Evaluation of Tropical Biomass Burning Emissions using a Global Aerosol Model: Implications for Climate and Air Quality

Friday, 19 December 2014: 8:00 AM
Carly Reddington1, Dominick V Spracklen1, Alexandru Rap1, Paulo Artaxo2, Luciana Varanda Rizzo3, Andrea Arana2, David A Ridley4, William Morgan5, Hugh Coe6, Ying Ying Toh7, Ken S Carslaw1 and Graham Mann1, (1)University of Leeds, Leeds, United Kingdom, (2)USP University of Sao Paulo, São Paulo, Brazil, (3)Universidade Federal de São Paulo, Departamento de Ciências Exatas e da Terra, Doadema, Brazil, (4)Massachusetts Institute of Technology, Cambridge, MA, United States, (5)University of Manchester, Manchester, United Kingdom, (6)University of Manchester, Manchester, M13, United Kingdom, (7)Malaysian Meteorological Department, Environmental Studies Division, Petaling Jaya, Malaysia
Tropical biomass burning (BB) emissions are a major global source of particulate matter (PM). However, high temporal and spatial variability in the emissions lead to major challenges for their quantification and representation in models. Previous modelling studies have found a persistent underestimation of aerosol optical depth (AOD) observed in BB regions, requiring fire emissions of PM to be scaled by a factor of ~2-5 to match observations. Here we evaluate a global aerosol microphysics model (GLOMAP) against long-term in-situ observations of PM and black carbon in Amazonia and SE Asia, in addition to AOD, to better understand tropical BB aerosol and improve estimates of the impact of BB on air quality and climate.

We performed simulations with GLOMAP for the period 1997–2012 using three different BB emission inventories: GFED3, GFASv1.0 and FINNv1. We find that in N.E. Brazil the model with GFED emissions is unable to capture the magnitude and seasonal variability in observed PM (R2=0.3, bias = -44%), even when emissions are scaled up by a factor of 3.4 (R2=0.3, bias = -26%). This is likely due to an omission of small fires in the GFED3 emissions. Agreement between model and observations in this location is improved by using an active fire based emission inventory (e.g. FINN, R2=0.6, bias = -7%), which better captures emissions from small-scale agricultural fires.

Comparisons with all collected observations in SE Asia and Amazonia show that the model captures the inter-annual and seasonal cycles of monthly mean PM (mean R2 GFED, 0.6; GFAS, 0.7; FINN, 0.6) and AERONET AOD (mean R2 GFED, 0.6; GFAS, 0.7; FINN, 0.8) reasonably well. However, the model is generally negatively biased for both PM and AOD, the smallest bias being with the FINN emission inventory. By scaling FINN emissions up by 50% (much less than found from previous studies using only AOD), we find that we can adequately capture the magnitude of dry season PM across Amazonia and AOD in SE Asia. To capture the magnitude of AOD in Amazonia an additional scaling factor may need to be applied to FINN emissions injected above the surface. To evaluate the vertical profile of modelled BB aerosol we use aircraft observations from the SAMBBA field campaign. Overall, our results could have important implications for quantifying the air quality and climate effects of BB aerosol.