Comparison of Trend Preserving Statistical Downscaling Algorithms Toward an Improved Precipitation Extremes Projection in the Headwaters of Blue Nile River in Ethiopia

Abstract

Projected changes in precipitation extremes can greatly impact the natural environment. Hence, the precipitation extremes must be precisely estimated with an appropriate bias correction algorithm to provide reliable information for the formulation of climate change impact adaptation and mitigation strategies. However, there is a lack of studies that discuss the effect of bias correction algorithms on the reproduction of precipitation extremes in the Blue Nile River Basin. This study compared three commonly used bias correction algorithms: the quantile mapping (QM), detrended QM (DQM), and quantile delta mapping (QDM). The QDM and DQM algorithms outperformed the standard QM bias correction algorithm in preserving the raw climate models projected relative changes of precipitation extremes. The performance differences between the standard QM and other bias correction algorithms (DQM and QDM) were more pronounced in the projection of extreme daily precipitation. Conversely, the projection of dry and wet spells was less sensitive for the choice of the bias correction algorithm. In general, the climate change impact analysis with the QDM algorithm revealed the increase in the frequency and severity of precipitation extremes. Moreover, the results showed the increase (decrease) in the maximum length of dry (wet) spells; indicating the increase in the severity of the meteorological droughts in the future that could potentially reduce the rain-fed agricultural productivity of the region.