POTENTIAL OF NEAR REAL TIME SATELLITE RAINFALL PRODUCTS IN MONITORING AND PREDICTING GEOHAZARDS IN THE CARIBBEAN

摘要

Hydro-meteorologically related geohazards frequently bring heavy damages to many tropical countries. The effects of these disasters can be devastating to developing countries, particularly small island nations in the tropic. Monitoring and predicting geohazards can save lives, property and help to devise plans to minimise the effects of the disasters. However, lack of meteorological data and other environmental information has made it difficult to predict meteorological information in these countries, particularly, rainfall and moisture conditions. This study examined the use of near real time satellite rainfall and moisture information from Tropical Rainfall Measuring Mission (TRMM) in monitoring and predicting geohazards in the island of Trinidad, in the Caribbean. As there was no well distributed rainfall measuring network or systematic collection of rainfall and flood records, inventory of floods and disasters were extracted from newspaper articles. In order to examine the association between rainfall and floods, satellite rainfall data were collected for before and after floods. The data were used in calculating the average intensity and duration of the rainfall events resulting in floods. Pre-processing steps of rainfall data were performed in ArcGIS using the Spatial analyst tool while statistical analysis of a non-linear regression was performed in SPSS 17 to determine the power law function for intensity/duration. The results showed an intensity/duration threshold for flood was I=4.06D-°267 where I is in mm/hr and D is rainfall in hours. This threshold can be useful in monitoring and predicting geohazards in Trinidad. A similar approach can be used in other countries affected by hydro-meteorologically related geohazards such as floods. Despite calibration difficulties of TRMM data due to the absence of sufficient and reliable ground truth data, TRMM products such as TMPA and V3B42 show high potential for geohazard monitoring in developing countries where there is no quality rainfall data available. More research is needed in examining satellite rainfall data calibration with good quality ground data supported by other environmental satellite data.