Statistical Parameterization Expressing ENSO Variability and Reversibility in Response to CO2 Concentration Changes

Abstract

he degree of reversibility in the behavior of El Niño?Southern Oscillation (ENSO) in response to CO2 concentration changes is examined using the air?sea coupled Community Climate System Model, version 4. The model is run under simple and idealized climate change scenarios where atmospheric CO2 concentration is gradually increased to four times the preindustrial level and then gradually reduced at a similar rate along this trajectory. While the simulated ENSO amplitude is reduced when CO2 concentration increases, the amplitude is enhanced when CO2 decreases. This enhancement occurs with a relative El Niño?like warming of mean sea surface temperature. Most of the El Niño?like warming is attributed to a weakened cold water upwelling in the equatorial eastern Pacific. This weakened cooling is a result of a vertical ocean temperature gradient that is reduced in relation to a lag between subsurface temperature warming and surface temperature change. From these findings, a statistical parameterization expressing ENSO variability is developed in terms of the global-mean surface temperature and the vertical gradient of the global-mean temperature in the upper ocean. Plugging this parameterization into a statistical model of ENSO and combining the latter with a statistical spatial pattern of the ENSO-related temperature anomaly enables the reconstruction of surface temperature variability over North America. This approach can be useful when one assesses future climate changes under various CO2 emission scenarios using very simple models whose outputs are limited to a small number of global-mean variables.