The mid-Holocene East Asian summer monsoon simulated by PMIP4-CMIP6 and PMIP3-CMIP5: Model uncertainty and its possible sources

Yubin Wu,Yonggang Liu,Weijian Zhou,Jian Zhang
DOI: https://doi.org/10.1016/j.gloplacha.2022.103986
IF: 4.956
2022-11-14
Global and Planetary Change
Abstract:The simulated changes of East Asian monsoonal precipitation between mid-Holocene (MH) and pre-industrial (PI) can be either positive or negative depending on the climate model used. This model disagreement becomes even more severe in the Paleoclimate Model Intercomparison Project Phase 4 with Coupled Model Intercomparison Project Phase 6 (PMIP4-CMIP6) than in PMIP3-CMIP5. Here we investigate the possible sources of this model disagreement in North China (NC) and South China (SC). Among the numerous factors analyzed, we find that the PI-to-MH precipitation change over NC and SC are most affected by the uncertainties in the simulated PI East Asian summer monsoon (EASM) intensity (represented by the strength of southerly wind) and the relative PI-to-MH temperature changes between extratropical Eurasian continent and the western tropical Pacific, respectively. If a model overestimates the EASM intensity in PI, it then tends to underestimate its strengthening from PI to MH and the NC summer precipitation, and may even produce a drying there. This relation is stronger in PMIP3-CMIP5 than in PMIP4-CMIP6. If a model overestimates the PI-to-MH summer warming over the extratropical Eurasian continent and the cooling over the western tropical Pacific, it then tends to overestimate the westward extension of the western North Pacific subtropical high over SC, and to produce a negative precipitation change there. The first factor represents the uncertainty in the simulated PI climatology of models while the second is related to the uncertainty in the climate sensitivity of models, they are independent of each other in PMIP4-CMIP6 but positively correlated in PMIP3-CMIP5.
geosciences, multidisciplinary,geography, physical
What problem does this paper attempt to address?