An improved ENSO simulation by representing chlorophyll-induced climate feedback in the NCAR Community Earth System Model

Xianbiao Kang,Chuan Gao,Jieshun Zhu,Rong-Hua Zhang

doi:10.1038/s41598-017-17390-2

Abstract

The El Niño-Southern oscillation (ENSO) simulated in the Community Earth System Model of the National Center for Atmospheric Research (NCAR CESM) is much stronger than in reality. Here, satellite data are used to derive a statistical relationship between interannual variations in oceanic chlorophyll (CHL) and sea surface temperature (SST), which is then incorporated into the CESM to represent oceanic chlorophyll -induced climate feedback in the tropical Pacific. Numerical runs with and without the feedback (referred to as feedback and non-feedback runs) are performed and compared with each other. The ENSO amplitude simulated in the feedback run is more accurate than that in the non-feedback run; quantitatively, the Niño3 SST index is reduced by 35% when the feedback is included. The underlying processes are analyzed and the results show that interannual CHL anomalies exert a systematic modulating effect on the solar radiation penetrating into the subsurface layers, which induces differential heating in the upper ocean that affects vertical mixing and thus SST. The statistical modeling approach proposed in this work offers an effective and economical way for improving climate simulations.

Highlights

Substantial biases still exist in climate models used to simulate the El Niño-Southern oscillation (ENSO) in the tropical Pacific[1,2], which is the strongest interannual signals in the climate system[3,4,5,6]
A statistical model for interannual CHL variations can be derived from satellite data; the derived CHL anomalies (CHLAs) model is implemented into the National Center for Atmospheric Research (NCAR) CESM to represent oceanic biology-induced feedbacks (Fig. 1)
An singular value decomposition (SVD)-based statistical model for interannual CHL anomalies is derived to identify its relationship with SST23, which is incorporated into the NCAR CESM

Summary

Introduction

Substantial biases still exist in climate models used to simulate the El Niño-Southern oscillation (ENSO) in the tropical Pacific[1,2], which is the strongest interannual signals in the climate system[3,4,5,6]. An El Niño case in the model years 153 and 154 is selected for two more experiments that use the CESM designed to illustrate direct effects of CHLAs on SST differences and the underlying processes responsible for the differences in two runs; CHLclim–EN is a run that uses CHL as its seasonally varying climatology, and CHLinter–EN is a run that considers the interannually varying CHL effects These two runs are restarted from the same initial state at model year 153 and are time integrated for 2 years. Physical basis to derive CHLAs from SSTAs. A statistical model for interannual CHL variations can be derived from satellite data (the details are given in the Supplementary section); the derived CHLA model is implemented into the NCAR CESM to represent oceanic biology-induced feedbacks (Fig. 1). The NCAR CESM-based experiments are conducted to examine the bio-effect on and the underlying processes of ENSO simulations (Table 1 and see Methods in details)

Objectives

Methods

Results

Conclusion