B23C-1490
Frameworks, Schemes and Hierarchies for Reducing Uncertainties in Estimation of Carbon Dynamics in Tree-Grass Systems
The global area of savannas varies between 15 and 37 M km2, depending upon inclusion of grasslands, dry woodlands and temperate tree-grass systems. There are large uncertainties in the carbon budgets of savannas, influenced by several factors: 1) large spatial extent and diversity of vegetation structure and floristics; 2) scarcity of field measured soil and biomass carbon stocks and fluxes; 3) non-linear relationships between vegetation dynamics, carbon stocks, carbon fluxes and climate across the savanna rainfall gradient; and 4) variation from seasonally inundated to semi-arid moisture regimes. Current approaches to assessing carbon dynamics of savannas may be grouped as: a) global scale biogeochemical models (e.g. Century, Biome-BGC) that simulate nutrient and carbon dynamics in response, primarily, to climate; b) dynamic global vegetation models (e.g. LPJ, IBIS) that simulate competitive interactions among plant functional types; c) land use change models in which net fluxes are computed based on estimated carbon stocks in ‘natural' versus transformed landscapes; d) inversion approaches in which regional net terrestrial fluxes are inferred from observed atmospheric carbon dioxide concentrations; and e) simple light use efficiency productivity models. A number of key surface properties can be supplied by the latest quantitative remote sensing, although not all are available with the spatial and temporal coverage required: a) optical remote sensing - fractional cover of soil, non-photosynthetic vegetation, and photosynthetic vegetation; light use efficiency; biochemistry; plant functional types; ground fire area and intensity, canopy water status; 3D canopy properties; and radiation interception probabilities; b) microwave remote sensing (passive, active and interferometry) – inundation and flooding dynamics; vegetation biomass and structure; soil moisture; c) LiDAR remote sensing – tree heights, canopy structure and biomass, synergies with radar and multi-angle optical. The limitations of in situ data for calibration and validation and the need to link process and pattern dynamically through time and across scales suggest that combinations of modelling approaches and data sources will be needed to significantly reduce uncertainties. Process models may be enhanced by specific biochemical and structural remote sensing retrievals. Physical energy and water balance models may be aided by time series of temperature, soil moisture, rainfall and atmospheric data retrieved by remote sensing. Land use change modeling relies upon remotely sensed definition of land cover change and combination with ancillary data to define land use. Model-data assimilation and inverse methods require prior information on surface states and fluxes that may be available from remote sensing products. Models describing savanna carbon dynamics as one biome among many global biomes, often neglect key processes relevant to savannas, including the impacts of fire, herbivory and shifting agriculture on carbon. In modeling the global savanna biome there is a pressing need for new frameworks, schemes and hierarchies that link the diverse approaches with key processes and mechanisms that are unique to tree-grass systems and that combine energy and water balance dynamics, the physiological and population dynamics of vegetation communities, and the spatially explicit patch dynamics of human disturbance.
B23C-1491
Estimating Fractional Cover of Photosynthetic Vegetation, Non-Photosynthetic Vegetation and Soil in Savannas Using the EO-1 Hyperion and MODIS Sensors
Monitoring the fractional cover of photosynthetic vegetation (PV), non-photosynthetic vegetation (NPV) and bare soil (BS) in savannas is important for carbon and water modeling, grazing management, fire risk assessment and erosion control. We developed a simple method for resolving their fractional cover with hyperspectral imagery, by combining the Normalized Difference Vegetation Index (NDVI) which measures vegetation greenness and the Cellulose Absorption Index (CAI), which quantifies the intensity of the cellulose-lignin feature at 2.0-2.2 μm, and then applying a linear unmixing. We applied this method to three EO-1 Hyperion scenes acquired during the 2005 growing season in a site in northern Australia. Data from field measurements and from fire scar maps provided a means for qualitatively validating the results obtained. We then explored the potential of the MODIS-TERRA sensor for resolving vegetation fractional cover. We generated synthetic MODIS data from the Hyperion images and also used actual MODIS reflectance from the MOD09 product, concurrent with the Hyperion images. We found that the MODIS-TERRA sensor, despite not being able to quantify the cellulose feature directly, can be used for mapping fractional cover. This is due to the fact that vegetation, regardless of its photosynthetic status, has a lower reflectance at 2.1 μm (MODIS band 7) than at 1.6 μm (MODIS band 6), compared to soils, which have a relatively flat spectra at those wavelengths. We propose using the ratio of band 7 to band 6 together with the NDVI for resolving the proportions of PV, NPV and BS. We tested the method in 10 independent savanna sites across Australia where grass curing is continuously monitored and found very good agreement both in space and in time between observed and modeled fractional cover. Finally, we developed a prototype of an operational product based on the MOD43 product (Nadir BRDF-Adjusted Reflectance 16-Day composites) and discuss its strengths and limitations.
B23C-1492
Remote Sensing Time Series Analysis of Coupled Impacts of Disturbance and Management on Landscape Carbon Storage in the Okavango Delta, Botswana
Tropical and subtropical savanna processes are largely impacted by both disturbance and disturbance regimes, notably fire and grazing. In the Okavango Delta, Botswana, savanna-grassland ecotones are also influenced by a third type of disturbance/disturbance regime: flooding. The interspersal of seasonal and occasional wetlands with savanna, grassland, and riparian woodland interfluves culminates in an environment with interacting drivers of landscape change over space and time: fire, flooding, grazing, climatic oscillations, and differential human management practices. The environs of the Okavango are currently subject to four different land management regimes, most of which are now separated by staffed veterinary fences: open or communal lands, photography concessions, hunting concessions, and reserves / national parks. This work utilizes a suite of 85 Landsat TM and ETM+ images from 1989 to 2002 to assess fire magnitude and impact on landscape carbon storage as mitigated by 1) changing and variable rainfall and flooding and 2) the introduction of management zones aimed in part at controlling both proximate and ultimate drivers of fire dynamics and landscape change.
B23C-1493
Ecological Succession, Land use Changes and Soil Organic C Stock in a Lake Retreat Area (Main Ethiopian Rift Valley)
In the Main Ethiopian Rift Valley, ecological succession is related to continuous lake retreat (Nyssen et al., 2004). Human activities, through their impact on land use and cover, affect this ecological succession. Through a remote sensing study, we extricated ecological succession and human activity as causative factors for land use and cover changes (LUCC) and explored which impact this has on soil organic C (SOC) stock in lake retreat areas. Remote sensing data used include a Landsat MSS from 1973, a Landsat TM from 1986 and a Landsat ETM+ from 2000. A conventional type of classification was used whereby supervised classification of the 2000 image was supplemented by unsupervised classification of the older datasets. Due to decreased rainfall and water abstraction for intense irrigated agriculture and floriculture in its catchment, Lake Abijata lost 46 % of its area between 2000 and 2006. On the emerged lands, a good ecological succession was observed between 1973 and 1986, with clear evidence for: emerged land -> grassland -> Acacia bushes -> open woodland. Between 1986 and 2000, LUCC tendencies were totally reversed and woody vegetation decreased strongly, indicating increased human impact (Habtamu et al., 2007). Based on an analysis of the Landsat imagery, coupled with soil and land use studies, determinants for SOC stock were found. Firstly, SOC stock significantly differs between cultivated land and grazing land (3301 and 2626 g m-2) on the one hand, and woodland (4594 g m-2) on the other. The strongest explanation of SOC stock is related to the duration of emergence and hence of pedogenesis. Its proxy, elevation, explains much of the variability of SOC (R2 = 0.48). Using a multiple regression model involving elevation and IR reflectance, the SOC stock in the study area could be assessed at 2196 (+ - 1517) g m-2 SOC in 2000, against 3222 (+ - 1639) g m-2 in 1973 (Nyssen et al., 2007), which is related to the post-1986 reversing of ecological succession in the lake retreat areas. Habtamu Temesgen, Nyssen, J., Amanuel Zenebe, Mengistie Kindu, Mitiku Haile, 2007. Ecological succession and land use changes in a lake retreat area (Main Ethiopian Rift Valley). Journal of Arid Environments, submitted. Nyssen, J., Poesen, J., Moeyersons, J., Deckers, J., Mitiku Haile, and Lang, A., 2004. Human impact on the environment in the Ethiopian and Eritrean Highlands – a state of the art. Earth Science Reviews, 64: 273-320. Nyssen, J., Habtamu Temesgen, Mulugeta Lemenih, Amanuel Zenebe, Mitiku Haile, 2007. Soil organic C stock in a lake retreat area under increased human pressure (Main Ethiopian Rift Valley). Global Change Biology, submitted.
B23C-1494
Ecosystem carbon budgets and crop yields in a tropical savanna ecosystem as related to changes in climate and management
The tropical savanna in West Africa is a critical ecosystem that provides products for people and livestock. It is, however, vulnerable to land disturbances and sensitive to climate change. We selected the Bawku savanna ecoregion in northern Ghana to evaluate the responses of both natural and managed ecosystems to climate change and management scenarios. Our results from this study show that the conversion of forest to agricultural use resulted in a substantial reduction in ecosystem carbon stocks across this ecoregion during the past century. Climate changes would lead to a continuous decline in soil carbon stock of the entire ecoregion through the 21rst century; however, this result could be offset by increasing nitrogen fertilization on the cultivated lands. An increase in nitrogen fertilization could also significantly increase grain yields of all crops, except groundnut. The biomass production and soil carbon stocks within woodlands would depend heavily on cutting options and little on climate change. In summary, this study indicates that food security and agricultural sustainability in the Bawku ecoregion would rely heavily on the level of nitrogen fertilizer applied in the future.
B23C-1495
Continuing and Proposed Studies on the Management and Conservation of Biodiversity and Soil Carbon in Western Africa
At the farm level, a study developed to carry out a conservation project in Tami, Togo, showed a low organic matter content in soils, often around 1% and rarely up to 10 %. The highest values were found in a sacred forest within the farm, but most soils, occupied by a woody savannah with a marked agricultural influence, showed different degrees of erosion, low fertility and very low organic carbon content. In a survey of pastures, grazed fallows and abandoned farmland, pastures showed the highest organic matter content. In addition to grazing, plant species richness, measured as the number of species per sample, showed a positive relationship with soil organic matter, significant near the 90% level. The results indicate a strong influence of human activity on soil formation and distribution, and also on plant diversity. In addition, the soil properties under the permanent forest suggest a high potential of the soils of the region for improvement of both agricultural yields and as a carbon sink in the frame of global change policies. Further research in northern Togo and in the greater region will assist the development of quantitative models on ecosystem service indicators, relating biodiversity, carbon, land use, and climate on multiple scales to best inform policy and adaptation strategies. Toward this end, new research is proposed to support ecosystem modeling, management, and policy initiatives such as the clean development mechanism of the Kyoto Protocol.
B23C-1496
The Use of Bayesian Modeling to Assess the Impact of Altered Precipitation on Leaf-level Carbon Exchange in Four Desert Savanna Ecosystems
Savannas are complex ecosystems with diverse plant communities and spatially variable nutrient and carbon dynamics. In semi-arid regions, savannas are rapidly changing as a result of climate change and/or land-use, both of which have the potential to alter carbon cycling processes. To determine the potential impacts of climate change on savanna systems, it is critical to understand the processes governing vegetation dynamics across the diverse range of savanna ecosystem types. Because water is the primary driver of biological activity in these ecosystems, changes in precipitation frequency and magnitude may significantly affect plant community composition and ecosystem carbon cycling through effects on leaf-level carbon dynamics. Here, we utilized photosynthesis data and models to explore the underlying mechanisms responsible for changes in leaf-level carbon exchange under altered precipitation. Our objective was to determine whether dominant plants in four North American deserts exhibited a common photosynthetic response to precipitation manipulations. In the summer of 2005 and 2006, photosynthetic CO2- and light-response curves were measured on the dominant plant functional groups (grasses and shrubs) in the Great Basin, Mojave, Sonoran, and Chihuahuan deserts. We used a hierarchical Bayesian modeling framework to integrate the extensive field data with a biochemical-based photosynthesis model, yielding estimates of photosynthetic parameters (e.g. rate of daytime respiration, maximum rate of carboxylation, and maximum rate of electron transport). The modeling results indicated that, generally, plant photosynthesis parameters were conserved across all desert sites and plant species. There is, however, evidence that supplemental precipitation affected photosynthetic responses as some species differed in key biochemical parameters under this treatment. This result suggests that in these ecosystems changes in precipitation associated with climate change have the potential to alter species composition through plant controls on ecosystem carbon cycling. The hierarchical Bayesian approach that we employed facilitated the estimation of key leaf-level carbon exchange parameters for native desert plants from geographically distinct regions. Such integration of data and models is expected to improve estimates of leaf to ecosystem carbon flux processes that are critical to understanding the impacts of climate change on complex systems such as desert savannas.
B23C-1497
Seasonal Trends in Photosynthesis and Mesophyll Conductance in a Mediterranean Oak Savanna
Understanding how and why leaves respond to extreme temperature and drought stress is an important aspect of modeling carbon uptake in a Mediterranean oak savanna. We made leaf-level measurements of CO2 (ACi) and light response (APar) curves as well as measurements of mesophyll conductance (gm), leaf nitrogen (NL), and specific leaf area (SLA) on blue oak (Quercus douglasii) leaves throughout the growing season during two consecutive years. From the ACi curves, photosynthetic capacity was calculated as well as gm using a model developed by Ethier and Livingston (2004). The modeled values were then compared to those measured in the field. In both years we found the same Vcmax trend with a peak in the early growing season, before the drought begins and after light availability is sufficient. It has been hypothesized that seasonal variance in Vcmax is actually an artifact of seasonal variance in gm. We, however, found a seasonal trend in both Vcmax and gm. This is supported by our data showing a trend in NL parallel to that of Vcmax. Mesophyll conductance was comparable to stomatal conductance (gs), ranging from 0.03 to 0.26 and 0.03 to 0.14 mol m-2 s-1 respectively. Additionally, both conductances followed the same seasonal trend. This research shows the importance of including mesophyll conductance in models of carbon uptake, as mesophyll conductance is often just as limiting as stomatal conductance to photosynthetic carbon uptake in a Mediterranean oak savanna.
B23C-1498
Where Does The Carbon Go? Carbon Dynamics And Fire of a North Australian Tropical Savanna
The role of fire as one of the primary natural carbon cycling mechanisms is a key issue in considering global change feedbacks. In north Australia, the dominant ecosystem is tropical savanna and for mesic savannas within 100 km of the northern coastline, fire, storms and cyclones all impact carbon stocks. Fire is the most frequent disturbance agent as fires burn with a near annual frequency in these systems. We aimed to determine the annual net ecosystem productivity (NEP) from these savannas and the impact of fire on productivity. We established a long-term eddy covariance flux tower at Howard Springs, Australia and present here 5 years of data from 2001 to 2005. Fire has direct impacts through emissions but also has indirect effects through the loss of productivity due to reduced functional leaf area index and the carbon costs of rebuilding the canopy. The impact of fire on the canopy latent energy exchange was evident for 40 days while the canopy was rebuilt; however, the carbon balance took approximately 70 days to recover. The annual fire free NEP at Howard Springs was estimated at -4.3 t C ha-1 y-1 with a range of -3.5 to -5.1 t C ha-1 y-1 across years. We calculated the average annual indirect fire effect as 0.7 t C ha-1 y-1 using a neural network model approach and estimated average emissions of fine and coarse fuels as 1.6 t C ha-1 y-1. This allowed us to calculate a net biome production of 2.0 t C ha-1 y-1. We then partitioned this remaining sink and suggest that most of this can be accounted for by woody increment (1.2 t C ha-1 y-1) and shrub encroachment (0.5 t C ha-1 y-1). Given the consistent sink at this site, even under an almost annual fire regime, there may be management options to increase carbon sequestration by reducing fire frequency. http://www.arts.monash.edu.au/ges/research/climate/fire/index.php
B23C-1499
Insights from Tower-Flux, Physiological and Stable Isotope Measurements to Elucidate Climatic and Structural Controls on CO2 Exchange in an Encroaching Savanna
Woody encroachment, a land cover change occurring in savanna ecosystems around the globe, remains a large uncertainty in regional carbon budgets. In this study, we investigate the structural and climatic controls on the interannual and seasonal variability of net CO2 exchange in an encroaching savanna. At our study site, located on the Edwards Plateau in Central Texas, two woody plant species have encroached a former grassland over the past 20 years. Continuous eddy covariance and related meteorological measurements began in July 2004. In 2006 and 2007, two years with markedly different precipitation inputs, monthly leaf-level gas exchange on four functionally different plant species, soil respiration and chamber net ecosystem exchange (NEE) measurements were made to give an insight in the structural controls on the carbon balance. In March 2007, weekly measurements of carbon isotope ratios of ecosystem-respired CO2 (δ13CR) started as part of a national AmeriFlux-isotope network. This data provides an independent assessment of the structural controls (C3 vs. C4) on the carbon exchange in this savanna ecosystem. The physiological and eddy flux measurements indicate that woody encroachment significantly alters carbon dynamics in this savanna ecosystem. The drought resistance of juniper trees and their year-round photosynthetic activity allows this ecosystem to maintain carbon neutral in a dry year (2006). This structural control however is only second to the climatic control, as shown in the contrasting pattern of NEE fluxes between a dry (2006) and an extremely wet year (2007). Preliminary results of the δ13CR measurements reflect a shift from C3 to C3 + C4 photosynthetic activity in the spring of 2007 and a sustained C3 + C4 activity throughout the wet summer of 2007, corroborating the results obtained from flux and physiological measurements.
B23C-1500
Carbon Dioxide and Water Cycling in a Semiarid Savanna in Southern Arizona, USA
The consequences of recent woody plant encroachment on the carbon and water cycling of semiarid ecosystems are not well understood. In this presentation, we present measurements made from 2004 - 2006 using sap flow and eddy covariance techniques to examine the carbon dioxide and water fluxes that occurred over a semiarid savanna on the Santa Rita Experimental Range in southern Arizona, USA. Over the last one hundred years this site has been transformed from a desert grassland to a savanna with greater than 35% tree cover by the encroachment of the native woody plant, mesquite ( Prosopis velutina). We have found that mesquite, even when they were dormant above ground, readily redistributed water upwards and downwards in the soil profile via their roots. This redistribution had important ecohydrological consequences like extending the season over which photosynthesis occurred. During the study period the site experienced below normal precipitation especially during the winter and spring period, and the site each year appeared to be a net carbon source. The two decades that preceded our study had above average precipitation, and this possibly resulted in a great deal of carbon accumulation that is now being released due to the current drought that has truncated the growing season.
B23C-1501
Scrub-Oak Biomass Stimulation by CO2 Enrichment: Sustained 11 Years But Mediated by Precipitation and Contrasting Species Responses
Terrestrial ecosystems may mitigate rising atmospheric carbon dioxide concentration (CO2) through increased carbon uptake and sequestration in plant biomass. Elevated CO2 commonly produces initial stimulation of photosynthesis and growth, but due primarily to complex interactions with climate related factors (i.e. water, light and nutrients), uncertainty regarding long-term biomass response persists. After 11 years of CO2 enrichment (ambient and ambient + 350 ppm) using open-top chambers, aboveground biomass stimulation was sustained in a Florida scrub-oak ecosystem, yielding a 67% increase at final harvest in June 2007. The scrub oaks Quercus geminata and Quercus myrtifolia represented 85% of total ecosystem aboveground biomass but displayed contrasting responses to elevated CO2. Q. myrtifolia showed consistent increase in shoot biomass over the course of the study (128% stimulation by elevated CO2) while shoot biomass of Q. geminata was not significantly increased (+6% difference between treatments). Both species displayed long-term mean stimulation of net leaf photosynthesis to elevated CO2 under saturated light conditions: stimulation of photosynthesis in Q. myrtifolia was nearly twice that in Q. geminata (63% and 35%, respectively). Over the course of the study, Q. geminata consistently displayed photosynthetic acclimation via reductions in maximum carboxylation rate (Vcmax) and maximum rate of electron transport (Jmax) while Q. myrtifolia photosynthesis did not acclimate to elevated CO2. Inter-annual variation in Q. myrtifolia annual biomass increment was correlated with rainfall and elevated CO2 stimulation of absolute biomass accumulation was greatest in wet years. This effect was muted at the ecosystem level because CO2 stimulation of biomass in Q. geminata, which utilizes the water table to a greater extent than Q. myrtifolia, showed no relationship with rainfall. These advantages afforded to Q. myrtifolia by elevated CO2 produced a significant change in ecosystem composition, a trend which may be further compounded over time by this system's short fire return cycle.
B23C-1502
Biosphere-atmosphere exchange of CO2, water and energy in natural savannah in Burkina Faso (W Africa)
Tropical savannahs cover an area of 17 x 106 km2 worldwide. Their role in the global greenhouse gas budget still remains uncertain. In the past large parts of the former native African savannahs have been converted to agricultural land. To contribute to evaluating greenhouse gas exchange of tropical savannah ecosystems, especially in sub- Saharan Africa, biosphere-atmosphere exchange of N2O, CH4 and CO2 was quantified in natural savannah and in rainfed agricultural land in Burkina Faso by means of eddy covariance (EC) and chamber measurements. An EC tower was established in a nature reserve to determine the net ecosystem exchange of CO2 (NEEC) and energy fluxes from November 2004 to October 2006. A Soil-Vegetation-Atmosphere-Transport (SVAT) model MixFor-SVAT (Oltchev et al., 2002) was used to compute the heat exchange between canopy and atmosphere, to conduct a plausibility test to the measured fluxes, and to investigate the component fluxes. MixFor-SVAT is a multi- layer model that describes the vertical distribution of radiation, momentum, energy and matter fluxes within and above mono- or multi-specific tree stands at 30-min resolution. Water fluxes in the plants are computed using a non-steady-state approach (HUNT et al., 1991). The EC measurements revealed that the natural savannah acted as a small C source in the dry period, whereas large amounts of CO2-C were bound during the rainy seasons, particularly from June to September. The balance of the first year of our observations indicated a C uptake of 373 g m-2 of the ecosystem, which is comparable to deciduous forests in Europe. The CO2 fluxes showed clear diurnal patterns with the highest uptake rates at noon (up to 1 mg m-2 s-1 in July and August) and a permanent slight release to the atmosphere during night-time.
B23C-1503
Combining Eddy Covariance, Leaf Level Measurements and Modelling to Investigate Ecosystem Fluxes in Tropical Grasslands
Our project determined seasonal and spatial variations in ecosystem fluxes of tropical grassland ecosystems by investigating three prominent grassland types along a hydrological gradient in the Okavango Delta, Botswana.To identify the environmental factors that control CO2 and H2O exchange in tropical grassland ecosystems, we successfully combined eddy covariance measurements, leaf level measurements and remotely sensed data.Grassland ecosystems growing under the same climate showed profound differences in ecosystem fluxes as well as what regulated those fluxes on an ecosystem level. The analysis of the eddy covariance measurements revealed a pronounced seasonal and spatial variation with maximum net ecosystem exchange (NE) varying between -25μ mol -2 s-1 and -1μ mol -2 s-1 across sites and seasons. Without water limitation the main factor for the differences in NE between ecosystems was nutrient content per vegetation unit. This importance of nutrient content was also confirmed by our leaf level measurements. Seasonal differences in NE varied between sites and were driven by phenology or temperature and light limitation.Eddy covariance measurements for this project were predominantly campaign measurements. To determine annual course and sum of NE, we adapted the ecosystem model BETHY (Biosphere-Energy Transfer Hydrology Scheme) by parameter inversion in combination with remotely sensed fraction of absorbed photosynthetically active radiation for each site.
B23C-1504
A Comparison Study of CO2 Exchange Within the Sahelian/Sudanian Zone of Africa
The paper summarizes some of the major findings from CO2 flux campaigns ranging from Hapex Sahel (Niger 1992) to the presently ongoing Sudan experiment. In total four sites are selected, all with sandy soils, low annual rainfall (250-500 mm), sparse vegetation with a maximum volumetric water holding capacity of approximately 15 %. At all sites the net carbon exchange is measured by eddy covariance technique while the two major components, i.e. the soil respiration and the net carbon assimilation are estimated by combining measurements and modeling. For the soil respiration it is found that the fist rainfall events prior to the proper rainy season generate significance CO2-busts form the wet soil. In general it is found that the soil is strongly controlled by temperature and soil moisture but also available soil carbon should play a significant role. Using the annual maximum leaf area index as a proxy for the available carbon content it is found that the modeling of the soil respiration can be significantly improved . Except for the early seson peaks the maximum respiration level is around 3-4 ƒÝmol m-2 s-1 during the growing season decreasing to 2 ƒÝmol m-2 s-1 during senescence and reaching values close to zero during the dry season. For the CO2 assimilation there are large diurnal fluctuations with downward directed fluxes reaching a maximum level of -20 ƒÝmol m-2 s-1 and upward directed nighttime fluxes typically around 5 ƒÝmol m-2 s-l. The photosynthetic CO2 uptake is modeled by use of a mechanistic model based on published values of the Rubisco capacity. The combined photosynthesis /soil respiration model is used for estimating seasonal carbon budgets and to estimate the environmental control of the CO2 exchange. Besides radiation and LAI, the CO2 assimilation is found to depend on soil moisture and temperature The sensitivity analysis shows that at low soil moisture the assimilation is determined by the water stress function whereas for soil moisture contents above 8 vol. % there is only little control on the assimilation. The temperature dependence confirms that there is an optimal temperature at around 36 oC and the assimilation rate decreases for both increasing and decreasing temperature. It is finally discussed how the extent of the Sahelian/Sudanian zone makes it necessary to combine the flux station data with earth observation technique when aiming at regional estimates.
B23C-1505
Spatial Heterogeneity and Sources of Soil Carbon in Southern African Savannas
Soil organic carbon (SOC) is one of the largest and most dynamic reservoirs of C on Earth, with nearly twice as much C stored in SOC than in the biosphere and atmosphere combined. SOC storage in global tropical savannas constitutes approximately 56 Gt of C, which rises to 216 Gt of C (i.e., about 17% of the terrestrial non- agricultural SOC), when woodlands, shrublands, and desert scrub are included. Savannas cover about 20% of the global land surface, including about one-half of Africa, Australia and South America. The shared dominance of trees and grasses in savannas, the dominant physiognomy in southern Africa, add more complexity to soil C pool partitioning and dynamics than is found in landscapes with a single physiognomy. Here, the spatial variability of the soil C pool was investigated with particular emphasis on understanding the contribution to SOC from trees and grasses at two savanna sites of the Kalahari Transect, one wet and the other dry. Using a combination of stable isotope techniques and geostatistics, the results showed that spatial patterns of soil δ13 C exist and were related to the distributions of woody (C3) and herbaceous (C4) vegetation at both sites. Heterogeneity of the sources of SOC, as well as heterogeneity in the amount of SOC, was greater at the dry site relative to the wet site. At the dry site, the grasses were the major contributor to soil C whereas in the wet site, woody vegetation was the major contributor, regardless of the location with respect to woody canopies.
B23C-1506
Patterns and Implications of Plant-Soil C and N Isotopic Compositions in African Savanna Ecosystems
Southern African savannas are mixed plant communities where C3 trees co-exist with C4 grasses. Owing to differences in their morphology and physiology, trees and grasses have different access to nutrients and water and different efficiency in the use of these resources. It is still unclear how climate variables such as the mean annual precipitation may affect the relative efficiency of grasses and trees in the use of water and soil nutrients such as nitrogen. In this study, the foliar δ15N and δ13C were used as indicators of nitrogen uptake and of water use efficiency, respectively, to investigate the effect of the rainfall regime on the use of nitrogen and water by herbaceous and woody plants. To this end, patterns of foliar δ15N and δ13C for both C3 and C4 plants as well as patterns of soil δ15N and δ13C in canopy and intercanopy areas were investigated both in the dry and in the wet season along the Kalahari megatransect, where a distinct rainfall gradient exists on a homogeneous soil substrate. Foliar δ15N signatures increased as aridity heightened for both C3 and C4 plants in both seasons, although the magnitude of the increase was different for these two plant functional types. Soil δ15N also significantly increased with aridity. Foliar δ13C signatures increased with aridity for C3 plants in the wet season but not in the dry season, while in C4 plants the relation between foliar δ13C signatures and aridity was more complex and non-linear in both seasons. The consistent higher foliar δ15N for C3 plants suggests that C4 plants are superior competitor for N. The different foliar δ13C relationships with rainfall for the C3 plants and C4 plants may indicate that the C3 plants have an advantage over C4 plants when competing for water resources. The differences in water and nitrogen use between C3 and C4 plants likely collectively contribute to the tree-grass coexistence in savannas.
B23C-1507
Long-term Soil C and N Dynamics in Response to Enhanced Wind Erosion in Semiarid Grassland, Using CENTURY Model
Recent studies show that enhanced wind erosion changes soil carbon (C) and nitrogen (N) cycling in desert grasslands of southern New Mexico. However, long-term effects at the scale of decades to centuries are less known, especially under the conditions of drought, directional changes in climate, and land use pressures. Additionally, previous studies have focused on the isolated response of soil C and N, with little understanding of their interactions and differential response of other sub-pools. Using CENTURY, a process-based biogeochemical model, we evaluate the potential impacts of enhanced wind erosion on the long-term dynamics of C and N in the Jornada Experimental Range, southern New Mexico. We find that enhanced wind erosion does have a significant effect on long-term dynamics of C and N that are similar to the short-term dynamics observed in a field experiment at Jornada. The relationships between pools of C and N, levels of wind erosion and vegetation cover reduction as well as the mechanisms by which wind erosion changes C and N cycling in desert grasslands are discussed.
B23C-1508
Spatially Explicit Modeling of Grazing Effect on Soil Organic Carbon Change in the Green River Basin, Wyoming
Proper grazing can improve ecosystem production and enhance carbon sequestration, while overgrazing can lead to net emission of carbon into the atmosphere. It is challenging to quantify the impacts of grazing at a regional scale owing to the spatial and temporal changes of biophysical settings and land management practices. In this study, we quantified and evaluated impacts of grazing intensity on the dynamics of soil organic carbon (SOC) across the Green River Basin in southwestern Wyoming using a biogeochemical model (EDCM—erosion-deposition-carbon model) and remotely sensed data. We simulated responses of ecosystem carbon stocks (including SOC) and fluxes to various grazing scenarios. Results based on these simulations indicate that: (1) sagebrush-dominated shrublands accumulated less soil carbon than did grass-dominated area; (2) grazing could lead to a greater decrease in production and SOC in the shrubland than in the grassland under the same intensity and (3) a grazing intensity of 0.03 animal units per hectare could reduce SOC at a rate of 3.1 gC/m2/yr in shrublands, but SOC changed little in grasslands during 30 years of simulated responses. Our model simulations also show that conversion of shrubland to grassland can enhance soil carbon sequestration in the Green River Basin.