Mechanisms that control interannual variations of the wet season onsets over the Amazon and their predictability

Interannual variation of the wet season onsets over the Amazon has a strong impact on the bioproductivity of the rainforests, fire risk, agriculture and hydropower production of the Amazon, and contributes significantly to global atmospheric carbon fluxes. Previous studies have established that the tropical Sea Surface Temperature anomalies (SSTA) have strong influences on the wet season onset and can explain part of its variance. However, few studies have investigated what can explain the rest variance of the interannual variation of the wet season onset and how the relative importance of SSTA and physical processes can impact its predictability.

We identify several physical processes that could influence the wet season onset. A poleward shift of the subtropical jet partly induced by the wave trains generated from the warming of the tropical Pacific could inhibit the cold air incursion and result in late wet season onset. The low level zonal winds over the southern Amazon and the meridional winds in the southwest Atlantic directly affect the moisture transport to the southern Amazon. Higher Convective Inhibition (CIN), induced by a decrease of evapotranspiration (ET) and resulting in a less moist boundary layer, inhibits the occurrence of convection during the late dry season. Some threshold values of these processes in JJA could be used as a simple indicator for whether the following wet season will be delayed or not.

Two statistical methods are employed to investigate the predictability of the onsets in order to avoid that the interconnections of the late dry season conditions dampen the predictability. The predicted onsets have the maximum correlation R=0.81 with the observations and the explained variance is about 66%.