Evaluation and bias correction of satellite rainfall data for drought monitoring in Indonesia

摘要

The accuracy of three satellite rainfall products (TMPA 3B42RT, CMORPH andPERSIANN) was investigated through comparison with grid cell average groundstation rainfall data in Indonesia, with a focus on their ability to detectpatterns of low rainfall that may lead to drought conditions. Each of thethree products underestimated rainfall in dry season months. The CMORPH andPERSIANN data differed most from ground station data and were also verydifferent from the TMPA 3B42RT data. It proved possible to improve TMPA3B42RT estimates by applying a single empirical bias correction equationthat was uniform in space and time. For the six regions investigated, thisreduced the root mean square error for estimates of dry season rainfalltotals by a mean 9% (from 44 to 40 mm) and for annual totals by 14%(from 77 to 66 mm). The resulting errors represent 10% and 3% of meandry season and annual rainfall, respectively. The accuracy of these biascorrected TMPA 3B42RT data is considered adequate for use in real-timedrought monitoring in Indonesia. Compared to drought monitoring with onlyground stations, this use of satellite-based rainfall estimates offersimportant advantages in terms of accuracy, spatial coverage, timeliness andcost efficiency.