Skip to Content

3D Radiative Transfer Model Coupled to the Weather Research and Forecasting Model

Larry Di Girolamo, University of Illinois at Urbana-Champaign

Usage Details

Brian Jewett, David Raila, Gregory Daues, Kent Muqun Yang, Guangyu Zhao, Larry Di Girolamo, Jonathan Kim, Alexandra Jones, Daeven Jackson, Yizhao Gao, Dongwei Fu, Ming Su, Landon Clipp, Shashank Bansal, Maghav Kumar, Brandon Chen, Yat Long Lo, Hyokyung Lee, Yizhe Zhan, Yulan Hong, Soumi Dutta, James Limbacher, Arka Mitra, Puja Roy, Jesse Loveridge, Javier Villegas Bravo, Scott Sinno

Recent events such as the typhoon in the Philippines and Hurricane Sandy point to the importance of understanding and predicting Earth’s weather and climate. Research in weather and climate has massive societal benefits, and indeed has been one of the leading drivers for advancing supercomputing infrastructures. The least understood aspect of the weather and climate system is the impact of Earth’s clouds on solar and terrestrial radiative forcing of the Earth–Atmosphere system, and the subsequent feedback onto the dynamics of the system. Clouds play an exceptionally important role in the atmosphere, redistributing radiative energy from the sun as well as that emitted from the Earth and atmosphere. Currently, radiative transfer (RT) in atmospheric models is represented in a primitive manner (mostly in a single dimension—up/down) because of the exceptional computational expense. Computational resources have always been a constraint on how radiation can be explicitly represented in dynamic models of the atmosphere such as the Weather Research and Forecasting model (WRF), especially at LES (large eddy simulation) scales with grid spacing ranging from 10m to 100m. As the numerical representation of other physcial processes have advanced (e.g., cloud microphysical processes), the representation of radiative processes has remained markedly unsophisticated. This is true even at the finest LES resolutions where the approximations that allow for computationally inexpensive treatments break down even further.

These crude representations of radiative transfer can result in localized orders-of-magnitude error in radiative heating in the presence of clouds with complex morphology. The resulting errors, when allowed to feed back on cloud dynamics, feed back as errors in cloud size, lifetime, and physical properties. These changes in turn may impact weather and climate by, for example, affecting rainfall, surface atmospheric heating, and photolysis rates. While 3D radiative transfer is a solved problem (e.g., by way of a Monte Carlo solution), albeit a computationally expensive one, a full 3D RT model has never before been coupled to a cloud dynamics model. Here we propose to do just that. This coupled model, for the first time, will allow us to study and understand how errors from the crude RT approximations used in the past feed back on cloud properties and their evolution. We anticipate that the contributions from radiative cooling at cloud edge and other radiative effects could close the current gap between numerical results and observations, not only of dynamical properties but of micro- and macro-physical cloud properties. In carrying out this project, we will address longstanding significant computational challenges inherent in solving the radiative transfer problem in cloudy atmospheres.