Hydrological Response to Climate Change and Variability in East Africa

African countries' economic growth and sustainable development is being constrained by their capability to adapt to climate change, in particular regarding water availability and distribution. Hydrological processes exhibit a high degree of temporal and spatial variability, and their modelling is affected by issues of nonlinearity of physical processes, conflicting spatial and temporal scales, and uncertainty in parameter estimates. In addition, conventional hydrometeorological data has long suffered from data breaks due to changes in reporting methods and from gaps (missing information), especially in Africa.

This study focuses on modelling the impact of climate change and variability on the long-term distribution of water-balance components in East Africa through the evaluation of historical patterns and future projections.

The methodology uses an integrated Soil and Water Assessment Tool (SWAT) with a machine learning model to examine historical data, to capture nonlinear hydrological patterns, and to generate accurate projections. The modelling analysis is partitioned into two phases: (1) a land-phase module, where SWAT simulates processes from the event of raindrops onto the land surface to the stream, and (2) a climate-phase module, where Regional Climate Models (RCMs) will be used to produce time series for a set of climatic variables under different scenarios. Machine learning algorithms that closely align with RCM will be used to impact assessments on water resources. Based on this, we aim to predict the impact of drought related to the water resources for sustainable agricultural production potential across the region.

At the EGU General Assembly, we will present the spatio-temporal variability of seasonal water balance in East Africa, and the modelling framework for climate change projection and drought prediction.

 

Acknowledgments:  This research is supported by Eni S.p.A. through the Eni Award “Debut in Research: Young Talents from Africa”. We thank our company tutors Alessandro Nardella and Alessandra Bertoli for their invaluable guidance and technical expertise.