Remote sensing of transpiration and heat fluxes using multi-angle observations

Surface energy balance is a major determinant of land surface temperature and the Earthʹs climate. To date, there is no approach that can produce effective, physically consistent, global and multi-decadal energy–water flux data over land. Net radiation (Rn) can be quantified regionally using satellite retrievals of surface reflectance and thermal emittance with errors < 10%. However, consistent, useful retrieval of latent heat flux (λE) from remote sensing is not yet possible. In theory, λE could be inferred as a residual of Rn, ground heat (G) and sensible heat (H) fluxes (Rn–H–G). However, large uncertainties in remote sensing of both H and G result in low accuracies for λE. Where vegetation is the dominant surface cover, λE is largely driven by transpiration of intercellular water through leaf stomata during the photosynthetic uptake of carbon. In these areas, satellite retrievals of photosynthesis (GPP) could be used to quantify transpiration rates through stomatal conductance. Here, we demonstrate how remote sensing of GPP could be applied to obtain λE from passive optical measurements of vegetation leaf reflectance related to the photosynthetic rate independent of knowledge of H, Rn and G. We validate the algorithm using five structurally and physiologically diverse eddy flux sites in western and central Canada. Results show that transpiration and H were accurately predicted from optical data and highly significant relationships were found between the energy budget obtained from eddy flux measurements and remote sensing (0.64 ≤ r2 ≤ 0.85). We conclude that spaceborne estimates of GPP could significantly improve not only estimates of the carbon balance but also the energy balance over land.