Predicting and Positioning for Hurricanes

June 17th, 2005

Posted by: admin

Greetings! Since this is my first post to Prometheus I will introduce myself: I am a third-year graduate student of Roger Pielke Jr. working toward a MS in policy and meteorology and an MBA, both at the University of Colorado.

Following the recent Prometheus posts on Hurricanes (here and here), I want to bring up another issue that involves hurricane forecasting. In 1999, Roger Pielke Jr. wrote this article in Science which points out the differences between improving hurricane track forecasts and translating this improved forecast into measurable benefits for emergency managers and other decision makers and stakeholders. The gist of the 1999 article is that hurricane track forecasts since 1970 improved at the rate of about 1% per year while the length of coastline warned per storm increased from about 300 nautical miles (nm) in the late 60’s to about 400 nm during the 1990’s. Basically, the science of prediction improved, but the science and art of positioning government agencies and the public for hurricanes did not improve, at least by the metric of ‘miles of coastline warned’.

Fast-forward to 2005, and it seems like both the improved prediction and stagnant positioning trends are the same as they were in 1999. An article in the May 2005 issue of the Bulletin of the American Meteorological Society (BAMS) points to the success of the U.S. Weather Research Program’s (USWRP) goal of “…tropical cyclone track forecast guidance products with an improvement in accuracy of 20%…” The author of this article points out that the 20% improvement goal was “challenging”, and as I write on page 26 of my undergraduate thesis, specific and challenging goals are proven to lead to more successful results than general, or “do our best” goals. In this vein, I applaud the research effort that lead to the successful completion of bettering hurricane track forecast models by 20%.


I find it hard to believe that the recent BAMS article did not once mention how the improvement in track forecasting has impacted other societal metrics (miles of coastline warned, number of people evacuated, public/emergency manager confidence in forecasts etc). Although the author pointed out the short time needed in this project to transfer knowledge from research findings into the operational hurricane models, there is no mention of advancements in the societal positioning for hurricanes as a result of the improving trend in track prediction.

The field of meteorology has and will continue to suffer from a positioning problem. While the state of the science is improving and prediction errors are decreasing, large-scale research efforts generally still do not address transitioning improvements in the science to improvements for the public at large. Perhaps we can simply assume that more accurate hurricane track models will directly lead to improvements in the public warning system, but a September 2003 BAMS article shows that official hurricane track forecast errors are larger when watches and warnings are in effect than at times when there are no watches and warning (i.e. as the hurricane nears land and people, track forecast errors increase).

All of this points to the need for a more systematic focus on how meteorology influences and helps the public. The science of weather will undoubtedly show continuous improvement as it has for the past 50 years, but will the science be able to position itself to make the most of these improvements and help the nation (and the world) use more accurate meteorological information to save lives and money? I argue that meteorological science will help the nation most and will receive more positive attention if it focuses on positioning as well as predicting.

One Response to “Predicting and Positioning for Hurricanes”

  1. Russ Elsberry Says:

    A few comments:
    1) I was a little misled by your use of the word positioning in the title until I read the article. In the tropical cyclone world, positioning refers to the determination of the location of the circulation center of the cyclone.
    2) The BAMS article reported on the achievement of a research goal, and only pointed out that because of a program facilitating the transition of research to operations called the Joint Hurricane Testbed, it led to an advance in forecasting skill as well. The demonstration that this research and forecasting achievement may have led to a measurable improvement in societal response was not a part of the USWRP Hurricane Landfall program (however, your advisor certainly advocated at the beginning of the Hurricane Landfall program that such a societal impacts program be included–we never got the funding to incorporate that aspect). Whereas perhaps $43 B in damages might be attributed to the four hurricanes that struck FL during the 2004 season, I can only assume that the improved skill in forecasting contributed to reductions in deaths (and
    perhaps damages) via better warnings.
    3) I believe you have misinterpreted the Franklin et al. (2003) article in BAMS. They show that the rate of improvement over a period of time for the forecasts when the storm is threatening landfall may be slightly smaller than the overall rate of improvement. I say may be because the number of such landfall-threatening forecasts is small and varies greatly from year-to-year, especially at 72 h since some years have no events, so that constructing trend lines is risky. I agree with the authors that no reason exists from the
    landfall-threatening forecasts to be worse given that aircraft reconnaisance is continuous in those situations. The rate of improvement for forecasts over open ocean can be larger because the errors have traditionally been larger over those regions owing to a lack of data.
    4) I agree that a more systematic focus is needed on how improved forecasts can help the public. Perhaps if the USWRP Hurricane Landfall program had been able to put more focus on societal effects the program would not have been terminated.

    Russ Elsberry (formerly Science Coordinator, USWRP Hurricane Landfall program)