Public health impacts of the severe haze in Equatorial Asia in September–October 2015: demonstration of a new framework for informing fire management strategies to reduce downwind smoke exposure

In September–October 2015, El Niño and positive Indian Ocean Dipole conditions set the stage for massive fires in Sumatra and Kalimantan (Indonesian Borneo), leading to persistently hazardous levels of smoke pollution across much of Equatorial Asia. Here we quantify the emission sources and health impacts of this haze episode and compare the sources and impacts to an event of similar magnitude occurring under similar meteorological conditions in September–October 2006. Using the adjoint of the GEOS-Chem chemical transport model, we first calculate the influence of potential fire emissions across the domain on smoke concentrations in three receptor areas downwind—Indonesia, Malaysia, and Singapore—during the 2006 event. This step maps the sensitivity of each receptor to fire emissions in each grid cell upwind. We then combine these sensitivities with 2006 and 2015 fire emission inventories from the Global Fire Assimilation System (GFAS) to estimate the resulting population-weighted smoke exposure. This method, which assumes similar smoke transport pathways in 2006 and 2015, allows near real-time assessment of smoke pollution exposure, and therefore the consequent morbidity and premature mortality, due to severe haze. Our approach also provides rapid assessment of the relative contribution of fire emissions generated in a specific province to smoke-related health impacts in the receptor areas. We estimate that haze in 2015 resulted in 100 300 excess deaths across Indonesia, Malaysia and Singapore, more than double those of the 2006 event, with much of the increase due to fires in Indonesia's South Sumatra Province. The model framework we introduce in this study can rapidly identify those areas where land use management to reduce and/or avoid fires would yield the greatest benefit to human health, both nationally and regionally.