The Atlantic Hurricane Database Re-analysis Project:
Data access, analysis and visualization

An FY2001 proposal to the
NOAA High Performance Computing and Communication Program






 

Principal Investigator : Dr. Christopher W. Landsea
NOAA/OAR/AOML
Hurricane Research Division
4301 Rickenbacker Causeway
Miami, Florida 33149
chris.landsea@noaa.gov
Phone:  305-361-4357
Fax:    305-361-4402

Co-Principal Investigator : Nirva Morisseau-Leroy, MSCS
DBA/Application Developer
NOAA/OAR/AOML
Hurricane Research Division
4301 Rickenbacker Causeway
Miami, Florida 33149
Tel: (305) 361-4337
Fax: (305) 361-4402
Email: moriss@aoml.noaa.gov

Proposal Theme:  Collaborative, Visualization or Analysis Tools
FY2001:  $55 K NOAA
 

Kristina Katsaros, Director
Atlantic Oceanographic and Meteorological Laboratory
katsaros@aoml.noaa.gov
Hugh E. Willoughby, Director
NOAA Hurricane Research Division
Atlantic Oceanographic and Meteorological Laboratory
willoughby@aoml.noaa.gov


Executive Summary
     Ongoing efforts have produced a revised hurricane database (HURDAT) for the Atlantic basin over the last year.  HURDAT is a crucial, widely used dataset for a large variety of projects:  climatic change studies, seasonal forecasting, risk assessment for county emergency managers, analysis of potential losses for insurance and business interests, intensity forecasting techniques and verification of official and various model predictions of track and intensity. The re-analysis project, supported by NOAA's Office of Global Programs, has been able to extend the digital record back in time as well as to revise more recent years' hurricane archives by utilizing today's analysis techniques with the historic raw data that have been made available.  However, such support did not provide for making the revised hurricane database easily accessible to researchers and the general public.  This proposal remedies this deficiency through creative visualization and analysis tools that would facilitate increased usage of this newly created re-analysis product.  If funded, this project would help to fulfill NOAA's goal to disseminate a portion of its archived data as well as to assist in disaster monitoring.
 
 

Problem Statement and Relevance to NOAA HPCC Program Objectives
     The National Hurricane Center's (NHC's) North Atlantic hurricane database (or HURDAT) is widely utilized for quantitative studies of hurricanes.  The original database of six-hourly positions and intensities were put together in the 1960s in support of the Apollo space program to help provide statistical track forecast guidance for tropical storms and hurricanes (Jarvinen et al. 1984).  In the intervening years, this database has been utilized for a wide variety of uses:  climatic change studies, seasonal forecasting, risk assessment for county emergency managers, analysis of potential losses for insurance and business interests, intensity forecasting techniques and verification of official and various model predictions of track and intensity.  Unfortunately, HURDAT was not designed with all of these uses in mind when it was first put together and not all of them may be appropriate given its original motivation and limitations.

     HURDAT contains many systematic and random errors which need to be corrected.  Additionally, as our understanding of tropical cyclones has developed, analysis techniques have changed over the years at NHC, leading to biases in the historical database that have not been addressed.  Another difficulty in applying the hurricane database to studies concerned with landfalling events is the lack of exact location, time and intensity information at landfall.  Finally, recent efforts into uncovering undocumented historical hurricanes in the late 1800s and early 1900s led by Jose Fernandez-Partagas have greatly increased our knowledge of these past events, which had not been incorporated into the HURDAT database.

     Because of these many issues, a re-analysis of the Atlantic hurricane database has been needed.  The PI has been funded through the NOAA Office of Global Programs to lead a three year program to extend and revise the HURDAT database (Landsea et al. 2000).  Currently, the program is in the middle of its second year and the results of the first year are complete and in final format.  The main accomplishment of this first year is the extension of the database back to 1851 by digitizing and quality-controlling original work by Partagas and Diaz (1995a,b,1996).

     The challenge now that these results from the revised hurricane database are available is getting them to researchers and the general public that would be interested in obtaining them.  While plans from the original NOAA OGP grant were to simply make the data available as an ASCII file via ftp, this proposal instead offers to place the data into a searchable database that would allow for vastly increased possibilities for direct visualization and analysis.  We envision users of the database to be able to both obtain relevant sub-sets of the hurricane data to suit their individual needs as well as to plot tracks of hurricanes stratified by a wide variety of option (time of year, latitude/longitude region, intensity stage, etc.)  Such a project would also be useful for disaster monitoring and response by allowing comparisons of current hurricanes versus historical analogs.
 

Proposed Solution:
     As of early August 2000, the first year of HURDAT re-analysis is complete and approved by the National Hurricane Center (since they have the official charge of maintaining the dataset and must approve any changes/additions).   (Available on the Web at:  <http://www.aoml.noaa.gov/hrd/>). Revised data for 1851-1885 are now available for use.  (Also available is the entire HURDAT dataset from 1851-1999, though the period 1886-1999 has not yet gone through the "re-analysis".)  This presents the opportunity for creative solutions for information sharing on hurricanes. It is proposed

     The actual formatted HURDAT, while quite useful for developed programs such as those historically written in FORTRAN, does not allow for easy access for users in its current form.  Researchers (and the general public) may desire obtaining subsets of the dataset through a front-end typing a particular choice or by selecting from a menu, the subset of interest.  These subsets could include the following:  name, year, month, 10-day period, week, individual day, latitude/longitude box, country, state, county, city, formation location, intensity (tropical depression, tropical storm, hurricane, major hurricane) and decay location. Combinations of the above subsets would be possible as well.  The user could then download the obtained data into a specified file.

     Secondarily, users may wish to use JAVA-based scripts that allow for variable plotting of the tracks of the hurricanes.  It is envisioned that a tool could be developed that would allow for plotting of the tracks stratified into a wide variety of subsets as specified by the user.  Examples of expected stratifications would also be:  name, year, month, 10-day period, week, individual day, latitude/longitude box, country, state, county, city, formation location, intensity (tropical depression, tropical storm, hurricane, major hurricane) and decay location.  The obtained figure could then be downloaded as a .gif or other format file to be utilized directly by the user.
 

Analysis:
     With this proposal, users of hurricane data will have access to a powerful tool to assist with data analysis and visualization.  The ability to quickly stratify the data into useful subsets as well as the opportunity to plot the hurricane track of such subsets will likely greatly enhance the usage of this new dataset.  It is envisioned that not only would basic and applied researchers find good use with this tool, but that it would also enhance educational possibilities with easy to use stratification of the rather voluminous dataset. (The NOAA Hurricane Research Division gets an incredible amount of requests via phone, letter and email from school-children trying to obtain specific information about a hurricane or a group of hurricanes. With this new tool, we should be able to better outreach these meteorologists-in-training.)

     Without the proposal, the re-analyzed HURDAT will still be available to use for researchers.  However, it would be up to the individual investigators to attempt their own stratification and visualization. For the casual users and students, however, the HURDAT will likely remain largely inaccessible for easy use without this project.
 

Milestones and Deliverables

Web-based prototype that allows scientists and the general public to query the HURDAT data stored in an Oracle8I database Management Systems (DBMS). This prototype will be deployed over the AOML intranet.

Budget Summary :

HPCC Contribution:

Contract* Labor:

Through contract to University of Miami's Cooperative Institute for Marine and Atmospheric Science (CIMAS)

Total Requested from HPCC:

HRD Contribution:


References


Administrative Officer: Cathy Steward
Phone: (305) 361-4303
E-mail Address: Steward@aoml.noaa.gov

FMC Number: 940