Impact of the stratosphere on the sea surface temperature and ENSO based on HadGEM control runs comparing high top and low top model configurations

Numerous studies on the effects of the El Nino-Southern Oscillation (ENSO) on the stratosphere have been conducted in recent years. However, few of these have examined whether the use of an adequate representation of the stratosphere might affect simulations of the ENSO. In the present work, sea surface temperature data from two numerical model configurations, namely one with a well-resolved stratosphere, the "high top configuration" (Hadley Centre Global Environmental Model, HadGEM2-CCS), and the other without a well-resolved stratosphere, the "low top configuration" (HadGEM2-CC), are employed to study the impact of the stratosphere on the surface climate, especially on the ENSO. A pre-industrial control run is performed to eliminate interference from other factors, such as greenhouse gas warming and volcanic eruptions. Based on the present research, both model configurations function reasonably well and have shown little difference from each other when analyzng the global annual and seasonal mean sea surface temperatures, except for the Northern Atlantic Ocean region. A statistical analysis performed using the t-test method shows that the significant differences in the annual and seasonal mean sea surface temperatures in the Northern Atlantic region result from real signals rather than random noises. Furthermore, the configuration with a better representation of the stratosphere simulates the quasi-period of the ENSO and the seasonal phase-locking characteristics of El Nino more precisely. Therefore, it is probably advantageous to adopt climate models that resolved stratosphere for a more realistic representation of ENSO climatology and its possible variations under certain conditions.