Patterns of atmospheric circulation linking the positive tropical Indian Ocean dipole and southern African rainfall during summer

Abstract

This study applied eigenvector-based time-decomposition and spatial-decomposition techniques to examine mechanisms through which the positive Indian Ocean Dipole (IOD) sea surface temperature (SST) pattern can influence the patterns of summer rainfall over southern Africa. For the time-decomposition analysis, it was found that over the eastern subcontinent of Africa, south of the equator, positive IOD can induce a meridional dipole rainfall structure through a circulation-type (CT) that its variability can be directly independent of the El Niño Southern Oscillation. Spatial decomposition of austral summer rainfall anomalies also detects a dipole-like pattern of austral summer homogeneous region of rainfall anomaly over the eastern subcontinent of Africa, south of the equator, which can be triggered by the IOD in a delay of about 5 months (with the IOD signal leading). Thus, the signature of the atmospheric signal associated with the SST pattern of the IOD can persist into austral summer and modify austral summer circulation patterns. The large-scale circulation mechanisms through which the positive IOD triggers the meridional rainfall structure are: (i) strengthening of the pressure gradient between the northern and southern domains of the Mozambique Channel and (ii) enhancement of the amplitude of the CT associated with enhanced convergence and convection (i.e., cyclonic circulation) north of Mozambique Channel; thus, disrupting onshore moisture transport from the southwest Indian Ocean into the southeastern subcontinent of Africa where divergence and anticyclonic activity rather dominates. The resulting rainfall pattern is that the northeastern (southeastern) regions are wetter (drier) than normal. Under the shared socio-economic pathway highest emission scenario (ssp585), overall, the analysed climate models indicated that the relative frequency of occurrence of the CT related to the positive IOD is projected to increase during the January and March months.

Appendix

Figure A1

Classified circulation types in Africa south of the equator from ERA5. Nine PCs or modes of variability were retained. For a given retained mode, days with PC loadings above $\left|0.2\right|$, the threshold used to define signals are grouped, resulting in two classes. The CTs are the SLP composites of the days grouped in a given class. The CT related to the positive IOD is marked by the thick black frame.

Figure A2

Classified circulation types in Africa south of the equator from MPI-ESM under the ssp585 scenario. The CT related to the positive IOD is marked by the thick black frame.