Optimizing a process-based ecosystem model with eddy-covariance flux measurements: A pine forest in southern France

摘要: We design a Bayesian inversion method (gradient-based) to optimize the key functioning parameters of process-driven land surface model (ORganizing Carbon and Hydrology In Dynamic EcosystEms (ORCHIDEE)) against combination prior information upon eddy covariance fluxes. The calculates energy, water, CO2 fluxes their interactions on half-hourly basis, we carry out using measurements CO2, latent heat, sensible heat as well net radiation over pine forest in southern France. makes it possible assess reduction uncertainties error correlations parameters. designed an ensemble inversions with different set ups flux data time periods, order (1) identify well-constrained loosely constrained ones, (2) highlight some structural deficiencies, (3) quantify overall gained from assimilating each type or energy sensitivity optimal parameter values initial carbon pool sizes is discussed analysis posterior performed. Assimilating 3 weeks during summer improves fit diurnal variations, but merely seasonal variations. full year also cycle more than cycle. This points importance timescales when inverting high-frequency eddy-covariance data. show that photosynthetic such carboxylation rates are by water get increased values, correction corroborated independent at leaf scale. contrast, controlling maintenance, microbial growth respirations, temperature dependencies cannot be robustly determined. could not discriminate between respiration terms. At face value, all budget can safely determined, leading good model-data timescales.