Improve Land Surface Emissivity Parameter for Land Surface Models Using Global Remote Sensing Observations

Menglin Jin

Climate and Radiation Branch, NASA GSFC/
University of Maryland, College Park


Because land surface emissivity (e) had not been reliably mapped, global climate model's (GCM) land surface schemes conventionally set this parameter as constant, for example 1 as in the National Oceanic and Atmospheric Administration (NOAA) National Centers for Environmental Prediction (NCEP) model and 0.96 for bare soil as in the National Center for Atmospheric Research (NCAR) Community Land Model (CLM2). Accurate broadband emissivity data are needed as a model input to better simulate land surface climate. We demonstrate in this paper that the assumption of the constant emissivity induces errors in modeling the surface energy budget, especially over large arid and semi-arid areas where e is far smaller than unity. The only feasible solution to this problem is to apply broadband emissivity, directly estimated from satellite observations, into land surface models.

The Moderate Resolution Imaging Spectroradiometer (MODIS) instrument has routinely measured spectral emissivities (e) in six thermal infrared bands. The empirical regression equations have been developed in this study to convert these spectral emissivities to broadband emissivity required by land surface models. The observed emissivity data show strong seasonality and land-cover dependence. Specifically, emissivity depends on surface cover type, soil moisture content, soil organic composition, vegetation density and structure. For example, broadband emissivity e is usually around 0.96-0.98 for densely vegetated areas (leaf area index LAI>2), but it can be as low as 0.85-0.90 for bare soils (e.g., desert).

To examine the impact of variable surface broadband emissivity, we conducted sensitivity studies using offline CLM2 and coupled NCAR Community Atmosphere Model CAM2/CLM2. These sensitivity studies illustrate that large impacts of surface e occur over deserts, with change up to 1-2C in ground temperature, surface skin temperature, and 2m surface air temperature, and evident changes in sensible and latent heat fluxes.