Incident Angle Dependency of the Normalized Surface Cross Section during TRMM End of Mission Experiment

TRMM satellite implemented special 90-degree yaw operation during the descent toward the Earth in November 2014 as one of the TRMM end of mission (EOM) experiment (Takahashi et al. 2016). This experiment was aimed to observe a single target from various incident angles from nadir to 28 degrees. During this experiment total 880 minutes (that corresponds to 88000 scans) of data were obtained mainly over the south eastern Pacific. In this experiment, incident angle dependencies of various targets were obtained from a single target (e.g. 5-km-diameter area). It means that almost 88000 sets of data were obtained. Targets are the normalized radar cross section (NRCS) of ocean and land, brightband and incident angle dependency of rain retrieval algorithm. On the incident angle dependencies of the sea/land surface echo, data (NRCS vs. incident angle of each scan) are sorted by the echo power of the fixed incident angle (e.g. 16 degrees). The result shows that the higher NRCS at nadir direction corresponds to the weaker echo at the wider angle over ocean, then these characteristics can be expressed by the slope and intercept in NRCS-incident angle domain. Similar relationship cannot be seen the cases over land.

This angle bin dependency assumed to be caused by the sea surface roughness. The surface roughness changes by the surface wind speed and the surface roughness may be correlate with the microwave brightness temperature of the surface (it is appeared as the system noise). The observed angle bin dependency data (slope and intercept in the NRCS-incident angle domain) are compared with surface wind speed and the system noise. In addition, the effect of the sea surface temperature is evaluated.

First, the “intercept” and “slope” correlate very well each other. The relationship between sea surface wind speed and slope of NRCS against incident angle is correlated negatively. System noise data from TRMM/PR also shows positive correlation with wind speed whereas the SST data almost insensitive with NRCS. These results indicate that the assumption that the sea surface roughness is caused by wind is valid.

Horizontal distribution of low NRCS area is compared with the wind field although the 90-degree yaw experiment covered only limited area. The result shows that horizontal NRCS pattern is similar to the observed wind field pattern.