Published August 4, 2021 | Version v1.0
Software Open

Initial replication of Seager et al. 2019 for CMIP5/6 -- Cold tongue bias in CMIP6 spuriously causes move towards El Niño in response to global warming over historical period

  • 1. University of Cambridge / British Antarctic Survey

Description

Abstract

The equatorial Pacific, and the associated state of ENSO, is a crucially important piece of the climate system. Over the last couple of decades, state-of-the-art climate models have been predicting that the temperature gradient across the Pacific should weaken with increased greenhouse gas forcing. However, observational reanalysis products show that this has not occurred. Seager et al. 2019 [A] showed that you can explain the bias in CMIP5 with a simple model. The model reproduces the reanalysis product when forced by the reanalysis fields, and when mean relative humidity and wind speed are swapped out for CMIP5 multimodal mean, it reproduces the bias in the change in NINO3.4. We reassembled their model and made it easy to use. We show that the same bias exists for CMIP6, and it can be almost entirely explained in the same way. The model is relatively insensitive to the variation of the more poorly constrained parameters (e.g. Rayleigh damping, Newtonian cooling). Through variation of coupling parameters, we can show that the majority of the effect comes from the suppression of the sensitivity of latent heat flux over the cold tongue region. This model demonstrates that the bias in the change of the temperature of the NINO3.4 region is plausibly a consequence of the bias in the mean state. This in turn is associated with the larger scale double ITCZ cold tongue bias. We would expect the bias in the change in the ENSO equilibrium to propagate into all of the environmental risks with which there are teleconnections to ENSO. This highlights the need for further progress to be made in reducing the double ITCZ cold tongue bias in climate models.

[A] Seager, R. et al. Strengthening tropical Pacific zonal sea surface temperature gradient consistent with rising greenhouse gases, https://doi.org/10.1038/s41558-019-0505-x (July 2019.)

Description & Re-use

This model is composed of a Fortran ocean model, alongside a Python atmospheric and ocean flux model. The repository contains a Dockerfile to create the right environment with old Fortran compilers to run the model. The associated data needed to run the model should automatically download from DropBox when the code is used as detailed in the main `README.md` file.

Documentation

The full documentation is available here https://seager19.readthedocs.io/en/latest/ , and this goes through the code and results much more thoroughly.

Acknowledgements

This project would not have been possible without the help of Dr. Naomi Henderson (https://scholar.google.com/citations?user=TMFVrlsAAAAJ&hl=en), who both put the original code online and answered my stupid questions.

Files

sdat2/seager19-v1.0.zip

Files (85.3 MB)

Name Size Download all
md5:3078ddd1223f37ed2827b53f78851a43
85.3 MB Preview Download

Additional details

Related works

Funding

UKRI Centre for Doctoral Training in Application of Artificial Intelligence to the study of Environmental Risks (AI4ER) EP/S022961/1
UK Research and Innovation