High-latitude northern ecosystems are experiencing rapid climate changes, and represent a large potential climate feedback because of their high soil carbon densities and shifting disturbance regimes. A significant carbon flow from these ecosystems is soil respiration (R S , the flow of carbon dioxide, generated by plant roots and soil fauna, from the soil surface to atmosphere), and any change in the high-latitude carbon cycle might thus be reflected in R S observed in the field. This study used two variants of a machine-learning algorithm and least squares regression to examine how remotely-sensed canopy greenness (NDVI), climate, and other variables are coupled to annual R S based on 105 observations from 64 circumpolar sites in a global database. The addition of NDVI roughly doubled model performance, with the best-performing models explaining ,62% of observed R S variability. We show that early-summer NDVI from previous years is generally the best single predictor of R S , and is better than current-year temperature or moisture. This implies significant temporal lags between these variables, with multi-year carbon pools exerting large-scale effects. Areas of decreasing R S are spatially correlated with browning boreal forests and warmer temperatures, particularly in western North America. We suggest that total circumpolar R S may have slowed by ,5% over the last decade, depressed by forest stress and mortality, which in turn decrease R S . Arctic tundra may exhibit a significantly different response, but few data are available with which to test this. Combining large-scale remote observations and small-scale field measurements, as done here, has the potential to allow inferences about the temporal and spatial complexity of the large-scale response of northern ecosystems to changing climate.