The second part of ENSO, the Southern Oscillation Index, or SOI, is defined by monthly fluctuations in the air pressure difference between Tahiti and Darwin Australia. The SOI is generally a good indicator of the current ENSO status, as persistent negative values correspond to an El Nino and sustained positive values are associated with a La Nina. When ENSO is in a neutral state, the SOI tends to flip back and forth between negative and positive values. Over the past few months, this has been the case. The January SOI value was strongly negative, with a value of -12.8, but it was followed by moderately positive values, 9.0 and 0.9, in February and March respectively. April was marked by a flip back to negative SOI with a strongly low recording of -16.0, and May so far has been closer to average. Such fluctuations are typical of a neutral ENSO phase as has been the case. However, since there has been no consistent trend in the SOI recently, there is no way to tell the future phase in ENSO by it alone.
Wind patterns in certain parts of the globe can be good predictors to any upcoming changes in ENSO. The most important pattern is upper level zonal wind anomalies off the northeast coast of South America. During February and March, these anomalies often tell how far the upper branch of the Walker Circulation extends. Strong easterly winds were observed this past Winter, which favors a more westward Walker Circulation and consequencely cooler ENSO conditions later in the year.
Climatology
Observing patterns and relationships with previous ENSO events, and then comparing to the current ENSO, can prove to be very useful in ENSO forecasting. This year, we have found 2 correlations noteworthy enough to take into consideration for our 2004 ENSO forecast. Both involve the Quasi-Biennial Oscillation, or QBO, which is the next factor we will discuss in detail.
For the first climatological study, we looked at all of the El Nino events dating back to 1954 and the season in which they were initiated. Second, the phase of the QBO when they formed was recorded, and from there we examined the data to see if there was any correlation with the time of El Nino formation and QBO. We found 15 El Nino events since 1954. 11 out of the 15, 73.33% of these events developed when the QBO was easterly. Out of the 7 events that formed during Spring, the QBO was split fairly equally. However, the other 8 El Nino events that formed in the Summer and Fall were all when the QBO was in its east phase. No El Nino events formed in the Winter. This research strongly indicates that the easterly QBO is overall more conducive for El Nino formation, and all of the few El Nino events that did form during a westerly QBO were during the Spring.
In the next study, April to May ENSO conditions for years when the QBO had just transitioned to the west phase, as the case this year, were examined. We then found the years that most closely matched the April to May ENSO conditions in 2004, and finally looked at how the ENSO evolved throughout those selected years. We found 8 years in the past 50 years with similar April-May SSTAs in the ENSO regions as 2004. Out of the 8 years, none of them had an El Nino develop before or during the season. 1990, the only slight exception, had a finite warm bias. The remaining years saw a continuation of neutral conditions, a cool bias, or the development of a La Nina, with the majority having a cool bias. This data strongly argues against El Nino formation this Summer, and it does increase the probability of a relatively cool ENSO bias.
ENSO Forecast 2004
There is no clear consensus as to how ENSO will evolve throughout the hurricane season. DTAs and models seem to point towards a warm bias or weak El Nino event, and models lean towards neutral or a warm bias. However, upper level winds off South America support the opposite, a cool bias or weak La Nina. The 2 climatological results are also strongly in favor of cool-biased ENSO. With the obviously discrete indications, weight has to be given on each parameter. DTAs are fairly useful for more near-term changes in ENSO, but when neutral conditions prevail, modest warm and cool expanses often fade away easily. It generally takes a very large anomalous pool to end a neutral period and initiate a true ENSO event. The warm DTA pool that is present now is not that large, so it would take more to spark a true El Nino. However, it is still crucial enough to be judged as support for at least a warm bias during the season. As already mentioned, ENSO models tend to have poor accuracy at this time of the year, so they really do not need to be taken into much consideration at all. February upper wind anomalies off South America has a generally good lag connection to the phase of ENSO, as suggested by research performed by Dr. Gray et al of Colorado State University. Therefore, this factor will be given a fair weight. Climatology can be one the best parameters, but it can also be the worst as it assumes that something will always or never happen, which in the weather world is usually not the case. Nonetheless, 50 years of data is not that short of a time period, and both climatologically based correlations found are rather convincing. These 2 patterns will be taken into consideration.
Based on the lack of any overwhelming signal for either El Nino or La Nina, neutral conditions are forecasted through the climax of the Atlantic Basin hurricane season. There is also quite a bit of evidence for a cool ENSO bias, which appears more probable than a bias towards weak El Nino. Therefore, we anticipate slightly cooler SSTAs in the ENSO regions, though not cool enough to be considered an actual La Nina, towards the latter half of the season. With a cool bias in the forecast, a slight enhancing influence on major hurricane activity is possible. However, since no true ENSO event is expected, the overall effect from ENSO will be minimal.
III. Quasi-Biennial Oscillation
The Quasi-Biennial Oscillation is another one of the more important factors involving Atlantic Basin hurricane activity. By definition, the QBO is a periodic variation in the direction of stratospheric winds across the deep tropics. The two phases, easterly and westerly, generally last from 12 to 16 months, with the easterly phase often being slightly longer. The exact mechanism in which the QBO influences activity is not yet completely understood, but it is known that the QBO causes alterations in upper level vertical wind shear in the tropical Atlantic Basin. Climatological research shows that it particularly influences major hurricane frequency in the tropical Atlantic Basin, with a relatively insigificant impact on the total amount of named systems and hurricanes. The westerly phase of the QBO tends to increase the number of major hurricanes while easterly years usually have an average number of intense hurricanes.
Unlike ENSO, the QBO's alternation between phases is timely, making it very simple to predict transitions in advance. Last year, the easterly phase of the QBO peaked in August, with a value of -24. Since then, it rapidly increased and transitioned to the westerly phase in early February. The QBO was +10 as of April, and has continued to become increasingly positive. It is expected to peak during the season, but should not switch to the easterly phase until sometime in 2005. Therefore, the QBO is judged to be an enhancing factor for major hurricane activity in the Atlantic Basin this year.
IV. Atlantic Thermohaline Circulation
The Atlantic Thermohaline Circulation, or ATC, is a density driven circulation in the Atlantic Basin that undergoes cycles on decadal timescales. It has been proven that the ATC influences Atlantic Basin hurricane activity. When the ATC is in its warm or strong cycle, hurricane activity is increased, but its cool or weak cycle results in less hurricane formation. During the strong cycle of the ATC, the most notable consequence is warmer SSTAs across the tropical Atlantic Basin, which provides more energy for a tropical storm or hurricane to sustain deep convection. The opposite occurs when the ATC is weak. Additionally, lower vertical wind shear and sea level pressure anomalies, or SLPAs, are typically observed across the Atlantic Basin with a strong ATC.
From the mid 1920s through the late 1960s, the ATC was running strong for the most part, and consequencely, hurricane activity increased during that period. The ATC began to weaken in the early 1970s and it lasted through the early 1990s. A significant decrease in tropical cyclogenesis was noted. Since 1995, the ATC shifted back to the strong cycle, which explains the significantly above average hurricane seasons that have occurred over the past several years. However, even during the ATC's dominant phase, it does occassionally fluctuate for several months. This was the case in 2002, when the ATC temporarily became weaker, thus helping to limit hurricane activity that year. These small fluctuations are very hard to predict in advance, but it is easy to tell when they are happening by observing SSTAs in the eastern Atlantic over a monthtly period. When SSTAs are normal to below normal, that denotes that the ATC has probably momentarily fluxed to a cool phase. Earlier this winter, the ATC was running very strong, but has since toned down a little. Nonetheless, SSTAs in the eastern Atlantic near the Azores Islands have been about 1ºC above average this past 2 months. Because the ATC has more of a lag effect than present, future changes in its strength during the season should not have any significant implications. With all of that being said, the ATC is judged to be a fairly enhancing factor for this year's Atlantic Basin hurricane season.
V. North Atlantic Oscillation
The North Atlantic Oscillation, or NAO, is the fluctuation in 500MB heights in the northern Atlantic, primarily between western Europe, Iceland, and Greenland. Despite the fact that the differential in phases has an extremely limited impact on seasonal tropical cyclone frequency, climatology does suggest that the NAO has a major involvement in the steering pattern across the Atlantic Basin.
While the NAO will play a significant role in determing the steering pattern for the upcoming hurricane season it will, as always, be extremely difficult to forecast this far in advance. An attempt to determine whether the oscillation would be predominantly postive or negative during the majority of the 2004 hurricane season was determined primarily by climatology and current trends. The last time the NAO dropped significantly was in January 2004, when the index dropped to -1.5. Since then, it has leveled off some and has been steadily rising through mid-May. Based on the selected years that experienced a similar NAO pattern, a generally positive NAO is expected during the season. The NAO has much more variability than the other factors mentioned. Therefore, it must be noted that several negative NAO fluctuations are expected throughout the season.
VI. Selected Analog Years
Analog years are always very important to consider when producing long range forecasts. Today's available weather tools, such as long range forecast models, can only make predictions with a fair amount of skill so many weeks and months in advance. A decent-sized dataset of climate history or climatology provides either reassurance or a reason to reassess the conclusion of what has already been analyzed. After making an assessment for ENSO, QBO, ATC and the NAO, we have selected the years that we believe will best match the upcoming season. The analog years chosen for the 2004 hurricane season are 1953, 1961, 1966, 1980, and 1995.
| Selected Analog Years |
|---|
| Years | Named Storms | Hurricanes | Major Hurricanes |
|---|---|---|---|
| 1953 | 14 | 6 | 4 |
| 1961 | 11 | 8 | 7 |
| 1966 | 11 | 7 | 3 |
| 1980 | 11 | 9 | 2 |
| 1995 | 19 | 11 | 5 |
| 2004 Forecast | 14 | 8 | 4 |
However, none of the above years are a perfect match. 1953 had a warm biased ENSO, which could have helped limit the amount of activity it saw to some extent. 1961 and 1995 were both on the verge of true La Nina event, which, conversely, may have enhanced major hurricane frequency. 1966 and 1980 had the most similar ENSO to what is forecasted for 2004, but the NAO was predominantly negative in both years. Furthermore, the ATC in 1961 and 1995 was a bit stronger than what has been the case so far this year, whereas 1966's ATC was weaker, almost neutral. Since there is no obvious top or bottom analog among the others, each year will be evaluated fairly equally for the regional, landfall, and monthly predictions below.
VII. Regional Highlights
Once the evolution of ENSO, the expected peak of QBO, the strength of the ATC, and the 500MB pattern based on the NAO in the northern Atlantic are determined, our focus can finally shift to the steering pattern. The Atlantic Basin has been divided into seperate regions so that the amount of hurricane activity in each sector can be analyzed.
Mean Development Region
The Mean Development Region is the area between the Lesser Antilles and the western coast of Africa south of 20ºN. The lack of noticeable warming in the equatorial Pacific SSTAs combined with the westerly phase of the QBO should keep the upper level winds generally low. SSTAs in the Mean Development Region will be influenced by the ATC and NAO. A warm ATC favors a weaker subtropical ridge in the eastern Atlantic, thus less transport of cooler waters to the tropics and therefore warmer SSTAs in the Mean Development Region. A positive NAO supports the opposite, a stronger ridge and cooler SSTAs. With the NAO expected to undergo fluctuations, contrary to the more consistent ATC, slightly above average SSTAs are expected in the Mean Development Region. Also worth noting is the strength and mean position of the intertropical convergenze zone, or ITCZ. The ITCZ is essentially a zonal area of low atmospheric pressure and ascending air, usually located around 10ºN during the season. A slightly weaker subtropical ridge in the eastern Atlantic combined with warmer SSTAs favors more convection and vorticity within the ITCZ. It has also been suggested that the ITCZ is generally shifted further north away from the equator when the QBO is westerly and no El Nino is present. Therefore, the ITCZ should help enhance tropical cyclogenesis in the Mean Development Region this season.
Taking all of the above into account, the main indications are pointing towards a fairly active Cape Verde season. Out of our 5 analog years, the minimum of 3 named storms formed in 1953, and the most active year was 1995 in which a total of 9 storms formed. 1953's warm biased ENSO could have played a slight inhibiting role, whereas the very strong ATC in 1995 probably enhanced Cape Verde activity. Therefore, 4 to 6 tropical storms are expected to form, with 2 to 3 of those storms developing into hurricanes within the region. 1 of those hurricanes are expected to reach major hurricane status. The numbers presented are for the Mean Development Region only.
Western Atlantic
The western Atlantic Ocean should be almost as favorable as most of the other regions included in our forecast. The only question is SSTAs. Current SSTAs are relatively below normal in some portions of the region. Our climatological data also suggests that they should be warmer this month and throughout most of the season. We are most likely seeing a small fluctuation within a long term pattern. There is little evidence that suggests that such a large area of cool SSTAs will be present over the next 6 months. Nonetheless, the SSTA trend in this area will continue to be monitored. As for our numbers, most of the activity in the west Atlantic is expected to be moving out of the Mean Development Region. The positive NAO will enhance the strength of the subtropical ridge. The stronger ridge will help steer most systems at least halfway through the Atlantic Basin. 4 to 6 tropical storms, 2 to 3 hurricanes and 1 to 2 intense hurricanes are expected to originate and or move into the western Atlantic west of 65ºW.
Caribbean Sea
First, we will begin in the eastern Caribbean Sea. The eastern Caribbean Sea is frequently known as the "dead zone" to most experts. The region earned the nickname primarily due to strong westerly shear that often persists, which acts to tear tropical cyclones apart or prevent weaker disturbances from forming. A large upper level trough referred to as the Tropical Upper Tropospheric Trough, or TUTT, is the cause of the shear. The westerly QBO and neutral to somewhat cool ENSO should minimize the strength of the TUTT. In addition, the strength of the subtropical ridge should enhance the westerly steering flow in the lower latitudes. As a result, more systems coming out of the active Mean Development Region should make their way into the Caribbean Sea. Even without a strong TUTT, the estimated probability of a tropical storm forming in the eastern Caribbean Sea is only about 40%. However, 1 to 2 tropical storms are expected to enter the eastern Caribbean Sea. 1 of those is forecasted to be a hurricane, with a fairly strong probability of it being major hurricane strength.
The western Caribbean Sea is an area that is a lot more favorable for development. Upper level winds are usually less hostile, and with a weaker steering environment, broad areas of low pressure sometimes stall in the southwest Caribbean Sea until conditions allow them to develop. The area is expected to be very conducive for development in 2004. Considering that the western Caribbean Sea is a place known to spawn major hurricanes, the westerly QBO alone is enough to stir some concern. With a neutral ENSO and strong ATC also present, conditions should be excellent for development. In fact, the southwest Caribbean Sea is expected to be one of the warmest areas in terms of SSTAs in the Atlantic Basin, which correlates largely to lower SLPAs as well. The majority of storms passing through the western Caribbean Sea are expected to originate there, but a system moving into the area cannot be ruled out. A total of 2 to 3 tropical storms, 1 to 2 hurricanes, and 1 major hurricane are expected to form or move into the western Caribbean Sea this season.
Gulf of Mexico
Like the Caribbean Sea section, we have decided to split the Gulf of Mexico discussion into 2 different parts. To go along with the westerly progression, we will start with the eastern Gulf of Mexico. A rather interesting tidbit is that out of our 5 analogs, only 1 named storm formed in the eastern Gulf of Mexico. All of the classified tropical cyclones that passed through the area either came from the western Gulf of Mexico, Caribbean Sea, or western Atlantic. A potential explanation is that a weakness over the southeast was strong enough that any disturbance quickly moved inland without having time to organize. Regardless of the lack of originating tropical cyclones, the eastern Gulf should still be active and conducive for development. 2 to 3 tropical storms are expected to move into these waters, with 1 to 2 of these systems developing into hurricanes while passing over the warm SSTAs of the Gulf of Mexico and western Caribbean Sea. Based on our climatological research, 1 of those hurricanes being major appears to be highly probable.
Conditions in the western Gulf of Mexico are not expected to differ much at all from the conditions that are expected in the eastern Gulf of Mexico. Therefore, our forecasted numbers are not all that different. We are expecting 2 to 3 tropical storms to pass through the western Gulf of Mexico as well. However, the projected number of hurricanes and major hurricanes are slightly less. This goes along with our theory that most tropical systems will recurve at some point in their lifespan primarily in part to the westerly QBO as noted in the western Atlantic section. 1 hurricane is expected to move into, or originate in, the western Gulf of Mexico, with a fairly high probability of it being major, though still a slightly less of a probability when compared to eastern Gulf of Mexico numbers.
VIII. Local Landfall Data
One question that a lot of people are concerned about before the season starts if whether their local area will get hit or not. In this forecast, we have decided to break down the land masses and coastlines of the Atlantic Basin by region and analyze their risks this season. Most of this data is based on analog years and climatology, though recent 500MB pattern trends have also been investigated.
Texas
For the second straight season, the threat posed to Texas by tropical cyclones this season is uncertain. We are fairly confident that at least one tropical storm will affect the Texas coastline. The probability of a hurricane landfall is when things come into question. With a positive NAO in place, the chance of a hurricane, and even more so a major hurricane hitting Texas drops significantly. If Texas were to get hit, it would most likely happen when the NAO temporarily fluctuates negative. The estimated percentage of all of these ingredients coming together is about 40%. Therefore, we are not expecting a hurricane to strike Texas this season.
Louisiana
With all of the ingredients needed for an active Gulf of Mexico, you would think that the entire Gulf Coast would be at a high risk of a landfall. Our data suggests that this is not the case. Louisiana is expected to be the most inactive area along the Gulf Coast. Not 1 single named storm struck the Louisiana coastline in our analog years, as the state tends to get hit more frequently during an easterly QBO or El Nino. For 2004, any low latitude tropical cyclones in the Gulf of Mexico or western Caribbean are expected to stay to the south or recurve east of Louisiana. This season, no tropical cyclones are expected to impact the state.
Mississippi to Steinhatchee River (Entire Florida Panhandle)
As noted in regional highlights, the eastern Gulf of Mexico is expected to be a bit more active than the western half due to recurving systems. The reason behind the expected recurvatures is the positive NAO and the westerly QBO. The positive NAO will enhance the subtropical ridge while 500MB heights remain fairly low along the East Coast and the Southeast. The majority of tropical cyclones in the western Caribbean Sea and eastern Gulf of Mexico will likely recurve as a result of the expected steering pattern. The steering pattern that we expect to shape up would obviously pose a serious threat to the Florida panhandle. Therefore, this region is expected to get hit by a tropical system, with a high chance of it being hurricane status. The probability of a major landfall is a toss-up, but with a strong NAO in place, a major hurricane landfall along this region of the coast is favored.
Steinhatchee River to Key West (Western Florida Peninsula)
Southwest Florida is usually a target for tropical systems exiting the northwest Caribbean Sea, especially in October. But it appears that southwest Florida will not be in the bull's-eye this season. Most systems that make it west of southern Florida will likely continue heading westward or simply curve northeast into the panhandle. The threat from the east does not appear to be very imminent. However, with such a large amount of tropical cyclone activity expected in the eastern Gulf of Mexico, odds would suggest at least 1 tropical storm to curve south of the panhandle, such as Tropical Storm Henri last year. 1 tropical storm is expected to impact the western Florida peninsula in 2004.
Key West to Jacksonville (Eastern Florida Peninsula)
Ever since southern Florida was hit by Hurricane Andrew in 1992 and the noticeable increase in tropical cyclone frequency since 1995, most residents have been waiting for the next big hurricane to threaten the Florida peninsula. However, the probability of getting hit depends solely on the expected steering pattern in 2004, not the length of time that has passed since the previous landfall. Most of the Mean Development Region originating tropical cyclones should pass through the Caribbean Sea and head into the Gulf of Mexico or recurve just east of the Bahamas. Any threat posed to east Florida will likely originate rather close to home, such as the Bahamas. Climatological research shows that this region gets hit by hurricanes less when the QBO is westerly. Nonetheless, with a favorable southwest Atlantic in the cards, 1 tropical storm is expected to impact the eastern Florida peninsula. If it were not for westerly QBO conditions, eastern Florida would be under a much higher risk this year.
Jacksonville to South Carolina/North Carolina line
No tropical storms are expected to affect this portion of the East Coast. Westerly QBO winds favor recurvature of any approaching storm to the north or east. If this area were to get hit during a westerly QBO, it would probably be from a storm that developed close by, in the Bahamas or western Atlantic. Such storms are more common in cool ATC or El Nino years. Moreover, much of this region rarely gets hit simply due to the inward shape of the coastline, regardless of the steering factors.
North Carolina
If a part of North Carolina were to be affected by a tropical storm, it would likely be the Outer Banks. As mentioned above, westerly QBO favors more recurvatures from tropical cyclones that form in the Mean Development Region. In the past, storms that made landfall in North Carolina during a westerly QBO were also during a La Nina event, which is not the case this year. Therefore, a landfall from a Mean Development Region storm is unlikely, though it is quite possible that 1 will recurve far west enough that it briefly threatens the coast. But if the state were to have a real landfall, it would be from a storm that originated in the Bahamas or western Atlantic. Based on this information, 1 brush from a named storm is in the forecast for North Carolina this year.
Virginia to New York
This portion of the East Coast is rarely affected by direct landfalls. Most landfalls come once a particular storm has already brushed the Outer Banks of North Carolina. The expected setup of the subtropical ridge should limit the amount of East Coast threats and therefore we do not see this area being at a high risk of a landfall. There is only a slight chance that a southeast United States landfalling storm could still be intact in a weak state as it passes through the Mid-Atlantic. However, the low probability of this happening is slim enough that it will not be included in the forecast. No landfalls are expected from Virginia to New York.
New York to Maine
First off, we are not anticipating a landfall along this section of the coast. This forecast is based solely on statistics. However, an interesting observation was made based on our analogs and reanalysis data. The 500MB ridge over the western Atlantic is expected to be vertically aligned south to almost as far north as the Canadian Maritimes. While a trough is expected to protect the East Coast, the alignment of the ridge could allow for a near-miss to Cape Cod. Again, there are no signs that a hurricane will be able to push north straight into New England though. Any tropical system that makes it that far north would most likely take a slow bend to the north-northeast or northeast around the periphery of the subtropical ridge. So although a close call or threat is possible, this region should not see any landfalls in 2004.
Canadian Maritimes
As mentioned in the landfall discussion for the northeast United States, our research combined with the available reanalysis data suggests that the alignment of the subtropical ridge in the western Atlantic could force tropical systems to head a bit more northerly before curving out to sea. This pertains to tropical cyclones that make it as far west as the western Atlantic only. A stronger, more north-south oriented subtropical ridge could place the Canadian Maritimes under a moderate risk of a landfall. The Canadian Maritimes witnessed a landfall in 3 out of our 5 analogs, with 2 landfalls occurring in 1995. It is also interesting to note that 1995 is the only year that experienced 2 landfalls dating back to 1950. Based on climatology, there is only about a 28% chance of a landfall each season. However, 1 hurricane is expected to strike or brush the coast this year. This is the most reasonable forecast based on the forecasted overall 500MB pattern.
Bermuda
Due to the fairly small size of Bermuda and lack of direct hits, near-misses were also included for this discussion. Due to the expected number of north-south tracks in the western Atlantic, it is fairly reasonable to assume that Bermuda will experience several threats this season. Bermuda should expect at least 1 near-miss from a tropical storm. The probability of that tropical storm being a hurricane is about as equally high.
Bahamas
Interestingly, the Bahamas were not hit by any tropical storms or hurricanes from the east in all 5 of our analog years. Any landfalls occurred from storms originating within the Bahamas themselves or storms recurving out of the central and western Caribbean Sea. This fits well with all of our East Coast and western Atlantic statistics and reanalysis. Any tropical system that passes west of Puerto Rico north of 20ºN should recurve to the right before reaching the eastern Bahamas islands. As mentioned before in the regional highlights section, the western Atlantic will be generally favorable for development this season. The Caribbean Sea is also expected to be active. Therefore, 1 tropical storm tracking through the Bahamas is in the forecast.
Western Cuba
Western Cuba is typically hit or brushed by tropical storms and hurricanes that move into the Gulf of Mexico from the Caribbean Sea. With an active Caribbean and the primary landfall region expected to be the Florida panhandle, it would seem fairly likely that at least one landfall will occur in western Cuba. Surprisingly, west Cuba was only hit during 2 out of the 5 analog years. The majority of classified systems should remain south of Cuba or pass through the Isle of Youth. Note that near-misses were not calculated into this part of the forecast. However, based on the expected level of tropical cyclone activity surrounding the area, it would be foolish not to forecast 1 tropical storm landfall.
Eastern Cuba
Tropical cyclone landfall frequency for the eastern portion of Cuba is usually less than that of western Cuba. This is also reflected in our statistics. The probability of a landfall appears to be rather low this season. A track similar to Hurricane Georges in 1998 is not probable due to the expected alignment of the subtropical ridge mentioned above. There was 1 landfall and one near-miss during our analog years, but there is no clear indication that eastern Cuba is under the gun for a direct hit.
Jamaica
Most of the activity should remain south and west of Jamaica this season. Any recurvatures will likely pass over western Cuba, just west of Jamaica. Hurricane Allen in 1980 did graze the northern side of the island, but that particular system was the only storm that resembled a landfall. Going primarily by odds, a direct hit is not in the forecast.
Hispaniola
Hispaniola will likely end up being one of the more probable strike zones of 2004. On the other hand, there is also a chance that the main storm track will be north or east of the island. Hispaniola was impacted in 60% of our analog years, with a total of 3 classified tropical systems, 2 of those storms being major hurricanes. The 2 seasons that did not have a direct landfalls did have several near-misses. Therefore, 1 tropical cyclone is expected to impact Hispaniola. Based on the steering pattern that we see shaping up, there is a decent chance that this cyclone could be a significant hurricane.
Puerto Rico
Puerto Rico did not experience a direct hit in any of our analogs, regardless of the fact that it was basically within the main storm path. In addition, there were several near-misses. It almost seems as if Puerto Rico should have been hit and that they got lucky. It is hard to say whether or not the island's luck will run out this season. A direct landfall does not seem to be in the cards, but unfortunately, Puertorriquenos will likely be on their toes during the heart of the season.
Lesser Antilles
Due to the alignment of the subtropical ridge, the favorable upper wind pattern in the Mean Development Region, and statistics based on climatological data, the Lesser Antilles will likely be a main landfall target this season. The Lesser Antilles easily have the greatest likelihood of being impacted by a major hurricane this season out of all of our selected landfall regions. The islands were affected by a major hurricane in the majority of our analog years, and at least 1 tropical storm during each of the 5 analogs. Based on this information, 1 to 2 tropical storms, 1 to 2 hurricanes, and 1 major hurricane are expected to affect the Lesser Antilles this year.
Brownsville to Central Mexico
Shifting our attention back to the west, we will once again focus on the western Gulf of Mexico. As noted earlier, the forecasted steering pattern does not represent a pattern in which we would expect to see tropical cyclones moving straight into the Mexican coastline out of the central Gulf of Mexico. More in the way of recurving tropical cyclones should be expected. Any threat posed to central Mexico would likely originate on either side of the Yucatan Peninsula. There was only one case in which a hurricane coming from the Florida Straits made it as far west as Mexico, and that was Hurricane Inez in 1966. Any system that would impact the coast would probably be weak in nature. No landfalls are expected.
Central Mexico to Northwest tip of Yucatan Peninsula
All storms posing a threat to this area will most likely originate east of the Yucatan Peninsula. As a result, only weak tropical cyclones should affect this portion of the Mexican coast. 1 tropical storm is expected to make landfall.
Northern and Eastern Yucatan Peninsula
The eastern side of the Yucatan Peninsula is another section of the Atlantic Basin that will either be under the gun or very lucky this season. The area was hit by a tropical cyclone in 3 out of 5 of our analog years. A major hurricane impacted the peninsula in 2 of the 3 landfall years. The Yucatan Peninsula also had a near-miss by Hurricane Inez in 1966, a year in which no other system posed a threat to the coast. However, based on statistics combined with what we expect to see in 2004, we are leaning more towards a weak tropical cyclone, such as a tropical storm, making landfall. We believe that Inez was simply an odd exception after looking at the overall scheme of things.
Honduras and Nicaragua
Honduras and Nicaragua are affected mainly by homegrown tropical cyclones in the southwest Caribbean Sea. During some occasions, a storm may develop or move out of the southern Caribbean Sea and track far enough south that it runs into Honduras or Nicaragua. That scenario is not expected to take shape in 2004. Most of the tropical cyclone activity that will develop in the southern Caribbean Sea should pull north for the same reason that storms in the eastern Gulf of Mexico will likely recurve. If there is a tropical cyclone in the southern Caribbean Sea, it should gradually gain enough latitude so that by the time the cyclone may get caught in the westerly steering flow, it will already be north of the Honduran coast. No landfalls are expected, though this area was one of the more difficult landfall regions to get a good handle on.
Costa Rica to Venezuela
Our final landfall region is extreme southern Central America and the Caribbean coast of Venezuela. Climatologically speaking, this area is not a favored area for tropical cyclone landfalls. There is no evidence that this region is at a greater risk of a landfall this season than it would be during any other given season. Therefore, no landfalls are expected.
VIV: Month By Month Breakdown
The Atlantic Basin hurricane season officially spans 6 months, as it begins on June 1 and ends on November 30. Unlike landfall and regional forecasting, it is very hard to venture out and predict tropical cyclone activity for each month by observing the recent patterns alone. Therefore, we looked primarily at climatology and data from analog years for this section of the forecast.
June
On average, a named storm in June forms in the Atlantic Basin once every 2 to 3 years. Therefore, putting all research aside, it is pretty much a coin toss to guess whether June will get a named storm or not. However, we have found a few interesting climatological correlations. For example, out of all of the June storms that formed in the Gulf of Mexico, only 3, with 2 being in the same year, occurred when the QBO was westerly. On both of those instances, a moderate El Nino was in place. So a June storm in the Gulf of Mexico is statistically very unlikely this year. Next we looked at all the June storms that originated in the western Atlantic, offshore the United States east coast or near the Bahamas. We found again that easterly QBO seems to be more favorable, with 81% of the storms occurring in the easterly QBO phase. The few that were in a westerly QBO were also during a moderate La Nina event. Based on these statistics, a named storm in the western Atlantic is unlikely this June. Since the Mean Development Region generally does not become favorable for development until late July, this leaves us with the Caribbean Sea. In our 55 years of research, only 3 named storms developed in the Caribbean Sea during June. However, all 3 occurred in a westerly QBO, and 2 of them were in our analog years, 1966 and 1995. Using analogs, this makes the probability 40%. Considering that there have not been many in the first place, we will call for no Caribbean Sea June storms. Altogether, no named storms should form in the Atlantic Basin in June this year.
July
By July, conditions normally improve slightly from June, however, overall the atmosphere is still fairly hostile in the Atlantic Basin. In a typical season, about 1 named storm forms in July. Our climatological research indicates that the Gulf of Mexico will probably not see a tropical cyclone in July either, given the fact that only 1 formed when the ENSO and QBO matched up similar to this year. The western Atlantic and Mean Development Region look more promising, however. All of our analogs except 1953, the year with warm bias ENSO, had a named storm originate in the Mean Development Region. 2 of our analogs, 1966 and 1995, saw 2 named storms form in the western Atlantic, whereas the rest had none. Based on this information, we anticipate 1 to 2 named storms, with 1 originating in the Mean Development Region.
August
Typically, Atlantic tropical cyclone activity experiences a fairly sharp increase during the month of August, particularly after August 15. By averaging the amount of August activity in our analogs, we expect a near average August this year. The development of at least 1 August Mean Development Region storm looks likely based on our analog data. Only one year, 1961, did not see Cape Verde development in August. It should be noted that 1961 saw no tropical cyclones form at all during August, which is highly unusual, especially for an active season. In our analog years, the named storms that developed in the Mean Development Region tended to intensify to hurricane or even major hurricane status. Based on the analog data, however, there is only a very small chance of a named storm forming in the Gulf of Mexico during August, though there is a somewhat better chance of a storm originating in another region crossing into this area. On the other hand, more in line with the Mean Development Region, the western Atlantic looks to be a favorable area for development in August this year based on analog years and climatology. All in all, 2 to 3 named storms, 1 to 2 hurricanes, and 1 major hurricane are forecasted to occur in August 2004.
September
September is climatologically the most active month of the season. Based on our analogs, hurricane activity should be above average this month in 2004. All of our analogs had at least 2 named storms first originate in the Mean Development Region in September except 1953, the warm bias ENSO year. The majority of these systems eventually became hurricanes and even major hurricanes. Each of the analog years also had at least 1 named storm form in the western Caribbean Sea or southern Gulf of Mexico during this month. None of the years saw a September tropical system develop in the high latitudes, which is not surprising considering those types of storms are more common in cool ATC or warm ENSO years. Another thing to take into consideration is that a lot of the years in past that had late starts, with a few exceptions, tended to see higher than normal hurricane activity in September. Recent examples include 1998, 2000, and 2002. Considering that we expect no named storm in June, this provides us with more confidence of an active September. Based on all of this information, 4 to 5 named storms, with 3 to 4 being hurricanes, and 2 major hurricanes. Most of these storms should develop in the Mean Development Region, with 1 or 2 forming in the western Caribbean Sea or Gulf of Mexico.
October
Activity is generally slower in October than it is in August and September, but more active than June and July. Mean Development Region hurricane formation is almost always shut down for good by the end of the month, though it is not uncommon to see development there in the first half. Our analog year October activity is a bit diverse, so dissecting each year is necessary. Interestingly, 1953 actually saw 2 named storms form in the Mean Development Region in October, though both dissipated early. 2 other tropical storms formed in the western Caribbean Sea. 1961 saw a named storm form in the western Atlantic and a strong major hurricane in the western Caribbean Sea. 1966 did not see any named storms form in October, and 1980 only saw 1 hurricane in the far northeastern Atlantic Basin. Like 1953, 2 weak named storms formed in the Mean Development Region in October in 1995, whereas a major hurricane also developed in the western Caribbean Sea and another hurricane in the subtropical latitudes. Averaging all of this out, 2 to 3 named storms with 1 being a hurricane is in the forecast for October, which is near average. There is also a fair chance that the hurricane will be a major one in the western Caribbean Sea, an occurrence observed in both 1961 and 1995.
November
Usually activity significantly calms down by November, and a storm forms in November about once every 3 years, just like June. Based on our research, there is not that much correlation with any of the major factors and November tropical cyclones except that lower frequency is observed in El Nino events. However, what we found with analogs is very convincing. First off, we scratched 1995 for November consideration as the QBO switched to easterly in November that year. This left 4 analogs for this month. 2 years, 1953 and 1966, had a named storm, whereas the other 2 years, 1961 and 1980, had 2 named storms. Out of the 6 storms total, 4 were hurricanes. Most of the tropical cyclones occurred in the high latitudes of the Atlantic Basin away from land, though 2, both in the seasons with 2 November storms total, did form in the western Caribbean Sea. Based on this data, 1 to 2 named storms are in the forecast for November with 1 being a hurricane.
X: Conclusion
The recent upswing in tropical cyclone activity in the Atlantic Basin is showing no signs of ending. A strong ATC, combined with the westerly phase of the QBO and neutral ENSO conditions should allow the above normal trend in tropical cyclogenesis to continue into the 2004 season. A noticeable aspect that has been lacking during the recent upswing in activity has been major hurricane landfalls in the southeast United States. However, that does not mean that our luck will continue. It is just a matter of time until we witness another major landfall somewhere along the coast of the United States. Some evidence suggests that our luck may run out as early as this summer, as noted in our forecast for the Florida panhandle. Regardless of our expectations for 2004, you should always be prepared for a landfall well in advance if you live in a hurricane prone area. It should also be noted that even if 2004 turns out to be inactive, it only takes 1 major hurricane landfall to make the season a devastating one. A well known example of this is the 1992 hurricane season, which was well below average, yet that was the year that category 5 Hurricane Andrew slammed into southern Florida and resulted in about $30 billion in damage.
Based on all of the research and data investigated above, a season of significantly above average hurricane activity is anticipated.
If you have any questions or comments about this forecast, please write to us at webmasters@independentwx.com.
