Impacts of Mt Pinatubo volcanic aerosol on the tropical stratosphere in chemistry–climate model simulations using CCMI and CMIP6 stratospheric aerosol data

摘要

To simulate the impacts of volcanic eruptions on the stratosphere,chemistry–climate models that do not include an online aerosol module requiretemporally and spatially resolved aerosol size parameters for heterogeneouschemistry and aerosol radiative properties as a function of wavelength. Forphase 1 of the Chemistry-Climate Model Initiative (CCMI-1) and, later, forphase 6 of the Coupled Model Intercomparison Project (CMIP6) two suchstratospheric aerosol data sets were compiled, whose functional capabilityand representativeness are compared here. For CCMI-1, the data set was compiled, which hinges on the measurements at four wavelengthsof the SAGE (Stratospheric Aerosol and Gas Experiment) II satelliteinstrument and uses ground-based lidar measurements for gap-fillingimmediately after the 1991 Mt Pinatubo eruption, when the stratosphere wastoo optically opaque for SAGE II. For CMIP6, the new dataset was compiled, which excludes the least reliable SAGE II wavelength anduses measurements from CLAES (Cryogenic Limb Array Etalon Spectrometer) onUARS, the Upper Atmosphere Research Satellite, for gap-filling following theMt Pinatubo eruption instead of ground-based lidars. Here, we performedSOCOLv3 (Solar Climate Ozone Links version 3) chemistry–climate modelsimulations of the recent past (1986–2005) to investigate the impact of theMt Pinatubo eruption in 1991 on stratospheric temperature and ozone and howthis response differs depending on which aerosol data set is applied. The useof SAGE-4 results in heating and ozone loss being overestimated inthe tropical lower stratosphere compared to observations in the post-eruptionperiod by approximately 3 K and 0.2 ppmv, respectively. However, lessheating occurs in the model simulations based on SAGE-3, because theimproved gap-filling procedures after the eruption lead to less aerosolloading in the tropical lower stratosphere. As a result, simulated tropicaltemperature anomalies in the model simulations based on SAGE-3 forCMIP6 are in excellent agreement with MERRA and ERA-Interim reanalyses in thepost-eruption period. Less heating in the simulations with SAGE-3means that the rate of tropical upwelling does not strengthen as much as itdoes in the simulations with SAGE-4, which limits dynamical upliftof ozone and therefore provides more time for ozone to accumulate in tropicalmid-stratospheric air. Ozone loss following the Mt Pinatubo eruption isoverestimated by up to 0.1 ppmv in the model simulations based onSAGE-3, which is a better agreement with observations than in thesimulations based on SAGE-4. Overall, the CMIP6 stratosphericaerosol data set, SAGE-3, allows SOCOLv3 to more accurately simulatethe post-Pinatubo eruption period.