B41F-01
Improving predictions of coastal air-sea CO2 fluxes
Coastal ocean air-sea CO2 fluxes are regionally and seasonally intense, and may significantly impact estimates of net continental fluxes or bias estimates of terrestrial carbon fluxes in regions where there is cross-shore airflow. Unfortunately the large natural variability of the coastal oceans means that the coastal oceans are observationally undersampled, and significant uncertainties regarding net carbon cycling in coastal margins limit the application of mass-balance approaches, placing significant limits on our ability to constrain net coastal fluxes. The Pacific coast of North America epitomizes these issues, with extreme variability in air-sea pCO2 differences and cross-coastal airflow patterns. We present new measurements of surface-water pCO2 in previously undersampled regions of the Pacific coast, based on mooring- and ship-of-opportunity sampling programs, which have signifcantly increased the coverage of observations in Pacific coastal waters. To this new data combined with historical observations we apply our previously- developed empirical algorithms based on remote-sensing observations to fill spatial and temporal observational gaps. From these we create seasonal and annual maps of CO2 flux distributions, and net air-sea flux estimates. We compare our predictions with observations to determine where, when and why the algorithms perform poorly or well.
B41F-02
Land and ocean biospheres coupling inferred from observations and modelling - implications for future precidtions
Remote sensing instruments (e.g., SeaWiFS) have provided about a decade-long continuous record of terrestrial biospheric parameters and surface ocean biology. Here I explore the coupled mode change over the land and ocean, using the SeaWiFS seasonal and monthly average biosphere data for the period of September 1997 - August 2006 (http://oceancolor.gsfc.nasa.gov). Relatively strong short-term negative NDVI tendencies are observed over large parts of Africa, the Gulf, Asia, Australia, and southern South America between first two and last two years of SeaWiFS measurements. In contrast Chl-a increases during the same period are predominantly observed in the Indian Ocean (IO), East Pacific (EP), West Pacific (WP), and in patches over the Southern Ocean. Important to note here that the regions of increased (decreased) Chl-a are generally adjacent to the land regions with negative (positive) NDVI tendencies. To reveal the coupled mode variability in NDVI and Chl-a Principal component analysis (PCA) is performed on the SeaWiFS biosphere dataset. PCA of NDVI and Chl-a jointly suggests existence of stationary/dipole modes of variability between several neighbouring land-ocean regions. Other satellite and analyzed products, e.g., SST, precipitation (PCP), absorbing aerosols suggest that the observed land-ocean biospheres are coupled beyond that can be explained by purely dynamical mechanisms. It is proposed here that not only the responses of biological observations to physical parameters, e.g., SST and Chl-a or NDVI and PCP, but also the biogeochemical coupling through the atmospheric constituents (e.g., aerosols from land) would be needed in fully interacting mode to explain SeaWiFS-like observations. Note the present coupled models do not treat the emission, transport and deposition of atmospheric constituents interactively in coupled carbon cycle simulations Some of these mechanisms are discussed in a recent publication (P. K. Patra, SOLA, 3, 77-80, 2007). I will use the modelling results using biogeochemical elemental cycling and atmospheric-CO2 inverse modelling (Patra et al., J. Geophys. Res., 112, G02012, 2007) for discussing implications of the observation based land-ocean coupled system for future projections of carbon exchange between the earth's surface and the atmosphere.
B41F-03
Carbon Cycling in Lake Superior: Impact on Upper Midwest Regional Carbon Balance
Understanding the regional surface-atmosphere exchange of carbon dioxide of inland water bodies is important for accurately quantifying and scaling regional continental carbon budgets. It is widely recognized that many lakes are net sources of CO2 to the atmosphere. However, our ability to predict CO2 fluxes from any particular lake remain rather limited. In a landscape rich in lakes and wetlands such as the upper Midwest Great Lakes region, assumptions that CO2 fluxes from lakes are negligible seem disingenuous and bear potential for adding error to estimates of regional carbon flux. The Laurentian Great Lakes cover 25 percent of the land area of the 8 Great Lakes states, and CO2 emission and seasonal cycling from them may be comparable to local terrestrial ecosystems. CO2 fluxes from Lake Superior are of particular interest because they may directly impact CO2 observations at nearby AmeriFlux towers. DOC inputs to lakes are considered a major controlling factor on lake CO2 concentrations. Recent findings estimated the turnover time of dissolved organic carbon (DOC) in Lake Superior to be about 8 years. We have collected lake water samples, analyzed above lake CO2 concentrations, and started coupling an ecosystem-carbon module to an existing high resolution hydrodynamic model of Lake Superior in an attempt to estimate these fluxes and their spatial and temporal variability. Here we present initial results from this effort and place these reseults in context with regional terrestrial CO2 fluxes. Municipal water intakes were analyzed for water properties. The measured values for alkalinity and pH in the Lake Superior samples were within the range previously reported for the lake. Additionally, CO2 concentrations measured above Lake Superior were shown to be elevated above what would be expected on land, with a gradient of increasing concentrations with increasing wind travel distance from shore. Both of these initial measurements suggest that Lake Superior can be a significant source of CO2. Modeling results reveal the complex flows and patterns that will need to be considered to estimate whole-lake annual CO2 emissions. http://atlantic.aos.wisc.edu/~superior/wiki/
B41F-04
Carbon Fluxes in the Underobserved Tropics
Two papers appearing in 2007 introduced new constraints to the perpetually-underdetermined atmospheric carbon dioxide surface flux inversion problem, but came to significantly different conclusions about the magnitude of tropical land fluxes. Stephens et al. (Science 316, 2007) used a new compilation of aircraft profiles to argue that most atmospheric transport models have unrealistic vertical transport, and that those models which agree with vertical CO2 profiles yield very small "flux polarity": a modest northern land uptake of carbon and effectively neutral net tropical land exchange. In contrast, the joint ocean-atmosphere inversion of Jacobson et al. (GBC 21, 2007) uses JGOFS ocean interior data to tightly constrain long-term average air-sea exchange of CO2. This strong constraint on oceanic fluxes allows atmospheric CO2 gradients to inform us more directly about terrestrial fluxes. We find relatively high flux polarity, with a tropical land source that is consistent with estimates of flux due to deforestation and land-use change. The discrepancy between these two papers for tropical and southern land flux is over 2 PgC yr-1. To resolve these radically different perspectives on surface fluxes of carbon dioxide, we have explored the use of model subsets as suggested by Stephens et al., and we have created a new joint inversion using a more advanced time-dependent atmospheric analysis. Preliminary results suggest that both groups are in part right: while flux polarity is indeed diminished by considering vertical CO2 gradients in the atmosphere, the tropical terrestrial biosphere remains a significant source of carbon to the atmosphere.
B41F-05
The Expanding NOAA Tall Tower Network for Monitoring Carbon Dioxide and Related Gases
Four new tower sites will have been added to the NOAA Earth System Research Laboratory's tall tower network by the end of 2007, bringing the total number of sites to seven. In addition to continuous monitoring of CO2, CO, and meteorological parameters, daily flask sampling using automated samplers has been implemented at several sites. The flask samples are analyzed for a large suite of species including CH4, N2O, SF6, δ13CO2, COS, a suite of halocarbons, and some hydrocarbons. O3 was measured at 10m and 400m above ground level at the Texas tower site during the summer of 2006 and has proven to be a useful indicator for urban and power plant pollution events. Radon-222 is measured at two of the sites and provides an indicator of surface influence. Several of the towers are directly beneath or near NOAA aircraft profiling sites. We will present an overview of results from the growing network, including an analysis of seasonal cycles and variability on synoptic time scales of CO2 and other gases, vertical gradients in CO2, and comparisons with results from the NOAA CarbonTracker CO2 data assimilation system. http://www.esrl.noaa.gov/gmd/ccgg/towers/index.html
B41F-06
The influence of resolution of meteorology, biogeochemical models and fossil fuel emissions on forward and inverse modelling of CO2 exchange over Europe using the network of tall towers.
Tall towers allow us to observe the integrated influence of carbon exchange processes from large areas on the concentrations of CO2. The signal received shows a large variability at diurnal and synoptic timescales. The question remains how high resolutions and how accurate transport models need to be, in order to discriminate the relevant source terms from the atmospheric signal. We will examine the influence of the resolution of (ECMWF) meteorological fields, antropogenic and biogenic fluxes when going from resolutions of 2° to 0.2° lat-lon, using a simple Lagrangian 2D transport model. Model results will be compared to other Eulerian model results and observations at the CHIOTTO/CarboEurope tall tower network in Europe. Biogenic fluxes taken into account are from the FACEM model (Pieterse et al, 2006). Results show that the relative influence of the different CO2 exchange processes is very different at each tower and that higher model resolution clearly pays off in better model performance.
B41F-07
Carbon Cycle Steady State Assumption Impacts in Biogeochemical Modeling Inverse Parameter Retrieval and Implications for Regional Net Ecosystem Fluxes Estimates
We analyze the impacts of the steady state assumption on inverse model parameter retrieval from biogeochemical models. An inverse model parameterization study using eddy covariance CO2 flux data was performed with the Carnegie Ames Stanford Approach (CASA) model under conditions of strict and relaxed carbon-cycle steady state assumption (CCSSA), in order to evaluate both the robustness of the model’s structure for the simulation of net ecosystem carbon fluxes and the assessment of the CCSSA effects on simulations and parameter estimation. Net ecosystem production (NEP) measurements from several eddy- covariance sites were compared with NEP estimates from the CASA model driven by local weather stations climate inputs as well as by remotely sensed fraction of photosynthetically active radiation absorbed by vegetation (fAPAR) and leaf area index (LAI). The parameters considered for optimization are directly related to above and belowground modeled responses to temperature and water availability, as well as a parameter (η) that relaxed the CCSSA in the model, allowing for site level simulations to be initialized either as net sinks or sources. A robust relationship was observed between NEP observations and predictions for most of the sites through the range of temporal scales considered (daily, weekly, biweekly and monthly), supporting the conclusion that the model structure is able to capture the main processes explaining NEP variability. Overall, relaxing CCSSA increased model efficiency and decreased normalized average error. Inter site variability was a major source of variance in model performance differences between fix (CCSSAf) and relaxed (CCSSAr) CCSSA conditions. These differences were correlated with mean annual NEP observations. A set of ancillary parameters were tested as alternatives to η, yielding statistically significant lower model performance results. η was found to be a key parameter in the optimization exercise, generating the highest model efficiency losses when removed from the initial parameter set. Differences between ε* and Q10 estimates under CCSSAr and CCSSAf conditions were correlated with η, suggesting parameter compensation effects in steady state approaches. Parameter errors were significantly lower under CCSSAr. Overall, the importance of model structure evaluation in data assimilation approaches is thus emphasized. Furthermore, using the parameter estimates and their covariances we compute and analyze spatially distributed carbon balance estimates for parts of Europe and associated uncertainties. Uncertainties introduced by wrong model assumptions (e.g. CCSSA) clearly exceed the purely statistical uncertainties that come out of conventional data- assimilation schemes.