Atmosphere‐biosphere exchange flux of carbon dioxide in a tallgrass prairie modeled with satellite spectral data

The estimation of the rate of net CO2 uptake of vegetated land surfaces is essential for studies of global carbon cycle. The present paper demonstrates the use of spectral reflectance data from satellite remote sensing to model net CO2 flux (NCF) of a tallgrass canopy at the Konza prairie, Kansas. A bidirectional reflectance canopy model was used to estimate seasonal changes in canopy leaf area index (LAI) from surface reflectances remotely sensed by SPOT 1 and Landsat 5 satellites. The radiation model was also coupled with leaf conductance‐photosynthesis models to scale up stomatal conductance and NCF from individual leaves to canopy level according to radiation distribution inside the canopy. The satellite‐data‐driven model was able to closely simulate the seasonal change in LAI as well as the short‐term variation of canopy LAI caused by the dry period during late July and early August in the area. Modeled canopy stomatal conductance (gc) and NCF agree with measurements within 0.16 cm s−1 and 0.28 mg m−2 s−1, respectively, during the growth season from late May to late August. In October both measured and modeled NCF turned to small negative values as canopy photosynthesis diminished and predicted LAI approached zero. In addition to data scatter, some of the differences between modeled and measured gc and NCF may be attributed to uncertainties in seasonal changes of plant physiological status that were not detected by satellite data; some of the differences were caused by inadequate description of the dependence of nighttime CO2 flux of soil respiration on near‐surface turbulent mixing.