SENSITIVITY OF MONTHLY 3-DIMENSIONAL RADAR-ECHO CHARACTERISTICS TO SAMPLING FREQUENCY

Estimates of any precipitation characteristics based on temporally spa rse observations entail uncertainty because of the natural variability of rainfall in space and time. This study measures the sampling-relat ed uncertainties of monthly mean reflectivity profile and surface rain fall distribution. Radar and rain gauge data collected during the 1993 /94 monsoon season at Darwin, Australia, are used to show the sensitiv ity of monthly three-dimensional radar-echo and precipitation characte ristics to the frequency of observation. The data are partitioned into convective, stratiform, and anvil components according to the horizon tal and vertical structure of the echoes. The analyses of this study r eveal the expected trend that the uncertainties of estimated precipita tion characteristics using infrequent observations scale with rainfall amount. The results have implications for climatological studies usin g spaceborne observation platforms revisiting a given area intermitten tly. The Tropical Rainfall Measuring Mission (TRMM) satellite radar, w hich will revisit a given 500 km by 500 km region approximately twice daily, will likely encounter significant problems in estimating the ve rtical profile of radar reflectivity in the tropics. Monthly mean refl ectivity statistics (based on observations within 150 km of the Darwin radar) exhibit a sampling-related uncertainty of about 20 % in both r ain and snow. In addition, the radar signal of the TRMM satellite will be highly attenuated below the 0 degrees C level, and the precipitati on radar will be insensitive to reflectivity less than about 20 dBZ. T herefore, the spaceborne radar will have an obscured view of the verti cal precipitation structure. Reliable reflectivity statistics based on TRMM satellite radar data may be obtained primarily within an altitud e range of about 5-7.5 km-an altitude range though that is important f or cloud electrification because of the mixed-phase precipitation proc esses taking place there. The sampling uncertainty, signal attenuation , and radar sensitivity vary with precipitation type. Moreover, estima tion of the convective rain fraction will be compromised by uncertaint ies in the echo classification as well as a choice of Z - R relation. These results imply the importance of information collected by ground validation site radars to improve upon TRMM satellite estimates of pre cipitation characteristics and the derived vertical profile of latent heating.