Predictions of the response of ecosystem respiration to warming in the Arctic are not well constrained, partly due to the considerable spatial heterogeneity of these permafrost-dominated areas. Accurate calculations of in situ temperature sensitivities of respiration (Q 10 ) are vital for the prediction of future Arctic emissions. To understand the impact of spatial heterogeneity on respiration rates and Q 10 , we compared respiration measured from automated chambers across the main local polygonized landscape forms (high and low centers, polygon rims, polygon troughs) to estimates from the flux-partitioned net ecosystem exchange collected in an adjacent eddy covariance tower. Microtopographic type appears to be the most important variable explaining the variability in respiration rates, and low-center polygons and polygon troughs show the greatest cumulative respiration rates, possibly linked to their deeper thaw depth and higher plant biomass. Regardless of the differences in absolute respiration rates, Q 10 is surprisingly similar across all microtopographic features, possibly indicating a similar temperature limitation to decomposition across the landscape. Q 10 was higher during the colder early summer and lower during the warmer peak growing season, consistent with an elevated temperature sensitivity under colder conditions. The respiration measured by the chambers and the estimates from the daytime flux-partitioned eddy covariance data were within uncertainties during early and peak seasons but overestimated respiration later in the growing season. Overall, this study suggests that it is possible to simplify estimates of the temperature sensitivity of respiration across heterogeneous landscapes but that seasonal changes in Q 10 should be incorporated into model simulations.

The northern permafrost region underlies 15% of the global soil area, and the estimated organic carbon (C) pool in the upper 3 m in this region accounts for ∼33% (1,035 ± 150 Pg C) of global belowground organic C storage (