Study predicts a power outage during a heat wave will kill over 13,000 people.
Climate change disasters keep coming. Not a week goes by without one. Wildfires may be getting all the attention, such as the recent fire in Hawaii that devastated Lahaina, or forest fires in British Columbia and Quebec. Deaths from heat waves, by contrast, don’t get that much attention, especially in cities and circles that can avail themselves of air conditioning.
But what if the power goes off? That is the question posed by researchers in a study published in Environmental Science & Technology’s Spring 2023 issue—before Phoenix (one of the cities modeled in the study) suffered 31 consecutive days of temperatures over 110°F by dropping to a mere high of 108°F on Monday, July 31. There was no dancing in the streets, however. More than a hundred people had died directly from the heat and more deaths were probably heat related. And temperatures over 110°F were expected to return later in the week.
A power outage during a heat wave requires no stretch of the imagination. Imagine a million people running their air conditioners full blast 24 hours a day and overloading the power grid. Imagine the Colorado River, which supplies hydroelectric power to much of the Southwestern U.S., running dry. Imagine power lines causing wildfires, which further stress grid infrastructure.
A City at Risk
Phoenix may not be the biggest city in the U.S. (it’s 5th) or be the hottest or have the biggest population of at-risk individuals unable to find shelter, but it ranks high enough in all categories to have the makings of the country’s biggest climate change disaster. In the final tally of heat-related deaths predicted in the three cities in the study, it wasn’t even close. Phoenix lost by a long shot. Residents of many U.S. southern cities survive summers by using air conditioning, but neither Detroit nor Atlanta is as hot or is so dependent on air conditioning. With no place to go, almost 1 percent of the population of Phoenix may die should a failure occur.
The study draws on data from the Three-City Heat and Electrical Grid Failure Adaption Study (3HEAT), where researchers from the Georgia Institute of Technology, University of Michigan and Arizona State University chose to model the biggest cities in their respective states: Atlanta, Detroit and Phoenix. Studying more cities would have taken more time. It may have been the shock of the high predicted death toll for Phoenix that added an urgency to publish the findings and ring the alarm for other cities. Seventy-five percent of the population in the U.S. shares weather zones with the three cities in the study. Shouldn’t we have similar studies of New Orleans, Houston and Las Vegas? Haven’t they the same long, hot summer?
Grid Anxiety
The decreasing reliability of U.S. power grids has the researchers most worried. The last few years have seen an alarming uptick in power outages. The U.S. allows regional jurisdiction of electricity grids. Texas, the only state with its own power grid, suffered a near-apocalyptic power outage in winter 2020.
The U.S. has had a 150 percent increase in the number of major blackouts (those lasting for more than an hour and affecting over 50,000 people) between 2015 and 2021. Most of them occur in summer months, when air conditioning use is at its highest and wildfires and droughts are most frequent, further stressing the electrical grid.
For example, the 2021 heat wave in the U.S. Pacific Northwest resulted in over 600 heat-related deaths and 3,500 people being admitted to hospitals.
Previous heat death models use a single high daily temperature taken at one location, usually at the airport, which fails to account for an increase in indoor temperature as a result of a power outage or variation in temperature in different parts of a city.
A Grim Calculus of Death
The researchers admit to a novel approach to predicting heat-related deaths: “estimating individually experienced temperature [IET] to approximate how personal-level heat exposure changes on an hourly basis, accounting for both outdoor and indoor exposure.” A multiday blackout could double the number of deaths in all three cities with power on, but the loss of power would raise IETs and kill almost 1 percent of the population of Phoenix and hospitalize 50 percent of the population.
IET accounts for people being in indoor and outdoor environments, and moving between the two, including air-conditioned vehicles using data from the American Time Use Survey (ATUS), which is conducted annually by the U.S. Bureau of Labor Statistics. A person indoors, whose temperature is rising because the air conditioning is not working, would experience an elevated IET. Previous mortality studies have relied on maximum outdoor temperatures only and have not accounted for high indoor temperatures from either there being no air conditioning or no power.
Outdoor temperatures were used as input to building models representing buildings typical to a particular parcel so that indoor temperatures in each building could be calculated. This was done for each parcel in the cities studied. A temperature differential between the temperature at the airport (T) and each parcel was determined for every hour of the day to give an analog temperature (T’).
The building modeling was detailed enough to distinguish single-story, two-story and multistory buildings (apartment complexes) as well as account for the variation in construction material and insulation typical to each city.
A “synthetic” population was created with points representing people placed at random locations in buildings until the entire population of each city was accounted for. Each synthetic person was assigned characteristics (age, sex, race and income) at random taken from census data and placed using ESRI ArcGIS software.
Which buildings had central air conditioning, window units or no air conditioning at all was predicted with a regression model prediction based on several factors, including home value, owner-occupied status (rent or own), housing age and structure.
Results
The study used a formula from a publication in the Lancet that estimated deaths from extreme temperatures in various cities worldwide and found the number of heat-related deaths in Atlanta, Detroit and Phoenix.
Detroit had by far the greatest number of deaths during a heat wave if the power stayed on because of Detroit’s “lower adaptive capacity for heat in the form of mechanical cooling.”
In Phoenix, however, afternoon temperatures can reach 45°C (113°F) in the late afternoon. The lowest daily high temperature over a 5-day heat wave modeled was 43°C (109°F) and daily minimum temperatures average 32°C (89°F).
“The heat-related mortality increase by about 700% relative to the Power On scenario, reflecting the extremity of heat exposures in a desert city in the absence of mechanical AC.”
A 5-day heat wave in Phoenix and a simultaneous power outage will have almost 1 percent mortality, or 917 deaths per 100,000 people. Given that the population of Phoenix is 1,446,000, we arrive at a total of 13, 260 heat-related deaths.
The study recommends these infrastructure changes: 1) enhancing and creating a more resilient power grid 2) planting more trees and 3) using high albedo roofing material.
References
- Brian Stone et al. How Blackouts during Heat Waves Amplify Mortality and Morbidity Risk, Environmental Science & Technology, May 23, 2023.
- Jack Healy. In Phoenix, Heat Becomes a Brutal Test of Endurance, New York Times, July 13, 2023.
- Carly D. Ziter et al. Scale-Dependent Interactions Between Tree Canopy Cover and Impervious Surfaces Reduce Daytime Urban Heat During Summer, Proceedings of the National Academy of Sciences, March 25, 2019.
