Short Duration Rainfall Estimation Using GMS IR and VIS Images
Flaviana D.Hilario
Ph.D.
CAB,PAGASA,DOST
1424 Quezon Ave., Quezon City, Phil.
Tel/Fax: 373 34 33
E-mail: cab@philonline.com.ph
Abstract
Point estimates of hourly, 3-hr, 6-hr and 12-hr rainfall from tropical cyclones were computed using 3-hr interval of infrared and visible images from the Geostationary Meteorological Satellite (GMS) of Japan. Regression models involving liner combination of IR and VIS was used to estimate short duration rainfall. Independent data were used to validate the models.
It was observed that high rainfall values are always associated with high IR brightness value but high IR brightness value is not always associated with high rainfall values. The total time period of the rainfall estimation appears to be an important factor in the accuracy of the estimation. In general, the estimated rainfall for the periods hourly, 3-hr, 6-hr and 12-hr showed overestimation for lower rainfall values and underestimation for higher rainfall values.
Introduction
Rainfall measurement is very important in many aspects of both operational and research meterology, climate monitoring, hydrology, and ecology. Unfortunately, the present network for rainfall monitoring by conventional means such as raingauges is deficient in many areas in the world, the more so if rainfall data are required very quickly. Many raingauge network provide data monthly or even less frequently, and with time-lags which further restrict their utility particularly for near real time applications. Satellite-derived rainfall estimates may supplement these data or, in some cases, it may be the only data available.
Areal average and volumetric rainfall are helpful in qualitatively determining how "wet" a tropical cyclone is compared to other tropical cyclones but they do not give the forecaster what is really needed; namely, a spatial distribution of rainfall. For this type of information, point rainfall estimation is needed. Techniques that fall in this category aare those of Robertson (1995), Griffith-Woodley Rainfall Technique(GWRT), GOES Precipitation Index (GPI) developed by Arkin and Meisner(1987), Abe(1990), and Goodman, Menzel and Curtin Method (1991). The main objective of this study is to develop a short-duration point-rainfall estimation technique using visible and infrared images from GMS
Data and Methodology
Satellite Data: the data set used in this study is infrared (10.5-12.5m) and the visible (.55-.75m) images from the Japanese Geostationary Meteorological Satellite (GMS). Figures 1 &2 show sample of infrared and visible images used in the study. Four visible images at 3 hour interval (0031, 0331, 0631, 0931) for each tropical cyclones were initially planned to be utilized in the study. However, only T.C. Ditang has 4 visible images while T.C. Akang has 3. The resolution which is finer than the IR images is 1.25 km at sub-satellite point. The amount of solar visible radiation reflected by the clouds depends on the position of the sun in the sky, the latitude of the station and the time of the year. To take these factors in consideration, the visible brightness values were normalized (Negri and Adler, 1987) by multiplying with 1/SQT(cosz), where z is the solar zenith angel.
One of the satellite parameters used in this study is the IRBV at the rainfall station. The other satellite parameters were average brightness value (AveBV) of different grid size ranging from 3x3 to 21x21 where the rainfall station is located at the center of each grid. For example , for a 5x5 grid, the average brightness value of the 5x5 grid(AveBV5) was computed by adding the brightness values inside the grid and then, divided by the total number of pixels, which is 25 in the case. Similar procedure was followed for the other grid size.
Another satellite parameter tested in the study is the variance (VAR) for each grid size. Variance which is the square of the standard deviation is a measure of the texture of the image. Again, similar procedure to that of the brightness value was used to determine the VAR for the other grid sizes.
