We have investigated global teleconnections of climate to regional satellite-driven observations for prediction of Amazon ecosystem production, in the form of monthly estimates of net carbon exchange over the period 1982–1998 from the NASA–CASA (Carnegie–Ames–Stanford) biosphere model. This model is driven by observed surface climate and monthly estimates of vegetation leaf area index (LAI) and fraction of absorbed PAR (fraction of photosynthetically active radiation, FPAR) generated from the NOAA satellite advanced very high-resolution radiometer (AVHRR) and similar sensors. Land surface AVHRR data processing using modified moderate-resolution imaging spectroradiometer radiative transfer algorithms includes improved calibration for intra- and intersensor variations, partial atmospheric correction for gaseous absorption and scattering, and correction for stratospheric aerosol effects associated with volcanic eruptions. Results from our analysis suggest that anomalies of net primary production and net ecosystem production predicted from the NASA–CASA model over large areas of the Amazon region east of 60°W longitude are strongly correlated with the Southern Oscillation index. Extensive areas of the south-central Amazon show strong linkages of the FPAR and the NASA–CASA anomaly record to the Arctic Oscillation index, which help confirm a strong relation to southern Atlantic climate anomalies, with associated impacts on Amazon rainfall patterns. Processes are investigated for these teleconnections of global climate to Amazon ecosystem carbon fluxes and regional land surface climate.