Modelling land surface interactions at forested ecosystems: results from the global validation of a 1-D SVAT model

Globally, the monitoring of the Earth’s energy system has developed into a very important and urgent research direction, especially in the face of climate change. Soil Vegetation Atmosphere Transfer (SVAT) models are fast becoming the preferred scientific tool employed in quantifying, at fine spatial and temporal resolutions, key parameters characterising the Earth system. In this context, there is a growing need to monitor and obtain a better understanding of forest dynamics and feedback processes due to their immeasurable importance within society and the Earth system. SimSphere is an example of a SVAT model, simulating a large array of parameters characterising various land surface interactions over a 24 hour cycle at a 1-D vertical profile. In this study, we performed an extensive validation of parameters predicted by the model which are of key importance in characterising the energy balance. Several distinct forest types were covered globally including Evergreen Needleleaf, Deciduous Broadleaf, and Evergreen Broadleaf forests. Model predictions of Latent Heat (LE), Sensible Heat (H) and Net Radiation (Rn) were compared against validated in-situ measurements acquired from observational monitoring networks (CarboEurope, AmeriFlux and OzFlux) to assess statistical significance. Results obtained confirmed SimSphere’s ability to accurately simulate, to a high statistical significance, terrestrial energy and water fluxes over the distinctly different forest types included in our study, particularly so on days with atmospherically stable conditions. Given the very small number of SimSphere validation studies, our work contributes decisively towards the models’ evolution and its correspondence to a real world system. This work is also very timely given that validation studies are a fundamental step in any models’ evolution before it can be confidently used in various ecosystems globally. This is vital, particularly so given that SimSphere’s use is currently expanding worldwide and at present been explored in developing globally distributed operational products, based on its synergy with EO datasets.