Simulation of the Central Indian Ocean Mode in CESM: Implications for the Indian Summer Monsoon System

Abstract

The simulation of the Indian summer monsoon and its pronounced intraseasonal component in a modern climate model remains a significant challenge. Recently, using observations and reanalysis products, the central Indian Ocean (CIO) mode was found to be a natural mode in the ocean‐atmosphere coupled system and also shown to have a close mechanistic connection with the monsoon intraseasonal oscillation (MISO). In this study, the simulation of the actual CIO mode in historical Community Earth System Model (CESM) outputs is assessed by comparing with observations and reanalysis products. The simulation of the Madden‐Julian Oscillation, a major component of tropical intraseasonal variabilities (ISVs), is satisfactory. However, the CIO mode is not well captured in any of the CESM simulations considered here. The force and response relationship between the atmosphere and the ocean associated with the CIO mode in CESM is opposite to that in nature. The simulated meridional gradient of large‐scale zonal winds is too weak, which precludes the necessary energy conversion from the mean state to the ISVs and cuts off the energy source to MISO in CESM. The inability of CESM to reproduce the CIO mode seen clearly in nature highlights the CIO mode as a new dynamical framework for diagnosing the deficiencies in Indian summer monsoon simulation in climate models. The CIO mode is a coupled metric for evaluating climate models and may be a better indicator of a model's skill to accurately capture the tropical multiscale interactions over subseasonal to interannual timescales.

Simulation of Central Indian Ocean Mode in S2S Models

“…Thus, the necessary condition for barotropic instability cannot be satisfied. Such results are consistent with Zhou et al (2018), which concluded that the bias in simulating the meridional structures of was a critical reason for the misrepresentation of the CIO mode in CESM.…”

“…The CIO mode is not sensitive to the domain of EOF analysis or to the data that are used to extract this mode. Previous studies have shown that the CIO mode (represented with the CIO mode index) has a high simultaneous correlation with MISO (Zhou et al, 2017a(Zhou et al, , 2018 but it does not have a significant linear dependence on ENSO or IOZDM (Zhou et al, 2017b). Therefore, it is expected that the CIO mode should contribute to improve the prediction of MISO as an independent predictand.…”

“…The MJO is usually assessed and monitored by daily Real‐time Multivariate (RMM) index developed by Wheeler and Hendon (2004). The CIO mode index does not have a high lead or lag correlation with either RMM index component (Zhou et al, 2018). The maximum correlation coefficient is just around 0.35, which is not statistically significant at 90% confidence level.…”

“…Although the CIO mode is recognized as an important driver of the subseasonal variability and a potential source of predictability during the ISM, Zhou et al (2018) found that the CIO mode is not well captured in historical simulations of the Community Earth System Model (CESM). The relation between U850 and SST anomalies associated with the CIO mode in CESM is opposite to that obtained from observations.…”

“…The bias in the CIO mode simulations is mainly attributable to the weak meridional shear of simulated zonal winds, which weakens the barotropic energy conversion from background zonal winds to subseasonal variabilities. Despite the reasonable MJO simulations in CESM (Boyle et al, 2015; Li et al, 2016; Subramanian et al, 2011; Zhou et al, 2018), most CESM ensembles examined in Zhou et al (2018) lacked the ability to reproduce MISOs and subseasonal rainfall during the ISM (Sperber et al, 2001; Waliser et al, 2003). However, it was found that the northward propagating MISOs were better simulated in the ensemble members which had a relatively better ability to capture the CIO mode.…”