partner login

BMS News

BMS Tropical Update 9/12/2018 10 AM CDT

Why is Florence Stalling and moving South?
The recent development of Florence stalling at landfall versus inland is a pretty big change in terms of the overall impact to the insurance industry and, frankly, I think the NHC best summed it up in their 5:00 p.m. forecast discussion Number 50 yesterday.

The steering currents have collapsed and could exacerbate any and all effects along hundreds of miles of coastline. This is not something the insurance industry wants to hear. It seems counterintuitive, and it is an unusual scenario. Florence will have nowhere else to go but south or along the coastline into South Carolina this weekend.

The best technical description might be from Doug Simonian, a meteorologist at www.nymetroweather.com.

So, what does this all mean? The 500mb layer, which is roughly 18,000 feet in altitude and known as our midlayer, is often used to look for steering currents that drive hurricane direction. This has bridged the steering currents to the north, which means Florence has nowhere to go but stall or move southwest. In the short term, it still seems like it is certain that Florence will touch the North Carolina coast near Cape Fear, NC and then drop south. After 72 hours, however, it’s anyone’s guess as to what direction Florence will go, which is why the NHC cone of uncertainty continues to balloon along the coastline.

The last four runs of the ECWMF ensembles guidance that shows a constant step south and slower landfall as Florence approaches the coastline. This is a remarkable turn of events as almost no ensemble members were forecasting a southward component of motion just 36h ago. Now almost every member does. Source: Weathernerds

 

Over the past day a blend of the best weather models called the TVCN has latched on to the idea that Hurricane Florence may skim the coast of the Carolina’s for 2-3 days before moving inland. The newest model run is the darkest blue. Source: Univ of Albany

The Good: Based on this forecasted track change, the overall inland impact from heavy rains and tropical storm-force winds has diminished greatly for North Carolina. Another potential outcome is that as the storm slows towards the coastline it will create upwelling, causing the storm to weaken and, therefore, unlikely to stay a major hurricane any time after Friday morning when it is near the North Carolina coastline. However, if Florence can stay over water, it will maintain its hurricane structure and could still be at Category 1 or 2 hurricane strength for another day or so. If it moves inland, expect Florence to quickly be downgraded to a tropical storm. Finally, another positive note of a slower moving storm towards landfall or along the coastline is that it would likely draw in dry continental air, thus weakening the storm as it stalls. There are still signs that stronger wind shear will also help weaken the storm as it nears the coastline.

The Bad: The overall winds will still be strong right along the coastline near southern North Carolina and, if the storm tracks along the coastline, it would just put that much more property exposure at risk to hurricane-force winds. The fact that hour upon hours of hurricane force winds could occur can be just as destructive as a short period of higher wind speeds.  This also shifts the heaviest rain and flooding along the entire coastline, and major flooding is still expected along the coastal inland areas, with heavier rains now forecasted over South Carolina and maybe even into Georgia.

Population Changes Along The Coastline
The insurance industry is well aware of the exponential growth that has occurred along the coastal areas of the U.S. The graphic below illustrates the population growth between Myrtle Beach, SC and Wilmington, NC in 1940, 1990 and 2018. This rapid increase in the coastal population will increase potential insured losses from both wind and storm surge if Florence decides to track along the South Carolina coastline this weekend versus moving inland as a weaker tropical storm. It should be noted that a large portion of the population in this region has not experienced a hurricane of this magnitude since Hugo in 1989, and very few likely lived in the area during Hazel in 1954, the most recent Category 4 hurricane to impact the region. Thankfully, these analogs are no longer relevant to Florence’s forecast.

Exposure change in region forecast to be affected by Florence. While much of the development has been along the coast, most areas inland are subject to major flooding. Source: Stephen M. Strader

 

BMS iVision allows clients to run risks to better understand various impacts from Florence. This is the newest 3 sec wind speed gust in MPH from Verisk Weather Solutions. Notice the expanding swath of high winds along the South Carolina coastline.

 

The newest rainfall guidance suggesting a large area of 20 inches of rain right along the coastline. Fortunately given the new track the impact in the mountains, where orographic lifting could have cause enhanced rain and mudslides is more limited with the newest forecast. Source NOAA/NHC

Storm Surge
There are known impacts of storm surge which are well modeled in the NHC Potential Storm Surge Flooding Maps, the best resource at this time. There is really no good news in this scenario, and the scenarios only get worse. Initially, the storm surge scenario still plays out with the concave shape of the North Carolina coast being a significant contributor to storm surge, which allows water to pool while being directed inland. Therefore, at this time, the worst of the storm surge will likely occur along the southern South Carolina coastline. If Florence stays along the coast, it means hours of onshore winds over multiple tide cycles creating even bigger problems, and if the storm moves down along the South Carolina coast, it will see some even higher water heights. But, the storm surge impacts are highly dependent on the final track.

It is very important to click here for the detail in the storm surge forecast. Source: NHC

To better understand the impacts upwelling might have on Florence’s intensity as it stalls near the coastline, data from Air Launched Autonomous Micro Observer (ALAMO) can be used. These floats are launched out ahead of the storm by hurricane hunters to get observation of the thermal energy in the ocean, which is a prime driver for hurricane strength. As the hurricane passes, the data is collected in real time. Last night, Florence went right over ALAMO floater #9136, which showed substantial vertical mixing of the water column with the water going from 29 degree c to 27.5 degree c. It is this same vertical mixing that will likely weaken Florence along the coastline.

Tornado Risk
There has been very little discussion around the tornado risk from a landfalling hurricane. Currently, the Storm Prediction Center is not indicating any increase in tornado potential. Just last year, five named storms were responsible for 119 reported tornadoes across the U.S. Hurricane Harvey had the largest number of tornado reports with 57, mostly clustered near the southeastern Texas coast as it stalled out. This could potentially happen as Florence stalls near the coastline over the next few days.

Plotted on the map are the tornadoes associated with the hurricanes and tropical storms that affected the U.S. in 2017. Source NOAA Storm Prediction Center

Insurance Loss Guidance
Given the change in forecast, any initial loss guidance has to be thrown out and new events need to be pulled as Florence will likely be a weaker storm. With the forecasted movement along the coastline, it will be challenging to find a scenario that would match a stochastic event out of catastrophe models. There might be events that moved slowly along the coastline in a northeast direction or even stalled like Hurricane Diana in 1984. However, those types of storms are entirely different than a storm that stalls and moves to the southwest along the coastline leaving the strongest winds along the coast versus the weaker winds, which would typically be observed in a northeast moving hurricane.

Other Tropical Troubles
A sneaky tropical storm is likely to be spinning up right before landfall near the southern Texas coastline. This Texas threat is not being well advised at this point, as all eyes are now on Florence. The NHC currently gives this tropical wave a 60% chance for genesis over the next 5 days. Its name would be Joyce if development does occur. At this time, I don’t expect it to become a hurricane prior to landfall, and it should mostly just be a rain event for coastal Texas. However, it’s important to watch.

Now Tropical Storm Isaac, which is currently a tropical storm 450 miles from Martinique, also needs to be watched. Although the NHC does not currently have this as a hurricane, some models are suggesting that Isaac could become a hurricane over the Caribbean. These models also show a risk for it to track into the Gulf of Mexico. This is something to watch moving forward and will depend on Florence’s final movements.

ECMWF ensemble forecast suggesting in the long range that Isaac could track up into the very warm waters of the Gulf of Mexico. Source Univ of Albany 

BMS Tropical Update 9/11/2018 12 PM CDT

T- 72 hours
Although yesterday I mentioned that Florence is not just a point on a map and will have a wide-ranging impact far away from the center the storm, the center does still matter. With Florence about 72 hours away from its expected landfall around 7:00 a.m. AST on Friday, there is still some uncertainty regarding the exact landfall location. Florence continues to fluctuate in intensity with winds of 130 mph, along with stronger gusts. The diameter of the most destructive winds, known as the radius of maximum winds, is currently only 10 miles, but will likely continue to grow. The plot below is from our  BMS iVision application with the Verisk Weather Solutions wind one minute sustained swath. This is only one possible forecast scenario, but it provides an idea of what the wind field could be. This will shift north or south based on the eventual track, with the strongest winds likely being on the right side of the track. If these maximum winds occur over a more populated area, the damage will be much greater. For example, the losses from last year could’ve been a different story if the strongest winds from Hurricanes Harvey and Irma hadn’t just missed the populated areas of Corpus Christi, TX, and Miami, FL, respectively. At this point, the forecast still holds that Florence will make landfall between Myrtle Beach and Morehead City, NC as a strong Category 3 hurricane with winds of 120 mph.

BMS iVision allows clients to run risks to better understand various impacts from Florence. This is the one minute sustained wind in MPH from Verisk Weather Solutions, which shows only a very narrow band (33 miles) of damaging wind speeds.

 

Starting to see increased ensemble members push the remnants of Florence into the eastern Ohio Valley later next week, which is quite far inland and will add to the flooding concerns. Source:  http://www.atmos.albany.edu/facstaff/tang/tcguidance/al062018/

 

Several hurricane intensity models suggest Florence is still expected to maintain at least Category 4 hurricane status for the next three days as it heads toward the North Carolina coastline. The map is from weathermodels.com and Ryan Maue shows there is plenty of ocean heat content.

 

North Carolina Wind Zones with the shaded areas representing the International Building Code Standard. Source: http://www.ncdoi.com/Media/Documents/WindSpeeds.pdf

If Florence comes ashore with winds of 140 mph as previously forecasted by the NHC, this would be above the International Building Code standards, which are the strictest right along the coast.  However, hurricane-force winds will likely extend up to 30-60 miles inland, with tropical storm-force winds extending even further. These wind gusts will be strong enough to cause power failures as far north as Virginia and, given that it has been decades since a major wind event has impacted this area, it’s likely that a lot of old, damaged and diseased trees will be cleaned up. These strong winds, combined with the wet soil conditions over the last month, may lead to a high likelihood of wind-related treefall damage.

Unlike areas of south Florida or south Texas, areas along the Southeast coast have a fairly high tree density which will compound potential damage from tree fall and power outages. Experience suggests even from minor hurricanes like Hurricane Hermine that made landfall in 2016 in the Florida Panhandle, higher tree density over an area that has not experienced a hurricane in a while can create major power outages.  Source: Max Galka

Flooding Potential
It appears that the forecasted track for the center of the storm and the NHC cone of uncertainty are getting closer together. The primary reason for the slow movement after landfall is a strong blocking ridge of high pressure to the north. This will cause Florence to potentially behave erratically this weekend, similar to what happened with Hurricane Diana off the Southeast coastline in 1984. The major issue with stalling is that a large amount of rain is forecasted to fall. The spread in model rainfall forecasts has increased since yesterday. It is likely that a maximum of 20 inches could fall, but this all depends on how long Florence stalls and what type of orographic influences occur in the more complex terrain further inland. The North Carolina state record of rainfall from a named storm is 24.06 inches from Hurricane Floyd in 1999.

Because of the wetter-than-average conditions recently, dams may not be able to contain the rainfall and areas that have never flooded before will likely flood. Unfortunately, in October 2016, Hurricane Matthew brought heavy rainfall and flooding to North Carolina. The story is the same after every major flood event – many of those affected were without flood insurance. North Carolina is a leader in floodplain mapping and risk communication, and the recent major flooding from Hurricane Matthew provided a reason for consumers to get flood protection. According to the following data, nearly 300 North Carolina towns, cities, and counties increased the number of NFIP policyholders since 2016. Seventy remained steady, and 130 saw decreased participation. Not surprisingly, coastal communities are the biggest participants in NFIP. Most notably, Fayetteville had a whopping 65% increase in the number of policies, albeit only to 1,290 (the 2010 Census documented 87,000 housing units in the city), which still shows a very large protection gap. Other large cities saw increases, including 10%+ increases in Charlotte, Durham, Raleigh, and Greensboro. If there is any good news here, it’s that more flood coverage has been provided and NFIP has reinsurance to aid in its expected losses. However, large protection gaps still remain.


Storm Surge
The NHC has started to issue watches and warnings this morning, and with that they have issued the first storm surge watch for Florence. The combination of a dangerous storm surge and the tide will cause normally dry areas near the coast to become flooded. Surge-related flooding depends on the relative timing of the surge and the tidal cycle, which can vary greatly over short distances. The NHC has now started to issue storm surge inundation maps and they should be watched closely. It’s likely that the current inundation levels of greater than nine feet above ground will likely grow.

NHC Potential Storm Surge Flooding Map

 

Insured Loss

There are several reasons why the insurance industry needs to understand the potential loss from Florence. Insurance companies are looking to understand the size of loss relative to their reinsurance placement. They also might be looking to buy top-up covers or 3rd or 4th event covers, if needed. There is also the need to understand where to stage claim adjusters or if third-party adjusters might be needed. The reinsurance industry is keen on triggers of ILW, Cat Bonds and retro needs. There is also an active LiveCAT market for the industry to buy or sell extra capacity. With the understanding that some catastrophe modeling companies usually don’t issue insurance industry loss guidance until about 48 hours before landfall, curiosity often takes people to GIS and the various stochastic track sets to see what potential losses might be. However, Florence’s expected stalling makes event loss selection more challenging, as many of these stochastic track events often move into the Northeast as climatology would suggest. Therefore, only losses from the Southeast should be used to limit losses that might occur from events that track further inland up into the Northeast. Also, given the general cone of uncertainty 85 hours before landfall, the loss ranges can be quite large. This is primarily driven by events that track over more urban coastal areas versus those that track over less populated areas. There are clearly some large loss scenarios with events that could hit Myrtle Beach, SC, and Wilmington, NC, which could easily reach into the billion dollar range. The most extreme loss scenarios at this time are from Charleston, SC, which is slowly becoming an unlikely scenario.

Simple exercise of narrowing down a stochastic track sent from a catastrophe modeling vendor. The events here are based on the model range of landfall, by looking at different storm variables at the time of expected landfall.  The light blue tracks are tracks that are not as realistic due to the recurving of the storm after landfall.

Florence is not like most hurricanes. Take the example of Hugo 1989 which has been used as an analog hurricane to Florence. Besides the completely different tracks the speed & rapid intensification right at landfall brought Hugo winds well inland. This will not be the case with Florence which should stalling and dump days of rain. Source: Brad Parovich

 

BMS Tropical Update 9/10/2018 12 PM CDT

Hate Being Wrong
Meteorologists hate being wrong, but with the magnitude of potential disaster that is ahead for the insurance industry this week, I hope I am. I have been working in the insurance industry for the relatively short period of 16 years, but I have also spent a lot of time researching the historical impact of past hurricanes on today’s exposure. I am concerned that what may occur later this week will likely be talked about for decades to come.

Track and Intensity Details
We are now past the point of having to focus on which model is performing best, as the majority of the model guidance is in agreement. It is now time to focus on the fact that Florence is going to make landfall, but the main forecast question now is what will Florence do as it nears the coastline? A blocking high pressure ridge is expected to slow Florence down or potentially even stall the storm. A stalled storm will bring a prolonged period of additional heavy rain (think Hurricane Harvey last year).

Although there will likely be some track shift on the exact location of the strongest winds at landfall, which is a very important detail for estimating insured losses, this detail won’t be known for another 48 hours or more. The overall intensity of the storm will likely influence the final track, and the intensity forecast is just now starting to play out. The loss potential is much smaller if the strongest winds are focused south of Myrtle Beach over the Tom Yawkee Wildlife area than if the strongest winds occurred over Myrtle Beach, SC, Charleston, SC or Wilmington, NC.

After weakening from its Category 4 start last week, Florence is now undergoing rapid intensification, with many of the intensity forecasts showing a range between Category 3 to Category 5 over the next four-day forecast period prior to a potential landfall.

This is all driven by an environment of low wind shear and significant upper-level outflow, which are perfect conditions for building intensity in a named storm. If you add in the fact that Florence will be moving over the very warm waters of the Gulf Stream, which are already as much as six degrees warmer than average, there should be no reason why Florence wouldn’t maintain major hurricane status until landfall. A Category 5 hurricane is becoming likely at some point in the next few days. The only formalities will be internal eyewall replacement that will allow for small fluctuations in strength, but it’s a double-edged sword: a slightly lower wind speed will probably mean a larger sized storm.

Another thing to watch before landfall is an increase in wind shear. Some models are showing this, which would help weaken Florence slightly before its potential landfall. This is something that will be well known in the next day or two, with Florence still likely to be a large damaging hurricane at landfall.

NHC Hurricane Florence Wind Speed Probabilities of at least tropical storm force. More details here.

 

It is too early to provide wind swath estimates, but the ensemble of tracks provide an idea of landfall location with the strongest winds to the right of the storm’s track.

 

Hurricane intensity guidance

 

Potential Impacts
Storm Surge
Should some of the intensity forecasts become verified, a Category 4 storm impacting the coastline of the Carolinas would be dangerous and problematic. Although the NHC hasn’t yet begun to issue official surge forecasts, overall the Carolinas’ coastline is particularly vulnerable because of low lying barrier islands, inland waterways and bays that can amplify surge. Because there is uncertainty in the track, the best tool to use right now is the NHC’s Maximum Of Maximum (MOM) storm surge layer, which is an attempt to quantify the potential impacts of storms of different strengths at different angles. This is done by running an ensemble of simulated Category 4 landfall scenarios for a region to get an idea of the maximum storm surge that can occur over various sections of the coastline. It should be noted that this is not a Florence forecast. I expect the NHC to begin issuing better storm surge guidance on Florence by tomorrow.

A Category 4 storm making landfall from the southeast along the South Carolina/North Carolina border could bring surges in excess of nine feet to coastal regions. These layers are in BMS iVision for clients to get a better understanding of maximum impact.

Waves
The force of water can’t be underestimated and, depending on track, waves as high as 15 feet are expected to crash along the coastline and add to the storm surge impact.

The right side of the storm will feature the largest storm surge and highest waves, which puts the area north of Cape Fear at highest risk, with a general predicted surge plus tide of 12-15 ft above mean sea level and 9-12 ft above high tide.

Rainfall and Flooding
Unlike Harvey (in flat terrain), Florence is coming onshore to hilly and mountainous terrain. This will magnify rainfall from orographic lift and runoff will channel into devastating flash/riverine flooding as well as liquefy mountain slopes. The amount of rain will surely impact infrastructure. We have seen time and time again from past flooding events that 10-15 inches of rain over two days is beyond the design capacity of most storm water infrastructure, which creates even more problems beyond flooding on roadways, rivers, creeks and streams.

Based on the government’s quantitative rainfall forecast product, the maximum amount of rainfall that is forecast is more than 15 inches! What makes this even worse is that more rain than usual has fallen during the last 30 days over most of the area where Florence’s heaviest rains are forecast to fall.

The number of inches of rain above (green/purple) or below (reds/yellows) normal over the last 30 days. In much of western Virginia, Maryland and eastern West Virginia, it has been wetter than normal.

With already wetter than normal conditions, the excessive rainfall from Hurricane Florence will…
• Run off more quickly than usual into rivers that are already higher than usual.
• Weaken the soil which could increase tree fall and increase insured loss.

Keep in mind that the models do not do a great job with forecasting rainfall over more complex terrain. In fact, total rainfall forecasts are difficult, period. As an example, days before Hurricane Harvey’s torrential rain began in Houston, the maximum amount forecasted for the region was 29 inches. However, the heaviest actual amount was 60.58 inches!
I expect many rainfall and river level records for the area to be broken, including many set from 1973’s Hurricane Agnes, which caused record flooding in the Mid-Atlantic region with less than 15 inches of rain over Virginia and West Virginia. Again, let’s hope that the government meteorologists are wrong with the forecasted rainfall amount of 15 inches from Florence.

NOAA rainfall forecast for Florence with up to 15 inches of rain forecasted over the next 7 days. Also, watch the tropical system along the Texas coastline as a rain maker.

Flood Maps
A quick reference for potential flooding might be the FM Global Flood mapping tool, which allows users to understand areas that are prone to flooding in great detail.  Another resource might be the real-time flood network for NC.

Power Outages
Given the potential amount of rainfall and hurricane-force winds reaching areas well inland, the insurance industry should expect high amounts of tree fall that will produce power outages. In the past, the Guikema Research Group at the University of Michigan has produced power outage maps. I contacted them and they will be running their model today, but www.poweroutage.us will be the source to watch as Florence makes landfall.

Building Codes Matter
The importance of strong, well-enforced building codes was clearly demonstrated with the hurricanes of 2017. The Insurance Institute for Business and Home Safety (IBHS) provides technical guidance to inform and improve model building codes, advocates for timely adoption of national model building codes and standards, and encourages uniform enforcement of these codes. In its “Rating the States” report, South Carolina scored among the highest states in the country and North Carolina was just slightly lower. However, as we saw with Hurricane Irma and Florida, (one of the highest scored states in the country), building codes are important to limit damage but are a minimum requirement. The expected winds from Florence will likely be at or above the codes that are enforced.

Insured Loss
The catastrophe modeling companies won’t be issuing any stochastic model guidance until roughly 48 hours before landfall so, currently, the best loss analysis would involve one’s own event selection or historical loss simulations. The range in historical loss simulation is quite large due to individual storm tracks relative to the exposed areas, which will ultimately influence Florence’s loss as well. On the higher end of the loss scenarios are Hugo 1989 and Hazel 1954, which each surpass $20B of insured losses for wind and storm surge. Other scenarios are below this $20B of insured loss, but with lower intensity and landfall with the highest winds impacting less populated regions. It should be noted that none of the historical storms stalled like Florence is expected to do and do not account for flood related losses. The forecasted impact to the insurance industry will continue to be refined over the next few days, and those details will be shared in future BMS Tropical Updates.

 

Other Issues Tropical Troubles
While U.S. interests are focused on Hurricane Florence, Hawaii is under threat from Hurricane OliviaHurricane Isaac is intensifying further south, with impacts to the Lesser Antilles by late in the week as a weak hurricane. Hurricane Helene is also in the Main Development Region and will likely recurve later this week. Another area to watch later this week is a tropical system that may move up into the Texas coastline. There is a chance that we could have as many as five simultaneous Atlantic named storms this week. The record for active simultaneous Atlantic named storms is four, which was set several times, and the record for named storms including tropical depressions is six, set in September of 1971.  Hopefully, we won’t see that occur, since the Basin is now starting to shut down any new development for the next several weeks.

BMS Tropical Update 9/6/2018 4 PM CDT

Gordon Recap

There are limited reports of damage near the location of Gordon’s landfall. Current damage reports in Louisiana, Mississippi, Alabama, and Florida indicate only minor coastal flooding, power outages affecting up to 29,000 people and damaged trees. The insured losses from Gordon should be relatively limited due to the overall small concentration of high winds which were mostly on the right side of the center of the storm as it made landfall near the Louisiana/Alabama boarder on Monday night. The maximum sustained wind speeds at landfall were 70 mph, which is just shy of Category 1 on the Saffir-Simpson hurricane wind scale. Because Gordon was not a Category 1 hurricane at landfall, the major risk modeling companies will not be issuing any loss guidance, but losses should be lower than after Hurricane Nate, which impacted the same general area in 2017. The most severe damage seems to be mostly confined to coastal islands like Dauphin Island, AL, which have strict frame building codes and likely high wind deductibles. But seeing even minor damage to some of these structures suggests the properties are very vulnerable to wind and surge, and that it’s only a matter of time before the landscape changes, similar to what has been observed on the Bolivar Peninsula in Texas after Hurricane Ike in 2008.

BMS iVision allows clients to run risks to better understand various impacts from Gordon. This is the 3 sec wind gust in MPH from Verisk Weather Solutions, which shows only a very narrow band of damaging wind speeds.

Florence’s Future
As I mentioned in Monday’s post, there is a high degree of uncertainty in the long-range track for Florence. This is the very reason why the National Hurricane Center (NHC) only issues a five-day forecast, and even its five-day forecast, which is historically one of the most skillful, still has an error of 198 miles. Only 60-70% of the time, the track of the storm can be expected to remain within the NHC cone of uncertainty. Therefore, it is far too soon to speculate what, if any, impact Florence may have on the U.S. East Coast next week.

What a difference one day makes. Animation of Hurricane Florence one day yesterday and today. Strong southwesterly shear has disrupted the circulation significantly.

Florence continues to fluctuate in intensity and, amazingly, it popped to a Category 4 hurricane yesterday in very hostile conditions. Today, however, those unfriendly conditions are impacting Florence’s strength, as it is currently down to a small Category 2 storm. As mentioned previously, the weaker Florence is, the more likely it will stay on a westward track. In fact, the animation below shows Florence tracking through an area of dry, dusty air known as the Saharan Air Layer (SAL), but still maintaining its intensity as it moves in a more northwest direction.

At the same time Florence is tracking through an area of high SAL, it is also fighting an area of high wind shear. One of the best sites to understand current wind shear and SAL is via the University of Wisconsin Cooperative Institute for Meteorological Satellite Studies.

Above is the latest wind shear analysis suggesting the wind shear that Florence has been fighting over the last day is affecting the structure of the storm. The wind shear is expected to subside in a few days, allowing Florence to strengthen.

The various forecast model ensembles are currently the best tool to understand the forecast uncertainty. In the short term, the forecasts are usually very good, with only an error of few hundred miles. In the case of Florence, what happens beyond five days is anyone’s guess. Currently, the spread from these various ensembles is HUGE, showing possible tracks from Florida to Newfoundland. However, over the last several days, there is growing confidence in the ensemble forecasts that Florence will track westward – a growing threat to the East Coast of the U.S. But, the ensembles are exhibiting a high spread in all directions. The source for much of the uncertainty is a possible inflection point that could occur on Sunday as a trough of low pressure passing to the north could provide an escape route for Florence. If the storm misses this trough of low pressure, the chance of eastern U.S. impact increases, as a strong ridge of high pressure blocks any possible escape to the north and allows Florence to track westward until late next week.

My favorite site to track the probabilistic ensemble spread is the University of Albany, which also has tools to track current model skill. Currently, the TVCN (Variable Consensus) model is the best forecast to watch. Click here for latest morning model run of ECMWF everyday (shown above).  Click here for latest afternoon model run of the ECMWF everday (Usually available after 4 pm CDT).  European Centre for Medium-Range Weather Forecasts (ECMWF).  Notice how the probability is high (red colors) up to 72 hour forecast.  Probabilities along the coastal U.S. are still below 30% (Blue Colors), but increasing.

 

One interesting short-term observation in regard to Florence is that it has persistently tracked on the north/east envelope of the ECMWF guidance. Models are indicating a left turn near 26 degrees north latitude tomorrow, which will be very important to the future track and might limit Florence getting caught in the passing trough of low pressure to the north. But, the track in the models has been biased to the west, which is why Sunday movement is key to the future insurance impact along the East Coast of the U.S and Bermuda.

The other tools that can utilized at this time are historical storm tracks and climatology. Never in the history of the Atlantic Basin track set has a named storm in Florence’s current location ever reached the U.S.

The past does not necessarily predict the future, but it’s worth noting that only one named storm out of 89 within 200 miles of Florence’s current position have ever affected the East Coast of the U.S. If Florence does, it would be rather unique.

The same analysis suggests that of the 220 historical storms that have tracked within 200 miles of the NHC five-day forecast location south of Bermuda, only a handful have reached the U.S., with even fewer being notable events like the 1933 Chesapeake – Potomac Hurricane.

Caption: There have only been five hurricanes out of 220 within 200 miles of Florence’s five-day forecast position that have impacted the U.S East Coast.

Given the large amount of uncertainty, now is the time to prepare. The Insurance Institute for Business and Home Safety has nine key things insured homeowners should do to limit overall damage.
https://disastersafety.org/hurricane/reduce-hurricane-damage-to-homes/

In summary, the current shear is forecasted to take its toll on Florence. I expect that once Florence tracks through the high areas of wind shear, the combination of much warmer waters in the path of the hurricane will strengthen and expand the storm. Expect Florence to be stronger in intensity and a much larger sized storm as it moves closer to Bermuda, the location that has the highest probability of impact given the various forecast ensembles.

The Parade Continues
The NHC continues to watch for development of soon-to-be named Helene, which is currently centered several hundred miles WSW of the Cabo Verde Islands and has a 90% chance of named formation over the next five days. The cloud mass now coming off of Africa has the potential to be Isaac with a 50% chance of named storm formation during the same timeframe. As I have been saying, these two system have a high likelihood of tracking west across the Atlantic, where conditions are more conducive for strengthening, and could threaten the Caribbean and coastal U.S. in two weeks. But, there are signs that after later next week any new activity should start to wane across the Atlantic Basin. The heart of this hurricane season might only be what develops over the next week as the end of the month looks quiet.

Tropical Update: July 25

With the peak of the Atlantic basin hurricane season (September 10) just 46 days away, I thought I’d provide a bit more context around what to expect in terms of tropical activity as we head into this part of the season. In the last seasonal update I provided on May 31, I highlighted several climate forcers that would likely drive this year’s activity and discussed possible track locations. So far, the general thinking has been accurate, and the recent activity can be used to expand on what the future might hold.

Believe it or not, despite the majority of forecasts calling for below normal activity for the 2018 season, the Atlantic basin activity is running above normal in terms of Accumulated Cyclone Energy (ACE). Year to date ACE is 221% of what would normally be expected by this time in the Atlantic Hurricane Season, and all of  this is from only two named storms (Beryl and Chris). Alberto was classified as a subtropical storm and did not contribute any ACE. Typically, the third named storm of the season forms by July 22, and Chris formed off the southeastern coast on July 6 of this year. The fourth named storm typically forms by August 8, and based on the long range forecast, there is a chance the season could fall back to below normal levels over the next two weeks.

You might have noticed that tropical activity in the Atlantic and East Pacific basins has gone eerily quiet, while the West Pacific basin had five active systems at the start of this week. This is primarily a function of large scale conditions that suppress convection from the current phase of the Madden Julian Oscillation (MJO) and/or a convectively coupled Kelvin wave, which tends to drive interseasonal named storm activity across the northern hemisphere. The nearly off the charts MJO is currently in between Phases 7 and 8, which would provide positive upward motion over the West Pacific basin making it easier for tropical convection to develop, hence the recent active systems over the West Pacific basin at this time.

As the MJO and a convectively coupled Kelvin wave move across the tropics, one should expect an increase in activity next week in the East Pacific basin as there are currently two areas being watched by the National Hurricane Center. Near the second week of August, the positive large scale upward motion should be positioned more over the Atlantic basin and will provide the next best chance for tropical development to occur.

 

In the meantime, no tropical development is forecasted for the Atlantic basin due to several climate forcers limiting named storm development at this time. Below, I provide an update to the major climate forcers outlined in the May 31 update.

African Dust (Currently Negatively Influencing Named Storm Development):
Although the recent African dust outbreaks are right in line with climatology, they have been pronounced, and in some instances have been immersing the Gulf Coast states with African Saharan dust. These masses of dry, dusty air that form over the Sahara and move over the tropical North Atlantic every 3-5 days are often 1-3 miles thick. As these dusty dry Saharan air masses move across the tropics, they can have significant impacts on tropical cyclones because the dry and intense air deprives the storm of moisture, and its winds interfere with a storm’s convection. The dust can also limit sunlight, which creates a more stable air mass.

This image from the Cooperative Institute for Meteorological Satellite Studies / University of Wisconsin-Madison is useful for monitoring the position and movement of dry air masses such as the Saharan Air Layer (SAL) and mid-latitude dry air.

High Wind Shear (Currently Negatively Influencing Named Storm Development)
One possible reason for the large amount of African dust could be due to a stronger African Easterly Jet which pushes the dust to the Americas. This stronger wind also creates wind shear (created when some horizontal winds are faster than others). This will limit the ability of clouds to grow vertically, which is needed for named storm development.
Over the past month, vertical wind shear has been above average across the Caribbean and near average across the tropical Atlantic Main Development Region (MDR). According to Colorado State University’s seasonal forecasts, July vertical wind shear in the Caribbean has historically had a stronger correlation with Atlantic hurricane activity than vertical wind shear over the Atlantic MDR.

This image is from the Cooperative Institute for Meteorological Satellite Studies / University of Wisconsin-Madison.  The presence of wind shear (Areas of Red and Yellow) is hindering named storm development. A low shear environment (Green Areas) is favorable for tropical cyclone development. Wind shear values below 20 kts are considered low enough for tropical cyclogenesis.  

El Niño – Southern Oscillation ENSO (Currently Negatively Influencing Named Storm Development):
The latest ENSO 3.4 region Sea Surface Temperatures (SST) anomaly is 0.4C, which is within the neutral ENSO range, but right on the edge of an El Niño classification of 0.5C. The atmosphere appears to already be responding to El Niño type conditions, which could be why the vertical wind shear has been above average across the Caribbean. The various climate model ensemble average calls for borderline weak El Niño conditions for the peak of Atlantic hurricane season (August-October), with the potential for weak to moderate El Niño during the winter months of 2018/2019. If these conditions develop as predicted, they would likely further hamper tropical development over the MDR and Caribbean Sea. However, El Niño historically does not have any influence on storms that form near the U.S coastline or in the Gulf of Mexico, where the bigger concern for this season is based on the SST.

The International Research Institute at Columbia University provides various ENSO forecasts.  The current forecasts are compiled from various climate models, and they  show the peak of the Atlantic hurricane season (ASO) generally show that that models are forecasting a weak El Niño

Atlantic SST (Currently Having a Mixed  Influence on Named Storm Development):
The Atlantic SST has a negative Atlantic Multidecadal Oscillation (AMO), which the insurance industry has traditionally associated with a lower landfall rate. What is even more amazing is that given the calendar has almost reached August, the zone that can support named storm development (water greater than 26C) over the MDR doesn’t start until 50 degrees West in longitude. And what is even more amazing is the water is warmer in the North Atlantic versus the water temperature in the MDR.  When we connect this with what has already occurred this season, Tropical Depression 2 (Beryl) first came to life after it passed 40 degrees West in longitude at 41.4 degrees West. It quickly became a hurricane before it weakened due to high wind shear over the Caribbean Sea and dry dusty air that got entrained into the system as Beryl strengthened. But yet, Beryl came back to life over the much warmer waters off the East Coast of the U.S. This is exactly the same area where Chris formed into a Category 2 hurricane. This area off the East Coast should be watched closely as the SST are more conducive for tropical development versus the MDR SST.

Current view of the Sea Surface Temperatures across the North Atlantic Ocean via www.TropicalTidbits.com

Highest Threat of Landfall
Both Hurricanes Beryl and Chris gave us some insight into what to expect this hurricane season. Chris formed over the warm waters off the North Carolina coastline. Beryl was originally a small named storm developing from an African wave, and even though it weakened over the Eastern Caribbean, its energy mass ended up recurving off the East Coast before redeveloping into a subtropical storm 290 miles north of Bermuda. It is this general flow that needs to be watched given the position of the Bermuda high, which could steer named storm activity near the East Coast as the Bermuda high pressure fluctuates in intensity and position. Therefore, the East Coast of the U.S. has the highest named storm landfall threat as we approach the peak of the season, even though many of the climate forcers suggest below normal activity overall.

This image outlines the possible flow of named storm tracks as we move into the peak of the Atlantic Hurricane Season.

Another area of concern that should be watched for late season named storm development is the Gulf of Mexico. The SST plot above shows very warm waters over this region, which is a result of the record heat that has been experienced in Texas over the last several weeks. Keep an eye on this area as we get into the later part of the Atlantic hurricane season.

BMS Atlantic Hurricane Seasonal Forecast

The Official Start to Atlantic Hurricane Season?

In theory, the Atlantic hurricane season started late last week with the naming of subtropical storm Alberto. However, interestingly, prior to 2002, subtropical storms were not given names, even though the National Hurricane Center would issue forecasts and warnings for them. Since 2002, such storms have been given names from the tropical cyclone list (Gustav in 2002 was the first). Taking that into consideration, and since no tropical storm days and no Accumulated Cyclone Energy (ACE) have been recorded yet this year, the season, perhaps, has not quite begun.

Regardless, June 1 is the official start of the season and, therefore, providing a bit more detail in regard to expectations for the 2018 Atlantic hurricane season and its impact on the insurance industry is timely. As I mentioned in the closing of the last BMS Tropical Update, there is no relationship between the date of the first Atlantic named storm formation and overall Atlantic seasonal hurricane activity.

Since 1950, there is no relationship between date of first Atlantic named storm formation and overall Atlantic seasonal hurricane activity. Source: CSU

Over 18 different companies and organizations produce seasonal forecasts for the Atlantic hurricane season and are tracked by the Barcelona Super Computing Center. Just last week, NOAA issued its outlook for the season, with a wide range of a 70% likelihood of 10-16 named storms for the 2018 season. When reviewing the historical record, 55% of all hurricane seasons fall into this range, however, which doesn’t provide much informational value to the insurance industry.

Seasonal hurricane forecasts generally come in one of two methodologies: statistical/dynamical models vs. educated guesses from people interpreting the models that produce analog years. The models can be trained and tested against previous years to provide either uncertainty or confidence. The educated guessing is prone to human bias, rules of thumb and big wins or losses. Although there is a general correlation between the greater number of named storms that occur in the Atlantic Basin and the higher chance of landfall, very few forecasts go beyond producing Basin-wide activity to predicting specific impacts or landfall forecasts. To illustrate this point, even if 16 hurricanes were to develop in a season, but none ever hit land, the hurricane season would be classified as above-average, but one that costs the insurance industry nothing in paid losses. Conversely, if three hurricanes develop and all hit land, it’s a below-average hurricane season that could cost the insurance industry millions. I believe there needs to be some attempt to connect the basin activity with landfalls in order to increase the value of utilizing such forecasts.

Where could storms track this season?

One of the most common climate forcers used for seasonal Atlantic Hurricane activity is the El Niño Southern Oscillation (ENSO) and, this year, there appears to be agreement that a weak El Niño event will take hold during the summer, which historically suppresses hurricane activity in the Main Development Region (MDR) of the Atlantic Basin due to higher wind shear. If the El Niño is weak, it could shift development of named storms further north and into the western part of the Atlantic basin near the U.S. East Coast thus increasing landfall chances. Historically El Niño, has little to do with activity in the Western Caribbean / Gulf of Mexico, and any activity in this region increases the chances of U.S. landfall.

Various models from IRI and NOAA CPC on the SST prediction in the Central Pacific ENSO region 3.4. Notice the ASO forecast suggest a weak El Niño should be developed with SST anomaly at 0.5C.

View from NOAA CPC ENSO analysis of subsurface temperature anomalies across the Central Pacific showing overall anomaly below the surface at 150m.  This warm water should rise to the surface over the next few months building the chances of an El Niño during the season.

The subsurface water temperature anomalies across the Central Pacific are warm, suggesting a developing El Niño similar to a Modoki-like Niño, which also points to higher landfall chances during a hurricane season.

Yet another climate forcer taking center stage in early season hurricane forecasts is the cool Atlantic Sea Surface Temperature (SST). In fact, the current North Atlantic SST anomaly pattern looks to be the opposite of an SST pattern which, historically, would be associated with an active Atlantic hurricane season, and could potentially be the coldest since the summer of 2015. However, in the plot below, the warmest SST relative to the average temperature is observed above 30 degrees north latitude and warms closer to the eastern coast of the U.S. This type of pattern suggests that development of stronger named storms might not occur this season in the MDR, but instead north and closer to the U.S. coastline or in the Bermuda Triangle Region of the Atlantic. It should be noted, however, that the MDR Atlantic SST will ultimately trend warmer; almost every year since 1995, the general SST in the MDR has warmed by August regardless of how cold it was in May.

Left is the current SST anomaly and the right the current SST. Reminder, named storms need at least 26C to maintain development.  Currently these types of temperatures are hard to find and likely not a great depth below the surface.

While acknowledging that global weather patterns can change toward the peak of the hurricane season, we can, at a minimum, surmise what the current named storm tracks might be based on the current position of the Bermuda – Azores high pressure center across the Atlantic Ocean. Named storms will often traverse the Bermuda – Azores, and this might provide some clues about the track of future hurricanes in the Atlantic. In its current state, we can see that storms will steer further west, increasing U.S. landfall chances. But if the high weakens to the east, storms may track off shore.

An enormous blocking atmospheric pressure ridge over the Atlantic Ocean stayed resilient for days, which prevented precipitation systems from propagating east. Source NASA GES DISC

Two other climate forcers to watch this Atlantic hurricane season will be Saharan dust and the Madden Julian Oscillation (MJO). Both of these climate forcers are difficult to predict seasonally, but understanding the phases during the season could help determine when storm development will occur. Saharan dust can be an inhibitor of Atlantic hurricane activity, but often moves away from Africa in waves.

Current level of dust off the coast of Africa. Breaks in this dust need to be watched over the given season. Source: http://tropic.ssec.wisc.edu/real-time/salmain.php?&prod=splitE&time

During the breaks in dust, the potential for the genesis of named storms is likely and, if combined with a positive phase of the MJO, could pose a threat to the insurance industry. In fact, I would speculate that the MJO was recently in a state that enhanced tropical convection, which assisted Alberto’s movement across the southwest Caribbean last week. If the MJO was in a favorable state, then we are just moving out of a positive phase for the Atlantic Basin, the next one not occurring for another 45 days. Therefore, this cycle is something to watch closely during the 2018 season, as both of these climate forcers could be the catalysts for named storm development.

The MJO can also considerably influence hurricanes in the Gulf of Mexico, Caribbean Sea, and tropical Atlantic. More hurricanes tend to occur in MJO phases 2 and 3 than in phases 6 and 7. Differences in major hurricane numbers and hurricane days in the main development region are a factor of 3. Source: https://journals.ametsoc.org/doi/pdf/10.1175/BAMS-D-12-00026.1

Summary:
It seems that, based on the information above, the forecasts from various organizations show a reduction in the overall numbers of named storms and hurricanes, thus indicating a less active season. However, even in quieter years like 1989 (Hugo) and 1992 (Andrew), there can be storms that have great impact on the insurance industry. The climate forcers and analog years (1934, 1951, 1996, 2002, 2006, 2009, and 2011) suggest less strong storms over the MDR, but as named storms or tropical waves track out of the MDR the could become stronger as they more reach northern latitudes.  The climate forcers and analog currently indicate named storms could track closer to the U.S. coastline increasing the chances of insured loss.    However, since 1920 there are only 16 years in which two or more hurricane made landfall in one season that have resulted in two or more landfalls the next season.

Given the state of the MJO and the cooler Atlantic SST, I don’t expect a lot of new named storm activity for the month of June. But, during the first part of the July, the MJO could return to a more favorable state over the Atlantic Basin, which will allow for a reassessment of future hurricane landfall possibilities.

Resources:
Barcelona Supercomputing Center has a nice summary of the various Atlantic Hurricane Forecasts
http://www.bsc.es/seasonalhurricanepredictions
NOAA Climatology Products – Avg Start Dates, Return Periods, Develop Origin by Date, Strike Density
http://www.nhc.noaa.gov/climo/
CSU U.S. Landfalling Hurricane Probability Project
http://www.e-transit.org/hurricane/welcome.html

2017 losses, long-term weather trends and their insurance consequences

Catastrophic weather events in 2017 left the re/insurance business with a conundrum. Record-breaking insured catastrophe losses exceeded any previous year. However, the impact of those unprecedented losses on re/insurance pricing has been less than many had hoped for or expected. Claims of nearly $140 billion have not delivered a traditional hard market.

To solve the conundrum and learn the year’s lessons, we must view the storms, fires, and floods of 2017 in a long-term historical perspective. We should not use loss data alone to define weather and climate trends; we need to also understand the impact of the growth of global wealth, as well as patterns of weather extremes.

When we do so, we find that the record-breaking losses were not so enormous, and could have been worse.  About $92 billion of 2017’s insured catastrophe losses arose from three major hurricanes according to the latest Swiss Re Sigma Report. When indexed to 2017 values, Hurricanes Katrina, Rita, and Wilma in 2005 caused insured losses of $112 billion.

We also need to consider that extreme weather memories are typically short. Catastrophe models that simulate the impact of historic storms in today’s built-up environments show that  events such as 1926’s Miami Hurricane or the 1928 Okeechobee Hurricane would cause insured losses that surpass those of 2005 and 2017 hurricane losses. When other named storms and natural catastrophes that occurred worldwide during those years are included, the insured impact on our modern built environment would easily exceeds $140 billion.

Bay Shore Drive in Miami then 1926 after Hurricane and Now. Source NOAA and Google Street View

Meanwhile there is a perception that extreme weather patterns may be changing.  In some cases the data supports this perception.  It is likely that, since 1951, the number of statistically significant regional increases in heavy precipitation events is greater than the number of declines. Strong regional and sub-regional variations modify the trend, but in short, when it rains it often rains more. However, despite the increase in these extreme precipitation events, little evidence suggests the rainwater has led to an increase in floods, with lots of regional variations.

The U.S. Climate Extreme Index quantifies the observed changes in one-day precipitation extremes across the U.S. Since 1990, there has been an increase in the area of the U.S. that has experienced a one-day extreme rainfall.

Tropical cyclones seem not to be changing. After accounting for past shifts in observational capabilities over time, the best records that started in the 1970’s suggest, globally, there are currently no significant observed trends in the number of named tropical cyclones, but the data is suggesting that, when named storms form, they have become stronger in recent years.

2017 had 18 total landfalling named storms. 14 of these had a category 1 or 2 and four of the three major categories were in North America.

As for severe weather, poor data quality makes conclusions about long-term trends problematic. Observed trends in small spatial-scale phenomena such as tornadoes and hail can be made with only low confidence levels, but more recently the data is suggesting clearly a downward trend in the number of occurrences of major tornadoes.

With all this in mind, we should learn lessons from the 2017 experience.

  • Whether it was the major wildfires in California, flooding of Harvey, the extreme winds of Irma over south Florida, or one of the many damaging hail storms; exposure management is very important and there needs to be a refocus on this simple task to limit losses.
  • Valuable data will be collected from the 2017 events, especially from the high wind speeds from HIM. This data will fill critical gaps in the historical record and will be valuable both meteorologically and financially, as it will help insurers and customers by providing better loss models, and lessons for loss mitigation and resilience.
  • Building codes matter. When wind-speed data is combined with claims data, we get critical knowledge of how high winds damage different types of construction. We now have validation for some wind-tunnel testing which has been lacking due to the lack of hurricane landfalls over the last decade.
  • Hurricane forecast track and intensity still have a great deal of uncertainty. The 2017 landfalls raised questions about the value of early landfall loss projections, and how they could be used.
  • Effectively, the people who need flood insurance do not have it.  How many more flood events like Harvey and Katrina do we need before more action is taken?
  • Companies with the best business interruption (BI) insurance came out on top after an event, but few small businesses don’t have enough BI.
  • Demand for certain post-event services seriously tested supply. The lack of adjusters, for example, created a major bottleneck that led to very high loss adjustment expenses.
  • Given the long 10 years without major hurricane landfalling activity, 2017 was a great systems test for InsurTech.  It would appear no major lapses in these new technologies have occurred which should further expedite system enhancements in InsurTech.

California Columbus Day Firestorm

Although October is usually known for Atlantic hurricane activity and sometimes a second peak of severe weather that can occur as summer wanes, this month is also the height of the California wildfire season, which typically runs from spring to late fall. Unfortunately, it should be no surprise that this western wildfire season has been one of the worst on record – in May, I briefly mentioned that it could likely get ugly. So far, the U.S. budget for fighting wildfires has topped out at $2.35B, which does not include the recent fire-fighting efforts. Year to date, the wildfires have burned 8.5 million acres across 51,000 fires.

National Interagency Fire Center Stats https://www.nifc.gov/fireInfo/nfn.htm

California has had nearly five years of extreme drought followed by record rains last winter, producing a bumper crop of grasses and fine fuels on top of drought-dried, diseased and stressed heavier fuels. Fuel analysis ahead of the Columbus Day firestorm were at all-time record dry levels. Like many previous firestorms, when this situation is combined with the Santa Ana/Diablo wind events that occur every year and the right weather conditions, any fire ignitions that occur will cause fire explosions that race along the landscape.

Columbus Day firestorm cause
The Oakland Hills, CA fire of 1991 had been the seminal fire event that changed the insurance industry’s perspective on wildfire loss, much like Andrew did with hurricane loss. However, the Columbus Day firestorm will no doubt be viewed as a considerable event in a growing list of large wildfires that have impacted the insurance industry in recent decades. There are very few natural catastrophic events that result in such destruction as a wildfire, although the effects of an EF 5 tornado are similar. However, the damage from wildfires is often worse since everything is reduced to ash, with virtually nothing being recoverable. With EF 5 tornado damage, some personal belongings can usually be salvaged. This is often not the case in fires, which make them a different beast entirely.

 

The exact cause of most of the Columbus day fires are still under investigation, it is most likely they were either human triggered or a result of sparking from down power-lines due to the high winds that occurred.  No thunderstorms were in the area so that can be ruled out at this time.

The weather conditions that created the firestorm are often referred to Diablo winds, which follow the same type of pattern as Santa Ana winds – the northern and southern California areas simply use different names for the same weather phenomenon. The increased fire conditions start when a big ridge of high pressure sets up over the Great Basin in the inter-mountains west from Utah to Nevada. This causes air to flow from east to west across California, from high elevation to sea level. It is at this point that the first law of thermodynamics takes over. As the air is compressed when moving from a higher elevation to a lower elevation, it heats up. During this movement of air, the moisture in the air does not change much, but with a rising temperature, a large disparity between the temperature and the moisture in the air is created and pushes the relative humidity to very low levels. As the air gets compressed, it moves faster, forced over mountains and pushed through canyons.

How hot, dry downslope winds form, like the Diablos and Santa Anas. Source: The Washington Post

Leading up to Monday’s firestorm, the pressure at Reno, NV at 9:00 p.m. on Sunday, October 8 was at 1020 mb, and the pressure at Napa, CA was even lower at 1008 mb. This difference in pressure was the catalyst for the Diablo winds, which in some places gusted to 79 mph.

Notice how fast the winds picked up in the evening hours on October 8th. This is scary as most people were in bed when these fires broke out so they have little to no warning to leave.

As of Thursday morning, 22 wildfires were burning across northern California, which is an increase from the 14 that were originally reported on Monday. Looking ahead, until there is a major North American weather pattern change, the Diablo/Santa Ana wind events could occur every few days. With a La Nina building in the Pacific Ocean, the long-term prospects for rain are slim across the state, particularly in the south. A large subtropical ridge is forecasted to be in the area until at least next week, as moisture-laden storms rumble ashore over the Pacific Northwest and keep California dry. Currently, the National Weather Service has a Red Flag Warning in effect until Thursday the 12th until 5 pm PDT, which means that critical fire weather conditions are either occurring now or will be shortly. The combination of strong winds, low relative humidity, and warm temperatures can contribute to extreme fire behavior.

Although the wind is not expected to be as strong as observed earlier in the week, this warning is due to the increasing north to northeast wind and low humidity that will generate critical fire weather conditions again. New and existing fires could rapidly spread during this time period.

Measuring insured loss of Columbus Day Firestorm
The latest figures from the Associated Press suggest that the Columbus Day firestorm destroyed at least 3,500 homes and businesses. PCS has issued two separate catastrophe bulletins which cover the Atlas Fire near Napa, CA and the Tubbs Fire near Santa Rosa, CA. However, these are just two of the most destructive fires across the region. For example, the Canyon Fire 2 near Anaheim Hills is now 60% contained, but burned 26 structures, damaged another 36, and continues to threaten 3,500 structures. Therefore, the wildfire losses will be more far-reaching than PCS has reported, since it only designates losses of $25m or more. For the latest information on these fires and their containment levels, please see Cal Fire Incident Information, which gets updated as often as needed.

Where and how to get fire information

Current location of active fires according to Cal Fire

 

This is an area in North Santa Rosa around Coffey Park. As shown many of these structures were burned to the ground. Zillow suggest many of these structures are worth at least $400. Locations outside of town in the foothills are easily double in value. It should be noted this is just the home value not counting the total loss of the contents of these structures.

If the AP estimate of at least 3,500 homes burned is correct, then as a combined event, this will be one of the largest wildfire insured loss events in U.S. history, if not the largest. Below is a chart illustrating some of the most damaging North American wildfires and the estimated insured losses if they were to occur today.

List of the largest damaging wildfires in the USA in recent time, ranked by # of buildings destroyed. The Oakland hills fire of 1991 was 2,900 structures. So, in aggregate, the ongoing Columbus Day Fires now exceed 1991 Oakland Hills fire. Notice how many of the CA wildfire have occurred in October.  Note most fires with over 2,000 structures are over $2 billion in insured loss.

 

BMS iVision has a direct feed of the current fire perimeters. These perimeters use the IRWIN (Integrated Reporting of Wildland-Fire Information) system. Perimeters are collected in the field by a variety of means, including infrared flights, and by using a GPS unit to map the perimeter. BMS clients can use these maps to see if any risks are exposed to the fires.

With the understanding that the Tubbs and Atlas Fires had a large impact on the expensive Napa/Santa Rosa areas, it is assumed that the losses will be much higher than past fires such a the Oakland hills fire of 1991 because of the effects on commercial property. At least 14 commercial wineries, a Hilton hotel, a Kmart, a McDonalds and even the historic Fountaingrove Round Barn have been destroyed by the fires.   Many other commercial properties have been destroyed and if not destroyed will have business interruption. Smoke related clean up that could also work its way into the insured losses from this event even if the structure did not have damage.

Catastrophe Model can’t come fast enough
Many areas of California have a long and active wildfire history. Typically, about 10,000 wildfires are ignited in California every year. Of those, only about 20 cause property losses and fewer still cause losses large enough to be designated a catastrophe. But when conditions are right, the losses can be truly catastrophic, as we saw with the Columbus Day firestorm. Many areas of the state are characterized by narrow valleys surrounded by steep, hilly terrain. The interface between wildland areas and development, exposed residents and businesses is increasing the wildfire risk. In the last several decades, the combination of firefighting technology, fire suppression policies, environmental regulation, and development trends has led to increased fuel loads, greater occupancy of remote areas and greater potential for catastrophic wildfires. This is a trend occurring all across the country.

The risk for wildfire is increasing across all 50 states, with the costly wildfire outbreaks in Texas, Tennessee, and Colorado being recent examples, yet the insurance industry still lacks the access to a full probabilistic wildfire loss model for the entire U.S. For years, the insurance industry has had access to assess the wildfire hazard from various sources and, of course, there are several tools built into the catastrophic loss models that help with accumulation management. Some catastrophic modeling companies have developed loss models for California only as a result of the high losses that occurred after the devastating fires that hit southern California in October and early November 2007. However, limited updates have been made since the initial release.

The good news is that modeling companies are currently working on U.S.-wide probabilistic wildfire loss models. However, as we have shown, wildfires can be very complex, with embers traveling several miles and igniting new fires. The models that are in development will need to incorporate the newest landfire fuel databases that seem to change constantly with growing vegetation across the U.S. These models need to have fire-spread algorithms and account for human fire suppression decision-making through stochastic simulations of man-made fire breaks. They should also model the possibility that a wildland-urban interface fire will transform into an urban conflagration, such as we saw in the Santa Rosa area.

To understand wildfire behavior, the models will need to include historical data on average hourly wind speed and direction based on weather stations, but those might not address the varied microclimates that can occur in a complex terrain. Of course the models will need to account for a wide-range of residential and commercial constructions, including the presence of mitigating factors, such as fire-resistant roofing and siding materials, which are becoming more and more popular.

Several million dollar homes in the Fountaingrove area northeast of Santa Rosa were destroyed. Notice the lack of defensible space around many of these structures.

Also, as we have seen in past wildfires, defensible space and vegetation control have a huge impact on site-specific loss. In fact, fire-wise communities, resilience efforts and research by the IBHS are helping to minimize wildfire loss.

BMS Tropical Update 10/07/2017 11 AM CDT

Last night at 11:30 PM EDT, the NHC upgraded Nate to hurricane status about 18 hours ahead of when it was originally predicted to become a Category 1 storm. This is now the first time since 2012 that more than nine hurricanes have developed in one season. This is also the first time since 1893 that nine Atlantic named storms have consecutively reached hurricane status (Franklin-Nate). But, it should be noted that, obviously, vastly different systems are now in place than in the 19th century, which means weaker systems likely were missed before satellite and aircraft observation.

Since Nate is now a hurricane, there is no doubt it will continue to strengthen and be a Category 2 storm as it makes landfall around Gulfport or Biloxi, MS.  There is even a possibility that Nate could be a major hurricane at landfall if it undergoes rapid intensification over the next 12 hours. However, this is more difficult to predict with a fast-moving storm like Nate. Regardless, a stronger storm at landfall will increase the insured losses expected from this event, as yesterday’s estimates were based on a weak hurricane or tropical storm at landfall.

Nate continues to improve its satellite presentation as it cruises north/northwest this morning. With winds at 90 mph, Nate has maintained its gradual strengthening overnight. Nate is located 180 miles south/southeast from the mouth of the Mississippi River. Intense thunderstorms have wrapped all the way around the center, and distinct spiral bands have developed both northwest and southeast of the center. Strong outflow is noted in all directions, which indicates low wind shear. It will be a race against time to see how much Nate can intensify before it moves onshore and weakens.

Nate’s Landfall Impacts
The general landfall area continues to narrow with an expected location between New Orleans, LA and Mobile, AL with the consensus of a landfall location near Gulfport or Biloxi, MS. The main question continues to be: how much strengthening can happen in the next 12-15 hours before landfall?  Nate is already close to a Category 2 hurricane.

Wind Impacts
What we do know is based on observations from the hurricane hunter – the overall wind structure is quite lopsided. This is likely due to Nate’s fast forward motion of 26 mph. This will cause Nate to have a small core of hurricane-force winds only to the northeast of its center.

Cross section of Nate showing pressure and wind. The wind profile clearly shows that the stronger winds are to the right side of the storm with weaker winds on the left side of the storm center. This is likely due to Nate’s rapid forward motion.

There will be little, if any, hurricane force winds to the west of the storm which would be good for New Orleans and cities to the west side of the landfall location. Tropical storm-force winds will extend well east into Alabama, and possibly the Florida panhandle, as Nate makes landfall around 7 pm CDT.  This is also much sooner than expected based on Nate’s rapid forward motion.

Because of Nate’s swift forward motion east of its center, locations further inland could experience higher winds, versus a typically slower-moving storm, as Nate will be further inland before it weakens over land.

As with all landfilling hurricanes, weak tornadoes are the primary threat away from the center of the storm. However, due to the incredibly warm, moist nature of hurricanes, hail is not expected to be an issue.

BMS iVision allows clients to run risks to better understand various impacts from Nate. This is the 3 sec gust wind speed, which shows only coastal areas will see the strongest winds. However, tropical storm-force winds will be widespread, particularly along the eastern side of the storm. Given current fwd speed (26 mph), concerned about higher inland wind potential in southeast as well.

 

BMS iVision allows clients to run risks to better understand various impacts from Nate. This is the 1 minute wind speed, which shows only coastal areas will see the strongest winds. However, tropical storm-force winds will be widespread, particularly along the eastern side of the storm. Given current fwd speed (26 mph), concerned about higher inland wind potential in southeast as well.

Surge Impacts
As with any northward moving hurricane, the onshore winds will cut across a wide swath of the Northern Gulf Coast from southeast Louisiana to the Florida panhandle. These onshore winds will pile water up against the coastline, resulting in storm surge flooding. As I mentioned yesterday, the slope of the Gulf Coast is quite flat, meaning that it doesn’t take much storm surge to cause issues. The factor that could limit the amount of storm surge is Nate is an extremely fast-moving system, meaning that it won’t have prolonged winds along the shore, thus limiting the storm surge. Currently, Mobile Bay seems to be the target for the highest storm surge amounts, which could be close to nine feet. One unfortunate coincidence is that the moon is full and high tide will occur around midnight locally, which coincides with landfall and peak storm surge, maximizing coastal surge.  At this time this does not look to be a big surge event for New Orleans as higher impacts should be greater in MS and AL coastal locations.

NHC Potential Storm Surge Flooding Map. These will get updated here: http://www.nhc.noaa.gov/refresh/graphics_at1+shtml/091740.shtml?wsurge#contents

Rainfall Flooding Impacts
Rainfall will be heavy as this system moves onshore, especially near the center, but could extend well east of the center to the Florida peninsula. As with the surge, the fast motion of the storm precludes any extreme impacts from rainfall, but flooding should be expected whenever 3-8″ of rain occurs. Also note that the heavy rain will continue north into the Appalachians.

Forecasted rainfall over the next 5 days. This shows up to 8″ of rain is now forecasted for parts of the eastern slopes of the Appalachian Mountains. This could create isolated flash flooding.

Insured loss
Over the last few days, I have provided a few analogs that were from tropical storms like Lee in 2012 to Cindy in 2005 and even Isidore in 2002. All of these analog events resulted in insured losses of under $300 million. However, with Nate now likely a Category 2 at landfall, a comparison to a higher analog storm such as Ethel in 1960 seems more appropriate, which today would be closer to a billion dollar insured loss event. Fortunately, it’s highly unlikely Nate will be worse than Opal was in 1995, which today could be close to a 3 billion dollar insured loss event. The intensity at which Nate makes landfall will dictate the amount of insured loss, but an insured loss closer to a billion dollars is more credible with a stronger storm at landfall.

Next Tropical Trouble

A strong suppressed Kelvin wave is forecast to push across the Atlantic this week. After Nate, there is a good chance of a lull in activity until late Oct.  There is a chance that  something manages to spin up in the sub-tropics, which are less impacted by suppressed Kelvin wave passages.  Overall Nate looks to be the last U.S. landfalling hurricane at least for the next two weeks.

 

BMS Tropical Update 10/06/2017 12 PM CDT

Nate’s overall structure is finally beginning to improve, based on satellite imagery and observation from a hurricane hunter aircraft. Strong thunderstorms have developed over/near the center of the system, and organized outflow is developing in the upper levels. Banding structures are now beginning to organize around the system. This is a marked change from late yesterday afternoon when thunderstorms were disorganized and weak. The storm is now over the warm waters of the western Caribbean and further intensification is expected.

Nate’s track forecast continues to wobble around, but the general expectation remains consistent. The storm will either make landfall on, or barely miss, the Yucatan Peninsula today before tracking north-northwest into the Gulf of Mexico. A turn north and then northeast is expected on Saturday before a landfall somewhere between Lake Charles, LA and Panama City, FL around 1 am Sunday morning. The storm’s remnants will then race northeast and bring heavy rains to the Mid-Atlantic and parts of New England.

Nate’s Forecasted Intensity
The biggest question mark continues to be just how strong Nate will be at landfall. I focused on this in yesterday’s update, but it is worth discussing again since intensity is key to the overall insured impact at landfall.

As I have been showing, warm waters and low wind shear are the only two factors that suggest strengthening may occur. Dry air and land interaction with the Yucatan Peninsula are the two opposing factors that may cause Nate to be weaker as it moves into the southern Gulf of Mexico. It appears that the dry air will be the primary inhibiting factor in terms of Nate’s intensity prior to landfall in the U.S. The last several runs of the models show dry air wrapping into the system from the west and south as it moves closer to New Orleans tomorrow. If this dry air can wrap into the inner core of the system, it will keep a lid on Nate’s intensity. However, if an inner core can organize today before the dry air really kicks in, and if interaction with the Yucatan doesn’t disrupt this inner core, the storm may be able to continue strengthening despite the dry air.

The majority of ensembles suggest that Nate will have a pressure of 988 – 999 mb at landfall, which is what we would see in a weak hurricane or tropical storm.

Nate is expected to make landfall in 48 hours, so note the pressure forecast from the ensembles during this forecast time period.

Nate’s Forecasted Landfall

The last few model runs of the GFS suggest that land interaction with the Yucatan will be less of a factor as the track has shifted back to the east over the last 24 hours. This may cause Nate to be stronger than forecasted this afternoon, as it tracks close to the Yucatan but not over the Yucatan.

Notice over the last 12 hours a shift in Nate’s track (Green Line) versus what the NHC had forecasted (Red Line).  This may allow Nate to track between the Yucatan and Cuba and would limit land interaction.

 

Last night’s multi-model European, American, and Canadian ensemble’s strike probabilities for Nate.

The landfall location currently appears to be near Gulfport, MS, as the models shift back east away from a New Orleans landfall.  However, the range of possible landfall locations fall somewhere between Lake Charles, LA and Panama City, FL just before midnight on Sunday morning. I would not be surprised if the models continue to shift a landfall location to the Florida Panhandle.

Landfall Impacts
Nate’s impact will extend well away from the center regardless of the inner core’s organization, especially to the east side of the storm. I expect to see strong thunderstorm activity in Nate’s outer bands as it makes landfall, with the heaviest storms moving onshore along Florida’s entire west coast, hundreds of miles east of the center. Gusty winds, heavy rains, and tornadoes will be concerns in these outer bands.

This simulated satellite forecast from the ECMWF shows strong thunderstorm activity in Nate’s outer bands moving onshore in Florida.

Even the weakest of hurricanes can cause storm surge along the northern Gulf Coast. Note in the image below the wide expanse of onshore winds from New Orleans all the way east to the Big Bend of Florida. Even if these onshore winds barely reach tropical storm-force, they will pile up water along the coast. Due to the very flat slope of the Gulf Coast, even a small rise in water level is enough to cause flooding concerns.

The extent of the storm surge will depend on the storm’s intensity, but coastal locations vulnerable to storm surge flooding will likely see at least some level of storm surge. Mobile Bay, AL, Gulf Port, MS, Biloxi, MS and Pensacola, FL are all expected to see at least three feet in storm surge, with some isolated locations expected to see six feet.  Also worth noting that landfall will be close to High Tide, which means water levels will be even higher usual.

NHC Potential Storm Surge Flooding Map. These will get updated here: http://www.nhc.noaa.gov/refresh/graphics_at1+shtml/091740.shtml?wsurge#contents

Several inches of rain may also cause freshwater flash flooding, but the storm’s fast forward motion should preclude any devastating flooding like we saw with Harvey. The map below shows total rainfall for the next 5 days across the U.S. Note the swath of higher totals along Nate’s forecasted path all the way up into New England, where the storm’s remnant low will enhance rainfall associated with a cold front.

Forecasted rainfall over the next 5 days. This shows up to 8″ of rain is now forecasted for parts of the eastern slopes of the Appalachian Mountains. This could create isolated flash flooding.

 

BMS iVision allows clients to run risks to better understand various impacts from Nate. This is the 3 sec wind gust in MPH, which shows only coastal areas will see the strongest winds. However, tropical storm-force winds will be widespread, particularly along the eastern side of the storm.  Given current fwd speed (21 mph), concerned about higher inland wind potential in southeast as well.

 

Potential For Insured loss

The insured loss will depend on the intensity at which Nate makes landfall. It is likely that only minor insured losses will occur if Nate makes landfall as a Category 1 hurricane along the Gulf Coast (under $1B in insured loss). The best analog in recent memory would be Hurricane Cindy 2005 which made landfall over the Mississippi River delta as a Category 1 hurricane, tracking north-northeast with a second landfall just west of Biloxi, MS.  Cindy caused roughly $200M (in 2016 dollars) in insured loss across Louisiana, Mississippi and Alabama. Another analog might be Isidore 2002, which, in 2016 dollars, would have caused $279M in insured loss. Arlene 2005 made landfall as a tropical storm in a similar area but did not cause a PCS loss event.