The dominant modes of the long rains interannual variability over Tanzania and their oceanic drivers

Mitigating the impacts of droughts and floods over East Africa (EA) may largely be possible through accurate predictions of the long rains (March-May). Though the close interannual variability between the long rains, Indian and Pacific Oceans has been widely studied, Tanzania's long rain variability remains unclear. This paper examines the Tanzanian long rains dominant modes through an empirical orthogonal function (EOF) analysis, and their relationship with winter sea surface temperature (SST) for the period 1951-2018. Correlation and composite analyses were applied to explore the associated mechanisms. The first EOF (EOF1) mode is characterized by negative anomalies over Tanzania. The second EOF (EOF2) mode depicts positive (negative) anomalies over northern (southern) Tanzania. The third EOF (EOF3) mode reveals positive (negative) anomalies over eastern (western) Tanzania. EOF1 is closely related to the SST anomalies (SSTAs) in the southeastern Indian Ocean (SEIO). When the SSTAs are positive, they subsequently linked to upper (lower) level convergence (southeasterly winds) over Tanzania. Associated with this convergence entails a sinking motion, leading to a negative rainfall anomaly. The southeasterly winds indicate an early onset of the southeast monsoon, leading to normal rainfall been recorded. EOF2 is linked to the SSTAs in the central Pacific (CP). Positive SSTAs are related to reversed Walker circulation (WC) over northern Tanzania, while the upper-level convergence was observed around southern Tanzania. This meridional dipole vertical motion pattern enhances (suppresses) rainfall anomalies. EOF3 is associated with mega El Nino-Southern Oscillation (ENSO)-like SSTAs. The SSTAs are related to eastward shift of WC over the Indian Ocean, and zonal cell over the western Indian Ocean with rising limb is located east of Tanzania. The ascending motion enhances rainfall over eastern Tanzania. Predictions based on the SEIO, CP, and mega ENSO index may significantly support long rains prediction efforts and provide early warning.