Evaluation of volcanic aerosol impacts on atmospheric water vapor using CMIP3 and CMIP5 simulations

The uncertainty in estimates of volcanic forcing and in the way general circulation models (GCMs) implement volcanic forcing affects the accuracy of the simulated atmospheric water vapor response to volcanic eruptions, which in turn influences the design of hydrologic infrastructure. This paper aims to investigate whether and to what extent the models can simulate the volcanic impacts on atmospheric water vapor. While a Kolmogorov‐Smirnov (K‐S) test on atmospheric water vapor from GCMs, reanalysis products, and observations all imply that volcanic aerosols have a statistically significant impact on atmospheric water vapor, and the areal extent of global atmospheric water vapor affected by volcanic eruptions is oversimulated by GCMs. Additionally, the spatial and temporal patterns of the volcano‐related water vapor variability are not well simulated by GCMs. Furthermore, the strong negative correlation between aerosol optical depth and water vapor residual is observed only in the equatorial Atlantic and Indian Oceans, and parts of the western Pacific Ocean rather than across the entire globe as simulated by volcanic forcing included Coupled Model Intercomparison Project Phase 3 (CMIP3) multimodel average (or over nearly the entire tropics by volcanic forcing included the BCC‐CSM1.1). Although the uncompleted removal of El Niño–Southern Oscillation signal may compromise the results, it is still safe, after performing the control run of K‐S test, to draw the conclusion that while volcanic forcing does exert a significant impact on atmospheric water vapor, the underlying relationship is considerably more complex than that evident in GCM simulations.