Ryan Truchelut ( ret08@fsu.edu ) and Robert E. Hart

1. Ryan Truchelut (ret08@fsu.edu) and Robert E. Hart Florida State University, Department of Earth, Ocean, and Atmospheric Science Results: Ways Reanalysis Can Be Used to Improve TC Climatology Background Prior to the advent of the satellite era in roughly 1966, an indefinite number of tropical cyclones (TCs) remained undetected in tropical basins globally. This bias has led to difficulties in interpreting long-term trends in TC activity. Earlier research identified previously unknown Atlantic Basin potential cyclones in the pre-satellite era (Truchelut and Hart 2011) through compositing the mean thermodynamic structure of historical TCs in the NOAA/CIRES 20th Century Reanalysis (Compo et al., 2011) and identifying similar signatures in the reanalysis that did not correspond to known cyclones. Synoptic verification using historical surface observations of wind speed and sea level pressure showed the technique effectively identified around 1.5 such cases per year for the 1951-1958 Atlantic Basin hurricane seasons. Methodology This research expands the scope of the original fully manual methodology by developing a filtering algorithm that dramatically improves the efficiency and speed with which candidate events are identified. The process of the algorithm, shown below, incorporates key synoptic fields from the reanalysis to discriminate plausible “missing TC” event candidates from the set of all reanalysis points. This scheme was subsequently applied to the Atlantic, Pacific, and Indian Ocean Basins for the years 1891-1979. Methodology, continued The points not filtered by the algorithm were extended into tracks of discrete candidate events. Historical synoptic data from the ICOADS database (Woodruff et al, 2011) were mapped each six hours for each event. An example is below. The wind speed, direction, and sea level pressure data were used to classify each event. Further results are shown at left. • Use 1: Finding Missing TCs • Tracks for individual candidate events can be constructed from reanalysis data • Using the ICOADS ship report database, these events can be classified using NHC criteria (Landsea et al., 2008) by their likelihood of being a “missing TC” • So far, 1951-1966 Atlantic seasons have been observationally verified, with about 1.25 “missing TCs” per year located • Results shared with HURDAT project team Global Identification of Previously Undetected Tropical Cyclone Candidates in NOAA/CIRES 20th Century Reanalysis Data Use 2: Spatial Patterns of Candidate Events Prior to Satellite Era Summary and References The algorithm demonstrated success in locating credible candidate events in the Atlantic and other basins, as measured by observational verification. While limited by the resolution of the reanalysis, the algorithm is able to quickly identify potential missing TCs with an accuracy nearing that of the manual technique. These findings suggest reanalyses are a useful tool to efficiently add new information to the TC climatological record and are a promising basis upon which to improve our understanding of long-term trends in TC activity. Compo, G. P. et al, 2009: The Twentieth Century Reanalysis Project. QJRMS, Landsea, C. W., et al 2008: A Reanalysis of the 1911-20 Atlantic Hurricane Database. J.Climate. Truchelut, R. E., and R. E. Hart (2011), Quantifying the possible existence of undocumented Atlantic warm‐core cyclones in NOAA/CIRES 20th Century Reanalysis data, Geophys. Res. Lett.,38 Woodruff, S.D., et al (2011), ICOADS Release 2.5: Extensions and enhancements to the surface marine meteorological archive, Int. J. Climatol., • Map plots the number of observationally unverified candidate events per 2° reanalysis gridpoint • Though algorithm filters known TCs, candidate events are clustered in favored development regions • Closely resembles long-term TC climatology both spatially and in basin-relative frequency, increasing confidence that the algorithm can find credible candidate events globally from reanalysis datasets • Use 3: Pre-Satellite Era Trends • Breaking down the number of candidate events per year by basin reveals patterns that add to our understanding of TC trends • In general, number of candidate events is roughly in line with each basin’s: • Relative level of historical activity • Observational density and density changes in the period of study • Length of the climatological record (varies by basin) • Prior efforts to revise climatology