Get access

Estimating parameters in a land-surface model by applying nonlinear inversion to eddy covariance flux measurements from eight FLUXNET sites

Authors


Yingping Wang, tel. +61 3 9239 4577, e-mail: yingping.wang@csiro.au

Abstract

Flux measurements from eight global FLUXNET sites were used to estimate parameters in a process-based, land-surface model (CSIRO Biosphere Model (CBM), using nonlinear parameter estimation techniques. The parameters examined were the maximum photosynthetic carboxylation rate (inline image) the potential photosynthetic electron transport rate (jmax, 25) of the leaf at the top of the canopy, and basal soil respiration (rs, 25), all at a reference temperature of 25°C. Eddy covariance measurements used in the analysis were from four evergreen forests, three deciduous forests and an oak-grass savanna. Optimal estimates of model parameters were obtained by minimizing the weighted differences between the observed and predicted flux densities of latent heat, sensible heat and net ecosystem CO2 exchange for each year. Values of maximum carboxylation rates obtained from the flux measurements were in good agreement with independent estimates from leaf gas exchange measurements at all evergreen forest sites. A seasonally varying inline image and jmax, 25 in CBM yielded better predictions of net ecosystem CO2 exchange than a constant inline image and jmax, 25 for all three deciduous forests and one savanna site. Differences in the seasonal variation of inline image and jmax, 25 among the three deciduous forests are related to leaf phenology. At the tree-grass savanna site, seasonal variation of inline image and jmax, 25 was affected by interactions between soil water and temperature, resulting in inline image and jmax, 25 reaching maximal values before the onset of summer drought at canopy scale. Optimizing the photosynthetic parameters in the model allowed CBM to predict quite well the fluxes of water vapor and CO2 but sensible heat fluxes were systematically underestimated by up to 75 W m−2.

Ancillary