Homogeneous tropical cyclone intensities from the Community Dvorak Analysis (CoDA) project

Tropical cyclone intensity, distribution and frequency significantly impact life and property. Understanding these aspects of cyclones, however, requires a uniformly constructed dataset. The International Best Track Archive for Climate Stewardship (IBTrACS) project has provided a centralized collection of TC best track data. Despite efforts to minimize discontinuities in IBTrACS, inhomogeneities remain caused by changes in instrumentation, personnel, forecast agencies and operational procedures, which has led to a catalog of storms that is incongruous. To ameliorate this issue, efforts are underway in the North Atlantic to perform a reanalysis of storm intensity, however; no global effort is currently in progress.

NCDC is uniquely positioned to champion a global reanalysis given the datasets and expertise on hand. IBTrACS and HURSAT, a collection of all satellite brightness temperatures from polar and geostationary satellites for the storms in IBTrACS, will serve as the starting point in the development and archive of a homogeneous record of satellite-derived tropical cyclone intensity. Using these resources, we are planning to design an open, community-driven, manual satellite-based reanalysis employing the widely-used Dvorak Techinque to subjectively estimate tropical cyclone intensity for the entire HURSAT period of record (1978-2009).

Rather than employ an objective algorithm, working group members (forecasters, researchers, students, citizen scientists) will be presented with a Java GUI of the Dvorak technique that will allow maximum inter-operability and allow intensity estimates to be made for the 3000+ individual storm images. As each member “classifies” a storm into an intensity category, all selections made during the process are recorded to a database.

The web-based member interface is the key to this project's success. Members, registered under pseudonyms, will be given storm imagery such that it will not be obvious which actual storm is being analyzed. Calibration storms will provide members with a storm of a known intensity their bias will be calculated. Active members will have access to storms they have processed as well as see their results compared to others.

Based on each member's performance against calibration storms, their resulting classifications will be normalized to derive a satellite-based intensity estimate. This project aims to remove the temporal and spatial problems in the current best track record. While it will not replace the current best track, it will enhance the best track record by providing a stepping stone toward a global reanalysis. The result will be a refined TC intensity record that will provide insight into TC variability, intensity, and uncertainty.