资源环境科技发展态势分析平台

Daily precipitation concentration in the Amazon basin (AB) is characterized using concentration index (CI), which is computed from HYBAM Observed Precipitation (HOP) data set, for 1980-2009 period. The ability of four satellite precipitation data sets (TMPA V7, TMPA RT, CMORPH and PERSIANN) to estimate CI is evaluated for 2001-2009 period. Our findings provide new information about the spatial irregularity of daily rainfall distribution over the AB. In addition, the spatial distribution of CI values is not completely explained by rainfall seasonality, which highlights the influence of different weather systems over the AB. The results of rainfall concentration indicate that the distribution of daily rainfall is more regular over northwest (northern Peru) and central Andes. Conversely, Roraima region and a large area of Bolivian Amazon register the highest irregularity in the daily rainfall. Bolivian Amazon also represents regions where the large percentage of total rainfall arises from extreme events (>90th percentile). Heavy rainfall episodes over Roraima region are induced by humidity influx come from Caribbean region, while heavy rainfall events over Bolivian Amazon and Andes region are induced by the northwards propagation of cold and dry air along both sides of Andes Mountains, but only propagate in all tropospheric levels for the Andes. The results also show that PERSIANN and TMPA7 data sets better estimates the daily precipitation concentration for whole AB, but with a relative error 8%. CI estimated from satellites does not agree well with HOP over the Andes and northern Peruvian Amazon. On the other hand, the temporal variability of CI can partly be detected using CMORPH and TMPAV7 data sets over the Peruvian Andes, and central and southern Brazil. Errors in CI estimating might be related to inaccurate estimation of daily rainfall. Finally, we conclude that satellite-based precipitation data sets are useful for analysing rainfall concentration in some regions of AB.