Interdependent effects of climate variability and forest cover change on streamflow dynamics: a case study in the Upper Umvoti River Basin, South Africa

Abstract

Streamflow dynamics are sensitive to both climate variability and land use change. However, estimating their separate and combined effects remains a research challenge. In South Africa, streamflow dynamics are important not only for irrigated agriculture but also for many rural communities that depend on streamflow for domestic water supply. In this paper, we analysed the effects of pine, wattle and eucalyptus plantation cover change vis-a-vis the effects of inter-annual climate variability on streamflow dynamics of the Upper Umvoti River in South Africa from 1994 to 2016. We modelled inter-annual variability in streamflow by precipitation, temperature and plantation cover using the Bayesian inference. We mapped plantation cover from Landsat satellite imagery. We found strong evidence for an interaction between temperature range and plantation cover net change on streamflow. Specifically, the plantation effect weakened under conditions of high-temperature range anomalies. We explain this interaction with a shift in soil water repellency and interception capacity within the plantation area under a changing temperature range, with important implications for the formation of surface runoff. Previous studies have assumed that the effects of climate variability and plantation cover change on streamflow dynamics are independent. Our results call this assumption into question. Hence, climate and land cover interdependencies should be accounted for in future statistical and process-based modelling studies.