Why Ending CSU Seasonal Hurricane Forecasts is a Setback for Tropical Meteorology


Back in late 1994, not too long after tropical storm Gordon dropped a ton of rain on us here in south Florida, I got a fax from my friend, Felix DeOsa, who worked at the National Hurricane Center.

The title of the article – “Outlook for 1995 Season – Bad”

That was my first exposure to Dr. William Gray and the Colorado State University seasonal hurricane outlook. Since then, every season like clockwork, I’ve looked to their forecasts to get an idea of what the upcoming Atlantic season could bring. Every season, the public tunes in at the beginning of the season, the media quotes Dr Gray and Phil, and their numbers become part of the larger public discussion about what could happen later in the year.

Are the Sea Surface Temps going to warm? Is an El Nino coming? Will the QBO be favorable?

Reading and rereading those CSU forecasts has taught me more about seasonal forecasting than any other single source of information over the last 19 years. However, my free information ride is over. Tuesday afternoon, the following message ran across my Twitter feed:

Philip Klotzbach ‏@philklotzbach
CSU 2013 forecast verification posted: http://tropical.atmos.colostate.edu/Forecasts/2013/nov2013/nov2013.pdf
Will be final forecast issued by CSU unless additional funding is forthcoming.

Gray and Klotzbach have had their work funded by private insurance agencies and other sources throughout most of the last 20 years. However, the dearth of Atlantic hurricane landfall activity over the last 8 years has reduced public concern for the hurricane problem. If hurricanes don’t hit the US, funding sources within the US don’t care. I wrote about these challenges for our work in a previous post earlier this fall.

These seasonal forecasts have been very good, overall, explaining more of the year to year variance than climatology alone. However, when the forecasts miss, especially on the low side (when the season is less active than expected) the media and other people, some who don’t seem to understand variability or probability are quick to jump in with how “bad” these forecasts are.

Ending the hurricane seasonal outlooks from CSU will be disruptive to the science of tropical meteorology for three huge reasons:

1. The relationship between El Nino and the Atlantic hurricane season was discovered by Dr. Gray. Given how much of the current year to year hurricane variance in the Atlantic is unexplained, having less resources focused on discovering relationships between tropical weather indicators and seasonal activity could harm our ability to understand these trends in the longer term.

2. Other, perhaps less skillful sources for seasonal hurricane forecasts will dominate the national discussion, further eroding public confidence in the process when results don’t meet the forecasts.

3. Few sources verify their forecasts as stringently as Dr Gray and Phil do. Every year, they closely assess their forecasts and note what worked and what didn’t, release those results to the public, and openly discuss what went right and what was wrong. . Few private weather sources do the same, and none are as detailed and brutally honest as the CSU evaluations.

The biggest critics of these seasonal outlooks do not seem to grasp that we won’t get better at understanding the factors that impact seasonal activity if we don’t actively study them. There is tremendous value in knowing if a season will be active or not…from the insurance and reinsurance industries to homeowners along the coast (and everywhere in between). Meteorologists will have less tools in their bag to discuss the seasonal numbers with their clients and viewers.

However, personally, I am most disappointed by the loss of a valuable tool in my ongoing education of how the tropical atmosphere works. There is already so little data available – having two less really, really smart people working on the hurricane problem will be a setback for the science. Hopefully, some entity will step in and fund this critical work in 2014 – and keep the process alive.


About Michael Watkins

Mike Watkins is the founder of Hurricane Analytics, a private organization specializing in data visualization and predictive analytics, with a special focus on tropical meteorology. They analyze complex meteorological data and communicate that information in easy-to-understand terms, to help clients prepare and anticipate the disruptive impact of Atlantic hurricanes.