Assessing precipitation extremes (1981–2018) and deep convective activity (2002–2018) in the Amazon region with CHIRPS and AMSU data

Abstract

The frequency and spatial distributions of precipitation extremes (PEs) and deep convective clouds (DCC) across the Amazon region were assessed using satellite-derived data. For PEs, CHIRPS dataset for the period 1981–2018 were used to calculate a set of absolute, threshold, duration, and percentile-based threshold indices defined by the Expert Team on Climate Change Detection and Indices. DCC occurrence was assessed based on the Advanced Microwave Sounding Unit data for the period 2002–2018. In northern Amazon (north of $$5^\circ \hbox {S}$$ ) PEs and DCC are more frequent ( $$\ge 60\%$$ frequency) during February–June. Averaged trends over these months have shown increase in daily rainfall above 20 mm of near 3 days over the 1981–2018 period, and an increase of 2 consecutive wet days (P $$\ge 1\,\hbox {mm}$$ ) in the same period. South of $$5^\circ \hbox {S}$$ prevalence of PEs and DCC is largely observed during November–March ( $$\ge 60\%$$ frequency), whereas the longest persistence of dry days is observed during June–August. Though all PE trends point to an intensification of rainfall in November–March, only consecutive dry days in winter (JJA) and spring (SON) show significant trends, pointing to an increase of 7 days over the 38-yr winters. Rainfall extremes over the entire Amazon region were found to be moderate to strongly correlated with the mean vertically integrated moisture divergence, and in southern Amazon also to upper level divergence and upward vertical velocity. Increased frequency of DCC were found over the whole basin ( $$\sim 18\% \,\, \hbox {yr}^{-1}$$ ), in contrast to decreased convective overshooting (up to $$\sim 15.4\% \,\, \hbox {yr}^{-1}$$ ).