Climate variability in Malawi, part 1: dry summers, statistical associations and predictability

Abstract

AbstractClimate variability and predictability over tropical southern Africa is studied using a Malawi summer rainfall index. Thermodynamic and circulation patterns are examined and statistical associations are analysed. Links with El Niño–southern oscillation (ENSO) and quasi‐biennial oscillation (QBO) phase are significant and consistent with those found for southern Africa. Correlations are mapped with respect to the Malawi rainfall index, and the regional expression of ENSO emerges quite clearly. Negative r values for sea‐surface temperature (SST) in the west Indian Ocean contrast with positive r values in the east Atlantic and Agulhas region. The thermodynamic pattern ‘anchors’ a standing Rossby wave, comprised of northwest–southeast‐oriented centres of action across the Mozambique Channel and southern Africa, and east of Madagascar to East Africa. A meandering of the sub‐tropical jet stream is an expression of the standing Rossby wave, leading to vertical motions over southern Africa. The pattern develops at least one season in advance, and predictive indices are extracted in the July–September season for use in a multivariate model to predict the Malawi summer rainfall. The three‐area SST index, formulated to capture the ENSO‐modulated Rossby wave pattern, is the most influential predictor, followed by air pressure over the east Indian Ocean and the stratospheric zonal wind anomaly (QBO). Skilful results are found, with a 55% hindcast fit and two‐thirds of tercile categories correctly forecast in independent tests. Copyright © 2002 Royal Meteorological Society.