Many studies utilize reanalysis datasets (NCEP/NCAR, ERA40, JRA25) to deduce trends in measures related to tropical cyclone activity on a variety of time scales. However, relatively few studies have actually addressed the robustness of these datasets on a storm-by-storm basis, which is required to say anything about integrated measures on monthly, annual, and decadal time scales. Thus, the main concern is whether the reanalysis datasets add any meaningful intensity information to such calculations as power dissipation.

We show that the intensity representations of tropical cyclones in the current reanalysis datasets are insufficient to deduce trends in integrated quantities, which necessarily reflect the dominance of lifecycle and frequency over intensity (as in the power dissipation calculation). Each respective deconvolution is compared with that from the Best-Track dataset, with and without recommended wind-speed corrections due to observing system changes.

Further, attribution of increasing trends in integrated quantities of TC activity from reanalysis datasets have been correlated and attributed to rising SST and surface temperatures, also from reanalysis products. With the aforementioned lack of robustness in reanalysis TC representation, it is not apparent that documented best-track intensity increases are reflected in the reanalysis trends.