We propose to use novel remote sensing-based indicators, more closely related to photosynthetic processes than to greenness, to parameterize phenology modules of Land Surface models and thereby improve their simulations.
We will use the FLUXNET 2015 database to validate not only the annual GPP and energy simulations by the Land Surface models, but also their seasonality and response to drought.
While a multitude of remote sensing-based GPP products using NDVI as a driver is available, their agreement with the field observations is weak . Here, we will replace NDVI by a novel indicator to produce a new generation remote sensing-based GPP product.
To date, MOD 17 is the only freely available remote sensing-based product for biomass production. However, it translates GPP into NPP by subtracting a climate-dependent Ra, whereas recent empirical studies have revealed that this conversion depends much more on nutrient availability and human land management than on climate [3-4]. We will therefore develop a new algorithm, completely avoiding the modelling of Ra but taking into account soil nutrient availability and land use intensity, to provide a relatively independent (and probably more realistic) remote sensing-based product for biomass production. We will use the global biomass production database  available in house to test the accuracy of these novel products.