2007 Volume 63 Issue 4 Pages 171-183
A simple empirical model was developed to calculate the carbon fluxes on a regional scale by integrating the observed ground truth data at a black spruce forest in interior Alaska in 2005 and original NOAA/AVHRR data. The satellite-derived variables of normalized difference vegetation index (NDVI) and land surface temperature (LST) were related with measured leaf area index (LAI) and measured CO2 flux (NEE) using a sub-empirical model, CBAT. In order to scale up the observed fluxes, gross primary production (GPP) and ecosystem respiration (Reco) were separately determined using NDVI and LST. The parameters and relationships were determined by applying the observed ground truth data in 2005. Comparing to the observed dataset, the diurnal and seasonal variations were calculated reasonably. The model satisfactorily reproduced Reco as an hourly base, but GPP tended to be an overestimation caused by the eliminated effect of vapor pressure deficit (VPD). The GPP, Reco and NEE over Alaska's black spruce forests were estimated as 2330, 1920 and -410 g CO2 m-2, respectively, during the growing season in 2005. Seasonal variations of estimated carbon flux distributions reflected heterogeneous ecosystem conditions.