The present study examines the ability of Coupled Model Inter‐comparison Project phase 6 (CMIP6) models in representing the dominant modes of tropical Indian Ocean (TIO) sea surface temperature (SST) variability on the interannual and decadal time scale. Historical simulations from 27 CMIP6 models are assessed against Extended Reconstructed SST over the period of 1854 to 2014. Spectrum analysis reveals that many models reproduce interannual and decadal variability of TIO SST but underestimate the amplitude of variability with some disparity in the periodicity. All models can reproduce the dominant basin‐wide mode of interannual and decadal variability of TIO SST reasonably well. Skill score analysis of TIO SST variability reveals that KACE‐1‐0‐G has highest skill, followed by FGOALS‐f3‐L, EC‐Earth3‐Veg‐LR, ACCESS‐ESM1‐5, CanESM5‐CanOE on the interannual timescale and FGOALS‐f3‐L, CanESM5‐CanOE, KACE‐1‐0‐G and CanESM5, respectively, showed highest skills for decadal variability. It is found that variations in radiation and latent heat flux are primarily responsible for interannual variability in TIO SST, the basin‐wide warming, in the observations. Taylor diagram analysis reveals that all the models exhibit better skill for the radiative flux; however, skill for the latent heat and momentum flux varies from model to model. It is important to note that the models in which the latent heat flux and zonal wind are better represented have produced better TIO SST variability compared to other models. A higher discrepancy in latent heat and zonal momentum flux leads to improper wind‐evaporation‐SST and wind‐circulation‐SST feedback, which in turn restricts the model skill. Besides, model that has realistic central and eastern Pacific SST variability show better skill for TIO SST variability in both interannual and decadal time scales. The present study advocates that better representation of latent heat flux and zonal wind in coupled models is important for the accurate simulation of interannual and decadal variability in TIO SST.

中文翻译：

---

CMIP6模型对热带印度洋海表温度变化的技能评估

本研究检验了比较模型间比较项目第6阶段（CMIP6）模型在年际和年代际尺度上代表热带印度洋（TIO）海面温度（SST）变异的主导模式的能力。根据1854年至2014年期间的扩展重建SST对27个CMIP6模型的历史模拟进行了评估。频谱分析显示，许多模型再现了TIO SST的年际和年代际变化，但低估了变化幅度，但在周期性上存在一些差异。所有模型都可以很好地再现TIO SST的年际和年代际变化的主导盆地范围模式。TIO SST变异性的技能得分分析显示，KACE-1-0-G具有最高的技能，其次是FGOALS-f3-L，EC-Earth3-Veg-LR，ACCESS-ESM1-5，年际时间尺度上的CanESM5-CanOE和FGOALS-f3-L，CanESM5-CanOE，KACE-1-0-G和CanESM5表现出最高的年代际可变性技能。在观测中发现，辐射和潜热通量的变化是造成TIO SST年际变化的主要原因，即整个盆地的变暖。泰勒图分析表明，所有模型都表现出更好的辐射通量技术。但是，潜热和动量通量的技巧因模型而异。重要的是要注意，与其他模型相比，能够更好地表示潜热通量和纬向风的模型产生了更好的TIO SST变异性。潜热和纬向动量通量的较高差异会导致不正确的风蒸发-SST和风循环-SST反馈，反过来限制了模型技能。此外，具有现实的太平洋中部和东部SST变异性的模型在年际和年代际尺度上都表现出更好的TIO SST变异性技能。本研究主张，在耦合模型中更好地表示潜热通量和纬向风对于TIO SST年际和年代际变化的准确模拟非常重要。