Characterizing Errors in 1D Solar Radiative Transfer Solutions as We Move to Cloud-Resolving Models

A primary goal of the Energy Exascale Earth System Model (E3SM) project is to develop a high-resolution, fully-coupled Earth system model for climate simulation and prediction. One of the E3SM future versions aims to develop a global cloud-resolving model (GCRM, ~3 km horizontal resolution). At these scales, clouds in neighboring columns will strongly contribute to the local heating and photochemical rates. Such affects clearly challenge the 1D approach used for solar heating modules in E3SM and almost all Earth system models. These use cloud statistics within the column atmosphere to create a set of independent column atmospheres (ICAs) and then solve the radiative transfer equation in 1D assuming horizontally homogeneous layers for a limited number of ICAs. With GCRMs, the location of clouds in neighboring grid cells a few kilometers away can alter the radiation within a grid cell that appears 'clear' in a 1D overhead sense. We must develop an understanding, through model development and measurements, of how to simulate solar radiation in ESMs when we have knowledge of the neighboring cloud fields. For example, the European Centre model is currently developing a solar heating model that accounts for the sides of clouds. This work applies a 1D solar radiation model to surface measurements that include identification of within-sight, neighboring clouds to quantify typical errors in 1D models according to the sun-cloud geometries, and thus establish a basis for developing hybrid1D-3D radiative transfer solutions to work within GCRMs.

The long-term, ground-based Atmospheric Radiation Measurement (ARM) facilities record a comprehensive set ofco-located,radiatively important variables (profiles of cloud, aerosol, temperature, water vapor, surface albedo, trace gases, and radiative intensities at the surface) at different locations representing various cloud regimes. The primary ARM data provide 1D profiles as well as sky images from the surface that record the clouds overhead and nearby. They provide a primary validation test for surface fluxes in solar heating modules; but through the sky images, they also provide a quantitative measure of how clouds not overhead can alter surface heating rates. For example, high clouds clearly visible in the sky images (and thus altering the surface radiation) at zenith angles of 50 degrees are more than 10 km away, 3+grid cells on a GCRM scale. In this study, we will combine the ARM data with an offline version of Solar-J (a readily adaptable benchmark E3SM solar radiative transfer module) to evaluate errors caused by the 1D assumption under different cloud conditions. Our results will provide important guidance in developing the next generation of solar radiation models for climate and Earth system models as well as Numerical Weather Models.