partner login

BMS News

Wildfire Outlook and Future Problem Areas – August, 2019

While the global insurance industry seems to be eyeing every cloud mass in the Atlantic Ocean for the development of the next possible named storm and what might unfold in terms of overall activity and landfalls, the industry should also pay attention to a different type of weather forecast. Some similarities can be drawn between the lower 48-state fire season and the Atlantic hurricane season in that they are both off to fairly quiet starts. There is curiosity within the industry as to whether the back half of both of these seasons will be hyperactive.

With a total of over $33B of North American wildfire insured loss since 2016, it’s clear that wildfire has taken on new importance in the insurance industry. The good news is that much of North America is exiting the core of the fire season and, so far this summer, there has been a welcome wildfire reprieve. In California, only 24,579 acres have burned this year as opposed to 621,784 acres last year (CAL FIRE Only stats to Aug 22). The other bit of good news is that California has only about 6.73% of the state under abnormally dry conditions, with no drought conditions being reported. This is a vast improvement from one year ago when 47.19% of the state was reporting drought conditions.

California Department of Forestry and Fire Protection (CAL FIRE) combined Statistics including the U.S. Forest Service for the state of California.

There is still much of the fire season remaining, including the most dangerous time of year when strong, dry offshore winds peak during the months of September, October and November and have, historically, been associated with the majority of California’s worst fires. The Santa Ana and Diablo winds which, when combined with abundant fine fuels, can lead to elevated fire activity, even if the current fire risk is low. In fact, after a fairly wet winter, the abundance of fine fuels, such as annual grasses, has increased, which adds to the fuel load across the region. As these fuels dry out, there is significant risk if a fire starts and high winds are present.

The National Interagency Fire Center is one of the official sources for wildfire assessments and outlooks. According to its forecast, the areas of greatest concern continue to be the lower and middle elevations across California, the northern and western Great Basin, and portions of the Pacific Northwest where the robust grass crop has cured and can become an easy source of fire.

Since the halfway point of the western wildfire season has passed, there are some signals that point towards an active, compressed, season across the west as the southwestern monsoon becomes more active. While this will effectively end the fire season across the southwest, lightning-induced fire activity is expected to increase elsewhere. And, since humans cause the majority of wildfires, risk is always present. When accidentally or intentionally set fires are combined with wind events, which have been largely absent thus far, but will intensify in frequency by mid to late month as dry frontal passages become more common, the risk of wildfire will also increase. Wind events and low humidity will contribute to a likely increase in fire occurrences into the start of November.

During the late fall or even early winter, when a high-pressure system forms over the deserts of the Great Basin, the high pressure circulates clockwise, pushing air westward toward the lower-pressure areas of the coast, where it drops from high elevation to sea level. This causes the air to become compressed and it heats up, and its relative humidity drops. As the air drops from high elevation to sea level it can be forced through passes and canyons. This can cause wind gusts of 40 to 60 mph or even stronger. Image Source: InsideClimateNews

Unlike a hurricane, which often provides the insurance industry with a few days’ notice to determine expected loss, devastating fires often occur at a moment’s notice. The best thing the insurance industry can do is practice good accumulation management now, acknowledging areas that could become the next Paradise, California. Recent work by Pamela Ren Larson and Dennis Wagner at azcentral.com concluded that there are as many as 526 locations with a higher wildfire potential than Paradise, CA. In the absence of a newer, sophisticated catastrophe model, the insurance industry can still use Larson and Wagner’s very thoughtful analysis to determine the overall risk to a portfolio via a simple risk score. Although their analysis looked at the hazard, it also tried to understand the overall human risk, which life and casualty lines should contemplate.

The insurance industry may also want to consider analyzing areas outside of the western U.S. that could become the next Gatlinburg, Tennessee. For example, the 1.1 million acres of Pine Barrens in New Jersey, which is mostly rural despite the proximity to the sprawling metropolitan cities of Philadelphia and New York City, could result in a major insurance industry loss. There are vulnerable locations along the wildland-urban interface where fires are common. At some point, the weather conditions will be right to create a worst-case scenario, which could impact towns like Tabernacle or Barnegat Township in New Jersey. Thankfully, the peak of the wildfire season in New Jersey is during the spring months.

General Wildfire Risk Hazard Across the State of New Jersey. Source: Rolling Stone Will-Americas worst Wildfire Disaster Happen In New Jersey

Another recent concern is the increased wildfire risk that has been created by Hurricane Michael across Georgia and the Florida Panhandle. According to the National Association of State Foresters, a total of 92 million tons of timber, or roughly 4 million truckloads of timber, were destroyed by Michael’s fierce winds. The volume of dead and downed fuels will likely contribute to an increase in the number, intensity, and duration of wildfires over the next three to ten years. According to the Association, there are typically 4.87 tons per acre of available fuel in Florida. Currently, the average is up to 58 tons per acre – a ten-fold increase. In a catastrophic area, there are over 100 tons per acre. In fact, analysis done by the Southern Wildfire Risk Assessment Portal puts 49 communities at risk in the catastrophic wildfire zone and 194 communities at risk in the severe wildfire zone. Again, much like New Jersey, the highest likelihood of wildfire across this region is in the spring.

While the wildfire season will likely end up being below normal in terms of acres burned for the western U.S., it only takes one fire under the right conditions to make a memorable season. Significant wildfires can occur even when fire conditions are not extreme. Therefore, it is beneficial to understand wildfire risk accumulation and determine what is a comfortable level of loss by simple aggregation analysis or using new sophisticated wildfire models. These models have suggested that recent wildfire seasons are on the order of a 30 to 60-year return period, so none of the recent U.S. wildfires are considered tail events. Instead, these events have low return periods of less than 100 years, which reflects increasing U.S. wildfire risk in large part due to development of the wildland-urban interface. Insurance companies should understand this risk due to the proliferation of wildfire-prone areas throughout the U.S.

BMS Tropical Update – August 13th

It’s been a while since the last BMS Tropical Update on July 11. That update focused on Tropical Storm Barry, which eventually made landfall along the central Louisiana coastline as a minimal Category 1 hurricane. Even though it was a disheveled mess of a storm, it still caused nearly $100 million of insurance industry loss in Louisiana and set an Arkansas state rainfall record of 14.58” near Murfreesboro, Arkansas. Luckily, this large amount of rain did not fall along the Mississippi River in southern Louisiana, which was at already historic high levels along many sections, including levy-protected New Orleans. With Barry’s landfall, it joined 12 other July hurricane landfalls in the Gulf of Mexico since 1900. However, as we look forward to the remaining months of the hurricane season, it’s worth noting that there is no correlation between hurricane activity occurring before August and how much activity will be seen during the remainder of the season.

By now you may have seen the various updated Atlantic hurricane season forecasts, which, for the most part, continue to call for above-normal activity. As I have stated several times, the overall number is not what is important, but, rather, the steering currents that influence the track(s). Contrary to popular belief, however, most named storms have fairly regular and well-defined tracks because of the location and orientation of the Bermuda Azores high pressure, which ultimately determines the tracks of most named storms. The difficulty in predicting a storm track occurs either when the typical climatological steering wind flow is replaced by a less common, large-scale flow or, even more importantly, when rapid changes occur in the strength and orientation of the steering current, such as a bypassing mid-latitude trough, which is really only well-forecasted 5 to 7 days in advance.

Above is the current steering flow across the Atlantic Ocean. Note the current weakness across the Bermuda Triangle. Absent any mid-latitude system, this would likely be the path for named storms out of the deep tropics.

The average date for the formation of the season’s third named storm is August 13, and the average date for the second hurricane is August 28, so there is nothing unusual about having a calm spell this time of year like we’ve seen recently. There have been several tropical waves that have propagated from the African coastline, and a few of these have been watched by the National Hurricane Center. Tropical Depression Three, which formed for less than 24 hours off of the eastern coast of Florida, provided a glimpse of where named storms could track this season if the current North American and Atlantic pressure patterns hold into the peak of the hurricane season. However, first let’s determine the factors to consider for the remainder of the hurricane season and then worry about the steering current once the storms begin to form. In the beginning of the season, it was suggested that named storms would form closer to the U.S. coastline rather than in the Main Development Region, and track closer to the eastern coast of the U.S. with the overall season seeing more back half activity.

Although many of the headlines suggest that an above-normal season is yet to come, these headlines should probably be taken with a grain of salt. Take the NOAA forecast, for example. There is a 45% probability of an above-normal season, which obviously means there is a greater chance of a normal to below-normal season at 55%. In fact, I don’t think there is much confidence in the August activity forecasts, even though they are usually the most accurate when it comes to the overall Atlantic Basin activity.

There are a number of reasons why there is a lack of confidence in the forecast:

Unfortunately, even with the more reliable August forecast, there are still a lot of variables for the remainder of the season. Perhaps the best guidance would be to view the analog years, which serve as a guide for potential activity and possible tracks for the remainder of the season.

Above are the various analog years (1991, 2012, 2014, 2015) which may best indicate the current conditions and possible track of storms this season. Depending on the overall timing of development, analyzing these years may provide a guide as to the general track storm could take this season, but the timing of mid-latitude weather systems will determine the steering level winds if they reach the U.S. coastline. Example: Isaac or Sandy 2012 (Remember Sandy could have also turned out to sea)

This week we may see the remnants of an old stalled front off of the eastern coast of the U.S., providing a chance at tropical cyclogenesis closer to home. Hopefully, there won’t be a need for too many BMS Tropical Updates over the next 30 days, will be keeping an eye on how any potential events could impact the insurance industry.