The double‐ITCZ problem in CMIP6 and the influences of deep convection and model resolution. (10th January 2023)

• Record Type: Journal Article
• Authors: Ma, Xinyu; Zhao, Shuyun; Zhang, Hua; Wang, Wuke
• Is Part Of: International journal of climatology. Volume 43:Number 5(2023)
• Page Start: 2369
• Page End: 2390
• Publication Date: 2023-01-10
• Subjects: CMIP6 -- deep convection -- double‐ITCZ bias/problem -- model resolution; Climatology -- Periodicals; Climat -- Périodiques; Climatologie -- Périodiques; 551.605
• DOI: 10.1002/joc.7980
• Languages: English
• ISSNs: 0899-8418

Abstract

Abstract: The double Intertropical Convergence Zone (ITCZ) bias is an outstanding bias in many climate models. This work assesses the annual‐mean double‐ITCZ problem in the models participating in the Coupled Model Intercomparison Project Phase 6 (CMIP6) based on several quantitative indices. Within the 46 CMIP6 models, 9 models from mainland China are evaluated as a group to verify the effort of model development from one perspective. The double‐ITCZ bias and its large intermodel spread still exist in CMIP6 models. The overall performance of the models from Chinese mainland is similar with all CMIP6 models. It is found that the top‐five models with relatively low double‐ITCZ biases can effectively restrain the frequency of deep convection and related sea surface temperature (SST) bias in the southeastern Pacific dry subsidence region, which highlights the necessity of improving convective physics in climate models. Impacts of model resolution on the double‐ITCZ problem are examined by comparing the high‐ and low‐resolution groups in CMIP6 and High Resolution Model Intercomparison Project (HighResMIP) historical experiments, respectively. Increased resolution in atmospheric models is found to be able to reduce the positive precipitation bias over the tropical southern Atlantic and improve the simulation of deep convection frequency and convective precipitation ratio there. However, the double‐ITCZ bias over the Pacific is not improved significantly by increased resolution. Abstract : Model biases (mod‐obs represents model results minus observation) of the (a) Pacific and (b) Atlantic southern ITCZ indices (SIs, unit: mm·day −1 ); the (c) global, (d) Pacific, and (e) Atlantic tropical precipitation asymmetry indices (TPAIs, unit: 1); the (f) global, (g) Pacific, and (h) Atlantic equatorial precipitation indices (EPIs, unit: 1) of different CMIP6 models. MMEM, L5MM, H5MM, and MCMM denote the multimodel ensemble medians of all CMIP6 models, the top‐five models with the lowest errors, the top‐five models with the highest errors, and nine models from Chinese mainland, respectively. Models are ordered according to the absolute values of different indices' biases (from smallest to largest), and hatched bars highlight the models from China mainland.