Estimation of Regional Surface Resistance to Evapotranspiration from NDVI and Thermal-IR AVHRR Data

Abstract

Infrared surface temperatures from satellite sensors have been used to infer evaporation and soil moisture distribution over large areas. However, surface energy partitioning to latent versus sensible heat changes with surface vegetation cover and water availability. We tested a hypothesis that the relationship between surface temperature and canopy density is sensitive to seasonal changes in canopy resistance of conifer forests. Surface temperature (Ts) and canopy density were computed for a 20 ? 25 km forested region in Montana, from the NOAA/AVHRR for 8 days during the summer of 1985. A forest ecosystem model, FOREST-BGC, simulated canopy resistance (Rc) for the same period. For all eight days. surface temperatures had high association with canopy density, measured as Normalized Difference Vegetation Index (NDVI) (R2 = 0.73 ? 0.91), implying that latent heat exchange is the major cause of spatial variations in surface radiant temperatures. The slope of Ts and NDVI, σ, was sensitive to changes in canopy resistance on two contrasting days of canopy activity. The trajectory of σ followed seasonal changes in canopy resistance simulated by the model. The relationship found between σ and Rc (R2 = 0.92), was nonlinear, expected because Rc values beyond 20 s cm?1 do not influence energy partitioning significantly. The slope of Ts and NDVI, σ, could provide a useful parameterization of surface resistance in regional evapotranspiration research.