An Empirical Study of TRMM-LBA Rainfall

The Measurement of rainfall is always one of the most difficult challenges in meteorology. A substantial effort was made to acquire the vast amount of valuable rainfall data during Tropical Rainfall Measuring Mission - Large Scale Biosphere Atmosphere in Amazonia (TRMM-LBA). These high-resolution data are essential for the rainfall and rainfall-related studies. However, the data quality sometimes is not completely satisfactory. The rainfall data from tipping bucket gaingauges in TRMM-LBA is not an exception. In this study, the problems of raingauge data and their possible causes are discussed. The quality-controlled data are processed into one-minute rain rates. These one-minute data are used in this study to provide rainfall statistical properties.

Statistics of rain rates from 40 gauges during the period between late December 1998 and middle March 1999 has been calculated. A typical rain event lasts 41.3 minutes and results in 4.35 mm of rain per gauge with the maximum rain rate of 20.61 mm/hr. The percentage of time raining is 5.41% of the total observation time. The average one-minute non-zero rain rate is 6.91 mm/hr. The empirical distribution shows rain rates that are equal to or below 2 mm/hr account for 40.10% of the time raining, but only 6.36% total rainfall. On the other hand, rain rates greater than 10 mm/hr are observed 15.07% of the time raining, but account for 66.43 % of total rainfall. The main feature of rain rate distribution is positive-skewed. The median is larger than the mode but smaller than the mean. The lognormal and other theoretical distributions are selected to describe rain rates based on the Kolmogorov-Smirmov test (Atlas et al.; 1990,Wilks, 1995)

The distribution of rain intensities over the day shows a striking diurnal cycle with maxima in the early afternoon, evening and midnight This diurnal variation is manifested largely by its principal rainfall forcing mechanisms. Diurnally, most rainfall occurs in the afternoon and evening, which is consistent with the development of isolated convection due to dynamical and thermodynamic forcing (Garreaud and Wallace, 1997). The maximum entropy spectrum analysis is employed to reveal the dominant diurnal and other cycles.

The results of this empirical study will be presented.