Extended Data Fig. 7: Uncertainty in projected annual smoke PM2.5 concentration (Panel A) and mortality (Panel B).
From: Wildfire smoke exposure and mortality burden in the USA under climate change
The red dashed line shows the main estimate. The solid bar shows the 10th and 90th percentile, and the black line shows the 2.5th and 97.5th percentile. Uncertainty from “climate projection” is calculated using the percentiles of 28 GCMs. Uncertainty from “climate-fire model” and “fire-smoke model” is calculated using bootstrap procedures performed on the climate-fire and fire-smoke models. Uncertainty from “emission downscaling” shows the range of smoke and mortality estimated across 40 different downscaling estimates. Uncertainty from “smoke-mortality function” is calculated using bootstrap procedures performed on the smoke-mortality response functions. The combined uncertainty is quantified using a Monte-Carlo simulation approach, where we randomly sample 500 combinations of GCMs, climate-fire models, downscaling estimates, and fire-smoke models, and smoke-mortality functions.
