These databases are the culmination of eight-plus years of research. I used a wide variety of sources to compile this data. The Atlantic Basin data comes primarily from National Weather Service archives as well as data reconstructed by NOAA’s Hurricane Research Division. Prior to HRD’s reanalysis project, sea level pressure data in NOAA’s original HURDAT database was sparse. Historical issues of the Monthly Weather Review proved invaluable in reconstructing this data. While the data provided by the MWR is far from complete, it offers a great deal more than what was given in the original HURDAT. HRD’s reanalysis has filled in a lot of the blanks, uncovering previously unknown data and, when necessary, using mathematical algorithms to approximate minimum SLPs not directly measured. The official Atlantic database dates back to 1851, however my full database dates to 1871 because prior to that point, data becomes progressively more sparse. Prior to 1871, some storms are entered as just a single point. The US Hurricanes database extends all the way back, however.
The Eastern Pacific data also comes from NOAA sources but is not subject to any reanalysis, and therefore subject to a larger margin of error. I’ve used the MWR to help improve the data, however SLP data is still very sparse. Western Pacific data is far more complete because, as in the Atlantic, the United States Air Force conducted reconnaissance missions into storms from airbases in Guam until 1988 (and in recent years, Japan has conducted limited recon missions). This wasn’t done in the Eastern Pacific until very recently, and even now only on a limited basis. Multiple different agencies forecast the Western Pacific. Up until 2000, the American Joint Typhoon Warning Center was the primary warnings office for the Pacific Ocean west of the dateline. These duties are now controlled by the Japan Meteorological Agency. Opinions, and therefore data, sometimes differ between the two offices. Since the end of recon in 1988, the JMA has provided fairly accurate pressure data. JTWC, however, has an extensive archive dating back to 1945.
It should be noted that this research is not complete and is constantly being improved. It is meant to be a handy resource for historical tropical cyclone data. Also, all landfall data prior to 1983 is interpolated using the Best Track and reanalysis data with the exception of all US hurricane landfalls, which are exact. This assumes a straight path between six-hourly points. The quality of this assumption varies and tends to decrease the farther the landfall point is from the nearest six-hourly point. A more detailed study is forthcoming, but typical errors are believed to be ≤5 miles. Errors are much more significant for storms moving parallel to the coast, where the slight wobbles common in tropical cyclones can have a significant effect on landfall location. These wobbles tend to not be as significant for storm tracks perpendicular to the coast. Thus US East Coast storms, for example, have a higher error range than Gulf of Mexico storms. Future efforts will be made to improve this data. Landfall data is not currently provided for the Western Pacific.
This material should not be considered a reiteration of the official record. It is based on my objective analysis of the data available to me and I at times deviated from the official record if I felt there was good reason to do so. Such action is never taken lightly and is only done with significant evidence from multiple reliable sources. I alone am responsible for the content of these databases.
Also, the dates given for each storm’s lifespan reflect storm local time, not GMT. For storms that existed in multiple time zones, the time zone of cyclonegenesis is used. This is done to give a better understanding of when the storm actually existed, reduce temporal inconsistencies, and prevent the user from constantly having to convert from GMT. Under this system, differences in observed local time and recorded time are seldom more than three hours, as opposed to six for GMT, and often times, no time conversion is needed. Changes in GMT conversion with Daylight Savings are also eliminated.
Go To:
Eric's Weather Library 