Upscaling latent heat flux for thermal remote sensing studies: Comparison of alternative approaches and correction of bias