I. 2.
A COMPARISON OF PASSIVE MICROWAVE MEASUREMENTS TO ESTIMATE RAINFALL
AND ITS 3-D STRUCTURE AS INFERRED VIA RADAR
O. Massambani , A. J. Negri,.J. L. de Oliveira and N. J. Ferreira
The passive microwave technique for the estimation of rainfall have experienced a considerable progress over the last decade due to the ongoing observational studies and the increased understanding of the transfer of microwave radiation through clouds and hydrometeors. Kummerow and Giglio (l994) presents an complete overview of the recent contributions as well as points out the great relevance to account for the vertical structure of the precipitation in order to properly simulate up welling radiance's by means of a radioactive transfer schemes. Simpson et al. 1988, points out also that one of the major objectives of Tropical Rainfall Measuring Mission (TRMM) is the quantitative mapping of the rainfall over the globe and the vertical structure of the rain systems. The objective of this paper is to report on the measurements performed for two events of an isolated convective precipitation system and a widespread precipitation associated to a frontal system as observed via radar over the sub-tropical region of São Paulo-Brazil in order to identify a relationship between the rainfall estimates from the brightness temperature measured from space and the vertical structure of those precipitating systems.
The microwave data used in this study come from the SSM/I instrument on board the Defense Meteorological Satellite Program (DMSP) F11 polar-orbiting satellite. A full description of the SSM/1 may be found in Hollinger et al. (l987). The 85 GHz data was processed using the Goddard Scattering Algorithm (GSCAT) using the relationship between brightness temperature and rainfall rates as proposed by Adler et al. (l991) and Negri et al. (l994). As the São Paulo Weather Radar is located close to the coast and the surveillance area covers part of the continent and part of the ocean, the linear relationships between R (mm/h) and Tb (K) were applied over the continent and over the ocean. To effect a smooth transition in rainfall rate between the ocean an the land, a linear interpolation was used. Instantaneous rain estimates were made for each overpass of the SSM/1 over the region of São Paulo and accumulated in 0. 15 degrees bins which provides a slightly larger than 15 km by 13 km resolution for the 85 GHz channel. From the data collected of the swath of about 1400 km, lt. was selected and remapped each pixel for the radar coverage region.
The São Paulo Weather Radar located at 23.3614 degrees S and 45.118 degrees W, was constructed at McGill University (Canada) and it is in operational use since 1988 performing full volume scans every 10 minutes. Electronic calibration of the receiving system is performed frequently to keep reliable reflectivity data set. The data acquisition system records reflectivity values up to 240 km range from an antenna program of 20 elevations and processes real time data for hydrometeorological nowcasting purposes for the Great São Paulo Metropolitan Region. The events chosen for this study were selected from matching the radar data bank and the SSM/1 data bank and identifying simultaneously observed cases. The observed convective event was recorded on March 07 (IO-.20 LT) and the widespread system was recorded on September 30 (II:IO LT), 1991, Cappi maps at various heights and EchoTop maps were processed from the recorded 3-D at those individual volume scans.
Figure 1a and 1b shows the horizontal distribution of the rainfall intensities as estimated for the convective event. The rainfall distribution at 3, 4, 6 and 8 km CAPPI are shown with the EchoTop map and the SSM/1 rainfall estimates. Frame a shows the 2x2 km radar resolution data and frame b shows the mean averaged rainfall intensities computed for the satellite resolution. These results indicates an excellent agreement in the location of the precipitation field, however shows that SSM/I estimates describes a much smaller area
Fig. 1 - Rainfall maps estimated from 85 GHz brightness temperature of SSM/I data and CAPPI maps at the indicated heights and ECHOTOP map computed for the March 07th event. a) 2x2 m radar resolution b) mean averaged rainfall at the satellite resolution
coverage, being not able to discriminate the very light rain intensities. lt. also can be seen from these figures that the most intensive rainfall core is in good agreement to the highest echo top location. The scatter diagram depicted in Fig. 2a and 2b, shows the averaged rainfall rates for each pixel position at 3 and 4 km height. It indicates that the algorithm is quite powerful to determine the rainfall field. Even though there is a positive trend, it can be seen that the rain intensities estimated from the algorithm based on the brightness temperature indicates a saturated level associated to the 12 mm/h and the radar has determined rain intensities above 35 mm/h. The 4 Km CAPPI level correlation shows a much better adjustment. The labeled points represents different average procedures for the radar data. The mean values seems to shows better adjustment. For the widespread event the results are presented in Fig. 3a and 3b which shows to have the same characteristics as observed for the convective event. The rain pattern shows also a general agreement, but it can be seen that over the eastern radar region the agreement is not as good. However the location of the most intensive cores seems to have being captured by both instruments.
From the construction of a mask based on the satellite rainfall pattern, it was possible to compute the average echo top height for all pixel of given rain intensity as estimated from the SSM/I data. The relationship between those parameters has led to an excellent correlation between them reinforcing the indication of the direct influence of the vertical development of the cloud. lt. was also interesting to observe that both temperature profiles were very similar, mostly above 3 km height as measured from radio-soundings launched at those days. lt. seems to indicate that the brightness temperature at 85 GHz is representing a strong dependence on the type and content of the ice crystals and probably liquid water at each given height. These results shows that the algorithm is capable to capture the major signature of the rain intensity, but in order to account for the light rain lt. should include the signal from lower microwave frequencies. These two cases should be used to simulate the cloud microphysical properties using their cloud vertical structure and the rain distribution to compute the brightness temperature and so allow to inspect if we can efficiently estimate the rainfall intensities using only a unique relationship based on passive microwave radiometry.
A one degree beam width S-band dual-polarization Doppler radar, due to be installed in 1996 by the Meteorological System for Parana (SIMEPAR) in the Parana State, south of Brazil, should contribute to Ground Validation Program of TRMM in addressing some of these aspects.
Fig. 2 - Scatter diagrams for radar rainfall (CAPPIs at 3 and 4 km) and rainfall from the SSM/I data.
Fig. 3 - Rainfall maps estimated from 85 GHz brightness temperature of SSM/I data and CAPPI maps at the indicated heights and ECHOTOP map computed for the Sept 30th event. a) 2x2 km radar resolution b) mean averaged rainfall at the satellite resolution
Fig. 4 - Scatter diagrams for radar rainfall CAPPIs at 3 and 4 km) and rainfall from the SSM/I data.
Fig. 5 - Scatter diagram for EchoTop heights and the SSM/I rainfall instantaneous estimates for the March O7th and Sept. 30 events.
This work was supported by FAPESP and by UCAR during the visit of the first author to the TRMM Office at GSFCINASA.
Adler, R.F., H.-Y.M. Yeh, N. Prasad, W.-K. Tao, and J. Simpson (l991) "Microwave simulation of a tropical rainfall system with a three-dimensional cloud model. J. Appl. Meteor., vol. 30, p. 924-953
Hollinger, J., R. Lo, G. Poe, R. Savage and J. Peirce (l987) "Special Sensor Microwave/lmager user's guide. Naval Research Laboratory, Washington, DC, 177 pp-
Kummerow, C. and Giglio, L. (l994) "A Passive Microwave Technique for Estimating . Rainfall and Vertical Structure Information from Space. Part I: Algorithm Description" J. Appl. Meteor., 33, 3-17
Negri, A.J., R.F. Adler, E.J. Nelkin and G.J.Hufmann (l994) "Regional Rainfall Climatologies Derived from Special Sensor Microwave Imager (SSM/I) Data" Bull. Am. Meteo. Soc., Vol. 75, No. 7, p. 1165-1182