"Estimating Economic Damage from Climate Change in the United States"

This page contains resources for "Estimating Economic Damage from Climate Change in the United States" by Hsiang, Kopp, Jina, Rising, Delgado, Mohan, Rasmussen, Muir-Wood, Wilson, Oppenheimer, Larsen, Houser (Science, 2017).

This analysis was produced by the Climate Impact Lab, a consortium of researchers from UC Berkeley, Rutgers, University of Chicago, and Rhodium Group, along with our research partners at Princeton University and RMS.

Data Visualization

An interactive data visualization at the Climate Impact Lab website is at http://www.impactlab.org/map/ where visitors can interact with both the climate and economic data produced through our analysis.

Data download

Note: This data is provided for non-commercial research and educational purposes.

County-level damages by sector (CSV) - This file contains median damages for each sector shown in Fig 2 of the main article (agriculture, mortality, energy, low-risk labor, high-risk labor, coastal damages, property crime, violent crime, total damages).  These impacts are the central estimate for average annual damage during 2080-2099 under a business-as-usual scenario (RCP8.5). UPDATE: This file has been updated (6/26/17) to include avg county income and population in 2012, as well as to replace 4 missing values.

Total economic damages for each county by likelihood (CSV) - This file contains total damages for each county as a fraction of county income, similar to Fig 2I in the main article except, it also contains ranges of total damages likelihood intervals (as shown in Fig 5C in the main article) defined using the standards set by the IPCC. The file contains median estimates (50 percentile), as well as the likely range (17-83 percentiles, which is the 66% chance central interval), and the very likely range (5-95 percentiles, which is the 90% chance central interval). All values are percent of county income. UPDATE: This file has been updated (6/26/17) to include avg county income and population in 2012, as well as to replace 4 missing values.

Total economic damages distributions for each income decile of counties (CSV) - This file contains descriptions of total damages for US counties grouped by their income decile (it is the data used to make the "Inequality of damages in the USA" figure below). For each decile, the file contains median estimates (50 percentile), as well as the likely range (17-83 percentiles, which is the 66% chance central interval), and the very likely range (5-95 percentiles, which is the 90% chance central interval).

Full data archive for researchers (zenodo.org)

Figures for press download

Featured
Median total damages
Median total damages

County-level annual damages in median scenario for climate during 2080-2099 under business-as-usual emissions trajectory (RCP8.5). Negative damages indicate economic benefits. Map corresponds with Figure 2I in main article.

Inequality of damages in the USA
Inequality of damages in the USA

Range of economic damages per year for groupings of US counties, based on their income (29,000 simulations for each of 3,143 counties). The poorest 10% of counties are the leftmost box plot. The richest 10% are the rightmost box plot. Damages are fraction of county income. White lines are median estimates, boxes show the inner 66% of possible outcomes, outer whiskers are inner 90% of possible outcomes. This figure is a simplified version of Figure 5C in the main article.

Climate Damage Function for the USA
Climate Damage Function for the USA

Estimated total direct damage to US economy per year in response to global mean temperature changes (horizontal axis). The red line is average losses across 29,000 simulations. The shaded region marks the 90% confidence interval across simulations (there is a 90% chance a simulation lands in this region). This graph corresponds with Figure 5A in the main article.

Damage function for USA (Fahrenheit only)
Damage function for USA (Fahrenheit only)

Estimated total direct damage to US economy per year in response to global mean temperature changes (horizontal axis). The red line is average losses across 29,000 simulations. The shaded region marks the 90% confidence interval across simulations (there is a 90% chance a simulation lands in this region). This graph corresponds with Figure 5A in the main article.

Median damages by sector
Median damages by sector

County-level annual damages in median scenario for climate during 2080-2099 under a business-as-usual emissions trajectory (RCP8.5). Negative damages indicate benefits. From left to right, top to bottom: percent change in yields, averaged for maize, wheat, soybeans, and cotton; changes in all-cause mortality rates, across all age groups; change in electricity demand; change in labor supply of full-time equivalent workers for low risk jobs where workers are minimally exposed to outdoor temperature; same as previous except for high risk jobs where workers are heavily exposed to outdoor temperatures; change in damages from coastal storms and sea level rise; changes in violent crime rates; changes in property crime rates; median total direct economic damage across all sectors. Maps corresponds with Figure 2 in main article.

Floodplain for New York, NY
Floodplain for New York, NY

Simulations of areas inundated by 1-in-100 year floods under median sea level rise under business-as-usual (RCP8.5) warming, accounting for forecast changes in tropical cyclone activity. The analysis covered the entire US coast, this image is for New York, NY. These correspond with Figure 4B and 4D in the main article.

Floodplain for Miami, FL
Floodplain for Miami, FL

Simulations of areas inundated by 1-in-100 year floods under median sea level rise under business-as-usual (RCP8.5) warming, accounting for forecast changes in tropical cyclone activity. The analysis covered the entire US coast, this image is for Miami, FL. These correspond with Figure 4B and 4D in the main article.

Floodplain for Norfolk VA
Floodplain for Norfolk VA

Simulations of areas inundated by 1-in-100 year floods under median sea level rise under business-as-usual (RCP8.5) warming, accounting for forecast changes in tropical cyclone activity. The analysis covered the entire US coast, this image is for Norfolk, VA. These correspond with Figure 4B and 4D in the main article.

Floodplain in Houston, TX
Floodplain in Houston, TX

Simulations of areas inundated by 1-in-100 year floods under median sea level rise under business-as-usual (RCP8.5) warming, accounting for forecast changes in tropical cyclone activity. The analysis covered the entire US coast, this image is for Houston, TX. These correspond with Figure 4B and 4D in the main article.

Floodplain for Washington, DC
Floodplain for Washington, DC

Simulations of areas inundated by 1-in-100 year floods under median sea level rise under business-as-usual (RCP8.5) warming, accounting for forecast changes in tropical cyclone activity. The analysis covered the entire US coast, this image is for Washington, DC. These correspond with Figure 4B and 4D in the main article.

Combining output from global climate models
Combining output from global climate models

44 climate models (outlined maps) and model surrogates (dimmed maps) are weighted so the distribution of the range of 2080-2099 global mean surface temperature changes match the probability distribution of estimated changes based on climate sensitivity estimates (blue-grey line) under a business-as-usual emissions trajectory (RCP8.5). Image corresponds with Figure 1A in the main article.