Preview of the 2014 Atlantic Hurricane Season

Atlantic Basin Forecast

On average 10-11 tropical storms (named storms) develop in the open Atlantic annually during the June 1 through November 30 hurricane season. As indicated in Figure 1, about half reach hurricane strength (74 mph) and approximately one quarter become major hurricanes (Saffir-Simpson Category 3 or greater). These average storm figures have inched up in the last 15-20 years because of elevated Atlantic sea-surface temperatures (SSTs).

As Figure 1 shows, storm activity tends to be rather constrained until midsummer, with a steady uptick in storm frequency starting in August. On a climatological (long-term) basis, the most active part of the season for tropical cyclones occurs in mid September, then steadily decreases through late October and November. This pattern holds true for hurricanes and major hurricanes as well, with tailing off occurring slightly sooner for these more-intense events.

Figure 2 presents the 2014 forecasts for named storms (tropical storms) and hurricanes from eight organizations, along with long-term averages (indicated by the dotted lines). A few common threads emerge from the numbers:

• Almost all categories of forecasts are below the storm and hurricane averages.
• A few of the forecast ranges cross into above-normal values.
• The relative percentage of tropical storms that develop both into hurricane and major hurricanes (most agencies are forecasting 1 to 2 major hurricanes) is in line with historical ratios.

Putting the 2014 Forecasts into Perspective

Hurricane season forecasts do have utility and can help put expectations into perspective, but because preseason storm predictions require anticipating climate and weather signals weeks to months in advance, they carry considerable uncertainty. Historically, the actual level of storm activity rarely matches the forecast level. As shown in Figure 3, between 1995 and 2013 the actual storm count minus the forecast storm count for the Atlantic ranged from eight more storms than forecast to seven fewer storms than forecast.

Even when environmental conditions are accurately forecast, day-to-day weather variability makes it extremely difficult to correlate basinwide storm activity with storm location and track, landfall risk, damage, and insured loss. A highly active season in terms of storm counts does not necessarily translate to high damage and losses; likewise, a season with a below average storm count does not always translate to low damage and losses—such as we saw in 1992, a low-activity, El Niño year, when Hurricane Andrew came ashore and devastated parts of Florida and Louisiana.

Despite the innate challenges, seasonal forecasts today do benefit from more sophisticated methodologies, such as numerical rather than solely statistical prediction, and make use of dramatically improved climate models. We advocate being aware of the trends these forecasts indicate for a given season, but not getting too attached to the absolute forecast values. Ultimately, what really matters is not storm counts, but whether storms make landfall and result in damage and insured loss.

Exploring the Reasoning—and Uncertainty—Behind the Forecasts

Forecasters assess numerous indicators to gauge the conditions for storm formation and strengthening, chief among them SSTs and wind shear. Storms need an ample supply of energy to form and strengthen, and humidity from warm SSTs provides that energy. Likewise, storms will not develop in the highly disturbed environment that results from wind shear.

In the short-term, when tracking a storm in real-time, SSTs and wind shear are fairly predictable; uncertainty increases, however, as the forecast time lengthens. Other factors influencing storm development, strength, track, and even landfall—such as atmospheric steering currents (which influence storm track and the probability of landfall)—are even less predictable in advance of the season.

SST anomalies, linked to the ocean's enormous capacity to retain heat, help us understand the potential impacts of SSTs. In Figure 4, the image on the left shows April 1, 2014, SSTs, while the image on the right presents a prediction of SSTs for July-September of this season. The Main Development Region (MDR) in the Atlantic—the area in which much of the Atlantic storm activity will originate—revealed some pockets of warm SST anomalies in April, but conditions appear to be very close to normal across the Atlantic later in the season. In the eastern equatorial Pacific—the area to watch for potential El Niño conditions—the cool signal from last year was replaced by widespread neutral SST conditions in the April image, with clear signs of a warm anomaly building in the forecast image, on the right. (See the callout box, "2014 Pacific Basin Hurricane Season Forecasts.")

As seen in Figure 5, Atlantic SSTs—and cyclone activity—correlate with a phenomenon called the Atlantic Multidecadal Oscillation (AMO), the source of some very warm Atlantic SSTs in recent seasons. During warm SST periods, a higher percentage of tropical storms mature into hurricanes and major hurricanes.

Figure 6 presents AMO data from NOAA's Earth System Research Laboratory from 1948 to 2014, with warm periods in red and cold periods in blue. The AMO index, which has been below zero, or cool, since January 2014, may shift phase into a multidecadal cool period and warrants monitoring over the next few years.

El Niño and La Niña—which refer to the warming (El Niño) or cooling (La Niña) of SSTs in the eastern equatorial Pacific Ocean—impact storm formation and storm weakening/strengthening. El Niño tends to decrease wind shear (with a resultant increase in storm activity) in the Pacific, and increase wind shear (with a decrease in storm activity) in the Atlantic, while the opposite occurs with La Niña conditions. According to the Climate Prediction Center (CPC) and International Research Institute for Climate and Society (IRI), the chances of El Niño increase during the remainder of 2014. As shown in Figure 7, the probabilistic El Niño forecast exceeds 65% by summer (see time period JJA) and peaks near 80% during the late fall and early winter.