Global Precipitation Monitoring using PERSIANN System

摘要

Global six-hour 0.25° rainfall estimates generated from the PERSIANN (Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks) system are introduced. The PERSIANN system estimates rainfall based on longwave infrared imagery of clouds from geostationary satellites; parameters of the system are constantly updated when passive microwave-based estimated rain rates from low-orbital satellites are available. Evaluation of PERSIANN estimates using daily TRMM ground validation data over the Melbourne Florida (from September 1998 to January 1999) shows that the correlation coefficient is 0.68, while root mean square error and bias estimates are 5.58 mm/day, and –0.08 mm/day, respectively. Further development of merging PERSIANN estimates with gauge observations is ongoing. Preliminary results of the merging procedure indicate that bias in the PERSIANN estimates can be effectively improved