Average 20 year old weather nerd. Plymouth State University Meteorology, Class of 2018. NOAA Hollings scholar. Summer 2016 intern at NWS Boston.
By: MAweatherboy1, 9:26 PM GMT on March 19, 2016
Hello, and welcome to my 2016 Atlantic Hurricane Season prediction blog! As always, it seems hard to believe that we are again approaching the time of year where the thoughts of meteorologists turn from snowstorms and Arctic air outbreaks to severe weather and the impending hurricane season. A few disclaimers before I begin. Firstly, my forecast is in no way official nor representative of the thoughts of any official forecasting agency. These are my thoughts alone. Secondly, seasonal forecasting is a very inexact science, so even experienced professionals will tell you that errors may be significant, and that revisions may be necessary as the season gets closer and as it gets underway. Finally, with seasonal hurricane forecasts, it is especially important to stress that while there is some skill in predicting overall seasonal activity, it is impossible to predict what impact individual systems may have on land areas beforehand. Anyone living in an area vulnerable to hurricanes should always have a personal/family plan in place to address the possibility of how to prepare for a hurricane impact, and they should monitor the National Hurricane Center throughout the season for the most up to date information on individual storms and their potential impacts.
Now, onto the forecast. For several months, I have occasionally contemplated what may await the Atlantic for this hurricane season. I never developed a very strong feeling, rather the only thing that has been apparent to me is this: The 2016 season will be significantly more difficult to make a forecast for than the 2014 and 2015 seasons were. 2014 and 2015 were both very clear candidates for below average hurricane seasons due to a number of factors, such as a weakening of the positive Atlantic Multidecadal Oscillation (AMO) that has dominated the Atlantic’s “active era” since 1995, as well as the development of one of the strongest El Nino events on record last season, which reduces Atlantic hurricane activity. Indeed, both of these seasons turned out to be below average, as most forecasters predicted. This season, the signals are less clear. The main factors I will discuss in creating this forecast are the El Nino Southern Oscillation (ENSO) and the sea surface temperature (SST) configuration in the Atlantic basin. These are usually the two most important factors in determining seasonal activity, and I will discuss them in detail along with a couple other measures that can be used to gauge seasonal activity.
There has been much buzz, and rightfully so, about the very strong El Nino event of the past year, which intensified into one of the strongest, if not the strongest, ever recorded. El Nino is known to make Atlantic hurricane seasons less active by increasing wind shear across parts of the basin, which prevents disturbances from organizing into stronger cyclones. Last year’s seasonal forecasts skewed heavily below average based on the high likelihood of this event, and they were correct, although it can be argued that the 2015 season did somewhat outperform given the circumstances. El Nino also contributed to a record setting East Pacific hurricane season, culminating in Hurricane Patricia, whose 215mph maximum winds were the highest ever measured.
Now, the question is how the ENSO will shape up for this season. It is not a clear picture. As of now, strong El Nino conditions remain present across the equatorial Pacific. However, it is almost certain that El Nino has peaked and is now into its decline phase. Observations of sub-surface conditions in the Pacific indicate that while warmer than normal water continues to dominate the surface, a large and growing pool of below normal water is present further below the surface, and is spreading eastward throughout the Pacific. This is in much the same way as the current El Nino event began last year. Usually, these subsurface anomalies build eastward and gradually rise to the surface. These observations would suggest significant cooling of the sea surface in the Pacific, perhaps even the development of a La Nina event, the “counterpart” of El Nino. Indeed, La Nina events very often follow strong El Nino events. Finally, most climate models, which attempt to simulate atmospheric and oceanic conditions months into the future, also agree that a transition towards neutral or La Nina conditions is likely this summer. A noteworthy exception is the CFS model. This model, the “flagship” of the United States climate modeling division, was one of the first and most accurate models in predicting the development and extreme intensity of the current El Nino event. While the CFS shows some cooling in the Pacific over the next couple months, it shows El Nino re-strengthening into the summer and fall, and a few other models show some leveling out or renewed warming of the Pacific around that time.
I posted awhile back on the main blog about how I feel El Nino is likely set for a rapid decline. It is not uncommon for strong ENSO events to fade out in a relatively short period of time. Already, water temperatures are on a slow but steady decline over the benchmark Nino 3.4 region, a large swath of the equatorial Pacific used in gauging the official intensity of El Nino. Waters off the coast of South America have shown signs of cooling as well. Within the next 1-2 months, I expect that decline to accelerate as the last of the massive pool of warm water that helped set off this El Nino is exhausted, and cool anomalies begin to surface. My expectation is that for the heart of hurricane seasons, ENSO conditions will generally be near neutral, perhaps cool neutral or weak La Nina conditions by August or September. As a result, ENSO will likely have a lesser than normal impact on the Atlantic hurricane season. I hesitate to say it will be a net positive, because besides the ENSO regions, there remains an expansive region of warm water farther north in the Pacific, towards 10-20N latitude, which if it remains in place may favor enhanced East Pacific convection, which can cause shear across the western Atlantic. It is also very possible, however, that these warm anomalies largely fade out with the fading El Nino. Monitoring changes in this area and in the equatorward regions will be critical in the coming months. Springtime is notorious for being difficult to forecast ENSO in, and there seems to be higher than normal uncertainty this year on how the decay of El Nino and potential rise of La Nina will play out.
Figure 1: The death of El Nino; note how the last of the warm anomalies associated with this historic El Nina event are surfacing and fading out, as cool anomalies associated with a likely pending La Nina move in underneath.
Figure 2: Mid-March 2016 ENSO model plume forecast. Each line represents a different model's forecast of ENSO conditions. The spread is massive, larger than normal even for this difficult to forecast for time of year, but there is a strong clustering of models showing neutral or weak La Nina conditions by mid-summer.
While Pacific sea surface conditions are important, we are after all concerned with the Atlantic basin, so barring an extreme ENSO event in the Pacific like what we saw last year, it’s usually the Atlantic SST pattern that matters more. Often talked about is the Atlantic Multidecadal Oscillation, somewhat analogous to the Pacific Decadal Oscillation. It refers to a prevailing pattern of warm or cool SST anomalies, mostly along a “ring” of water around the edges of the North Atlantic basin, and in particular the far north Atlantic. This oscillation is not very well understood, and its period may vary dramatically, from 15-40 years. From 1995-2012, the Atlantic had a sustained period of remarkable activity, with some of the most active hurricane seasons on record, such as the record breaking 2005 season. This was largely fueled by primarily above normal SSTs over the Atlantic’s “Main Development Region”, which is warmer than normal in a positive AMO phase. From 2013-2015, however, this strongly positive AMO has weakened, and SSTs in the MDR have generally been only near average in the past 1-2 years, while regions north and south of the MDR have been significantly warmer than average. This imbalance- a neutral or cool MDR with surrounding regions of warmth to the north and south, is unfavorable for tropical cyclone formation, particularly with regards to tropical waves moving off of Africa, as limits rising motion, convergence, and instability across the MDR. An accurate forecast of future MDR sea surface conditions greatly improves the chances of an accurate hurricane season forecast.
Over the early winter, there was a good deal of speculation that the positive AMO of the late 1990s and first decade of the 2000s was about to come roaring back, with some climate models, especially the CFS, showing an almost ideal positive AMO developing, with very warm SSTs in the MDR. However, this has simply not materialized. Instead, the AMO has remained in a generally neutral/negative state, with SSTs in the MDR only near or slightly above average. In addition, an expansive area of warm anomalies exists in the mid-latitude waters, which as mentioned is generally detrimental to storms forming in the deep tropics if the water in the deep tropics is cooler than average. I believe, as I have for some time now, that it is highly likely that the AMO has flipped to a negative phase, which will persist for at least the next 10-20 years. That’s not to say there won’t be active hurricane seasons sprinkled among this period, but it will be the opposite of the recent +AMO era in which active seasons were the norm, and quiet ones the exception. Based on the persistence of the recent Atlantic SST pattern, my instinct and best forecast at this time is for only a continuation of the status quo, with a SST pattern generally unfavorable to storm formation in the deep tropics, especially east of the Caribbean Sea. Again, that is not to say there won’t be storms there, but activity will likely be near to below normal east of the Caribbean due to the SST configuration, as well as a continued lack of instability over that part of the tropical Atlantic, which has plagued the basin for years and appears likely to continue doing so. As with ENSO conditions, I would rate confidence in the Atlantic SST forecast as no better than average and perhaps below average at this time. It will be important to monitor any changes in the months leading up to hurricane season.
Here’s a quick visual of the current state of affairs in the Atlantic. It’s a pretty interesting, dynamic setup. Note that while the “warmth” in the Caribbean isn’t particularly impressive, it’s less of a factor either way because the SSTs there are always plenty warm for storms to form. Likewise, much of the MDR remains above normal, but not by a lot, and it isn’t the strong, concentrated, low-latitude band of warmth you’d look for in an active MDR season. The extreme warm anomalies off the East Coast, which will likely persist for much of the season, are also of interest. While that doesn’t necessarily cause systems to develop (you still need some incipient disturbances), it means any storms developing in or moving into that area have a much higher than normal potential intensity. The Gulf of Mexico has also been consistently above normal.
This concludes my comprehensive discussion of Atlantic and Pacific SST patterns. There is certainly much more that goes into tropical cyclone formation. Wind shear patterns, atmospheric moisture content, and the strength and nature of incipient disturbances are all in fact more important to individual storm formation, but these characteristics are highly variable day-to-day and week-to-week, and can only be handled on a “case-by-case” basis during hurricane season. What I’d like to do now is break down what I think some of the more and less likely regions for activity this season are. As I've said, the formation of individual storms cannot at all be predicted more than a few days in advance, let alone a few months. However, it is possible to highlight areas of higher or lower than normal risk.
Less active: Atlantic MDR, east of the Caribbean; I’ve already talked quite a bit about my thought process here, with the generally near normal MDR SSTs unlikely to inspire a strong Cape Verde season. While some degree of activity is likely here, as in any year, I expect it to be at least a little below average. Instability in the tropical Atlantic also continues to lag very far behind climatology, something that has limited MDR activity for the past three years.
More active: Caribbean Sea and Gulf of Mexico. The Caribbean Sea also constitutes part of the MDR. Caribbean SSTs have been solidly and consistently above normal through the winter, generally by half a degree or a little more. Extreme wind shear conditions, fueled largely by El Nino, plagued the Caribbean last year, and this often fertile region for tropical cyclone development and intensification was turned into a graveyard, with multiple storms and disturbances dying out or failing to organize due to the wind shear. Similarly, the Gulf of Mexico also experienced above normal shear, preventing nearly every disturbance over it from developing, save for Tropical Storm Bill. Like the Caribbean, the Gulf has seen mainly above normal water temperatures through the winter. While not a certainty, I see these areas as possible higher activity zones this season. With El Nino weakening, especially deeper into the season, shear should return to normal levels over these regions, which is often quite low at the peak of the season. Combined with warm waters, this sets up a potentially favorable environment for tropical cyclones. It could create an opportunity for tropical waves or weak storms that have struggled through the less than ideal waters east of the Caribbean to blossom, not to mention any “homegrown” systems that develop from disturbances other than African waves. With El Nino weakening, I also have some concern for later season developments in the Caribbean, as by October we will likely be in a La Nina, which could allow low shear conditions to last longer into the fall than normal.
It’s interesting to note that some climate models, including the CFS, actually show significant dry anomalies over the Caribbean for the peak of hurricane season. These models seem to indicate a stronger than normal and somewhat northward displaced Inter-Tropical Convergence Zone (ITCZ), a band of convection that circles most of the globe year-round in response to trade wind convergence near the equator. It shifts northward and southward as the seasons change, peaking in northerly latitude around late summer in the Northern Hemisphere, corresponding with the peak of hurricane season. A strong and/or northward displaced ITCZ could favor systems being less likely to move into the Caribbean, as their more northward initial position lessens the chances that they will be steered westward all the way to the Caribbean, which is particularly true if they develop into tropical cyclones early on. As of now, however, I don’t really buy this scenario. Taking the CFS as an example, it’s very easy to spot a hypothetical “prevailing path” for tropical cyclones using precipitation anomaly forecasts. This can be confirmed with sea-level pressure forecasts, as the projected strength and placement of the “Bermuda High” can tell a lot about where tropical cyclones will track. Unfortunately, it is very difficult to forecast these features months in advance. The CFS forecast is predicated on it correctly predicting several features in succession, such as the northward displaced ITCZ, a strong but eastward displaced Bermuda High, and an area of warm SSTs it shows north of the MDR. In addition, given it's likely far less than perfect ENSO forecast, I'm hesitant to trust it on much of anything right now. As of now, I will go against these predictions, as I think the CFS is suffering from multiple issues largely stemming from poor prediction of ENSO conditions and Atlantic SST conditions. However, I will give some respect to its forecast of the prevailing sub-tropical high pressure pattern, which will be shown in graphical forecasts below.
Here's a look at how I think activity will pan out spatially. "ACE" refers to "accumulated cyclone energy", a metric of activity often considered superior to the traditional number of storms, hurricanes, and major hurricanes. It takes into account storm strength and duration, so, for example from last year, a long lasting major hurricane like Joaquin will produce significantly more ACE than a brief one like Danny.
Next, some possible prevailing storm tracks. Much of this isn't really groundbreaking- it basically just says the storms could go anywhere-, but my emphasis is that I think that while the A-B (Azores-Bermuda) High will be displaced somewhat north/east like the models show, it likely won't be as big a factor as usual, since many storms will wait until they are further west to develop, along with homegrown storms.
Now, a look at what the CFS is showing for August, and some of my thoughts on it. **Note (added 4/4): This is pre-upgrade. However, even with the upgrade, the CFS doesn't look right to me for the peak of the season.
Actual Numbers and total ACE: Now for some concrete values. Keep in mind that we’ve already had one hurricane this year, Alex, so it will count towards this season’s total and is factored into my forecast. Also keep in mind that the formation of Alex in January has no bearing at all on activity for the rest of the season. With that said my numbers are as follows:
*11 Named Storms
*2 Major Hurricanes
*Ace 75-90% of normal
So-called "analog" seasons can also be used to help determine activity in the coming season. Finding past years with a close match to the current year sometimes yields tremendous results in a seasonal forecast. But it is far from a perfect science, and no two years, no two weather patterns, are the same. I could go pretty deep into finding the best analogs- there's decades of records to pick from. However, I'll keep it fairly brief, as the analog approach is not a favorite of mine for hurricane season forecasting. The obvious places to look, given the state of the ENSO, are for the small handful of years which featured strong El Nino events that gave way to La Nina events. Some have picked 1998 as a good analog, as it followed the collapse of the 1997 El Nino. However, the Atlantic SST analog is not a great match, and 1998 was likely more active than this year will be. 1983 could also be looked at, as it followed the demise of the 1982-83 El Nino. That El Nino decayed slower, and likely had some impact on the 1983 season, which was the least active on the reliable records. This is a better match to the Atlantic SSTs, but may not be as good of a match to the ENSO picture. The actual answer likely lies somewhere in between these years, perhaps something like 1988, which followed the strong 1987 El Nino, and saw a change to La Nina during the spring and summer. That year featured numbers close to what mine are (it had 12/5/3), and also illustrates my prevailing storm tracks quite well. SST-wise, it's not an ideal match- using August as a comparison month, the extreme cold anomalies in the Pacific west of Mexico and parts of central America are suspect- but it highlights what I think the MDR will look like quite well, shows the warm Caribbean, and has the warm mid-latitudes with the cold waters to the far north in the Atlantic.
Here's the August 1988 SST map, and the storm tracks in 1988. On SSTs, the East Pacific is iffy as I said, likely too cold on this map, but I think the Atlantic is fairly accurate. The waters around the Bahamas and really the whole band around 20N should be warmer than this, and the 1988 storm tracks map doesn't quite give credence to what I think is a sizable threat for the East Coast this season.
I expect a slightly below average hurricane season owing to a continued lack of instability in the Atlantic, and a generally unfavorable sea surface temperature pattern associated with a near neutral AMO. However, I believe the odds of tropical cyclone impacts to land masses such as the Caribbean islands and the United States are at least as high as, if not higher than, normal. The primary reason for this is that while conditions over the MDR east of the Caribbean are likely to be less than ideal, the dissipation of El Nino should allow for more favorable conditions in terms of shear and moisture in the Caribbean and Gulf of Mexico. This both improves the chances for late-developing tropical waves to flourish in these regions, as well as increases the chances for "homegrown" developments to occur.
Keep in mind that, as I have mentioned, forecast confidence, at least for me, is below average this year. It is still quite early to be speculating on a season which doesn’t officially begin for over two months. I chose the past week to put this forecast together simply because it was spring break week, and I had some time to focus on work outside my academics. I may issue another update on my thoughts in May as hurricane seasons approaches. And as always, no matter what the year or what the preseason predictions are, if you live in a hurricane prone location, always be prepared In advance. Have a plan, know what to do and when to do it, and check in frequently with official forecasts throughout the season as storms start to develop. I look forward to blogging with you all over the summer!
Thank you very much for reading! Feel free to leave comments or questions below. Enjoy the spring season!
The views of the author are his/her own and do not necessarily represent the position of The Weather Company or its parent, IBM.