Pages

22 August 2022

Highlights From "Hazardous Heat"


BACKGROUND
New research from First Street Foundation analyzes the the prevalence of increasing extreme temperatures and dangerous heat wave events throughout the contiguous United States, with a key finding being the incidence of heat that exceeds the threshold of the National Weather Service’s (NWS) highest category for heat, called “Extreme Danger” (Heat Index above 125°F) is expected to impact about 8 million people this year, and grows to impact about 107 million people in 2053, an increase of 13 times over 30 years. This increase in “Extreme Danger Days” is concentrated in the middle of the country, in areas where there are no coastal influences to mitigate extreme temperatures.

The First Street Foundation Extreme Heat Model (FSF-EHM) was built using datasets from the US Federal Government, augmented with publicly available and third party data sources, and existing research and expertise on heat modeling. The model estimates localized heat risk at a 30-meter resolution across the United States today and 30 years into the future, creating a high- precision, climate-adjusted heat model that provides insights at a property level. Its analysis combines high-resolution measurements of land surface temperatures, canopy cover, impervious surfaces, land cover, and proximity to water to calculate the current heat exposure, and then adjusts for future forecasted emissions scenarios. This allows for the determination of the number of days any property would be expected to experience dangerous levels of heat.

METHODOLOGICAL OVERVIEW

The model makes use of data which has been shown to be strongly correlated with air temperatures in past research. Satellite derived land surface temperature (LST) estimates were obtained and compared with actual observed air temperatures at the same location in order to gain an understanding of the relationship. LST is distinct from air temperature, but the values tend to be strongly correlated. For each station, the complete June, July, and August monthly temperature records for the years between 2014 and 2020 are used, as summertime land surface temperatures show the best differentiation among smaller areas.

LST data were acquired at a 30-meter spatial resolution from the Landsat 8 Collection 2 surface temperature product for the years 2014 to 2020. These data are used to create a single image representing the summertime land surface temperature across the Contiguous United States by using the median land surface temperature value for each 30-meter pixel. The values used to create these median estimates come only from satellite images with less than 15% cloud cover.

The resulting model was validated with out-of-sample observed station data, where station records were split into different groups to train and test the model for multiple iterations. Additionally, using the results from the model, the 30-meter air temperature predictions are validated by comparing the estimates against an independent temperature map for a city (Durham, North Carolina) created by the Climate Adaptation Planning and Analytics (CAPA) Heat Watch (CAPA/NIHHIS, 2021), where volunteers collected temperature measurements across the city at a hyper- local scale. The comparison with this temperature map is particularly useful as the mapped area contained none of the climate station data used in training the FSF-EHM.

NATIONAL OVERVIEW

The model highlights the local impacts of climate change by identifying the seven hottest days expected for any property this year, and calculates how many of those days would be experienced in 30 years. The most severe shift in local temperatures is found in Miami-Dade County where the 7 hottest days, currently at 103°F, will increase to 34 days at that same temperature by 2053. Across the country, on average, the local hottest 7 days are expected to become the hottest 18 days by 2053.

In the case of extreme heat, the model finds 50 counties, home to 8.1 million residents, that are expected to experience temperatures above 125°F in 2023, the highest level of the National Weather Services’ heat index. By 2053, 1,023 counties are expected to exceed this temperature, an area that is home to 107.6 million Americans and covers a quarter of the US land area. This emerging area, concentrated in a geographic region the Foundation calls the “Extreme Heat Belt,” stretches from the Northern Texas and Louisiana borders to Illinois, Indiana, and even into Wisconsin.

Across the country, Dangerous Days – days exceeding the 100°F threshold from the National Weather Service – occur more commonly in the southern half of Contiguous United States and impact a greater number of properties in Florida and Texas. Currently, the top 20 counties across the United States expected to experience the greatest number of Dangerous Days annually are located in Texas, California, Arizona, and Florida. Topping the list this year with 109 days above the heat index temperature of 100°F is Starr County, TX. The other top four counties with over 100 Dangerous Days are in Texas and California.

No comments:

Post a Comment