The results indicate that leaf senescence has been delayed over time and in response to temperature, although low-latitude sites show significantly stronger delays in senescence over time than high-latitude sites. While temperature alone may be a reasonable predictor of the date of leaf senescence when examining a broad suite of sites, it is important to consider that temperature-induced changes in senescence at high-latitude sites are likely to be constrained by the influence of photoperiod. Ecosystem-level differences in the mechanisms that control the timing of leaf senescence may affect both plant community interactions and ecosystem carbon storage as global temperatures increase over the next century.

Nutrient limitation is pervasive in the terrestrial biosphere, although the relationship between global carbon (C) nitrogen (N) and phosphorus (P) cycles remains uncertain. Using meta-analysis we show that gross primary production (GPP) partitioning belowground is inversely related to soil-available N : P, increasing with latitude from tropical to boreal forests. N-use efficiency is highest in boreal forests, and P-use efficiency in tropical forests. High C partitioning belowground in boreal forests reflects a 13-fold greater C cost of N acquisition compared to the tropics. By contrast, the C cost of P acquisition varies only 2-fold among biomes. This analysis suggests a new hypothesis that the primary limitation on productivity in forested ecosystems transitions from belowground resources at high latitudes to aboveground resources at low latitudes as C-intensive root- and mycorrhizal-mediated nutrient capture is progressively replaced by rapidly cycling, enzyme-derived nutrient fluxes when temperatures approach the thermal optimum for biogeochemical transformations.

Peatland measurements of CO 2 and CH 4 flux were obtained at scales appropriate to the in situ biological community below the tree layer to demonstrate representativeness of the spruce and peatland responses under climatic and environmental change (SPRUCE) experiment. Surface flux measurements were made using dual open-path analyzers over an area of 1.13 m 2 in daylight and dark conditions along with associated peat temperatures, water table height, hummock moisture, atmospheric pressure and incident radiation data. Observations from August 2011 through December 2014 demonstrated seasonal trends correlated with temperature as the dominant apparent driving variable. The S1-Bog for the SPRUCE study was found to be representative of temperate peatlands in terms of CO 2 and CH 4 flux. Maximum net CO 2 flux in midsummer showed similar rates of C uptake and loss: daytime surface uptake was -5 to -6 lmol m -2 s -1 and dark period loss rates were 4-5 lmol m -2 s -1 (positive values are carbon lost to the atmosphere). Maximum midsummer CH 4 -C flux ranged from 0.4 to 0.5 lmol m -2 s -1 and was a factor of 10 lower than dark CO 2 -C efflux rates. Midwinter conditions produced near-zero flux for both CO 2 and CH 4 with frozen surfaces. Integrating temperaturedependent models across annual periods showed dark CO 2 -C and CH 4 -C flux to be 894 ± 34 and 16 ± 2 gC m -2 y -1 , respectively. Net ecosystem exchange of carbon from the shrub-forb-Sphagnummicrobial community (excluding tree contributions) ranged from -3.1 gCO 2 -C m -2 y

Boreal peatlands contain approximately 500 Pg carbon (C) in the soil, emit globally significant quantities of methane (CH ), and are highly sensitive to climate change. Warming associated with global climate change is likely to increase the rate of the temperature-sensitive processes that decompose stored organic carbon and release carbon dioxide (CO ) and CH . Variation in the temperature sensitivity of CO and CH production and increased peat aerobicity due to enhanced growing-season evapotranspiration may alter the nature of peatland trace gas emission. As CH is a powerful greenhouse gas with 34 times the warming potential of CO , it is critical to understand how factors associated with global change will influence surface CO and CH fluxes. Here, we leverage the Spruce and Peatland Responses Under Changing Environments (SPRUCE) climate change manipulation experiment to understand the impact of a 0-9°C gradient in deep belowground warming ("Deep Peat Heat", DPH) on peat surface CO and CH fluxes. We find that DPH treatments increased both CO and CH emission. Methane production was more sensitive to warming than CO production, decreasing the C-CO :C-CH of the respired carbon. Methane production is dominated by hydrogenotrophic methanogenesis but deep peat warming increased the δ C of CH suggesting an increasing contribution of acetoclastic methanogenesis to total CH production with warming. Although the total quantity of C emitted from the SPRUCE Bog as CH is <2%, CH represents >50% of seasonal C emissions in the highest-warming treatments when adjusted for CO equivalents on a 100-year timescale. These results suggest that warming in boreal regions may increase CH emissions from peatlands and result in a positive feedback to ongoing warming.

We analysed work-related musculoskeletal injuries (WMSI) in two modern dance companies to determine whether injury rates decreased and patterns altered compared to previous 3-yr and 6-yr audits (0.48 and 0.25/1000-hrs exposure respectively). In this prospectively designed 15-yr cohort study, data were collected in 30-dancer Company-1 and 12-dancer Company-2. In-house physical therapists tracked WMSI and time-loss-injuries for 159 dancers (42 dancers/yr). 15-yrs were grouped into five 3-yr blocks for comparison with prior audits. Negative binomial logistic regression analyses were conducted with exposure-hrs converted to the natural log and used as the offset variable. Block and company were categorical predictors for dependent variables: WMSI, time-loss-injuries, trauma-injuries and overuse-injuries (p < 0.05). 69% of dancers reported WMSI; 45% sustained at least one time-loss-injury. Company-1, with greater annual exposure, was 1.6-times more likely to sustain time-loss-injuries (p = 0.016, CI = 1.095-2.422) and 5.6-times more likely to sustain time-loss overuse-injuries (p = 0.003, CI = 1.812-17.327). Compared to Block-1, WMSI and time-loss-injuries decreased in Blocks-2, 3, and 5 (p ≤ 0.027). The ratio of time-loss overuse to trauma-injuries was reversed, with trauma-injuries accounting for over 80% of injuries by Block 5. Time-loss-injuries averaged 0.16 injuries/1000-hrs, lower than rates in ballet and sports. Decreased injury rates and changed injury patterns demonstrate efficacious injury management and prevention programming.