| dc.description.abstract | Base on the decade-long and highly successful Tropical Rainfall Measuring Mission（TRMM）,it is now possible to provide more quantitative comparisons between ground-based radar and the spaceborne TRMM precipitation radar (PR) with greater certainty. Therefore, this research select the significant precipitation events in the southwestern area of Taiwan in 2008, and compare the PR reflectivity data with simultaneous RCCG plan position indicator reflectivity data. In addition, rain gauge data are used to compare the rainfall derived from PR and RCCG separately, and try to improve the rainfall estimation ability of RCCG. In the aspect of accuracy for PR reflectivity, we introduce the NCAR SPOL data as a ground truth to validate the performance of PR retrieval algorithm.
Comparison results suggest that the reflectivity distribution of PR and RCCG are quite close at lower levels (<5km), but the RCCG’s reflectivity is overestimated at higher levels (>5km). Analyzing the probability distribution frequency of reflectivity at different height, it reveals that the distribution pattern of reflectivity is more closer at three kilometer height. Therefore, the data at 3km will be used to analyze and intercompare in this research.
The results of identifying precipitation type for the reflectivity larger than 18 dBZ agree well between PR and RCCG, the average agreement can reach 83%. Besides, for the different type of rainfall over different surface,such as convective,stratiform rain over land and ocean, the PDF of reflectivity at 3km height is quite close between PR and RCCG. It’s worth mentioning,the offset between RCCG and PR is linear-correlated with the reflectivity magnitude,and not a constant traditionally.
In the part of rainfall estimation,using Z-R relationship to estimate rainfall often underestimate the heavy rainfall rate in the past. In this research, we first tune the reflectivity distribution by PR-RCCG statistical relationship. Second, using the PR-tuned RCCG reflectivity estimate rainfall again, and adjust the heavy rainfall part by linear regression with rain gauge in 2008. After the upper two steps, we use two significant precipitation events in different years to verify the method. The results show that it indeed improve the ground-based radar’s ability in rainfall estimation.
| en_US |