The official record of tropical cyclone activity in the Atlantic basin is contained in the Hurricane Best Track file. This database, compiled by Jarvinen et al. (1984) and updated annually by forecasters at the U.S. National Oceanic and Atmospheric Administration (NOAA)'s National Hurricane Center (NHC), covers a very long period of time, from 1886 to present. As such, it spans many changes in observing technology and operational guidelines and policies.

Because of its comprehensiveness, this record should serve as an outstanding source for tropical cyclone climate studies. However, as has been pointed out by several researchers, such long-term databases are prone to non-meteorological problems resulting, for example, from the above-mentioned inconsistencies in technology and policy.

There have been numerous published studies that have relied upon the Hurricane Best Track file. However, problems with this database could significantly reduce its usefulness as a tool in climate research and cast doubt on the findings based upon the dataset. Thus, it is important to identify and explain any discontinuities that are contained in the period of record.

This study investigates trends in tropical storms, hurricanes, and intense hurricanes in the Atlantic basin from 1886 to 2003 that are contained in the Hurricane Best Track file. By employing a range of approaches, including a wavelet analysis, discontinuities in tropical cyclone activity are documented. Each break point is then closely examined and compared against a timeline of technological and operational changes.