Evaluation of six process‐based forest growth models using eddy‐covariance measurements of CO2 and H2O fluxes at six forest sites in Europe

摘要: Reliable models are required to assess the impacts of climate change on forest ecosystems. Precise and independent data essential this accuracy. The flux measurements collected by EUROFLUX project over a wide range types climatic regions in Europe allow critical testing process-based which were developed LTEEF project. ECOCRAFT complements with wealth plant physiological measurements. Thus, it was aimed study test six growth against European types, taking advantage large database parameters. reliability both parameter values itself not under discussion study. provided researchers sites, possibly local corrections, used minor gap-filling procedure avoid loss many days observations. model performance is discussed based their accuracy, generality realism. Accuracy evaluated goodness-of-fit observed daily net ecosystem exchange, gross primary production respiration (gC m2 d1), transpiration (kg H2O d1). Moreover, accuracy also systematic unsystematic errors. Generality characterized applicability different Reality comparing modelled responses production, radiation temperature. results indicated that: Accuracy. All showed similar high correlation measured carbon data, low prediction errors at one or more sites case several when water fluxes considered. Most fulfilled criteria sufficient for ability predict exchange between forests atmosphere. Generality. Three could be applied deciduous coniferous forests. Furthermore, four boreal temperate conditions. However, no severe water-limited conditions encountered, year-to-year variability tested. Realism. fulfil criterion realism that relationships phenomena (carbon exchange) environment described causally. Again able reproduce measurable variables such as (GPP), environmental driving factors Stomatal conductance appears most process causing differences predicted those accurately describe annual totals GPP, transpiration. As conclusion, available produce accurate estimates Europe. This considerable fulfils requirement balance should further expanding time space. In addition, behaviour level indicate testing, special emphasis modelling stomatal realistically