Development of a processor in BEIS3 for estimating vegetative mercury emission in the continental United States

Abstract

We have developed a regression-based processor for estimating vegetative mercury emission within the framework of Biogenic Emission Inventory System Version 3.11 (BEIS3 V3.11). In this development, we incorporated the 230 categories of USGS landcover data to generate the vegetation-specific mercury emission in a 36-km Lambert Conformal model grid covering the continental United States (CONUS). The surface temperature and cloud-corrected solar radiation from a Mesoscale Meteorological model (MM5) were retrieved and used for calculating the diurnal variation. The implemented emission factors were either evaluated from the measured mercury flux data for selected tree species, wetland and water, or assumed for the tree species without available flux data. Annual simulations using the 2001 USEPA MM5 data were performed to investigate the seasonal emission variation. From our sensitivity analysis using three sets of emission factors, we estimated that the vegetative mercury emission in the CONUS domain ranges from a lower limit of 31 ton yr −1 to an upper limit of 140 ton yr −1, with the best estimate at 44 ton yr −1. The modeled vegetative emission was mainly contributed from southeast US. Using the best estimate data, it is shown that mercury emission from vegetation is comparable to that from anthropogenic sources in summer (nearly half of the total emission). However, the vegetative emission decreases greatly in winter, leaving anthropogenic sources as the major emission source (>90% in winter months). Modeling assessment indicates that including vegetative emission (44 ton yr −1) can force an increase of ambient mercury concentration of up to 0.2 ng m −3 in summer midday, but has little impact on dry deposition of mercury. Additional emission factors can be implemented in the model once further mercury flux data become available.