Observation and Model Comparison on Precipitation response to Volcanic Aerosols in the Asian Monsoon Region

Abstract:

Disagreement between observation and models were shown on the volcanic sulfate aerosols’ effect on precipitation in Asia monsoon region. Here, we investigate it by classifying two groups of historical volcanism between AD 1300 and AD 1850 to 2, 1, and 0.5 times Pinatubo sulfate injection into the northern hemisphere (NH) stratosphere based on IVI2 and Crowley2013 volcanic reconstructions, then comparing precipitation response of BCC-CSM1 and CCSM4 model outputs under past1000 scenario to IVI2 volcanic group, and that of MIROC-ESM and MPI-ESM-P to Crowley2013 group with tree-ring reconstruction data MADA. In both groups, Superposed Epoch Analysis (SEA) of MADA and four model outputs show a drying trend over Asia monsoon regions after the NH injections and drier with larger sulfate magnitude, with a 1 or 2 year time lag in MADA comparing to the model outputs, this may result from the biological response of tree ring data that lag behind the meteorological forcing of model outputs. On the other hand, different responses to Southern Hemisphere (SH) only injections were found between the two groups as well as MADA and model outputs. Most of the results were found significant at 90% or even 95% significance level with a 10,000 Monte Carlo resampling procedure. Spatial variation of MADA show a significant drying effect in central Asia in year 1, and then move westward in year 2 and 3 after 2, 1×Pinatubo eruptions of IVI2, while a significant wetting effect in northwest Asia but drying effect in south Asia were shown in Crowley2013 group. However, model outputs did not show spatial variation, with a pattern drier in northwest than in southeast Asia along the years after the eruptions in both volcanic groups. Thus, observation and model outputs are well consistent on precipitation response to NH aerosol injections, but models may need large improvement on the response to SH aerosol injection as well as the spatial variation. Besides, opposite precipitation response to SH injection of the two volcanic groups also indicate the discrepancy between the two reconstructions, which may lead to discrepancy in different models used as volcanic forcing.