14 April 2014

Profiling Clouds' Inner Life: Subgrid modeling pinpoints cloud transformation to uncover true reflective power



In the global scheme of things, clouds are the small stuff. And clouds’ inner workings are even smaller. Scientists at Pacific Northwest National Laboratory found that neglecting the small stuff has consequences for understanding how clouds contribute to heating or cooling the planet. When global models average cloud cooling effects over a large area, they mask the cooling variability within a grid that eventually leads to changes in the clouds’ characteristics. This research leads to uncovering the amount of cooling or warming enabled by clouds, and understanding the true energy balance of the planet.



Marine boundary layer clouds are produced by interactions between radiative cooling from the cloud tops combined with turbulent cloud-mixing processes. PNNL researchers used a high-resolution, large-eddy simulation model to reveal the small spatial details of the cloud-top radiative cooling. They used those results to simulate detailed cloud radiation-turbulence interactions, as well as scenarios where the radiation was smoothed over the horizontal model grid area mimicking a climate model grid cell.


Next, the team compared these two types of simulations, finding that when small-scale variability is neglected, the model accelerates the transition of marine stratocumulus to trade cumulus clouds resulting in more sunlight warming the ocean instead of being reflected back to space by the white clouds.


This study was designed to reveal the impact of subgrid radiation variability of marine clouds and determine its overall impact for the Earth’s energy balance.



Calculating the right amount of sunlight beaming down to warm the ocean means understanding the kind of clouds parked between the sun and the ocean. When climate models do not accurately depict the characteristics of these clouds, perception of the planet’s energy balance can be lopsided. Getting accurate understanding of the type and amount of clouds over the ocean is important for climate change projections. Because of their tendency to average out the effects over an entire spatial grid cell, global climate models may not capture the type of cloud and its real-world climate-influencing properties. This study accounts for the small-scale radiative energy and turbulent cloud interactions that will lead to more accurate model simulations of real-world clouds.



Low clouds cover a significant portion of the world’s oceans and play an important role in the Earth’s energy budget. Subgrid-scale interactions between turbulence and radiation are potentially important for accurately simulating marine low clouds in climate models. To better represent these clouds in climate models, a team of DOE researchers at Pacific Northwest National Laboratory examined the possible impact of subgrid-scale turbulence-radiation interaction by artificially removing the interaction in a large-eddy simulation of marine low clouds. They found that removing the interaction accelerates the transition from the stratocumulus to trade cumulus cloud regime, which could partly explain a bias towards early transitions commonly found in climate models. The team also found possible directions for improving the treatment of turbulent mixing and turbulence-radiation interaction in climate models. The impacts of subgrid-scale turbulence-radiation interactions can be included in turbulence parameterizations, such as CLUBB that include detailed handling of subgrid features like temperature variance. 

Heng Xiao
Xiao, H, WI Gustafson Jr., and H Wang.  2014.  "Impact of Subgrid-scale Radiative Heating Variability on the Stratocumulus-to-trade Cumulus Transition in Climate Models."  Journal of Geophysical Research Atmospheres, doi:10.1002/2013JD020999.


The U.S. Department of Energy (DOE) Biological and Environmental Research Earth System Modeling program and a DOE Early Career grant supported this research. Computational support provided through the Environmental Molecular Sciences Laboratory (EMSL) and PNNL Institutional Computing.