Methane dynamics regulated by microbial community response to permafrost thaw. 4,5,16 . The net effect is that the high methane-emitting fen contributes 7 55 times the greenhouse impact per unit area as the palsa. This thaw progression is also associated 56 with an increase in overall organic matter lability, including a decrease in C:N and an increase in 57 humification rates 9 . We hypothesized, consistent with previous studies of in situ bog and fen 58 systems [17][18][19] , that thaw progression also facilitates a shift from hydrogenotrophic to acetoclastic 59 CH 4 production. 60We used the distinct isotopic signatures of different microbial CH 4 production and 61 consumption pathways to directly relate changes in CH 4 dynamics across the thaw gradient to 62 underlying changes in the microbial community. Methane produced by hydrogenotrophic 63 methanogens generally has lower  13 C and higher D ( 13 C = -110 to -60‰ and D = -250 to -64 170‰) relative to that produced by acetoclastic methanogens ( 13 C = -60 to -50‰ and D = -400 65 to -250‰) 19,20 . If methanotrophic microbes then oxidize CH 4 , lighter molecules are 66 preferentially consumed, leaving the remaining CH 4 13 C-and D-enriched relative to the original 67 CH 4 pool (see expected patterns in Extended Data Fig 1) 19 . Greater fractionation is associated with hydrogenotrophic methanogenesis, and was 85 found in the thawing Sphagnum site (average  C = 1.053 ± 0.002). Significantly less 86 fractionation (p=0.002) associated with more acetoclastic production or with consumption by 87 oxidation was found in the fully thawed Eriophorum porewater (average  C = 1.046 ± 0.001). 88Here, increases in acetoclastic production, not oxidation, best explain isotopic shifts because 89 lower  C and higher  13 C-CH 4 are accompanied by significantly lower D-CH 4 (Extended Data 90 Fig. 1, p< 0.001) 19 . This is consistent with the pattern of isotopes in CH 4 emissions as well as 91 incubations of Stordalen peat 9 and studies showing bog-to-fen shifts from hydrogenotrophic to 92 acetoclastic methanogenesis [17][18][19] . 93The CH 4 flux and isotope results provide compelling but indirect evidence for changes in 94 CH 4 -cycling microbial communities with permafrost thaw. These microbiological changes could 95 be shifts in activity of particular community members or changes in community composition. We 96 examined the role of community composition through 16S rRNA gene amplicon sequencing. All 97 known methanogens belong to a small number of archaeal lineages within the Euryarchaeota 23 . 98As expected, the shift from CH 4 -neutral intact permafrost palsa to CH 4 -emitting wetland 99 corresponded to a substantial increase in the relative abundance of methanogenic archaeal 100 lineages (Fig. 1c, Extended Data Table 2,3). In the aerobic palsa and surface Sphagnum habitats, 101 methanogens were found in low relative abundance (average <0.6%), while the anaerobic 102 environments of the Eriophorum and deeper (below the water table) Sphagnum habitats harbored 10...

Terrestrial ecosystems currently offset one-quarter of anthropogenic carbon dioxide (CO2) emissions because of a slight imbalance between global terrestrial photosynthesis and respiration. Understanding what controls these two biological fluxes is therefore crucial to predicting climate change. Yet there is no way of directly measuring the photosynthesis or daytime respiration of a whole ecosystem of interacting organisms; instead, these fluxes are generally inferred from measurements of net ecosystem-atmosphere CO2 exchange (NEE), in a way that is based on assumed ecosystem-scale responses to the environment. The consequent view of temperate deciduous forests (an important CO2 sink) is that, first, ecosystem respiration is greater during the day than at night; and second, ecosystem photosynthetic light-use efficiency peaks after leaf expansion in spring and then declines, presumably because of leaf ageing or water stress. This view has underlain the development of terrestrial biosphere models used in climate prediction and of remote sensing indices of global biosphere productivity. Here, we use new isotopic instrumentation to determine ecosystem photosynthesis and daytime respiration in a temperate deciduous forest over a three-year period. We find that ecosystem respiration is lower during the day than at night-the first robust evidence of the inhibition of leaf respiration by light at the ecosystem scale. Because they do not capture this effect, standard approaches overestimate ecosystem photosynthesis and daytime respiration in the first half of the growing season at our site, and inaccurately portray ecosystem photosynthetic light-use efficiency. These findings revise our understanding of forest-atmosphere carbon exchange, and provide a basis for investigating how leaf-level physiological dynamics manifest at the canopy scale in other ecosystems.

Abstract. Stomatal conductance influences both photosynthesis and transpiration, thereby coupling the carbon and water cycles and affecting surface-atmosphere energy exchange. The environmental response of stomatal conductance has been measured mainly on the leaf scale, and theoretical canopy models are relied on to upscale stomatal conductance for application in terrestrial ecosystem models and climate prediction. Here we estimate stomatal conductance and associated transpiration in a temperate deciduous forest directly on the canopy scale via two independent approaches: (i) from heat and water vapor exchange and (ii) from carbonyl sulfide (OCS) uptake. We use the eddy covariance method to measure the net ecosystem-atmosphere exchange of OCS, and we use a flux-gradient approach to separate canopy OCS uptake from soil OCS uptake. We find that the seasonal and diurnal patterns of canopy stomatal conductance obtained by the two approaches agree (to within ±6 % diurnally), validating both methods. Canopy stomatal conductance increases linearly with above-canopy light intensity (in contrast to the leaf scale, where stomatal conductance shows declining marginal increases) and otherwise depends only on the diffuse light fraction, the canopy-average leafto-air water vapor gradient, and the total leaf area. Based on stomatal conductance, we partition evapotranspiration (ET) and find that evaporation increases from 0 to 40 % of ET as the growing season progresses, driven primarily by rising soil temperature and secondarily by rainfall. Counterintuitively, evaporation peaks at the time of year when the soil is dry and the air is moist. Our method of ET partitioning avoids concerns about mismatched scales or measurement types because both ET and transpiration are derived from eddy covariance data. Neither of the two ecosystem models tested predicts the observed dynamics of evaporation or transpiration, indicating that ET partitioning such as that provided here is needed to further model development and improve our understanding of carbon and water cycling.

Abstract. For the past decade, observations of carbonyl sulfide (OCS or COS) have been investigated as a proxy for carbon uptake by plants. OCS is destroyed by enzymes that interact with CO2 during photosynthesis, namely carbonic anhydrase (CA) and RuBisCO, where CA is the more important one. The majority of sources of OCS to the atmosphere are geographically separated from this large plant sink, whereas the sources and sinks of CO2 are co-located in ecosystems. The drawdown of OCS can therefore be related to the uptake of CO2 without the added complication of co-located emissions comparable in magnitude. Here we review the state of our understanding of the global OCS cycle and its applications to ecosystem carbon cycle science. OCS uptake is correlated well to plant carbon uptake, especially at the regional scale. OCS can be used in conjunction with other independent measures of ecosystem function, like solar-induced fluorescence and carbon and water isotope studies. More work needs to be done to generate global coverage for OCS observations and to link this powerful atmospheric tracer to systems where fundamental questions concerning the carbon and water cycle remain.

Gross ecosystem productivity (GEP) in tropical forests varies both with the environment and with biotic changes in photosynthetic infrastructure, but our understanding of the relative effects of these factors across timescales is limited. Here, we used a statistical model to partition the variability of seven years of eddy covariance-derived GEP in a central Amazon evergreen forest into two main causes: variation in environmental drivers (solar radiation, diffuse light fraction, and vapor pressure deficit) that interact with model parameters that govern photosynthesis and biotic variation in canopy photosynthetic light-use efficiency associated with changes in the parameters themselves. Our fitted model was able to explain most of the variability in GEP at hourly (R = 0.77) to interannual (R = 0.80) timescales. At hourly timescales, we found that 75% of observed GEP variability could be attributed to environmental variability. When aggregating GEP to the longer timescales (daily, monthly, and yearly), however, environmental variation explained progressively less GEP variability: At monthly timescales, it explained only 3%, much less than biotic variation in canopy photosynthetic light-use efficiency, which accounted for 63%. These results challenge modeling approaches that assume GEP is primarily controlled by the environment at both short and long timescales. Our approach distinguishing biotic from environmental variability can help to resolve debates about environmental limitations to tropical forest photosynthesis. For example, we found that biotically regulated canopy photosynthetic light-use efficiency (associated with leaf phenology) increased with sunlight during dry seasons (consistent with light but not water limitation of canopy development) but that realized GEP was nonetheless lower relative to its potential efficiency during dry than wet seasons (consistent with water limitation of photosynthesis in given assemblages of leaves). This work highlights the importance of accounting for differential regulation of GEP at different timescales and of identifying the underlying feedbacks and adaptive mechanisms.