Original 28storms.com forecast

Although this review could have been written months ago, it seemed best to wait until the National Hurricane Center released the final data for every tropical cyclone. As anticipated, the NHC made several adjustments which not only altered the final storm totals but cleared up questions about some landfall intensities. So rather than writing this report in late November and having to make amendments later, our forecast can now be evaluated with revised and more accurate information.

The 2011 Atlantic hurricane season finished above the long-term average, with 19 named storms, 7 hurricanes, and 4 major hurricanes. A side-by-side comparison of the observed totals with our June 1 forecast is below:

The number of named storms was somewhat underestimated — the season had several weak, short-lived tropical storms. However, the number of hurricanes was just one fewer than predicted, and the number of major hurricanes exactly matched what was forecast. Thus, the forecast of total activity is judged to be an overall success.

One storm, Irene, made multiple landfalls as a hurricane and as a strong tropical storm. Three other tropical cyclones (Arlene, Harvey, and Maria) made landfalls as strong tropical storms. Below, the tracks and landfall points of these storms are superimposed on our original landfall risk zone map from June 1:

 
As evident, our risk assessment proved to be quite accurate. Every hurricane impact — Irene in Puerto Rico, the Bahamas, and North Carolina — occurred in an elevated risk zone. Additionally, Arlene, Harvey, and Irene made landfalls as strong tropical storms in Veracruz state, Belize, and St. Croix, respectively. These landfall points were all in an elevated risk zone. Only two strong tropical storm impacts happened outside the elevated risk zones: Irene in New Jersey and New York and Maria, borderline extratropical, in Newfoundland.

There were no other significant tropical cyclone landfalls during the season. Rina made landfall in the Yucatan Peninsula as a moderate tropical storm, while Harvey and Nate hit the southwestern Mexican Gulf coast as weak tropical storms. Lee, while a major rain and flood event in the eastern United States, hit Louisiana as just a weak subtropical storm. Don came close to hitting south Texas as a tropical storm but collapsed just offshore.

The forecast was based on assumptions regarding four major factors, as stated in the original writeup:

- Neutral ENSO trending cool by the fall, with SSTAs roughly uniform across the equatorial Pacific
- Negative PDO
- Warm SSTAs across the tropical Atlantic and cooler SSTAs in the subtropical Atlantic
- Downwelling easterly QBO

All of this materialized, although ENSO cooled more quickly than predicted. Examining these factors allowed for an outlook on the dominate longwave pattern during the bulk of the season, which became the main basis for the landfall risk assessment:

In the Pacific, the cooling ENSO and negative PDO will likely result in persistent anomalous upper-level ridging over the subtropical Pacific and into the southwest United States [R1]. Anomalous troughing aloft is expected to be common in the northeast Pacific, Alaska, and British Columbia [T1]. These anomalies should force abnormal ridging and troughing over central-eastern Canada [R2] and central-eastern United States [T2], respectively … Based on the conflicting signals, upper-level heights across the subtropical Atlantic will probably stay average in the west to slightly below average in the east. The net result in the tropical Atlantic will likely be slightly above average heights aloft.

This forecast pattern verified well in August-September, the peak two months of the season. The placements of the anomalous ridging and troughing, labeled in the excerpt as R1, R2, T1, and T2, are likewise marked in the below August-September 500mb geopotential height map. (The downstream relatively enhanced ridging in the western Atlantic was not expected to be anomalous, hence it is not marked.)

The 28storms.com 2011 Atlantic hurricane season forecast, while not perfect, generally succeeded in both predicting the total amount of activity and highlighting the elevated landfall risk zones. The analysis of the factors and the resulted pattern forecast were also largely correct, which strengthens our confidence that the successes were the result of an effective methodology more than simply luck. To be sure, future forecasts will need to consistently demonstrate similar levels of skill.