On the use of the eddy covariance latent heat flux and sap flow transpiration for the validation of a surface energy balance model