Carbon exchange variability over Amazon Basin using coupledhydrometeorological-mixed layer PBL-carbon dioxide assimilation modeling system forced by satellite-derived surface radiation and precipitation

A hydrometeorological model is modified to include a simple slab model of the mixed layer for the estimation of CO2 fluxes in Amazonia. Three carbon assimilation models are examined for use in the FSU hydromet model, of which the NCAR LSM module is chosen because respiration rates are provided and CO2, latent and sensible heat fluxes are coupled through stomatal resistance. Initial calculations of NEP show a necessity for modeling canopy-boundary layer interactions to reproduce observed morning effluxes at forest tower sites at Manaus and Jaru. CO2 concentrations in five layers in and above the canopy are modeled with associated fluxes. Sensible heat fluxes are consistently overestimated until canopy heat capacity is taken into account. Estimations of canopy heat storage are found using observed differences between net incoming radiation and latent and sensible heat fluxes, or observed total residual energy. Calibration of fluxes at three tower sites is conducted using modeled total residual energy at the forest sites and modified photosynthesis parameters at the pasture site. The forcing parameters of downwelling solar radiation (K ↓) and temperature are found to exert the most influence over modeled CO2 fluxes at the tower sites. Model application over the basin shows that while vegetation type is the primary factor controlling CO2 fluxes area-wide, K ↓ is the primary forcing variable that produces spatial and temporal variability of CO2 fluxes. Modeled CO2 fluxes show mean monthly uptake values in the range of 1-3 mumol m-2 s-1 and diurnal progressions of large coherent areas of CO2 effluxes over the forest, progressing from SE to NW in December, and from NE to SW in June. Inspection of area-wide modeled fluxes near tower sites shows that the use of ECMWF winds and temperatures creates a spurious nocturnal stability that produces much larger morning efflux magnitudes than observations suggest. Comparison of CO2 fluxes at nearly 20,000 forest points within Amazonia with overall forest mean values using eight months of model output suggest that flux observations from five strategically placed towers used in conjunction with existing towers at Manaus and Jaru would be sufficient to reproduce representative area-wide CO 2 flux variability.