Modelling above-ground herbage mass for a wide range of grassland community types

Whereas it is recognized that management of plant diversity can be the key to reconciling production and environmental aims, most grassland models are tailored for high-value grass species. We proposed to adapt a mono-specific grass model to take into account specific features of species-rich permanent grasslands, especially over the reproductive phase. To this end, we used the concept of plant functional type (PFT), i.e. the grouping of plant species according to plant traits determined by the response of plant species to different management practices (land use and fertilization) and characterizing of agronomic properties of the corresponding species. In the model, weather and nutrient availability act upon rates of biophysical processes (radiation capture and use, plant senescence). These rates are modified over times due to PFT-specific parameters determined experimentally which represent the different strategies of plant species regarding growth. The integration of these parameters into the model made it possible to predict herbage biomass accumulation rate under different management practices for a wide range of plant communities differing in their PFT composition. The model was evaluated in two steps, first by analyzing separately the effects of PIT and an indicator of nutrient availability on herbage accumulation and then by conducting a sensitivity analysis. it was validated using two independent datasets; a cutting experiment running over the whole growing season to examine the consistency of the model outputs under different cutting regimes, and a monitoring of meadows and pastures in spring over a whole growth cycle to assess the model's ability to reproduce growth curves. Although a good fit was observed between the simulated and observed data, the few discrepancies noticed between field data and predicted values were attributed mainly to the potential presence of non-grass species. More specifically, we noticed that nutrient (mainly nitrogen) availability is the main driver of plant growth rate, and that PFT determines the times at which this rate changes in relation to the phenological characteristics of species present. We concluded that integration of the PFT concept into the initial mono-specific growth model is especially suited to evaluating the consequences of management practices on species-rich permanent grasslands to meet feed production targets. (C) 2008 Elsevier B.V All rights reserved.