Our objective was to gain a detailed understanding of how photosynthetically active radiation (PAR), vapor pressure deficit (D) and soil water interact to control transpiration in the dominant canopy species of a mixed hardwood forest in northern Lower Michigan. An improved understanding of how these environmental factors affect whole-tree water use in unmanaged ecosystems is necessary in assessing the consequences of climate change on the terrestrial water cycle. We used continuously heated sap flow sensors to measure transpiration in mature trees of four species during two successive drought events. The measurements were scaled to the stand level for comparison with eddy covariance estimates of ecosystem water flux (Fw). Photosynthetically active radiation and D together explained 82% of the daytime hourly variation in plot-level transpiration, and low soil water content generally resulted in increased stomatal sensitivity to increasing D. There were also species-specific responses to drought. Quercus rubra L. showed low water use during both dry and wet conditions, and during periods of high D. Among the study species, Acer rubrum L. showed the greatest degree of stomatal closure in response to low soil water availability. Moderate increases in stomatal sensitivity to D during dry periods were observed in Populus grandidentata Michx. and Betula papyrifera Marsh. Sap flow scaled to the plot level and Fw demonstrated similar temporal patterns of water loss suggesting that the mechanisms controlling sap flow of an individual tree also control ecosystem evapotranspiration. However, the absolute magnitude of scaled sap flow estimates was consistently lower than Fw. We conclude that species-specific responses to PAR, D and soil water content are key elements to understanding current and future water fluxes in this ecosystem.

Summary• Quantitative assessment of carbon (C) storage by forests requires an understanding of climatic controls over respiratory C loss. Ecosystem respiration can be estimated biometrically as the sum ( R Σ ) of soil ( R s ), leaf ( R l ) and wood ( R w ) respiration, and meteorologically by measuring above-canopy nocturnal CO 2 fluxes ( F cn ).• Here we estimated R Σ over 5 yr in a forest in Michigan, USA, and compared R Σ and F cn on turbulent nights. We also evaluated forest carbon-use efficiency ( E c = P NP / P GP ) using biometric estimates of net primary production ( P NP ) and R Σ and F cn -derived estimates of gross primary production ( P GP ).• Interannual variation in R Σ was modest (142 g C m − 2 yr ; 71% from R s , 18% from R l , and 11% from R w . Hourly R Σ was well correlated with F cn , but 11 to 58% greater depending on the time of year. Greater R Σ compared with F cn resulted in higher estimated annual P GP and lower annual E c (0.42 vs 0.54) using biometric and meteorological data, respectively.• Our results provide one of the first multiyear estimates of R Σ in a forested ecosystem, and document the responses of component respiratory C losses to major climatic drivers. They also provide the first assessment of E c in a deciduous forest using independent estimates of P GP .

This study compares carbon sequestration rates along two independent tidal mangrove creeks near Naples Bay in Southwest Florida, USA. One tidal creek is hydrologically disturbed due to upstream land use changes; the other is an undisturbed reference creek. Soil cores were collected in basin, fringe, and riverine hydrogeomorphic settings along each of the two tidal creeks and analyzed for bulk density, total organic carbon profiles, and sediment accretion. Radionuclides 137 Cs and 210 Pb were used to estimate recent sediment accretion and carbon sequestration rates.Carbon sequestration rates (mean˘standard error) for seven sites in the two tidal creeks on the Naples Bay (98˘12 g-C m´2¨year´1 (n = 18)) are lower than published global means for mangrove wetlands, but consistent with other estimates from the same region. Mean carbon sequestration rates in the reference riverine setting were highest (162˘5 g-C m´2¨year´1), followed by rates in the reference fringe and disturbed riverine settings (127˘6 and 125˘5 g-C m´2¨year´1, respectively). The disturbed fringe sequestered 73˘10 g-C m´2¨year´1, while rates within the basin settings were 50˘4 g-C m´2¨year´1 and 47˘4 g-C m´2¨year´1 for the reference and disturbed creeks, respectively. These data support our hypothesis that mangroves along a hydrologically disturbed tidal creek sequestered less carbon than did mangroves along an adjacent undisturbed reference creek.

We designed a CO ; -controllecl cuvette and stripping system to trace a 14 CO 2 pulse-label from photosynthetic assimilation by wetland plants (in this study Orontium aquaticum L.) to its release as 14 CH 4 by microbial respiration. The system maintained cuvette CO 2 concentrations to within ±5 Pa of the set-point, and it allowed continuous recovery of CH 4 produced by aceticlastic methanogenesis vs. that produced by CO 2 reduction. Radiocarbon activity in the soil dissolved inorganic C pool peaked on the first day then declined slowly. We did not detect radiocarbon activity in soil solution pools of several low molecular weight organic acids (acetate, formate, lactate, and propionate), but the label was detected in the bulk dissolved organic C pool. We argue that radiocarbon will be useful for investigating the contribution of root exudates to methanogenic metabolism, but data interpretation will require separation of the relative contributions of CO 2 reduction and aceticlastic methanogenesis to overall I4 CH 4 emissions. Processes such as CH 4 oxidation and acetogenesis must also be considered in quantitative estimates of photosynthetic support of methanogenesis.