Projecting wildfire emissions over the south-eastern United States to mid-century

Abstract

Wildfires can impair human health because of the toxicity of emitted pollutants, and threaten communities, structures and the integrity of ecosystems sensitive to disturbance. Climate and socioeconomic factors (e.g. population and income growth) are known regional drivers of wildfires. Reflecting changes in these factors in wildfire emissions estimates is thus a critical need in air quality and health risk assessments in the south-eastern United States. We developed such a methodology leveraging published statistical models of annual area burned (AAB) over the US Southeast for 2011–2060, based on county-level socioeconomic and climate projections, to estimate daily wildfire emissions in selected historical and future years. Projected AABs were 7 to 150% lower on average than the historical mean AABs for 1992–2010; projected wildfire fine-particulate emissions were 13 to 62% lower than those based on historical AABs, with a temporal variability driven by the climate system. The greatest differences were in areas of large wildfire impacts from socioeconomic factors, suggesting that historically based (static) wildfire inventories cannot properly represent future air quality responses to changes in these factors. The results also underscore the need to correct biases in the dynamical downscaling of wildfire climate drivers to project the health risks of wildfire emissions more reliably.

Evaluating wildfire emissions projection methods in comparisons of simulated and observed air quality

“…The carbon bond 05 gas-phase mechanism (cb05tucl) used in our simulations includes updates to toluene chemistry, homogeneous hydrolysis rate constants for N 2 O 5 , and updates to the chlorine chemistry (Sarwar et al, 2011;Whitten et al, 2010). The aerosol mechanism, AERO6, includes a primary organic aerosol aging scheme (Simon and Bhave, 2012) and an improved representation of fugitive dust; primary speciated emissions needed to model dust are based on Reff et al (2009).…”

Evaluating wildfire emissions projection methods in comparisons of simulated and observed air quality

“…Thus, there is a critical need in Southeastern land and air quality management to consider both these drivers to plan effectively for protecting the public and the environment. This has motivated the recent development of 5 methodologies that include these drivers in projections of wildfire activity (Prestemon et al, 2016) from the present to 2060, and their use in assessing not only current, but also future air quality (Shankar et al, 2018). Prestemon et al (2016) estimated annual areas burned (AAB) over 13 states in the U.S. Southeast using county-level projections of the major wildfire drivers in the Southeast (climate, population and income, and land use).…”

“…Their median projection of aggregate wildfire activity in the Southeast over hundreds of iterations in a Monte Carlo framework shows an increase, albeit by a small amount (4%), from 2011 to 2060 due to the combined influences of these climate and socioeconomic factors (Prestemon et al, 2016). Shankar et al (2018) leveraged the AAB projections of Prestemon et al (2016) to estimate wildfire emissions over a Southeastern modeling grid at 12-km x 12-km spatial resolution suitable for air quality impact assessments. They applied the methodology in that first 15 study to project daily wildfire emissions, which include the influence of both climate and socioeconomic changes in selected years from 2011 to 2060.…”

“…Prestemon et al (2016) used statistically downscaled outputs of nine general circulation model (GCM) realizations to provide the needed meteorological inputs at a fine spatial resolution (12-km x 12-25 km) for their statistical model estimates of AAB at the county-level. These meteorological inputs, however, do not include all the variables needed for an air quality simulation; nor are the available variables at the temporal resolution (hourly) needed for such simulations (Shankar et al, 2018). Thus, the use of AAB estimated with statistically downscaled climate inputs in air quality studies requires additional mesoscale meteorological modeling.…”

“…In this study, we examine the model performance of retrospective air quality (AQ) simulations using wildfire emissions (Shankar et al, 2018) that are a function of changes in climate and socioeconomic factors, with both the statistical and the dynamical climate downscaling methods for the underlying AAB estimates. We compare the AQ model results using these two emissions estimates to those with a standard wildfire inventory compiled from observed daily fire activity without 10 considering changes in climate and socioeconomic factors.…”