2019 Hurricane Outlook

Subtropical Storm Andrea

On May 17th, the National Hurricane Center (NHC) began forecasting the formation of an area of low pressure south of Bermuda which had the potential of becoming a short-lived tropical or subtropical storm before conditions became hostile for development. The next day, a large area of showers and thunderstorms developed over the southwest Atlantic in association with a decaying frontal boundary which gradually became better organized over time. On May 19th, a broad area of low pressure formed and an air force reconnaissance aircraft was tasked with investigating the system on May 20th due to this increased organization. This mission was able to identify a closed circulation and a small swath of gale force winds removed from the center. These finding in addition to the known proximity of an upper level low warranted the issuance of advisories by the NHC on Subtropical Storm Andrea at 22:30 UTC on May 20th. This made 2019 the fifth consecutive year in which a named storm formed prior to the official start of hurricane season in the North Atlantic. Andrea peaked in intensity during the following hours and began to weaken as dry air intrusion from the increasingly hostile environment began to disrupt the storm ultimately causing the storm to become a remnant low later the next day.

The North Atlantic hurricane season occurs from June 1st to November 30th and averages around twelve named storms, six hurricanes, and three major hurricanes based on the 1981-2010 mean. These tropical cyclones typically form in regions of copious moisture, strong instability, light upper level winds, and sea surface temperatures of 26.5 ⁰C or greater. In this hurricane outlook, I will discuss the multitude of variables that can influence tropical cyclone activity in the Atlantic basin as well as illustrate a general idea of what could be expected from this hurricane season given these variables.

I. AMO

The Atlantic Multidecadal Oscillation (AMO) refers to the variance of sea surface temperatures (SST) compared to average across the North Atlantic Ocean. The AMO follows a general 65-80 year period that has an amplitude close to 0.4 ⁰C that has been observed for over 160 years. A warm AMO phase has been associated with above average tropical cyclone activity due to the above normal warmth observed in the Atlantic’s Main Development Region (MDR) where a majority of Atlantic tropical cyclones form. The cool phase is associated with below average tropical cyclone activity in the North Atlantic due to the MDR being cooler than normal.

In 1995, the Atlantic Multidecadal Oscillation made a notable transition from a cool phase to a warm phase which has been attributed to starting an active period in Atlantic tropical cyclones. The AMO remained predominantly in the warm phase throughout the remainder of the 1990s and through the 2000s coinciding with copious above normal hurricane seasons such as the record-breaking 2005 season. However, this changed in 2014 and 2015 with the onslaught of a warm ENSO event and a persistent positive North Atlantic Oscillation (NAO) winter pattern which caused the tropical Atlantic to cool significantly. The strong cool signal that occurred over this time period began speculation that it could be the beginning of a more permanent shift of the AMO to a cool phase. However, it was noteworthy that the MDR warmed during each summer to near average. 2016 served to quiet some of this speculation as the tropical Atlantic warmed through the summer giving way to a warm AMO and the first above average hurricane season since 2012. This warm state continued in 2017 as the tropical Atlantic warmed to record temperatures through the late spring and summer fueling the hyperactive season during that year. However, when comparing these two years to other warm AMO years, it was apparent that the subtropical Atlantic warm pool and North Atlantic cool pool present during this period was anomalous to the traditional warm AMO SST profile.

In 2018, persistent ridging caused significant upwelling across the tropical Atlantic resulting in the coldest AMO measured since 1995. This upwelling was not just a detriment to surface temperatures, but reduced the depth of the thermocline across the tropical Atlantic. The thermocline is what is referred to as the layer of subsurface water where temperatures rapidly decrease with depth and is typically used as an indicator to determine depth of warm surface ocean waters. When the thermocline is shallow, like we have observed since 2018, it is much easier for cold deep ocean waters to mix to the surface during periods of ocean surface turbulence, or upwelling. Through the first five months of 2019, the SST profile across the North Atlantic has resembled much more of a negative AMO configuration like 2018. However, the winter pattern has been much less conducive for upwelling and thus the below average anomalies across the MDR have been relatively meniscal. The most anomalous portion of the May SST configuration is the spatial coverage of warmer than normal anomalies off of the Southeast United States which has helped to exaggerate the negative Atlantic tripole in appearance (Figure 4). Nonetheless, the MDR SSTs remain near normal which increases the probability that changes occur during hurricane season like what has been observed during the last several years. Some hints of these changes are now underway in the Atlantic with recent cooling of subtropical warmth near the Southeast US as well as the gradual warming of the MDR and northeast Atlantic. This has provided a more transitional look to the North Atlantic SST configuration which, with enhanced African Monsoon activity, would tend to enhance warming across the Tropical Atlantic through the summer and into fall. Another observation to note in the Tropical Atlantic is the recent downwelling that has gradually occurred during the month of May (Figure 5). This increasing warmth in the subsurface is important to future warming of the region and would serve to increase the depth of the thermocline to where ocean surface turbulence would have less impact on the sea surface temperatures as we approach the peak of trade wind season.

Regardless, the continued evolution of this SST configuration over the next six to eight weeks will be crucial to TC favorability in the basin during peak season. This is because July is the climatological peak of trade winds in the Tropical Atlantic which tends to enhance Saharan Air Layer (SAL) activity. Furthermore, during above average trade seasons like 2018, this can promote upwelling which cools the MDR greatly. Therefore, if sufficient warming does occur across the MDR and is not impeded by July trade winds, then we would expect TC activity in the MDR to be enhanced by warmer than normal SSTs during peak hurricane season.

II. ENSO

The El Niño Southern Oscillation (ENSO) is defined as the variance of the SSTs in the eastern Equatorial Pacific Ocean. The period for ENSO is two to seven years and can vary upwards of 2 to 3 ⁰C in amplitude. This climate oscillation is one of the most influential in overall global climate and is associated with significant variations in precipitation in the continents neighboring the Pacific Ocean. ENSO is categorized by three different phases: La Niña, neutral, and El Niño. A La Niña event is considered the cool phase of the ENSO and is associated with weaker upper level winds across the tropical Atlantic due to the lack of convective enhancement of the Subtropical Jetstream (STJ) in the lower latitudes during the summer months. El Niño is known as the warm phase of the ENSO and typically favors upward motion across the eastern pacific and American continents. This upward motion not only enhances convection across the eastern Pacific which promotes a more active summer STJ, but also generally promotes downward motion across surrounding basins like the North Atlantic, Africa, and the maritime Continent.

Throughout the summer of 2018, ENSO warmed gradually as a series of downwelling oceanic kelvin waves propagated across the equatorial Pacific Ocean which enhanced warming across the Central and Eastern Pacific. This resulted in the development of El Niño during the fall of 2018 lasting through spring of 2019. However, this event has been low amplitude and has had a relatively minimal impact in the largescale global climate pattern. Additionally, it appears the lingering downwelling oceanic kelvin wave is beginning to abate and is being replaced by an upwelling kelvin wave in the subsurface which will tend to reverse much of the warming that has been observed over the prior months. This upwelling kelvin wave is likely to impact the eastern equatorial Pacific the most in the short term with cooling occurring as it begins to surface. However, there is a more notable and sustained region of warmth still present in the equatorial central Pacific which could assist in keeping warmer anomalies in this region for a longer period of time. With these observations in the subsurface, as well as the current trends in climate forecast modeling away from El Niño for this summer, it is becoming increasingly likely that ENSO will return to relatively neutral conditions during the summer and into the fall this year. Some uncertainty does exist in this forecast, however, depending on the rate of decay of the El Niño event as well as the largescale atmospheric variability that occurs with intraseasonal forcing such as the Madden Julian Oscillation (MJO). With strong MJO activity near the maritime continent, like what was observed during the first half of May, this enhances westerly winds across the western Pacific and can serve to initiate genesis of a downwelling kelvin wave. Currently, there are no signs of this occurring in the subsurface Pacific as the upwelling kelvin wave signal still remains notable in this region, but any development of such a downwelling kelvin wave could serve to return ENSO back to a warm state this fall. This is exemplified well in a scenario such as 2002 in which subsurface ocean temperature profiles look similar to 2019, however, due to genesis of a new downwelling kelvin wave, 2002 went onto feature a moderate El Niño event through the fall.

Based on the general consensus of climate modeling, a sustained or gradual decay of El Niño is the most likely solution. This is represented in the Climate Prediction Center forecast in which they give a 55% chance of El Niño continuing through peak hurricane season. Regardless of the specifics of whether El Niño is maintained through summer or not, weak El Niño events such as this one hold a much weaker correlation in the inhibition of North Atlantic tropical cyclone activity. In addition to this, the aforementioned retained warmth in the Central Pacific, even as the Eastern Pacific cools, could tend to enhance convective activity further west across the tropical Pacific and behave similarly to that of a Modoki El Niño especially during the summer months as decay progresses. In Modoki El Niño, convective activity is most favored across the central equatorial Pacific thus enhancing the Subtropical Jetstream (STJ) across the tropical Eastern Pacific which inhibits tropical cyclone formation much like what a traditional El Niño does to the North Atlantic. The lack of significant inhibition of TC activity in the North Atlantic during peak hurricane from ENSO would also put more emphasis on the importance of intraseasonal forcing and the favorability of SSTs in the North Atlantic when making a forecast for this hurricane season.

III. PDO & Tropical Pacific

Another aspect of the Pacific’s impacts on Atlantic tropical cyclone activity is the Pacific Decadal Oscillation (PDO). The PDO refers to the variation in SSTs in the subtropical Eastern Pacific. In the positive phase of the PDO, waters are warmer than normal across the eastern part of the subtropical Pacific near the west coast of the United States with cooler waters in the central North Pacific. In a negative PDO, the SST profile is the opposite with cooler than normal waters off the US west coast.

For the past few years, the PDO has been positive with an anomalous area of warmer water off the western US coast. In 2018, the PDO observed record warmth across the subtropical and tropical Pacific east of Hawaii. With significant warmth in this region contrasting against well below normal SSTs in the tropical Atlantic, upward motion was favored over the Pacific which contributed to record TC activity in the Eastern Pacific through the summer months. With enhanced upward motion over the Pacific basin, much like you would see in a warm ENSO event, this favored an invigorated and further south oriented Subtropical Jetstream (STJ). With the amplified STJ over the MDR, this became an impediment to tropical cyclone intensity and longevity in this region which resulted in below average activity across the Caribbean Sea. Since 2018, the positive PDO SST profile has weakened and become less traditional with cooler waters apparent near the western coast of Mexico while subtropical warmth remains across the North Central Pacific. This will likely favor a displacement of TC activity in the Eastern Pacific basin further to the west and serve as an inhibitor to activity in the far Eastern Pacific. With this displacement of tropical cyclones, it would also continue to promote the idea of a more Modoki El Niño type of climate background in which upward motion is favored closer to the International Dateline and thus Eastern Pacific tropical cyclone activity would be expected to be closer to average or even below average.

IV. AEJ & African Sahel

The African Easterly Jet (AEJ) refers to the region where easterly trade winds are enhanced due to the contrast in temperatures between the Sahel and the Gulf of Guinea during the summer. This jetstream is known for transporting African Easterly Waves (AEWs) across the African continent and into the Atlantic Ocean. AEWs that emerge into the Atlantic account for the majority of tropical cyclone genesis that occurs in the basin. The differences in placement of the AEJ can have significant effects on the overall capabilities tropical waves will have in the basin. When the placement of the AEJ is further north, AEWs are enhanced due to the increased latitudinal distance from the equator that allows for greater vorticity. Contrarily, a southerly displaced AEJ is associated with the suppression of AEWs which inhibits further development once they have emerged into the Atlantic. Good indicators on the favored placement of the AEJ as we approach summer are differences in western African rainfall and SSTs in the Gulf of Guinea. In years like 2017, the African Sahel was much wetter than normal which enhanced AEWs and in-turn tropical cyclone activity in the basin.

This wetter than normal pattern across the West African Sahel continued through 2018 which once again enhanced the AEJ and was a large factor in the tropical cyclone outbreak that occurred during peak hurricane season. This TC outbreak contained several tropical cyclones which formed off of Africa including storms such as Hurricane Florence. Furthermore, this TC outbreak portrayed the importance of intraseasonal forcing and its impact on AEWs. In 2019, the African Sahel observed its wettest winter in recorded history as evidence to surface precipitable water being at all-time highs throughout the first five months of the year (Figure 9). This indicates that the African Sahel is observing one of its wettest years on record for the third consecutive year and thus this will play a significant role in enhancing the AEJ and West African Monsoon throughout the summer. With this being said, it is vital to note how important the placement of intraseasonal forcing and overall climate background is on the African continent. Therefore, the propagation of the MJO and Convectively Coupled Kelvin Waves (CCKW) will be important to note during hurricane season as this will play a role in timing of when the African continent will be at its most convectively favorable to produce strong AEWs capable of becoming tropical cyclones over the Atlantic. This was an important factor in both the TC outbreaks observed during September 2017 and 2018.

In addition to the wetter than normal African Sahel, another factor that could enhance convective activity near the African continent is the Indian Ocean Dipole (IOD). The IOD is similar to that of aforementioned climate oscillations, however, it is an aperiodic oscillation of SSTs classified as positive, neutral, and negative phases. In the positive phase, warmer than average SSTs and greater precipitation is present in the western Indian Ocean with corresponding cooler waters in the eastern Indian Ocean which causes drier conditions near the Indonesia and Australia. Alternatively, a negative IOD features the opposite conditions with warmer waters and greater precipitation in the eastern Indian Ocean, and cooler and drier conditions to the west. During any given year, the IOD is not typically strong enough to be a considerable factor to the global climate background. Nonetheless, 2019 has a strong positive IOD signal which favors upward motion over the western Indian Ocean and eastern parts of Africa. This signifies yet another potential factor in favoring upward motion over Africa which would enhance the African monsoon and AEWs during peak hurricane season.

V. Seasonal Analogs

The base features as noted above when searching for seasonal analogs for the upcoming hurricane season were a warm neutral ENSO or weak El Niño, a neutral or quasi-positive AMO SST profile in the Atlantic, a positive IOD, a positive PDO, and an anomalously wet African Sahel.

The years that best fit this profile are 2003, 2004, 2017, 1961, 2002, and 2012. The average tropical cyclone activity of these six analogs are 15 named storms, 8 hurricanes, 4.3 major hurricanes, and a seasonal ACE value of 172.3. This would be considered as a well above average signal and another sign that the Atlantic is likely to be more favorable for tropical cyclone development this hurricane season. Within these analogs, 2002 is an outlier and would be considered as the lower end solution in which El Niño remains and the MDR does not warm resulting in slightly below average activity. Alternatively, 2017 would be considered as the higher end solution in which ENSO cools quicker and the MDR warms rapidly throughout the summer resulting in hyperactive tropical cyclone activity in the North Atlantic basin.

When looking ahead to the peak of hurricane season in these analogs, we see that the eastern equatorial Pacific cooled while the central Pacific remained warmer supporting the solution for a weakening or Modoki El Niño. Furthermore, the Tropical Atlantic warmed during the summer in most of these years which, in addition to the wetter than normal African Sahel, is a large reason for the abundant amount of TC activity in the MDR from strong AEWs emerging off of Africa during these years (Figure 12). Based on the changes presently observed in ENSO and AMO, it is likely that 2019 will generally follow these analogs as peak hurricane season approaches (Figure 13 & 14). As another note, the analog average for Eastern Pacific hurricane activity, excluding 1961 for accuracy, yields 15.6 named storms, 8 hurricanes, 3.6 major hurricanes, and a seasonal ACE of 89.8. This is considered to be below the climatological average in the Eastern Pacific which correlates with an anticipated above average North Atlantic hurricane season.

An additional analog case that can be evaluated for 2019 is 1876-1881 which followed a rather similar ENSO progression to that of the 2014-2019 and therefore 1881 could be used as an analog for how ENSO evolves through the remainder of the year (Tables 1 & 2). This analog also agrees with the aforementioned forecast for gradual decay of El Niño in the upcoming months as the equatorial Pacific returned to ENSO neutral by the peak of hurricane season during that year. Although this analog is great for getting a better understanding of the possible outcome with ENSO, it does not give a fair representation of the potential amount of tropical activity that could occur in the North Atlantic largely due to the lack of accurate records prior to the satellite era. With this being said, there were still a couple of long-lived hurricanes that made landfall on the southeast United States during the 1881 hurricane season.

With the expected gradual decay of the weak El Niño event, this would likely enhance favorability for tropical cyclone development in the North Atlantic this hurricane season. Additionally, the wetter than normal African Sahel, and subsequently the active AEJ, will enhance easterly waves as they emerge off of the African continent. This would lead to a higher abundance of tropical cyclones forming in the eastern Tropical Atlantic as observed in 2017 and 2018, respectively. The largest part of forecast uncertainty remains to be the changing state of the Atlantic SST profile, which appears to be trending toward a more positive AMO currently, but is still overall nearly neutral throughout. If further warming of the MDR were to occur this would be beneficial to tropical cyclone development by creating less limitations for AEWs as they emerge and propagate westward. The extent of this warming is crucial to monitor over the next six to eight weeks given the importance of timing with the climatological peak of trade winds across the tropical Atlantic which can serve to upwell surface waters as well as limit longwave radiation at the surface due to Saharan Air Layer (SAL) outbreaks. Therefore, the strength of the trade winds this July will be important to monitor as well since they will play an important role in how the Atlantic SST profile is shaped as peak season approaches. This is a large difference maker in the analog years selected as they all largely observed warming of the tropical Atlantic throughout the summer months which led to above average to hyperactive hurricane activity. Alternatively, if trade wind and SAL activity is above normal then cooling of the tropical Atlantic would occur like it did during July 2018. Even then it is still possible for the Tropical Atlantic to rebound through peak hurricane season, as 2018 did, which would limit the inhibition on tropical cyclone activity the cooler waters would have. In this scenario in which ENSO remains near neutral and the Atlantic does not warm, it is still likely that hurricane activity would be above average in large part due to the active AEJ. This is accounted for in the range forecast in which the lower end of the range accounts for little warming of the MDR while the upper end of the range illustrates the scenario in which the MDR warms and a more classic positive AMO is observed.

During the past few years of these hurricane outlooks being written, I have been keeping track of how these forecasts verify to evaluate areas of each forecast that may not have been up to par. When looking at this analysis closely, I noted a much larger margin of error in my Accumulated Cyclone Energy forecasts that were far worse than the errors for other numerical values such as number of named storms. This became the motivation to finding alternative ways to better estimate the ACE relationship with the other numerical variables in which forecast errors were typically lower. To do this, statistical linear regression methods were utilized to evaluate data from 1950 to 2018 in order to understand what kind of relationship numbers of named storms, hurricanes, and major hurricanes had with the amount of accumulated cyclone energy units. In the figures below are plots that give a better visual representation of the nature of these relationships (Figure 15).

As seen in these figures, the number of hurricane and major hurricanes had the most linear relationship and thus could be utilized using a simple slope formula to calculate general ACE values that occur with regards to the number of hurricanes and number of major hurricanes. Applying this to prior years in order to see how this relationship holds up generally, it is worth noting that the best performance with this formula lies away from the highest volume of data is located otherwise known as the climatological average (Table 3). Therefore, there is increased uncertainty closest to climatological average years due to the presence of more outliers. Some of these outliers can be identified when looking at the plot of ACE vs Number of Major Hurricanes in which we see a notable grouping of years with ACE above the 95% confidence interval for 2 major hurricanes. This is due to differences in the overall favorability of the basin that allow certain years, like 2018, to have higher seasonal ACE values due to the abundance of short-lived storms in the subtropics and long-lived storms that don’t attain major hurricane status like Hurricane Leslie and Hurricane Helene. Additionally, it is worth noting that the decrease in data points away from the mean also creates a higher uncertainty for years that deviate from the mean such as 2005. This can be seen in the projection of 2017’s seasonal ACE versus the number of major hurricanes in which the formula projects a value of 200 units when in reality 2017 observed approximately 225 units of seasonal ACE.