Identifying the Drivers of Modeling Uncertainties in Isoprene Emissions: Schemes Versus Meteorological Forcings

Abstract

AbstractIsoprene dominates global biogenic volatile organic compounds and has significant impacts on atmospheric chemistry. Global simulations of isoprene emissions show large uncertainties in both the trend and interannual variability. Here, we explored the uncertainties of simulated isoprene emissions during 1982–2015 using the Yale Interactive terrestrial Biosphere model, which is implemented with two emission schemes (PS_BVOC and Model of Emissions of Gases and Aerosols from Nature [MEGAN]) and driven by two long‐term reanalysis meteorology (WFDEI and Modern‐Era Retrospective Analysis for Research and Applications [MERRA]). Model evaluations show that both PS_BVOC and MEGAN schemes capture reasonable spatial patterns of the global isoprene emissions. The averaged fluxes from simulations using two meteorological forcings show isoprene trends of 0.69 Tg C m2 a−1 with PS_BVOC scheme but −0.22 Tg C m2 a−1 with MEGAN scheme. Such opposite trend is mainly because PS_BVOC considers both CO2 fertilization and inhibition effects to isoprene while the MEGAN scheme implements CO2 inhibition effect alone. Meanwhile, the averaged fluxes from simulations using two schemes yield interannual variability of 1.59% with WFDEI and 2.06% with MERRA reanalyses. Such discrepancies are mainly attributed to the larger variability in MERRA data. On the global scale, differences in schemes are the main driver of uncertainties in isoprene trends. In contrast, differences in meteorological forcings dominate the uncertainties in the interannual variability of isoprene emissions. Our sensitivity experiments reveal the key sources of modeling uncertainties, which are vital for the improvement of parameterizations and future projections of isoprene emissions.