Histogram matching of ASMR-E and TMI brightness temperatures

A microwave enhanced version of the Statistical Hurricane Intensity Prediction Scheme (SHIPS) is trained on a passive microwave sample consisting of tropical cyclone overpasses from SSM/I and TMI instruments utilizing the 19.35 and 85.5 GHz channels. To increase sample size and real time data availability, AMSR-E data will be applied in the future. The ASMR-E instrument has 18.4 and 89.0 GHz channels in place of the 19.35 and 85.5 GHz. While the physical interpretation of these channels is the same across all sensors, brightness temperature differences on the order of 10 K exists between AMSR-E and TMI over low precipitation regions. In order to use AMSR-E data in a model trained from 19 and 85 GHz data, the AMSR-E 18 and 89 GHz brightness temperatures are adjusted to simulate 19 and 85 GHz brightness temperatures. A histogram-matching technique is used to match AMSR-E brightness temperatures to a baseline histogram derived from TMI brightness temperatures. The comparison histograms are computed from 10 TMI and AMSR-E tropical cyclone overpasses that occur within 2 h of each other. A best-fit linear model is computed by matching brightness temperature pairs where the cumulative histogram probability is equal. The best-fit coefficients are then used to simulate 19 and 85 GHz brightness temperatures from AMSR-E. Initial results indicate an accuracy to within ± 1 K, which is within the accuracy needed to compute microwave predictors for use in SHIPS-MI.