Modelling the impacts of global multi-scale climatic drivers on hydro-climatic extremes (1901-2014) over the Congo basin

Abstract

The knowledge of interactions between oceanic and atmospheric processes and associated influence on drought episodes is a key step toward designing robust measure that could support government and institutional measures for drought preparedness to promote region-specific drought risk-management policy solutions. This has become necessary for the Congo basin where the preponderance of evidence from few case studies shows long-term drying and hydro-climatic extremes attributed to perturbations of the nearby oceans. In this study, statistical relationships are developed between observed standardised precipitation index (SPI) and global sea surface temperature using principal component analysis as a regularization tool prior to the implementation of a canonical scheme. The connectivity between SPI patterns and global ocean-atmosphere phenomena was thereafter examined using the output from this scheme in a predictive framework based on non-linear autoregressive standard neural network. The Congo basin is shown to have been characterized by persistent and severe multi-year droughts during the earlier (1901–1930) and latter (1991–2014) decades of the last century. The impacts of these droughts were extensive affecting more than 50% of the basin between 1901 and 1930 and about 40% during the 1994–2006 period. Analysis of the latest decades (1994–2014) shows that relative to the two climatological periods between 1931 and 1990, the Congo basin has somewhat become drier. This likely contributed to the observed change in the hydrological regimes of the Congo river (after 1994) as indicated by the relationship between SPI and runoff index (r = 0.69 and 0.64 for 1931–1990 and 1961–1990 periods, respectively as opposed to r = 0.38 for 1991–2010 period). Pacific ENSO influences large departures in precipitation (r = 0.89) but prediction skill metrics demonstrate that multi-scale ocean-atmosphere phenomena (R2 = 84%, 78%, and 77% for QBO, AMO, and ENSO, respectively) significantly impact on hydro-climatic extremes, especially droughts over the Congo basin.