Representation of the ENSO Combination Mode and its Asymmetric SST Response in Different Resolutions of HadGEM3

Previous studies have revealed a combination mode （C-mode） occurring in the Indo-Pacific region, arising from nonlinear interactions between ENSO and the western Pacific warm pool annual cycle. This paper evaluates the simulation of this C-mode and its asymmetric SST response in HadGEM3 and its resolution sensitivity using three sets of simulations at horizontal resolutions of N96, N216 and N512. The results show that HadGEM3 can capture well the spatial pattern of the C-mode associated surface wind anomalies, as well as the asymmetric response of SST in the tropical Pacific, but it strongly overestimates the explained variability of the C-mode compared to the ENSO mode. The model with the three resolutions is able to reproduce the distinct spectral peaks of the C-mode at the near annual combination frequencies, but the performance in simulating the longer periods is not satisfactory, presumably due to the unrealistic simulation of the ENSO mode. Increasing the horizontal resolution can improve the consistency between atmospheric and oceanic representations of the C-mode, but not necessarily enhance the accuracy of C-mode simulation compared with observation.

Indian Ocean SST modes and Their Impacts as Simulated in BCC_CSM1.1(m) and HadGEM3

The sea surface temperature anomalies （SSTAs） in the tropical Indian Ocean （TIO） show two dominant modes at interan- nual time scales, referred to as the Indian Ocean basin mode （IOBM） and dipole mode （IOD）. Recent studies have shown that the IOBM and IOD not only affect the local climate, but also induce remarkable influences in East Asia via teleconnections. In this study, we assess simulations of the IOBM and IOD, as well as their teleconnections, using the operational seasonal pre- diction models from the Met Office （HadGEM3） and Beijing Climate Center [BCC_CSMI.I（m）]. It is demonstrated that the spatial patterns and seasonal cycles axe generally reproduced by the control simulations of BCC_CSM1.1 （m） and HadGEM3, although spectra biases exist. The relationship between the TIO SSTA and E1 Nifio is successfully simulated by both mod- els, including the persistent IOBM warming following E1 Nifio and the IOD-E1 Nifio interactions. BCC_CSMI.I（m） and HadGEM3 axe capable of simulating the observed local impact of the IOBM, such as the strengthening of the South Asian high. The influences of the IOBM on Yangtze River rainfall are also captured well by both models, although this teleconnec- tion is slightly weaker in BCC_CSM 1.1 （m） due to the underestimation of the northwestern Pacific subtropical high. The local effect of the IOD on East African rainfall is reproduced by both models. However, the remote control of the IOD on rainfall over southwestern China is not clear in either model. It is shown that the realistic simulations of TIO SST modes and their teleconnections give rise to the source of skillful seasonal predictions over China.