Seasonal Hurricane Forecasts

North Atlantic

*The table values for these organizations represent the mean of the forecast range.

Colorado State predicts a slightly above-average hurricane season

A slightly above-average Atlantic hurricane season is on tap for 2012, according to the seasonal hurricane forecast issued August 3rd by Dr. Phil Klotzbach and Dr. Bill Gray of Colorado State University (CSU). The CSU team is calling for 14 named storms, 6 hurricanes, and 2 intense hurricanes, and an Accumulated Cyclone Energy (ACE) of 99, which is average. The forecast calls for an average chance of a major hurricane hitting the U.S., both along the East Coast (28% chance, 31% chance is average) and the Gulf Coast (28% chance, 30% chance is average). The risk of a major hurricane in the Caribbean is also average, at 39% (42% is average.) The CSU teams expects we will have a weak El Niño develop by the peak of this year's hurricane season in September, which will cut down on this year's activity by increasing wind shear over the Tropical Atlantic. However, there is considerable uncertainty in this outlook.

Analogue years

The CSU team picked five previous years when atmospheric and oceanic conditions were similar to what we are seeing this year: years that had weak to moderate El Nino conditions and approximately average tropical Atlantic sea surface temperatures. April - May, followed by August - October periods that were generally characterized by weak El Niño conditions and average tropical Atlantic SSTs . Those four years were 2009, a quiet El Niño year with only 3 hurricanes; 2006, a moderately quiet year with 10 named storms, 5 hurricanes, and no landfalling hurricanes; 1957, A quiet year with 8 named storms, 3 hurricanes, 2 major hurricanes, and major hurricane Audrey made landfall near the Texas-Louisiana border; 1953, a moderately busy year with 14 named storms, 6 hurricanes, and 4 intense hurricanes; and 1951, a low-number but intense year that saw 10 named storms, 8 of which were hurricanes, and 5 of which were major hurricanes.

How accurate are the June forecasts?

The June forecasts by the CSU team between 1998 and 2009 had a skill 19% - 30% higher than a "no-skill" climatology forecast for number of named storms, number of hurricanes, and the ACE index (below). This is a decent amount of skill for a seasonal forecast, and these June forecasts can be useful to businesses such as the insurance industry and oil and gas industry that need to make bets on how active the coming hurricane season will be. Unfortunately, the CSU June 1 forecasts do poorly at forecasting the number of major hurricanes (only 3% skill), and major hurricanes cause 80% - 85% of all hurricane damage (normalized to current population and wealth levels.) This year's June forecast uses a brand new formula tried in 2011 for the first time, so there is no way to evaluate its performance. An Excel spreadsheet of their forecast skill (expressed as a mathematical correlation coefficient) show values from 0.41 to 0.62 for their June forecasts made between 1984 and 2010, which is respectable.

Comparison of the percent improvement over climatology for May and August seasonal hurricane forecasts for the Atlantic from NOAA, CSU and TSR from 1999-2009 (May) and 1998-2009 (August), using the Mean Squared Error. Image credit: Verification of 12 years of NOAA seasonal hurricane forecasts, National Hurricane Center.

Comparison of the percent improvement in mean square error over climatology for seasonal hurricane forecasts for the Atlantic from NOAA, CSU and TSR from 2002-2011, using the Mean Square Skill Score (MSSS). The figure shows the results using two different climatologies: a fixed 50-year (1950 - 1999) climatology, and a 2002 - 2011 climatology. Skill is poor for forecasts issued in December and April, moderate for June forecasts, and good for August forecasts. Image credit: Tropical Storm Risk, Inc.

TSR predicts a near-average hurricane season

The British private forecasting firm Tropical Storm Risk, Inc. (TSR) calls for 14.2 named storms, 6.0 hurricanes, 2.9 intense hurricanes, and an Accumulated Cyclone Energy (ACE) of 106, which is slightly above average. TSR rates their skill level as 23 - 27% higher than a "no-skill" forecast made using climatology, though an independent assessment by the National Hurricane Center (above) gives them somewhat lower skill numbers, using a different metric than TSR uses. TSR predicts a 38% chance that U.S. landfalling activity will be above average, a 31% chance it will be near average, and a 31% chance it will be below average. TSR's two predictors for their statistical model are the forecast July-September trade wind speed over the Caribbean and tropical North Atlantic, and the forecast August-September 2012 sea surface temperatures in the tropical North Atlantic.

TSR projects that 4.2 named storms will hit the U.S., with 1.4 of these being hurricanes. The averages from the 1950-2011 climatology are 3.1 named storms and 1.5 hurricanes. They rate their skill at making these June forecasts for U.S. landfalls at 7 - 11% higher than a "no-skill" forecast made using climatology. In the Lesser Antilles Islands of the Caribbean, TSR projects 1.2 named storms, 0.5 of these being hurricanes. Climatology is 1.1 named storms and 0.5 hurricanes.

FSU predicts a slightly above-average hurricane season: 13 named storms

The Florida State University (FSU) Center for Ocean-Atmospheric Prediction Studies (COAPS) issued their fourth annual Atlantic hurricane season forecast, calling for a 70% probability of 10 - 16 named storms and 5 - 9 hurricanes. The mid-point forecast is for 13 named storms, 7 hurricanes, and an accumulated cyclone energy (ACE) of 122. The scientists use a numerical atmospheric model developed at COAPS to understand seasonal predictability of hurricane activity. The model is one of only a handful of numerical models in the world being used to study seasonal hurricane activity and is different from the statistical methods used by other seasonal hurricane forecasters such as Colorado State, TSR, and PSU (NOAA uses a hybrid statistical-dynamical model technique.) The FSU forecast has been the best one over the past three years, for predicting numbers of Atlantic named storms and hurricanes:

• 2009 prediction: 8 named storms, 4 hurricanes. Actual: 9 named storms, 3 hurricanes
• 2010 prediction: 17 named storms, 10 hurricanes. Actual: 19 named storms, 12 hurricanes
• 2011 prediction: 17 named storms, 9 hurricanes. Actual: 19 named storms, 7 hurricanes

Penn State predicts a near-average hurricane season: 11 named storms

A statistical model by Penn State's Michael Mann and alumnus Michael Kozar is calling for an average Atlantic hurricane season with 11.2 named storms, plus or minus 3.3 storms. Their prediction was made using statistics of how past hurricane seasons have behaved in response to sea surface temperatures (SSTs), the El Niño/La Niña oscillation, the North Atlantic Oscillation (NAO), and other factors. The statistic model assumes that in 2012 the current 0.35°C above average temperatures in the MDR will persist throughout hurricane season, the El Niño phase will be neutral to slightly warm, and the North Atlantic Oscillation (NAO) will be near average.

The PSU team has been making Atlantic hurricane season forecasts since 2007, and these predictions have done pretty well:

• 2007 prediction: 15 named storms, Actual: 15
• 2009 prediction: 12.5, named storms, Actual: 9
• 2010 prediction: 23 named storms, Actual: 19
• 2011 prediction: 16 named storms, Actual: 19

UK Met Office predicts a slightly below-average hurricane season: 10 named storms

The UK Met Office uses a combination of their Glosea4 model and the ECMWF system 4 model to predict seasonal hurricane activity. These dynamical numerical models are predicting a slightly below-average season, with 10 named storms and an ACE index of 90.

NOAA predicts an above average hurricane season: 14.5 named storms

As I discussed in detail in a May 24 blog post, NOAA was calling for 12 named storms, 6 hurricanes, 2 major hurricanes, and an ACE index 102% of normal in May. In August, given that El Nino had not fully developed, NOAA increased the forecast to around 15 storms, 7 hurricanes, and 2-3 intense hurricanes.

East Pacific

NOAA predicts an average Eastern Pacific hurricane season

NOAA's pre-season prediction for the Eastern Pacific hurricane season, issued on May 24, calls for a near-average season, with 12 -18 named storms, 5 - 9 hurricanes, 2 - 5 major hurricanes, and an ACE index 70% - 130% of the median. The mid-point of these ranges gives us a forecast for 15 named storms, 7 hurricanes, and 3.5 major hurricanes, with an ACE index exactly average. The 1981 - 2010 averages for the Eastern Pacific hurricane season are 15 named storms, 8 hurricanes, and 4 major hurricanes. So far in 2012, there have been two named storms. On average, the 2nd storm of the year doesn't form until June 25. We had a record early appearance of the season's second named storm (Bud on May 21.) Bud was also the strongest Eastern Pacific hurricane on record for so early in the year. Records in the Eastern Pacific extend back to 1949.

West Pacific

West Pacific typhoon season forecast not available yet

Dr. Johnny Chan of the City University of Hong Kong issues a seasonal forecast of typhoon season in the Western Pacific, but this forecast is not yet available (as of June 1.) An average typhoon season has 27 named storms and 17 typhoons. Typhoon seasons immediately following a La Niña year typically see higher levels of activity in the South China Sea, especially between months of May and July. Also, the jet stream tends to dip farther south than usual to the south of Japan, helping steer more tropical cyclones towards Japan and Korea. With the formation of Tropical Storm Mawar today east of the Philippines, the Western Pacific is exactly on the usual climatological pace for formation of the season's third storm.

Time series of the annual number of tropical storms and typhoons in the Northwest Pacific from 1960 - 2011. Red circles and blue squares indicate El Niño and La Niña years, respectively. Note that La Niña years tend to have lower activity, with 2010 having the lowest activity on record (15 named storms In 2011, there were 20 named storms. The thick horizontal line indicates the normal number of named storms (27.) Image credit: City University of Hong Kong.