Distinguishing the drivers of trends in land carbon fluxes and plant volatile emissions over the past three decades

Abstract

Abstract. The terrestrial biosphere has experienced dramatic changes in recent decades. Estimates of historical trends in land carbon fluxes remain uncertain because long-term observations are limited on the global scale. Here, we use the Yale Interactive terrestrial Biosphere (YIBs) model to estimate decadal trends in land carbon fluxes and emissions of biogenic volatile organic compounds (BVOCs) and to identify the key drivers for these changes during 1982–2011. Driven with hourly meteorology from WFDEI (WATCH Forcing Data methodology applied to ERA-Interim data), the model simulates an increasing trend of 297 Tg C a−2 in gross primary productivity (GPP) and 185 Tg C a−2 in the net primary productivity (NPP). CO2 fertilization is the main driver for the flux changes in forest ecosystems, while meteorology dominates the changes in grasslands and shrublands. Warming boosts summer GPP and NPP at high latitudes, while drought dampens carbon uptake in tropical regions. North of 30° N, increasing temperatures induce a substantial extension of 0.22 day a−1 for the growing season; however, this phenological change alone does not promote regional carbon uptake and BVOC emissions. Nevertheless, increases of LAI at peak season accounts for ~ 25 % of the trends in GPP and isoprene emissions at the northern lands. The net land sink shows statistically insignificant increases of only 3 Tg C a−2 globally because of simultaneous increases in soil respiration. In contrast, driven with alternative meteorology from MERRA (Modern Era-Retrospective Analysis), the model predicts significant increases of 59 Tg C a−2 in the land sink due to strengthened uptake in the Amazon. Global BVOC emissions are calculated using two schemes. With the photosynthesis-dependent scheme, the model predicts increases of 0.4 Tg C a−2 in isoprene emissions, which are mainly attributed to warming trends because CO2 fertilization and inhibition effects offset each other. Using the MEGAN (Model of Emissions of Gases and Aerosols from Nature) scheme, the YIBs model simulates global reductions of 1.1 Tg C a−2 in isoprene and 0.04 Tg C a−2 in monoterpene emissions in response to the CO2 inhibition effects. Land use change shows limited impacts on global carbon fluxes and BVOC emissions, but there are regional contrasting impacts over Europe (afforestation) and China (deforestation).