Representing Land Surface Heterogeneity: Offline Analysis of the Tiling Method

Abstract

he tiling method is used in many land surface models to represent the surface heterogeneity. Each grid box is divided into fractions of different types of land use with independent solutions of the surface energy budget. An area-weighted average of the energy fluxes is computed to couple with the atmosphere, assuming the air above the surface is well blended at a given height. In the framework of validation of the tiling method, the ECMWF land surface scheme has been tested in offline mode driven by meteorological forcing provided by the ECMWF Interim Re-Analysis (ERA-Interim). Two contrasting surfaces in the boreal region of southern Finland are considered: a Scots pine forest (Hyytiälä) and a small nearby lake (Valkea-Kotinen). The field observations are used to evaluate the land surface model simulations for both energy fluxes and reservoirs. The model is able to characterize the main difference between the two sites, which appears in the energy partitioning, explained by the lake?s large thermal inertia. In fact, while a large portion of the incoming solar radiation in the forest is released as sensible heat, the lake stores a substantial amount of energy in the water during late spring and summer and releases it in autumn. The different behavior of the fluxes both on annual and diurnal time scales confirms the benefit of a tiling mechanism in the presence of large contrast. However, it is also shown that the assumption of similar conditions at the blending height introduces errors.