B13D-01
Integrating Measures of Soil Respiration Across Spatial and Temporal Scales Along a Woody Plant Encroachment Gradient Using Traditional and Innovative Techniques
The physiognomic shift in ecosystem structure from a grassland to a woodland may alter the sensitivity of CO2 exchange of entire ecosystems to variations in growing-season temperatures and precipitation. Grasslands and woodlands are fundamentally different types of ecosystems that differ in both structure and function. Understanding ecosystem carbon flux dynamics and identifying whether regions or landscapes are sources or sinks for atmospheric carbon is especially important in light of climatic and vegetative cover change. This source/sink status may change greatly as the dominant vegetation shifts from a grassland to a woodland. One large component of ecosystem flux is the background efflux of CO2 from the soil (soil respiration, Rsoil). Rsoil has been shown to be a function of both biotic (vegetation cover type, litter quality, rooting depths) and abiotic (resource availability, temperature) factors, but the relative importance of these drivers has not been fully quantified. Such investigations are particularly of interest within the semiarid southwest where the major environmental drivers (temperature and available moisture) vary and covary throughout a growing season and vegetative cover change, in the form of woody plant encroachment into grasslands, is rampant. Regional Rsoil rates are often modeled using a single parameter, however, in semiarid regions that experience distinct periods of precipitation and drought, such parameterization is likely oversimplified. Therefore, determining when ecosystems are temperature sensitive and when they are not is vital for predicting future source/sink status of ecosystems. Our understanding of the spatial and temporal complexity of Rsoil has been limited by the methods available for field measurements. Within this study, we used a combination of the traditional soil-collar technique and soil CO2 sensors to obtain an extensive temporal and spatial estimation of Rsoil at each site along a woody encroachment gradient. Measures of Rsoil were made under grasses, shrubs, and in bare spaces so that the individual responses of multiple microhabitats could be analyzed within each site. Each of these three microhabitats differed significantly in terms of the magnitude of peak flux and the duration of activity in response to precipitation, suggesting that individual measurements within an ecosystem are also important for scaling studies. All three microhabitats followed a similar pattern throughout a growing season. Average Rsoil rates regressed against soil temperature showed no relationship in the months prior to the monsoon nor in the months after the monsoon, illustrating a restriction in temperature sensitivity due to lack of available soil moisture. Regressions of Rsoil against temperature within the monsoon months illustrated the differential timing of microhabitat response to temperature, with the sites under the mesquites experiencing the greatest Rsoil. These results suggest that as ecosystems shift in vegetative cover to woody plants and temperatures within the southwestern United States increase, ecosystem respiration rates may increase more rapidly than if the ecosystem were still dominated by grasses.
B13D-02 INVITED
Deconstructing Soil Respiration Response to Environmental Variables at Varying Temporal Scales
Soil respiration, (SR), a combination of autotrophic and heterotrophic respiration, is generally treated as a single dependent variable, influenced by environmental variables similarly at all time scales. However as a sum of two belowground processes, both of which may be influenced by differing environmental factors and at differing time scales, continued advancement in the development of predictive SR models needs to incorporate environmental influences at multiple time scales. Using an automated system, high frequency temporal SR, temperature and soil moisture measurements were collected in a well-drained deciduous stand at the Harvard forest, MA, and in a well drained coniferous dominated stand at the Howland forest, ME. Our objective was to utilize these datasets to tease out process based responses of SR to environmental variables at varying time scales. Coherence analysis is a unique tool for examining the correlation between factors that influence SR at varying temporal frequencies, thus isolating the factors important at each time-scale. Results of this analysis showed that SR had a significant coherence with soil and air temperature at seasonal and diel time scales and with moisture at synoptic scales of 2 days to 2 weeks. At the seasonal time scale, soil temperature was the dominant influence on SR. However when the diel variation in SR was compared to the diel variation in SR predicted from a soil temperature model, the predictive model underestimated the diel amplitude of observed SR. The residuals of observed versus predicted SR based on a soil temperature function were significantly correlated with lagged air temperature and lagged photosynthetically active radiation. This may indicate that the diel variation in root respiration is influenced by aboveground processes, which may affect the timing and amplitude of daily SR. At periods of days to weeks, SR response to precipitation events depended on the pre-soil moisture status, as well as the magnitude and duration of the precipitation event. Based on the identification of these three important temporal scales of variation, predictive models for SR at Harvard and Howland forest were developed using environmental variables representing processes at each of these temporal scales. Only with the use of high frequency temporal SR measurements that tease out the underlying processes of SR at their unique and varying time scales can adequate models for SR can be developed.
B13D-03
Effects of Fire and Drought on Soil CO2 Respiration in a Native Tallgrass Prairie
Climate variability, land management, and fire are all potentially important influences on grassland carbon cycling. Characterizing the effects of fire on grasslands, independent from climate variability, is difficult in experiments following a single field over time. To address this, we carried out an experiment on two adjacent 33 ha tallgrass-prairie pastures, burning one and using the other as a no-burn control at the USDA Grazinglands Research Laboratory near El Reno, OK, to distinguish the effects of interannual climate variation and fire on soil respiration and ecosystem carbon cycling. We found that while burning and water stress both had significant effects on soil respiration, the climatic driver (drought) was, by far, the larger factor in determining annual soil CO2 respiration from 2005 to 2007. Taking advantage of replicated, automated soil respiration chambers, we observed the treatment effects on hourly as well as seasonal or annual time scales. The effects that we directly attribute to fire can be divided into structural effects that produce prompt changes in soil respiration and biochemical effects that take longer to become evident and may persist longer.
B13D-04
Permafrost Thaw Stimulates Old Carbon Release and Alters Net Carbon Exchange From Tundra
At least 1218 Pg (billion tons) of soil carbon (C) are stored in permafrost soils in boreal and arctic ecosystems, almost twice as much C than currently contained in the atmosphere3. Permafrost thaw, and the microbial decomposition of previously frozen organic C, is considered one of the most likely positive feedbacks from terrestrial ecosystems to the atmosphere in a warmer world. Yet, the rate of release is highly uncertain but crucial for predicting the strength and timing of this C cycle feedback, and thus how important permafrost thaw will be for climate change this century and beyond. We report results from a tundra landscape undergoing permafrost thaw, where net ecosystem C exchange and the radiocarbon age of ecosystem respiration were measured to determine the influence of old C loss on ecosystem C balance. Sustained transfers of C to the atmosphere that could cause a significant positive feedback to climate change must come from old C, which forms the bulk of the permafrost C pool that accumulated over thousands of years. By partitioning respiration sources, we determined that areas that thawed over the past 15 years had 75% more annual losses of old C compared to minimally thawed areas, but had overall net ecosystem C uptake as increased plant growth offset these losses. In contrast, sites that thawed decades earlier lost an additional 25% more old C annually, which contributed to overall net ecosystem C release despite increased plant growth. These data document significant losses of soil C with permafrost thaw that, over decadal time scales, overwhelms increased plant C uptake at rates that could make permafrost a large biospheric C source in a warmer world, similar in magnitude to current C fluxes from land use change.
B13D-05
Identification of organic substrates used in heterotrophic respiration within thawing permafrost: Eight Mile Lake, Alaska
At the Eight Mile Lake Watershed (Healy, Alaska), ecosystem respiration exceeds gross primary productivity in extensively thawed permafrost sites. Hydrologic losses of dissolved organic- and inorganic carbon (DOC and DIC, respectively) from extensively thawed sites are 2 to 3 times greater than from minimally thawed tussock tundra and are an important component of the regional carbon balance. DOC and DIC in extensively thawed permafrost ranged from 40 to 60 mg C/L and 1 to 4 mg C-CO2/L, respectively and on an areal basis hydrologic fluxes of carbon were ca. 10 g C/m2. In order to gain better understanding of the organic substrates consumed in ecosystem respiration, we used tangential-flow ultrafiltration to separate colloidal DOC (5000 Daltons < DOC < 0.2 μm) from low molecular weight (LMW) compounds (DOC <5000 Daltons) in samples of surface- and soil-water draining extensively thawed permafrost plots. Samples were collected monthly between May and September 2008. After isolation we measured and report: 1) the radiocarbon content of the total, colloidal and LMW DOC fractions, 2) their relative heterotrophic bioavailability and 3) the radiocarbon content of CO2 produced by heterotrophic respiration of these DOC fractions.
B13D-06
Ecosystem Respiration in an Undisturbed, Old-Growth, Temperate Rain Forest
Old-growth forests are usually close to carbon neutral, and climate change may push them towards becoming net carbon sources. Ecosystem carbon exchange and its component fluxes, were measured in a temperate rainforest in South Westland, New Zealand. The forest, which receives over 3 m of rain a year, is dominated by 400 year-old podocarp trees, and is on a low nutrient, acidic, peat soil. Nighttime respiration measurements using eddy covariance were problematic due to katabatic induced CO2 drainage flows near the ground and low turbulence. Instead of the friction velocity filtering technique, we used the maximum eddy flux within a few hours of sunset to derive a function relating nighttime ecosystem respiration to soil temperature. The function was then used to calculate respiration for the nighttime periods. Soil respiration was measured at regular intervals during the growing season. Soil temperature was regulated by incoming radiation and changes in the soil heat capacity. The water table was typically only 160 mm below the ground surface. Soil respiration (mean = 2.9 μmol m-2 s-1) increased strongly with both an increase in soil temperature and an increase in the depth to the water table, and accounted for approximately 50% of ecosystem respiration. Changes in the water table depth caused by altered rainfall regime, evaporation and drainage are likely to have a significant effect on the soil respiration rate and carbon balance of this old-growth forest. Foliage and stem respiration were also measured and integrated to the canopy scale using a model. The model was then used to decompose ecosystem respiration measurements into its components. A combination of measured and modelled data indicates that the ecosystem is a net source for carbon (-0.34 kg C m&-2 yr-1).
B13D-07 INVITED
Semi-empirical Modeling of Biotic and Abiotic Factors Controlling Ecosystem Respiration Across Eddy-Covariance Sites
In this study we analyze the ecosystem respiration (RECO) data from 102 eddy covariance sites belonging to the FLUXNET network. The aim is to develop a semi-empirical model able to explain the temporal variability of RECO, the between-site variability within each plant functional type (PFT) and the variability between different PFTs (e.g. evergreen needleleaf, grasslands, etc.). For the temporal variability at each site we found that a semi-empirical model using only climate drivers as predictors failed to describe part of the variability in the data: a residual analysis showed that productivity had an additional effect on RECO since model residuals were correlated with gross primary production (GPP). We analyzed a range of different functional responses of the RECO to the GPP using the Akaike's information theoretic criterion for an objective model selection of the best model formulation. Although we included into the new model formulation the effect of GPP as driver of RECO we found a linear relationship between the reference respiration (R0), and maximum leaf area index (LAIMAX). On the basis of this result we may consider the LAIMAX as a possible factor explaining the site-to-site variability of (R0) within each PFT (cf. Reichstein et al. 2003, Global Biogeochemical Cycles), thus we included it into the model (TPGPP&LAI Model). The new extended model showed high cross-validated modelling efficiencies ranging from 0.51 to 0.86, indicating that using both abiotic factors (climate), recent productivity (daily GPP) and the general site productivity (LAIMAX) it is possible to describe both the spatial and temporal variability of RECO with a good accuracy. Via a residual analysis we identified however further variability in the data not described by the model that is related to phenology: immediately after the bud-burst and immediately after litterfall the model is not able to describe the increase in ecosystem respiration that we interpret as increasing autotrophic respiration in spring and increased heterotrophic respiration in fall as a result of litter decomposition, respectively. Moreover, the residuals analysis showed a systematic underestimation of model due to the decomposition of crop and grasses residues in the post-harvesting season or after cuttings. Finally an additional bias seems to be due to the nitrogen depositions in forested sites since the mean annual model residuals were correlated with the total nitrogen depositions (r=-0.36, p<0.01) indicating an overestimation of the model in sites with high nitrogen depositions which might have a limiting effect on soil respiration. In summary the main findings of this study are: i) above-ground productivity have an important effect on temporal variability of ecosystem respiration; ii) the general site productivity (using LAIMAX as an indicator) has an additional effect accounting for the spatial variability within different PFTs; including LAI in this model formulation opens interesting perspectives for regional estimates of RECO linking remote sensing and modelling; iii) additional variance not explained by the model may be explained by phenology and nitrogen depositions.
B13D-08
Seasonal and Long-Term Behavior of Soil and Autotrophic Respiration in the Ent Dynamic Global Terrestrial Ecosystem Model
To partition the contributions of autotrophic vs. soil respiration to total ecosystem net CO2 fluxes, we have developed new models of temperature adaptation of autotrophic respiration, and temperature and moisture responses of soil respiration, for simulation of CO2 fluxes in the Ent Dynamic Global Terrestrial Ecosystem Model (Ent DGTEM). Studies show that autotrophic respiration does not increase indefinitely with warmer climates due vegetation adaptation. In addition, seasonal cycles of soil respiration are strongly controlled by substrate as well as seasonal climate. Our temperature adaptation scheme is a simple adjustment of autotrophic respiration''s Q10 response to temperature relative to a running temperature mean. The soil temperature and moisture responses are based on analyses of field data, and replace the responses in the CASA'' (Bonan, 1996; Thompson et al., 1996; Fung et al., 2005) soil biogeochemistry model. We present results of simulations compared to Fluxnet data for grassland, boreal forest, temperate broadleaf forest, and tropical rainforest, demonstrating the model''s ability to predict net ecosystem exchange (NEE), seasonal respiration pulses following rainfall, and soil carbon stocks. We then apply these algorithms at the global scale driven by global observational meteorological data coupled with the land hydrology model of the NASA Goddard Institute for Space Studies'' general circulation model (NASA GISS GCM) to simulate global NEE. Results are forthcoming about the model''s ability to capture the diversity of behavior of the same plant functional type in different climate zones, and the predicted equilibrium carbon stocks compared to the global soil carbon dataset of the International Soil Reference and Information Centre- World Inventory of Soil Emission Potentials (ISRIC-WISE, Batjes, 1996).