A Service of Energy CentralEnergyBlogs.com Logo

By Ria Persad, CEO of StatWeather

Have you ever wondered how accurate the hurricane forecasts are that you see in the media and on television? A meteorologist can have a very plausible-sounding scientific explanation to their forecast, but what really counts to the general public is, “How accurate is this? What’s your track record? Should I believe you?”

This article tracks the accuracy of pre-season tropical and hurricane forecast numbers.  For example, when a forecaster says that it will be an “above average” or “very active” hurricane season and gives you the expected number of tropical storms, hurricanes, and major hurricanes for the upcoming season, how far off are these numbers compared to what actually transpired?

I turned to the Reference source page data from Wikipedia’s entries on “2013 Atlantic Hurricane Season”, “2012 Atlantic Hurricane Season”, and so on, assimilating 10 years’ worth of hurricane forecasts from a total of 7 weather forecasters:  NOAA, Colorado State University, TSR at University College London, North Carolina State University, UK Met Office, Florida State University, and WSI / The Weather Company.

When a weather forecaster presented a range of values (e.g., NOAA predicts 8-10 hurricanes), I took the average of the range in my error analysis.  I also used the forecasts that were released closest to June 1st, which is the official start of the Atlantic hurricane season and also the time when many industries are making preparations for the upcoming season.

The table below reflects how far off forecasters have been in predicting the numbers of tropical storms, hurricanes, and major hurricanes over the last 10 years (2004 to 2013).  For example, the table shows that on average, over the last 10 years, NOAA’s Forecasted Count for Tropical Storms was off by plus or minus 4.

In the table, the 50-Year Average and the 30-Year Average are tracked as if they were “forecasters.”  In other words, if you just went with a 50-Year Average or a 30-Year Average, would it have been closer to what actually happened, as opposed to going with a “real” weather forecaster? Smaller numbers in the table represent greater accuracy.

2004-2013 ERROR ANALYSIS (10-YEAR RECORD)

NOAA

CO State Univ.

Consensus of 7 Forecasters

50-Year Average (1950-2000)

30-Year Average (1981-2010)

Forecasted Number of TROPICAL STORMS (Average Absolute Error)

4

4.1

3.8

5.9

5.2

Forecasted Number of HURRICANES (Average Absolute Error)

3.2

3.3

3.1

3.3

3.2

Forecasted Number of MAJOR HURRICANES (Average Absolute Error)

1.8

1.8

1.8

1.9

1.9

The major takeaways from the data are

1)   The weather forecasters above are better than going with a climatological average when predicting the number of Tropical Storms (winds up to 73 mph).

2)   The weather forecasters above are about the same as going with a climatological average when predicting the number of Hurricanes (winds above 74 mph, Category 1 or 2).

3)   The weather forecasters above are about the same as going with a climatological average when predicting the number of Major Hurricanes (winds above 111 mph, Category 3, 4, or 5).

A consensus of forecasters can be slightly better than going with one forecaster alone.  But if the consensus are all thinking along the same lines, the advantage is very small indeed.

What I have learned from this data is that conventional hurricane forecasting science has gotten us “ahead of the curve” when anticipating the number of tropical storms in a given season, with clear skillfulness.

These methods do not get us “ahead of the curve” when it comes to predicting what proportion of tropical storms will then develop into hurricanes and major hurricanes.  For example, if you were to go with a hurricane climate average instead of the weather forecasters above, you’d have a 50-50 chance of beating the weather forecaster.  According to my battery of tests, weather forecasters were better than hurricane climate averages for 5 out of the past 10 years.  It is accurate to say that when it comes to predicting whether a season will have more or less hurricanes, your guess is as good as the weather forecaster’s.

The science of hurricane development and intensification has been tackled by organizations who use non-traditional methods of statistics and machine learning, astrophysics, dynamical systems, and physical oceanography. My company, StatWeather, employs some of these non-traditional methods which have, thus far, “beat the curve” at the expense of going against the grain and being named contrarian. In 2013, StatWeather publicly forecasted a below normal hurricane season, a minority viewpoint, which verified.  In 2012, StatWeather forecasted a strong hurricane season, again a minority viewpoint, which verified. In 2014, StatWeather is forecasting an active hurricane season, which is again regarded as contrarian.

Many 2014 hurricane forecasts are heavily dependent upon El Niño development.  Over the last 10 years, El Niño episodes have coincided with hurricane activity ranging from above normal to below normal activity.  StatWeather’s position is that El Niño is but one ingredient in the pie, and is not always the overriding component.

If you would like to experience long-range weather forecasts that will get you ahead of the curve, visit www.statweather.com or email service@statweather.com.

171 Views Comments 0 Comments Comments Add Comment Author BioAuthor Bio
ReportReport This Post as Foul/Inappropriate

On forecaster confidence and accuracy... As the head of a weather company, StatWeather, I often get the question, "How confident are you in that forecast?"  Techniques for determ...

Most energy professionals receive weather information stemming from government models.  The two government numerical weather prediction models most commonly employed in the U.S. energy markets...

The Quest for Speed and Accuracy - Being an energy trading floor meteorologist is one of the toughest jobs I know. I field many questions from people who would like to know exactly what is invo...

October is National Women in Business Month, and I would like to dedicate this article to Women in Energy, addressing the issues women face and how we can encourage women in this field. Energy...

I would like to take this opportunity to promote the following Risk Management and Energy Trading Course with a focus upon evaluating Weather Risk. http://www.marcusevansassets.com/HTMLEmail/VR...

Author:  Ria Persad - CEO, StatWeather   As the president of a weather company, I have compiled this list of questions that prospective customers have asked us to fulfill at some...

Author:  Ria Persad, CEO - StatWeather  (BIO) About 6 years ago, I traded commodities using my own money, because I was sick of losing money by entrusting it to 401(k) plans and broke...

Article by Ria Persad, President of StatWeather (Bio) Over the last year, I conducted a poll of approximately 3,500 weather and climate scientists on the matter of climate change. Just to clari...

Author:  Ria Persad, President of StatWeather  (Bio) When I worked as an Energy Trading Floor meteorologist, there was a period of time every morning when we would evaluate different...

Author:  Ria Persad, President of StatWeather - Bio In 1997, I had heard that there was a major shipwreck on an oil and gas exploration survey because of unanticipated weather conditions,...

Article by Ria Persad, President of StatWeather In 1997, it was challenging to forecast bad weather in the North Sea for oil and gas operations 5 days in advance. We looked at satellite al...

Article by:  Ria Persad, President, StatWeather  Occasionally I hear a long-range or seasonal forecaster say, “We are better than climate normals [climatology] 80% of the time....

Article By:  Ria Persad, President, StatWeather <click for bio> Whether you are forecasting the weather, stocks, elections, or sports, the question of when to go with a con...

My name is Ria Persad, and I am the President of StatWeather.  For the last 13 years, I have been tracking forecaster accuracy--both publicly available weather forecasters and private vendors....

 
Toolbox

Blog Editor
Search
Calendar
Recent EntriesRecent Entries
Recent CommentsRecent Comments
RSS


Sponsored Content

Copyright © 1996-2014 by CyberTech, Inc. All rights reserved.
Energy Central ® is a registered trademark of CyberTech, Incorporated.
CyberTech does not warrant that the information or services of Energy Central will meet any specific requirements; nor will it be error free or uninterrupted; nor shall CyberTech be liable for any indirect, incidental or consequential damages (including lost data, information or profits) sustained or incurred in connection with the use of, operation of, or inability to use Energy Central.
2821 S. Parker Rd. Ste 1105 Aurora, CO 80014
Contact: Phone - 303-782-5510 Fax - 303-782-5331 or service@energycentral.com.