With a Bachelors Degree in Environmental Sciences (2009), began tracking tropical storms in 2002 and is now a private forecaster.
By: Cavin Rawlins , 10:33 AM GMT on July 09, 2009
One of the most provocative questions that we ask every hurricane season would be what are the chances of Cape Verde Hurricanes (CVHs) affecting land areas in the Western Hemisphere. This question is often related to the East Coast of the United States but areas such as the Bahamas, Turks, and Northern Caribbean also is affected by whether a tropical cyclone re-curves or not. A recurving tropical cyclone is one that changes from a westerly to easterly component during the course of its life. The North Atlantic Oscillation (NAO) is probably the main factor in determining the mean longitude of recurvature (MLOR) each year.
The North Atlantic Oscillation
The North Atlantic Oscillation describes the difference between atmospheric pressures near the Icelandic Low near 60N and the Azores High near 30N. A negative NAO shows a weaker Icelandic Low and Azores High, while in the positive, a stronger/deeper of the two features occur. Over the past 58 years, patterns in the North Atlantic Oscillation provided forecasters with some level of accuracy in predicting landfalling CVHs. Probably the largest connection occurred between the NAO indices the December before the hurricane season as in Table 1.
Table 1. Unfavourable and Favorable steering patterns during the 8 recent hurricane seasons.
Table 1 shows the last eight hurricane seasons, the NAO oscillation the fall before, the recurvature pattern and notable Cape Verde Hurricanes during August, September and October, the peak of the Cape Verde Season (CVS). Years with negative NAO indices the fall before favour the landfall of CVHs the next hurricane season. The Hurricane Season of 2005, the most active on record, had 1 CVH during the peak of the season and it recurved out to sea. Conversely, other less active seasons with also 1 CVH, like 2008 and 2007 was very unfavourable. The 2002 hurricane season was relatively inactive and even below normal by ACE standards, yet the only CVH made it to the Gulf coast. The 2001 CVS was relatively more active (well above average) but both CVHs that year recurved out to sea.
Figure 1. Observed tracks of Cape Verde-type hurricanes in both phases of the NAO during the December before the hurricane season.
What Causes the NAO to have such an influence on hurricane tracks?
It is the two components of the oscillation, the Azores High and Icelandic Low, that act together to produce such variability in CVH tracks. The NAO phase the fall before normally reverses during the winter and spring (January-May) and returns again for summer and fall (June-December) which implies the NAO phase the previous fall normally indicates the NAO phase the next September.
In a negative NAO, the high is weaker but more sprawling and centralised, bridging further west along the US East Coast. The Icelandic Low is normally located over Eastern Canada/Davis Strait during summer and is a major component of mid-latitude (frontal) weather. During a negative NAO, the low is weaker than normally, indicating, weaker than normal mid-latitude weather and frontal troughs especially along the US East Coast. Therefore, we have a ridge extending further west and weaker troughs along the US East Coast, allowing the mean longitude of recurvature of CVH to be near 83.4W during negative events of the NAO.
In a positive NAO, roughly the opposite occurs, but the stronger than normally high is more developed in the North East Atlantic and does not bridge as far west as normal. You also have a deeper Icelandic Low, and more frontal activity, which in turn creates a favourable pattern. The mean longitude of recurvature during positive phases of the NAO is 61.5W.
Figure 2. 500 mb height anomalies for the month of September. Credit Australian Bureau of Meteorology (BOM)
Figure 2 shows the 500 mb height anomalies for the Western Hemisphere for two years, 2004 and 2006, which were oppositely, unfavourable and favourable years. If one was to direct their attention at the area from Eastern Canada to Greenland, one would notice the big difference in height anomalies. Higher than normal heights indicated a weaker Icelandic Low (2004), while lower than normal heights indicated a deeper Icelandic Low (2006). Similar results were produce in 2008 and 2007. The source of the data presented in this figure came from the BOM, which only goes back to 2002, but similar results were produce for 1998 and 2001. I was unable to provide an image since NOAA/ESRL website went down for maintenance during the construction of this report.
Data on the NAO go back to about 58 years and there are still some inconsistencies in this correlation due to smaller scale modulators of steering flow. An example would be 1992, which showed a positive phase the fall before, but a landfalling CVH, Andrew. In addition, there is only an average 1 Cape Verde Storm a year, accounting for a smaller percentage of all destructive hurricanes.
However, the inconsistencies though seemingly important, are relatively small. Other accurate correlations were the negative phase of the NAO the fall before 1989 (Hugo) and 1998 (Georges), two storms that affected my area.
Last fall the NAO was in negative phase, which is probably what we might see this season, since the pattern over the last several weeks, and the forecasts issued by the Climate Forecast System (CFS) indicate we may be entering a negative phase of the NAO. Now we can look at all the troughs leaving the Eastern United States now, but that the CVS starts next month and peaks September. This really implies a greater risk for the Western Hemisphere this year regardless of storm numbers. Another problem 2009 poses is that tropical waves may not start to develop until about 50-60W, closer to home.
The 2000-2009 decade will go down as the most destructive decade in hurricane history. The human impact from Hurricanes Katrina, Ike, Wilma, Charley, Ivan, Rita and Frances, within the top 10, easily tops 150 billion US Dollars. The most notable storm in that list was Hurricane Ike, which was a Cape Verde Hurricane in 2008, and third costliest Hurricane, only behind Andrew and Katrina.
Figure 3. Average indices for July, August, September and October from 1990 to the present. Credit: NOAA/ESRL
Figure 3 shows the combinations of above normal sea surface temperatures contributed the ongoing Atlantic Multidecadal Oscillation (AMO) and persistent Negative NAO phases contributed to the increase intensity and frequency of landfalling hurricanes this decade. In an article from USA today, Dr. William Gray quoted that: “The NAO has been weaker than normal since 1995, and we've had more intense hurricanes during that period.”
North Atlantic Oscillation Realtime Products
Recurvature and landfall of hurricanes and their relationship to the North Atlantic Oscillation
The tropics remain quiet this morning, with little to speak of. Most of the models develop an area of low pressure offshore the Southeast United States later tonight/early Friday but it is likely this feature will be non-tropical in nature. However, it will still be monitored for signs of subtropical development. Most of long-range models (MCM, EMCWF and GFS) show a low-pressure area in the Eastern Atlantic sometime next week. The GFS model has been forecasting the development of this feature for the past week and is the most aggressive of the three models. The source of development appears to stem from a vigorous monsoon trough, probably enhanced by a passing tropical wave. If the models continue to forecast development by Monday, the East Atlantic will be watched. Sea surface temperatures have significantly increased in the Tropical Atlantic and it seems models decrease shear enough during that time frame.
Comments will take a few seconds to appear.