Evaluation of biogenic emissions from three different vegetation distributions in South Korea

Map of plant functional type (PFT) is one of key information in estimating the bottom-up biogenic emissions. This study investigated the performance of three different maps of PFT on biogenic emissions in South Korea: i) PFT map from the MODerate resolution Imaging Spectroradiometer (MODIS) satellite sensor onboard Terra-/Aquasatellites (MPFT); ii) PFT map from the Environmental Geographic Information System (EGIS) provided by the Korean Ministry of Environment (EPFT); and iii) PFT map from the Forest Geographic Information System (FGIS) established by the Korea Forest Service (FPFT). These three different vegetation distributions were used as input information into the Model of Emissions of Gases and Aerosols from Nature (MEGAN) model. The biogenic emissions estimated by the MEGAN model simulations were then evaluated by comparisons between the Community Multi-scale Air Quality (CMAQ) simulated mixing ratios of biogenic species with ground-and aircraft-borne mixing ratios during the period of the Korea-United States Air Quality (KORUS-AQ) campaign. The analyses found that the simulation using the FPFT map, which has been regarded as 'ground truth', produced the best results in terms of several statistical metrics. For example, among the three experiments, the simulation with the FPFT map produced the best index of agreement (IOA) of 0.76 and the lowest normalized mean bias (NMB) of 5.61%, in comparison with ground-based isoprene observations. The simulations with the MPFT map also performed reasonably, showing low NMBs of-42.13% and-14.87% for the isoprene and the first -generation products of isoprene, respectively, in comparison with the DC-8 airborne observations. In addition, the applications of the MPFT map to the MEGAN modeling showed significant overestimations in the mixing ratios of biogenic species over the SoBaek (SB) mountain ranges. It was believed that this was due to the overestimated density of broadleaf deciduous temperate trees over the SB mountain ranges in the MPFT map. On the other hand, the EPFT map showed the worst performances, which might be related to crude classification of vegetation. This study also suggests two important possibilities. First, the current isoprene-related photochem-istry appears to be highly uncertain, based on large underestimations of the first-generation products of isoprene, in comparison with observations. Second, the emission factors of monoterpenes in the MEGAN model may be too small, based on large underestimations of monoterpenes, in comparison with observations