B31E-01 INVITED
Climatic Controls on the Carbon, Water and Energy Fluxes from North Temperate and Southern Boreal Forests
The Fluxnet-Canada Research Network was a multi-year, inter-disciplinary study of the influences of inter-annual climate variability and disturbance on the carbon, water and energy cycles of the north temperate and southern boreal forests. The network, which ran from 2002 to 2007, was built on work that began during the Boreal Ecosystem-Atmosphere Study (1994-1996), continued as the Boreal Ecosystem Research and Monitoring Sites program (1997 to present), and now continues as the Canadian Carbon Program (2007-2010). The network comprises an east-west transect of flux stations, each with flux towers in mature as well as younger forest stands following disturbance by fire or harvesting. We present results from an ongoing synthesis study of the climatic and biophysical factors that control interannual variability in the forest-atmosphere exchanges of carbon, water and energy.
B31E-02 INVITED
Ecohydrologic Hypotheses Revisited: Global Synthesis with Canopy-Scale, FLUXNET Observations
Global synthesis of canopy-scale evapotranspiration measurements shows smooth transition from water-limited to energy-limited control, broadly consistent with catchment-scale relations and of a surprisingly similar functional form to that proposed by Budyko. Annual evaporative index (evaporated fraction of annual precipitation) tends to be reduced by precipitation in the form of snow as well as seasonal hydrologic surplus. Unexpectedly, grasslands tend to have a higher ratio of actual to potential evapotranspiration than treed landscapes, particularly savannas and shrublands. Control of daily evapotranspiration by soil moisture and potential evapotranspiration is well described by a logistic function akin to the two-stage model of evaporation decay, except during spring thaw when soil moisture can be high but ecosystems evapotranspire relatively little. Sites with seasonal droughts following periods of hydrologic surplus demonstrate significant water storage capacity that sustains evapotranspiration during periods of water deficit. Taken together, a simple framework of supply- or demand- limited evapotranspiration is supported by global FLUXNET observations, however climate seasonality and variability, the timescale of analysis, as well as vegetation cover all exert additional control.
B31E-03 INVITED
A Global Assessment on the Effects of Incoming Diffuse Radiation on the Primary Productivity of Terrestrial Ecosystems
It is well known that an increase in the fraction of incoming radiation diffused by clouds or aerosol may enhance the light use efficiency (LEU) by generating a more uniform light distribution on the leaf area. An analysis of this effect at five Fluxnet sites has been published by Gu et al. (JGR, 2002), demonstrating the possibility to use flux measurements for the assessment of these phenomena at ecosystem scale. Despite the scientific relevance of this issue, a comprehensive assessment of the impact of the inter-annual variability in the incoming photosynthetic radiation (PAR) on the primary productivity is still missing. Such an analysis is needed to improve our understanding of the effect of aerosols and cloud cover on the spatial and temporal variability of the gross primary productivity (GPP). In addition, a better knowledge of the dependence of LUE on the composition of the incoming light is required to improve the parameterization of models based on satellite estimates of the fraction of absorbed PAR. The increasing availability of ecosystem carbon fluxes, together with direct measurements of incoming direct and diffuse PAR, allow an improvement of the analyses performed in the past, both in terms of ecosystem and spatial coverage and of scientific understanding. For this purpose the Fluxnet database has been used in order to perform a global analysis that covers about 200 sites and 900 year x site of flux measurements. The sensitivity of the different biomes at the diffuse fraction in the incoming PAR has been assessed by analyzing the variation of the ecosystem light response curves. Finally, the fraction of the inter-annual variability of GPP related to the variability of the direct/diffuse ratio has been estimated for the different biomes.
B31E-04
Latitudinal Patterns of Interannual Variability in Net Ecosystem Exchange
Interannual variability (IAV) in net ecosystem exchange of carbon (NEE) is a common phenomenon observed at almost all eddy covariance flux site worldwide. However, our understanding of the spatial patterns of IAV is extremely limited. In this study, we examined latitudinal patterns of IAV in NEE based on 207 years of data at 55 eddy flux sites consisting located from 3.02o south hemisphere to of 56.62o north hemisphere. . We computed standard deviation (STD) of yearly mean values of NEE at individual sites to represent the absolute IAV (AIAV) and coefficient of variation (CV) for relative IAV (RIAV). Our results showed that AIAV significantly decreased with latitude in mixed forests of deciduous broadleaf and evergreen needleleaf forest (MIX) and grasslands (GRS) (P<0.06). RIAV significantly increased with latitude in in deciduous broadleaf forests (DBF) (P<0.06) but decreased in GRS (P=0.06) and croplands (CRP) (P<0.06). The spatial patterns of IAV in NEE need to be further improved once more years of NEE data become available. http://bomi.ou.edu/luo
B31E-05
Semi-Empirical Modelling Of Biotic And Abiotic Factors Controlling Ecosystem Respiration Across Eddy-Covariance Sites
In this study we analyse the ecosystem respiration (RECO) data from 83 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 and the site-to-site variability within each plant-functional-type (PFT) and the variability between different PFT (e.g evergreen needleleaf, grasslands, croplands..). At site level daily RECO data, derived from eddy covariance measurements, were described with a semi- empirical model that has been developed for soil respiration by Reichstein et al. (2003), which uses air temperature and precipitation as predictors of respiration. While the model generally gave a good description of 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 analysed different functional responses of the GPP to the RECO and the best results, in terms of variance explained by the model, modelling efficiency and standard error of parameters estimates, were obtained with a simple linear model. Even though the model was able to explain the temporal variability of RECO for all sites, the high level of variability of model parameters estimates within each PFT was not easily generalizable into a single model parameterization. For all PFTs we found a linear relationship between the reference respiration (R0), and maximum leaf area index (LAIMAX). Considering the LAIMAX as possible factor explaining the intersite variability within each PFT, we included it into the model (TPGPP&LAI Model) as an additional predictor possibly accounting for spatial variability of RECO. Finally, model parameters for each PFT were estimated. The new extended model showed higher modelling efficiencies ranging from 0.51 to 0.86, indicating that both abiotic factors, recent GPP and the general site productivity (indicated by LAIMAX) influence RECO. Additional variance might be explained by site history such as disturbance. The TPGPP&LAI Model could be used for up-scaling ecosystem respiration from flux sites to continental and global scale, linking for example the model into the MODIS GPP/NPP data stream. http://www.disat.unimib.it/telerilevamento/
B31E-06
A Cross-Site Evaluation of Alternative FPAR Products for Use in Diagnostic Carbon Flux Models
Input requirements for diagnostic carbon flux models include estimates of FPAR (the fraction of incident photosynthetically active radiation that is absorbed by the canopy). These estimates are available from a variety of satellite-borne sensors, but in their raw form may contain significant artifacts associated primarily with clouds. In this study we evaluated the sensitivity of simulated gross primary production (GPP), ecosystem respiration (Re), and net ecosystem production (NEP) to alternative formulations of FPAR from the MODIS sensor. These alternatives included the original product (FPARorig), a version based on filling and interpolation (FPARntsg), and a version with a smoothing based on the TIMESAT algorithm (FPARts). Estimates of daily GPP, Re, and NEP from multiple eddy covariance flux towers over multiple years were assembled as reference data. In almost all cases, model output based on FPARntsg and FPARts reduced bias on an annual basis and RMSE for daily values relative to model runs using FPARorig. At a wet conifer site, FPARts maintained the highest and most consistent FPAR, a pattern consistent with the dense coniferous forest canopy there. Limitations of FPARts were apparent at a grassland site with an abrupt fall off in greenness associated with onset of the dry season. Model effectiveness in capturing the interannual variation in NEP was also enhanced in many cases with the adjusted FPAR products.
B31E-07
Global Variability of Light Use Efficiency: a Model Data Integration Approach
Estimating the present global gross primary production (GPP) and giving reliable future predictions is one of the major challenges of carbon cycle research. Light-use efficiency (LUE) algorithms are a potentially effective approach for monitoring global primary production (GPP) using satellite-borne sensors such as the Moderate Resolution Imaging Spectroradiometer (MODIS). Their advantage is that all the drivers can be easily derived from remote sensed data or from existing climate observation networks offering a number of opportunities for up- scaling approaches. These algorithms are applied at relatively wide geographical areas, which may subsume significant heterogeneity in vegetation LUEmax and, hence, introduce error. However, data on the variability of the LUEmax coefficient, are scarce, often derived from leaf-level or NPP measurements and sometimes conflicting. A solution to this problem is to estimate the this physiological parameter inverting the algorithm against eddy covariance derived measurements of GPP, that are now globally available in a consistent format. Hence, the objective of this study is to fill this gap by means of quantifying and compare the variability of LUEmax and its sensitivity to environmental conditions among plant vegetation types. For this purpose, we optimized the MOD17 algorithm against the GPP time series from the FLUXNET measurement network. In doing this, we made a Bayesian data model synthesis by means of the Metropolis-Hastings algorithm. The LUEmax values from the BPLUT lookup table in the MOD17 user's manual were used as prior. The uncertainties in flux data were characterized specifically site by site. The parameter estimation considerably increased LUEmax for vegetation types with a short leaf-life while for evergreen vegetation types a posteriori parameter values were lower than the a priori ones. These optimized a posteriori values open different research questions regarding the uses of this model at wide geographical scale. In particular, while a classification in plant functional types showed a biome dependent LUEmax variability, intra-vegetation variability is still hampering upscaling approaches. This issue has to be addressed in our future research. Furthermore multiple- constraint approaches considering water fluxes could contribute significantly to the reduction of uncertainties in model estimates.
B31E-08
The use of Fluxnet to improve processes and parameters in land surface models: a case study with the NCAR Community Land Model
The Community Land Model version 3 (CLM3) simulates land-atmosphere exchanges in response to climatic forcings. CLM3 has known biases in the surface energy partitioning as a result of deficiencies in its parameterizations. Such models however need to be robust for multi-decadal global climate simulations. Fluxnet now provides an extensive data source for investigating land processes since it encompasses a global range of ecosystem-climate interactions. Data from 22 Fluxnet sites are used to identify model deficiencies. Process-based knowledge from these observations can help to improve (a) physical processes and (b) reduce uncertainty of empirical parameters in the model. Including a prognostic aquifer and a new bare soil evaporation resistance formulation results in a significantly improved soil hydrology and energy partitioning. Terrestrial water storage increased by up to 300 mm in warm climates and decreased in cold climates. Nitrogen control of photosynthesis is revealed as another missing process in the model. These improvements increase the correlation coefficient of latent heat fluxes from a range of 0.5-0.6 to the range of 0.7-0.9. Additionally, RMSE of the simulated sensible heat fluxes decrease by 20-50 %. Remaining deficiencies (such as overestimated primary production during the wet season in the tropics) might be attributed to model parameter uncertainties once processes are reasonably well constrained. A flux tower data assimilation framework for land surface models based on the Ensemble Kalman Filter may be used at this point in order to reduce model parameter uncertainty. First results from using such an approach are presented as an outlook and should serve as a starting point for discussion of Fluxnet synthesis efforts in this direction. In summary, Fluxnet has proven to be a valuable tool to develop and validate land surface models prior to their application. Such a model inherits a realistic representation of ecosystem-climate relationships for many ecosystems and climate zones and it will therefore be more suitable for computationally expensive coupled global climate simulations.