The regional variability in tundra and boreal carbon dioxide (CO₂) fluxes can be high, complicating efforts to quantify sink‐source patterns across the entire region. Statistical models are increasingly used to predict (i.e., upscale) CO₂ fluxes across large spatial domains, but the reliability of different modeling techniques, each with different specifications and assumptions, has not been assessed in detail. Here, we compile eddy covariance and chamber measurements of annual and growing season CO₂ fluxes of gross primary productivity (GPP), ecosystem respiration (ER), and net ecosystem exchange (NEE) during 1990–2015 from 148 terrestrial high‐latitude (i.e., tundra and boreal) sites to analyze the spatial patterns and drivers of CO₂ fluxes and test the accuracy and uncertainty of different statistical models. CO₂ fluxes were upscaled at relatively high spatial resolution (1 km²) across the high‐latitude region using five commonly used statistical models and their ensemble, that is, the median of all five models, using climatic, vegetation, and soil predictors. We found the performance of machine learning and ensemble predictions to outperform traditional regression methods. We also found the predictive performance of NEE‐focused models to be low, relative to models predicting GPP and ER. Our data compilation and ensemble predictions showed that CO₂ sink strength was larger in the boreal biome (observed and predicted average annual NEE −46 and −29 g C m⁻² yr⁻¹, respectively) compared to tundra (average annual NEE +10 and −2 g C m⁻² yr⁻¹). This pattern was associated with large spatial variability, reflecting local heterogeneity in soil organic carbon stocks, climate, and vegetation productivity. The terrestrial ecosystem CO₂ budget, estimated using the annual NEE ensemble prediction, suggests the high‐latitude region was on average an annual CO₂ sink during 1990–2015, although uncertainty remains high.

Cryoturbated peat circles (that is, bare surface soil mixed by frost action; pH 3–4) in the Russian discontinuous permafrost tundra are nitrate-rich ‘hotspots’ of nitrous oxide (N₂O) emissions in arctic ecosystems, whereas adjacent unturbated peat areas are not. N₂O was produced and subsequently consumed at pH 4 in unsupplemented anoxic microcosms with cryoturbated but not in those with unturbated peat soil. Nitrate, nitrite and acetylene stimulated net N₂O production of both soils in anoxic microcosms, indicating denitrification as the source of N₂O. Up to 500 and 10 μM nitrate stimulated denitrification in cryoturbated and unturbated peat soils, respectively. Apparent maximal reaction velocities of nitrite-dependent denitrification were 28 and 18 nmol N₂O g_DW⁻¹ h⁻¹, for cryoturbated and unturbated peat soils, respectively. Barcoded amplicon pyrosequencing of narG, nirK/nirS and nosZ (encoding nitrate, nitrite and N₂O reductases, respectively) yielded ≈49 000 quality-filtered sequences with an average sequence length of 444 bp. Up to 19 species-level operational taxonomic units were detected per soil and gene, many of which were distantly related to cultured denitrifiers or environmental sequences. Denitrification-associated gene diversity in cryoturbated and in unturbated peat soils differed. Quantitative PCR (inhibition-corrected per DNA extract) revealed higher copy numbers of narG in cryoturbated than in unturbated peat soil. Copy numbers of nirS were up to 1000 × higher than those of nirK in both soils, and nirS nirK⁻¹ copy number ratios in cryoturbated and unturbated peat soils differed. The collective data indicate that the contrasting N₂O emission patterns of cryoturbated and unturbated peat soils are associated with contrasting denitrifier communities.

See also the Commentary on this article by Hättenschwiler et al., 223: 5–7.