USING UNFORCED VARIABILITY OF LOW CLOUD ‘HOT-SPOTS’ TO DEVELOP BETTER CONSTRAINTS ON EARTH’S CLIMATE SENSITIVITY

Developing better constraints on Earth’s Equilibrium Climate Sensitivity (ECS) is one of the
central goals of climate science, but despite decades of work, there is still large uncertainty in
Earth’s ECS. The response of low clouds to warming has been identified as the primary source of
this uncertainty and, while recent evidence suggests that low cloud cover is reduced in warmer
climates, uncertainty around the sign and, especially, the magnitude of the low cloud response is
responsible for much of the spread in climate models’ ECS.

To further refine our understanding of the link between low clouds and ECS, and to develop
potential constraints on the behavior of low clouds with warming, we propose studying the
unforced variability of low-level clouds and their governing meteorological conditions over the
global ocean in order to identify specific geographic regions (“hot-spots”) in which the
variability of low cloudiness is especially indicative of models’ response to warming. The
proposed project will combine data from simulations with comprehensive climate models,
including from CMIP6 and from perturbed physics ensembles with models developed at NOAA’s
Geophysical Fluid Dynamics Laboratory (GFDL), with observations to (1) evaluate model skill
in representing the meteorological conditions governing low cloudiness, (2) identify ‘hot-spot’
regions in which unforced low cloud variability is strongly related to the forced low cloud
response, and (3) combine metrics of model skill in simulating the governing meteorology and of
low cloud variability in the hot-spots with observations to develop emergent constraints on
Earth’s ECS. A number of steps will be taken to ensure that the emergent constraints are robust –
both physically and statistically – and not simply the result of data-mining. The proposed
analysis will consider variability on time-scales from monthly to the 2-5 year time-scales of the
El Nino Southern Oscillation, and will consider low cloud regions over the entire global ocean.
Comprehensively characterizing low cloud variability, and its relationship with the governing
meteorological conditions, in CMIP5/6 and GFDL models is an important additional benefit of
the proposed project.

This project fits squarely within the aims of the MAPP competition, and addresses Priority Areas
A, B and C, as we will quantify uncertainty associated with low clouds in CMIP6 models (and in
climate models developed at GFDL), develop new metrics for assessing climate models, and use
observations to assess the models and develop emergent constraints. Additionally, the project
will provide a more refined view of the causes of intermodel differences in ECS and a better
understanding of the relationships between model parameters and low cloud variability in the
GFDL models. More broadly, the project will help NOAA prepare for the potential impacts from
increased atmospheric CO2 concentrations by reducing uncertainty in Earth’s ECS and by
improving the models used by NOAA to forecast future climate states.