Publication issue des travaux du projet MORCE MED.

Les auteurs du présent article ont évalué la capacité du modèle ORCHIDEE (Organizing Carbon and Hydrology in Dynamic Ecosystems) à représenter les flux de carbone et de chaleur à diffentes échelles spatiales et temporelles, sur la région euro-méditerranéenne. Ils ont forcé ORCHIDEE par les sorties de deux modèles régionaux de climat (RegCM3 and WRF) sur la période 2002 à 2007, et comparé les résultats obtenus avec les observations météorologiques et les données de chaleur latente, de production primaire brute et d'échanges écosystémiques nets issues du réseau CARBOEUROPE.

ORCHIDEE est capable de reproduire correctement le cycle annuel de chaleur et de carbone, avec des erreurs comparables aux incertitudes de mesure. Ce n'est pas le cas en ce qui concerne les variations interannuelles.

Résumé en anglais

A validation of heat and carbon fluxes from high-resolution land surface and regional models

The performance of Organizing Carbon and Hydrology in Dynamic Ecosystems (ORCHIDEE), a model of surface hydrology, plant phenology, and vegetation dynamics, in reproducing field measurements of heat and carbon fluxes at various spatial and temporal scales is assessed. The model is forced by two high-resolution (30 km) regional climate models (RegCM3 and WRF) output in the Euro-Mediterranean region for the period 2002–2007. First a validation of the regional models surface climatology is conducted in comparison to gridded meteorological station data and to other state of the art regional models. Then annual cycles and interannual variability of latent and sensible heat, gross primary production, and net ecosystem exchange are compared to in situ experimental data provided from the CARBOEUROPE network. Six sites were chosen across the Euro-Mediterranean region, representing different forest environments. Results show that ORCHIDEE is able to reproduce the annual cycle of heat and carbon fluxes, with errors comparable with the measurement uncertainties. The interannual variations appear more problematic. While the variations of sensible heat are at least of the same sign as observed, latent heat is less well reproduced, while there is hardly any skill in reproducing carbon fluxes. The main difference between the two regional models used for forcing is an excessive precipitation produced on average by RegCM3. This causes errors in the latent heat and gross primary production of ORCHIDEE.