The utility of the ERA40 Cyclone Phase Space in Trend Diagnosis and North Atlantic Tropical Cyclone Reanalysis

As understanding of tropical cyclone (TC) evolution both during and beyond the tropical phase improves, forecasting and analysis techniques are adjusted accordingly. While these changes hopefully lead to more accurate forecasts, they introduce inconsistencies into best track datasets. The tropical cyclone reanalysis project was started in an attempt to remove such inconsistencies and biases (Landsea et al. 2004). It is important to understand biases within best track datasets before studies of long term trends can be meaningfully completed. Here, the strengths and limitations of the ECMWF reanalysis data (ERA40; Uppala 2005) are examined within the cyclone phase space (CPS; Hart 2003).

This process included using the CPS to quantify biases and evolving trends in North Atlantic TC representation within the ERA40. It is found that TCs are poorly resolved even beyond what is expected given the ERA40's grid size. By binning data into three temporal groups it is found that the introduction of satellite data results in a drastic improvement in the representation of ERA40 TCs. Not surprisingly, TC size seems to have the most profound effect since the ERA40 grid size is 1.125° (Uppala 2005). It is also found that location (specifically longitude) also has a marked effect on TC representation within the ERA40, but it should be noted that location is linked to data density.

Despite these inconsistencies, the ERA40 CPS can be used to scrutinize historical structural classification of some TCs, provided that the TCs are represented much better than the mean (usually by more than one standard deviation). Within this study, it is found that some TCs may require refined timing of extratropical transition.

Others may need refined structural classification at the beginning or in the middle of their HURDAT tracks. Finally, there are some TCs for which the ERA40 and HURDAT classifications agree throughout the duration of the TC's lifecycle. Within the study, cases from each of these subsets of TCs are presented and one potential addition to the best track dataset is examined. It should be noted that potential revisions are suggested only when evidence outside of the ERA40 CPS can be obtained since no singular source should be the basis of revisions to the best track dataset.