GOES Satellite Data in Rainfall Estimation

Geostationary Operational Environmental Satellite (GOES) visible and thermal infrared data, in conjunction with the Griffith‐Woodley method, are used to estimate the rainfall in Florida. Both the life history techniques (LHT) and streamline techniques (SLT) of the Griffith‐Woodley method are studied. The results show that the satellite data are useful for estimating both the volume and areal distribution of rainfall, which are very difficult to assess using the conventional rain‐gauge method. The satellite rainfall estimation is well correlated with the gauge rainfall measurement. The coefficients of determination $(\mathrm{r2}⊐)$ range from 0.76 to 0.82 in the LHT and from 0.56 to 0.60 in the SLT. The paired t‐test is able to differentiate the performance difference between the LHT and SLT, and the rainfall estimated by the LHT was significantly higher than that by the SLT. The LHT appears to be more suitable for estimating rainfall under Florida conditions, mainly because of the rainfall increment rate between the satellite estimation and gauge measurements is near 1.0 in the LHT but not in the SLT. A further study is strongly recommended concerning the possibility of improving the SLT application.