showing how modeling streamflow response under various climate scenarios can offer crucial insights into how rising temperatures and decreasing precipitation are likely to reduce water availability in the LRC.