Estimating the anthropogenic sea surface temperature response using pattern scaling

Abstract:

This study seeks to derive the sea surface temperature (SST) response to anthropogenic forcing from observations (1900-2005), using a simple method inspired from pattern scaling. As in pattern scaling, the spatial response of SST to anthropogenic forcing is assumed to scale with global-mean and annual-mean surface temperature. The long term aim of this work is to generate anthropogenically-forced SST and sea-ice patterns of change for recent past and near-term future, and use them to force atmosphere-land climate models for attribution and prediction purposes.

The present work uses a Monte Carlo framework based on two large initial condition ensembles of climate model simulations to quantify the method's ability to correctly isolate the anthropogenically-forced SST spatial response pattern from internal variability when derived from a single observation. We use our method to estimate the SST spatial pattern response to anthropogenic forcing in each member of the large ensemble, and base our quantification on their spatial correlations: If the method successfully removes all of the internal variability, the estimated spatial pattern response of each member should be identical across the ensemble.

We first treat each member of the large ensemble as a pseudo-observation and compare the different estimated spatial response pattern pairwise. We find that on average 16-36% of the spatial pattern response estimated in each member is common to the other ensemble members. We then treat one member as a pseudo-observation and the ensemble mean as the true anthropogenically-forced signal. Here, we find that on average 36-56% of the spatial pattern response estimated in each member is common to the true response pattern. We conclude that about 16-56% of the spatial pattern response derived from a single observation with our method results from anthropogenic forcing. These numbers increase with the length of the time period chosen to apply the method.

Finally, we show a first application of the ensemble of experiments obtained by forcing the atmosphere-land climate model CAM5 with our estimate of anthropogenically-forced SST derived from observations, and show that we are able to attribute the increase in winter snow cover observed during 1980-2010 in northwestern North America to SST internal variability.