Improvement of biogenic emissions estimation in the Canadian Lower Fraser Valley and its impact on particulate matter modeling results

As part of the effort to reduce uncertainties in air quality modeling in the Canadian Lower Fraser Valley (LFV) area, we made modifications to the landuse dataset, emission factors and environmental correction algorithms of Biogenic Emission Inventory System (BEIS2) implemented in Sparse Matrix Operator Kernel Emission (SMOKE) and studied the impact of these modifications on Community Multiscale Air Quality particulate matter (PM) modeling results. The landuse dataset Biogenic Emissions Land cover Database version 3 with 1 km spatial and 230 landuse-category resolution was enhanced with Canadian forest and crops data. This replaced the SMOKE/BEIS2 county-based, 127 landuse-category dataset. BEIS3 emission factors were used, with modifications in nitric oxide factors for agricultural categories, to replace BEIS2 emission factors. The surface shortwave downward radiation flux from MM5 was used to replace the flux calculated by Models3 Meteorological-Chemistry Interface Processor. The soil temperature from the Oregon State University/Eta land surface model coupled with MM5 was used to replace empirically computed soil temperature. The spatial distribution of modified biogenic emissions and comparison of air quality model results against observations in the LFV region showed that the modifications improved the model performance. On a domain-average basis, these modifications resulted in approximately 100% increase in biogenic nitric oxide, 50% increase in monoterpenes and 14–20% decrease in isoprene emissions for the period 31 July–7 August 1993. Analyses showed that the modifications to the landuse dataset and emission factors contributed most to the changes in emissions. From an episode-average perspective, these changes caused approximately 12–85% biogenic secondary organic aerosol (SOA) increases. The changes in anthropogenic SOA and fine inorganic aerosol were approximately −5% to 4% and −0.5% to 4%. The modifications also changed the modeled contributions from biogenic sources to the PM loading in the atmosphere.