Search

The integration of functional traits into vulnerability assessments is a promising approach to quantitatively capture differences in species sensitivity and adaptive capacity to climate change, allowing the refinement of tree species distribution models. In response to a clear need to identify traits that are responsive to climate change and applicable in a management context, we review the state of knowledge of the main mechanisms, and their associated traits, that underpin the ability of boreal and temperate tree species to persist and (or) shift their distribution in a changing climate. We aimed to determine whether current knowledge is sufficiently mature and available to be used effectively in vulnerability assessments. Marshalling recent conceptual advances and assessing data availability, our ultimate objective is to guide modellers and practitioners in finding and selecting sets of traits that can be used to capture differences in species' ability to persist and migrate. While the physiological mechanisms that determine sensitivity to climate change are relatively well understood (e.g., drought-induced cavitation), many associated traits have not been systematically documented for North American trees and differences in methodology preclude their widespread integration into vulnerability assessments (e.g., xylem recovery capacity). In contrast, traits traditionally associated with the ability to migrate and withstand fire are generally well documented, but new key traits are emerging in the context of climate change that have not been as well characterized (e.g., age of optimum seed production). More generally, lack of knowledge surrounding the extent and patterns in intraspecific trait variation, as well as co-variation and interaction among traits, limit our ability to use this approach to assess tree adaptive capacity. We conclude by outlining research needs and potential strategies for the development of trait-based knowledge applicable in large-scale modelling efforts, sketching out important aspects of trait data organization that should be part of a coordinated effort by the forest science community. Key words: vulnerability assessment, drought tolerance, fire tolerance, migration ability, intraspecific variation in trait, species persistence. L'utilisation des traits fonctionnels dans l'evaluation de la vulnerabilite est une approche prometteuse pour integrer de maniere quantifiable les differences de sensibilite des especes et leur capacite d'adaptation aux changements climatiques, ameliorant ainsi les modeles de repartition des arbres. Afin d'identifier dans un contexte d'amenagement les traits cles qui sont affectes par les changements climatiques, nous examinons l'etat des connaissances sur les principaux mecanismes - ainsi que leurs traits associes--qui regissent la capacite des arbres des forets boreales et temperees a persister ou a migrer. Nous avons tente de determiner si les connaissances actuelles sont suffisamment matures et disponibles pour etre utilisees efficacement dans le contexte des evaluations de vulnerabilite. En synthetisant les avancees conceptuelles les plus recentes et en evaluant la disponibilite des donnees, nous avons comme objectif principal de guider les modelisateurs et autres intervenants dans la selection de traits fonctionnels pouvant servir a caracteriser les differences dans la capacite des especes a persister et a migrer face a un climat en changement. Par exemple, malgre que l'on comprenne assez bien les mecanismes physiologiques qui determinent la sensibilite aux changements climatiques (ex., cavitation induite par la secheresse), un grand nombre de traits associes a ces mecanismes n'ont pas ete systematiquement documentes pour les arbres d'Amerique du Nord. Nous constatons egalement des differences dans les methodologies utilisees pour mesurer ces traits (ex., capacite de retablissement des xylemes), ce qui nuit a l'integration des traits dans l'evaluation de la vulnerabilite. Pour leur part, les traits traditionnellement associes a la capacite de migrer et de resister au feu sont generalement bien documentes; cependant, de nouveaux traits cles emergeant dans le contexte des changements climatiques demeurent peu documentes (ex., l'age de la production optimale de graines). De facon generale, le manque de connaissance entourant la variabilite intraspecifique des traits, ainsi que sur la covariation et l'interaction entre traits, est limitant dans nos evaluations de la capacite adaptative des arbres. Nous concluons en soulignant des besoins precis en matiere de recherche et en identifiant certaines avenues possibles pour le developpement des connaissances liees aux traits applicables dans des projets de modelisation a grande echelle. Nous soulignons finalement l'importance de certains aspects de la gestion de donnees de traits qui devraient faire partie de tout effort coordonne de documentation par la communaute scientifique du milieu forestier. Mots-cles : evaluation de la vulnerabilite, tolerance a la secheresse, tolerance au feu, capacite de migration, variabilite intraspecifique des traits, persistance de l'espece., 1. Introduction Recent years have seen a marked increase in efforts to assess potential effects of climate change on the distribution and abundance of forest plant species (Thuiller et al. [...]

Aim: Plant functional groups are widely used in community ecology and earth system modelling to describe trait variation within and across plant communities. However, this approach rests on the assumption that functional groups explain a large proportion of trait variation among species. We test whether four commonly used plant functional groups represent variation in six ecologically important plant traits. Location: Tundra biome. Time period: Data collected between 1964 and 2016. Major taxa studied: 295 tundra vascular plant species. Methods: We compiled a database of six plant traits (plant height, leaf area, specific leaf area, leaf dry matter content, leaf nitrogen, seed mass) for tundra species. We examined the variation in species‐level trait expression explained by four traditional functional groups (evergreen shrubs, deciduous shrubs, graminoids, forbs), and whether variation explained was dependent upon the traits included in analysis. We further compared the explanatory power and species composition of functional groups to alternative classifications generated using post hoc clustering of species‐level traits. Results: Traditional functional groups explained significant differences in trait expression, particularly amongst traits associated with resource economics, which were consistent across sites and at the biome scale. However, functional groups explained 19% of overall trait variation and poorly represented differences in traits associated with plant size. Post hoc classification of species did not correspond well with traditional functional groups, and explained twice as much variation in species‐level trait expression. Main conclusions: Traditional functional groups only coarsely represent variation in well‐measured traits within tundra plant communities, and better explain resource economic traits than size‐related traits. We recommend caution when using functional group approaches to predict tundra vegetation change, or ecosystem functions relating to plant size, such as albedo or carbon storage. We argue that alternative classifications or direct use of specific plant traits could provide new insights for ecological prediction and modelling. [ABSTRACT FROM AUTHOR]

We describe the Max Planck Institute Carbon Cycle Data Assimilation System (MPI-CCDAS) built around the tangent-linear version of the land surface scheme of the MPI-Earth System Model v1 (JSBACH). The simulated terrestrial biosphere processes (phenology and carbon balance) were constrained by observations of the fraction of photosynthetically active radiation (TIP-FAPAR product) and by observations of atmospheric CO2 at a global set of monitoring stations for the years 2005-2009. The system successfully, and computationally efficiently, improved average foliar area and northern extra-tropical seasonality of foliar area when constrained by TIP-FAPAR. Global net and gross carbon fluxes were improved when constrained by atmospheric CO2, although the system tended to underestimate tropical productivity. Assimilating both data streams jointly allowed the MPI-CCDAS to match both observations (TIP-FAPAR and atmospheric CO2) equally well as the single data stream assimilation cases, therefore overall increasing the appropriateness of the resultant parameter values and biosphere dynamics. Our study thus highlights the role of the TIP-FAPAR product in stabilising the underdetermined atmospheric inversion problem and demonstrates the value of multiple-data stream assimilation for the simulation of terrestrial biosphere dynamics. The constraint on regional gross and net CO2 flux patterns is limited through the parametrisation of the biosphere model. We expect improvement on that aspect through a refined initialisation strategy and inclusion of further biosphere observations as constraints. [ABSTRACT FROM AUTHOR]

Since 70% of global forests are managed and forests impact the global carbon cycle and the energy exchange with the overlying atmosphere, forest management has the potential to mitigate climate change. Yet, none of the land-surface models used in Earth system models, and therefore none of today's predictions of future climate, accounts for the interactions between climate and forest management. We addressed this gap in modelling capability by developing and parametrising a version of the ORCHIDEE land-surface model to simulate the biogeochemical and biophysical effects of forest management. The most significant changes between the new branch called ORCHIDEE-CAN (SVN r2290) and the trunk version of ORCHIDEE (SVN r2243) are the allometric-based allocation of carbon to leaf, root, wood, fruit and reserve pools; the transmittance, absorbance and reflectance of radiation within the canopy; and the vertical discretisation of the energy budget calculations. In addition, conceptual changes were introduced towards a better process representation for the interaction of radiation with snow, the hydraulic architecture of plants, the representation of forest management and a numerical solution for the photosynthesis formalism of Farquhar, von Caemmerer and Berry. For consistency reasons, these changes were extensively linked throughout the code. Parametrisation was revisited after introducing 12 new parameter sets that represent specific tree species or genera rather than a group of often distantly related or even unrelated species, as is the case in widely used plant functional types. Performance of the new model was compared against the trunk and validated against independent spatially explicit data for basal area, tree height, canopy structure, gross primary production (GPP), albedo and evapotranspiration over Europe. For all tested variables, ORCHIDEE-CAN outperformed the trunk regarding its ability to reproduce large-scale spatial patterns as well as their inter-annual variability over Europe. Depending on the data stream, ORCHIDEE-CAN had a 67 to 92% chance to reproduce the spatial and temporal variability of the validation data. [ABSTRACT FROM AUTHOR]

Terrestrial productivity in semi-arid woodlands is strongly susceptible to changes in precipitation, and semi- arid woodlands constitute an important element of the global water and carbon cycles. Here, we use the Carbon Cycle Data Assimilation System (CCDAS) to investigate the key parameters controlling ecological and hydrological activities for a semi-arid savanna woodland site in Maun, Botswana. Twenty-four eco-hydrological process parameters of a terres-trial ecosystem model are optimized against two data streams separately and simultaneously: daily averaged latent heat flux (LHF) derived from eddy covariance measurements, and decadal fraction of absorbed photosynthetically active radiation (FAPAR) derived from the Sea-viewing Wide Field-of- view Sensor (SeaWiFS). Assimilation of both data streams LHF and FAPAR for the years 2000 and 2001 leads to improved agreement between measured and simulated quantities not only for LHF and FAPAR, but also for photosynthetic CO2 uptake. The mean uncertainty reduction (relative to the prior) over all parameters is 14.9% for the simultaneous assimilation of LHF and FAPAR, 8.5% for assimilating LHF only, and 6.1% for assimilating FAPAR only. The set of parameters with the highest uncertainty reduction is similar between assimilating only FAPAR or only LHF. The highest uncertainty reduction for all three cases is found for a parameter quantifying max- imum plant-available soil moisture. This indicates that not only LHF but also satellite-derived FAPAR data can be used to constrain and indirectly observe hydrological quantities. [ABSTRACT FROM AUTHOR]

The vertical distribution of soil organic matter (SOM) in the profile may constitute an important factor for soil carbon cycling. However, the formation of the SOM profile is currently poorly understood due to equifinality, caused by the entanglement of several processes: input from roots, mixing due to bioturbation, and organic matter leaching. In this study we quantified the contribution of these three processes using Bayesian parameter estimation for the mechanistic SOM profile model SOMPROF. Based on organic carbon measurements, 13 parameters related to decomposition and transport of organic matter were estimated for two temperate forest soils: an Arenosol with a mor humus form (Loobos, the Netherlands), and a Cambisol with mull-type humus (Hainich, Germany). Furthermore, the use of the radioisotope 210Pbex as tracer for vertical SOM transport was studied. For Loobos, the calibration results demonstrate the importance of organic matter transport with the liquid phase for shaping the vertical SOM profile, while the effects of bioturbation are generally negligible. These results are in good agreement with expectations given in situ conditions. For Hainich, the calibration offered three distinct explanations for the observations (three modes in the posterior distribution). With the addition of 210Pbex data and prior knowledge, as well as additional information about in situ conditions, we were able to identify the most likely explanation, which indicated that root litter input is a dominant process for the SOM profile. For both sites the organic matter appears to comprise mainly adsorbed but potentially leachable material, pointing to the importance of organo-mineral interactions. Furthermore, organic matter in the mineral soil appears to be mainly derived from root litter, supporting previous studies that highlighted the importance of root input for soil carbon sequestration. The 210Pbex measurements added only slight additional constraint on the estimated parameters. However, with sufficient replicate measurements and possibly in combination with other tracers, this isotope may still hold value as tracer for SOM transport. [ABSTRACT FROM AUTHOR]

In current dynamic global vegetation models (DGVMs), including those incorporated into Earth System Models (ESMs), terrestrial vegetation is represented by a small number of plant functional types (PFTs), each with fixed properties irrespective of their predicted occurrence. This contrasts with natural vegetation, in which many plant traits vary systematically along geographic and environmental gradients. In the JS-BACH DGVM, which is part of the MPI-ESM, we allowed three traits (specific leaf area (SLA), maximum carboxylation rate at 25 °C (Vcmax25) and maximum electron transport rate (Jmax25)) to vary within PFTs via trait-climate relationships based on a large trait database. For all three traits, the means of observed natural trait values strongly deviated from values used in the default model, with mean differences of 32.3% for Vcmax25, 26.8% for Jmax25 and 17.3% for SLA. Compared to the default simulation, allowing trait variation within PFTs resulted in GPP differences up to 50% in the tropics, in > 35% different dominant vegetation cover, and a closer match with a natural vegetation map. The discrepancy between default trait values and natural trait variation, combined with the substantial changes in simulated vegetation properties, together emphasize that incorporating observational data based on the ecological concepts of environmental filtering will improve the modeling of vegetation behavior in DGVMs and as such will enable more reliable projections in unknown climates. [ABSTRACT FROM AUTHOR]

Over recent years, it has become increasingly apparent that climate change and air pollution need to be considered jointly for improved attribution and projections of human-caused changes in the Earth system. Exchange processes at the land surface come into play in this context, because many compounds that either act as greenhouse gases, as pollutant precursors, or both, have not only anthropogenic but also terrestrial sources and sinks. And since the fluxes of multiple gases and particulate matter between the terrestrial biota and the atmosphere are directly or indirectly coupled to vegetation and soil carbon, nutrient and water balances, quantification of their geographic patterns or changes over time requires due consideration of the underlying biological processes. In this review we highlight a number of critical aspects and recent progress in this respect, identifying in particular a number of areas where studies have shown that accounting for ecological process understanding can alter global model projections of land-atmosphere interactions substantially. Specifically, this concerns the improved quantification of uncertainties and dynamic system responses, including acclimation, and the incorporation of exchange processes that so far have been missing from global models even though they are proposed to be of relevance for our understanding of terrestrial biota-climate feedbacks. Progress has also been made regarding studies on the impacts of land use/land cover change on climate change, but the absence of a mechanistically based representation of human response-processes in ecosystem models that are coupled to climate models limits our ability to analyse how climate change or air pollution in turn might affect human land use. A more integrated perspective is necessary and should become an active area of research that bridges the socio-economic and biophysical communities. [ABSTRACT FROM AUTHOR]

Terrestrial carbon (C) cycle models applied for climate projections simulate a strong increase in net primary productivity (NPP) due to elevated atmospheric CO2 concentration during the 21st century. These models usually neglect the limited availability of nitrogen (N) and phosphorus (P), nutrients that commonly limit plant growth and soil carbon turnover. To investigate how the projected C sequestration is altered when stoichiometric constraints on C cycling are considered, we incorporated a P cycle into the land surface model JSBACH (Jena Scheme for Biosphere-Atmosphere Coupling in Hamburg), which already includes representations of coupled C and N cycles. The model reveals a distinct geographic pattern of P and N limitation. Under the SRES (Special Report on Emissions Scenarios) A1B scenario, the accumulated land C uptake between 1860 and 2100 is 13% (particularly at high latitudes) and 16% (particularly at low latitudes) lower in simulations with N and P cycling, respectively, than in simulations without nutrient cycles. The combined effect of both nutrients reduces land C uptake by 25% compared to simulations without N or P cycling. Nutrient limitation in general may be biased by the model simplicity, but the ranking of limitations is robust against the parameterization and the inflexibility of stoichiometry. After 2100, increased temperature and high CO2 concentration cause a shift from N to P limitation at high latitudes, while nutrient limitation in the tropics declines. The increase in P limitation at high-latitudes is induced by a strong increase in NPP and the low P sorption capacity of soils, while a decline in tropical NPP due to high autotrophic respiration rates alleviates N and P limitations. The quantification of P limitation remains challenging. The poorly constrained processes of soil P sorption and biochemical mineralization are identified as the main uncertainties in the strength of P limitation. Even so, our findings indicate that global land C uptake in the 21st century is likely overestimated in models that neglect P and N limitations. In the long term, insufficient P availability might become an important constraint on C cycling at high latitudes. Accordingly, we argue that the P cycle must be included in global models used for C cycle projections. [ABSTRACT FROM AUTHOR]

The vertical distribution of soil organic matter (SOM) in the profile may constitute a significant factor for soil carbon cycling. However, the formation of the SOM profile is currently poorly understood due to equifinality, caused by the entanglement of several processes: input from roots, mixing due to bioturbation, and organic matter leaching. In this study we quantified the contribution of these three processes using Bayesian parameter estimation for the mechanistic SOM profile model SOMPROF. Based on organic carbon measurements, 13 parameters related to decomposition and transport of organic matter were estimated for two temperature forest soils: an Arenosol with a mor humus form (Loobos, The Netherlands), and a Cambisol with mull type humus (Hainich, Germany). Furthermore, the use of the radioisotope 210Pbex as tracer for vertical SOM transport was studied. For Loobos the calibration results demonstrate the importance of liquid phase transport for shaping the vertical SOM profile, while the effects of bioturbation are generally negligible. These results are in good agreement with expectations given in situ conditions. For Hainich the calibration offered three distinct explanations for the observations (three modes in the posterior distribution). With the addition of 210Pbex data and prior knowledge, as well as additional information about in situ conditions, we were able to identify the most likely explanation, which identified root litter input as the dominant process for the SOM profile. For both sites the organic matter appears to comprise mainly adsorbed but potentially leachable material, pointing to the importance of organo-mineral interactions. Furthermore, organic matter in the mineral soil appears to be mainly derived from root litter, supporting previous studies that highlighted the importance of root input for soil carbon sequestration. The 210Pbex measurements added only slight additional constraint on the estimated parameters. However, with sufficient replicate measurements and possibly in combination with other tracers, this isotope may still hold value as tracer for a SOM transport. [ABSTRACT FROM AUTHOR]

Terrestrial productivity in semi-arid woodlands is strongly susceptible to changes in precipitation, and semi-arid woodlands constitute an important element of the global water and carbon cycles. Here, we use the Carbon Cycle Data Assimilation System (CC-DAS) to investigate the mechanisms controlling ecological and hydrogical activities for a semi-arid savanna woodland site in Maun, Botswana. Twenty-four eco-hydrological process parameters of a terrestrial ecosystem model are optimized against two data streams either separately or simultaneously: daily averaged latent heat flux (LHF) derived from eddy covariance measurement, and decadal fraction of absorbed photosynthetically active radiation (FAPAR) derived from Sea-viewingWide Field-of-view Sensor (SeaWiFS). Assimilation of both LHF and FAPAR for the years 2000 and 2001 leads to improved agreement between measured and simulated quantities not only for LHF and FAPAR, but also for photosynthetic CO2 uptake. The closest agreement is found for each observed data stream when only the same data stream is assimilated. The mean uncertainty reduction (relative to the prior) over all parameters is 16.1% for the simultaneous assimilation of LHF and FAPAR, 9.2% for assimilating LHF only, and 7.8% for assimilating FAPAR only. Furthermore, the set of parameters with the highest uncertainty reduction is similar between assimilating only FAPAR or only LHF. The highest uncertainty reduction is found for a parameter describing maximum plant-available soil moisture for all three cases. This indicates that not only LHF but also satellite-derived FAPAR data can be used to constrain and indirectly observe hydrological quantities. [ABSTRACT FROM AUTHOR]

Plant litter stocks are critical, regionally for their role in fueling fire regimes and controlling soil fertility, and globally through their feedback to atmospheric CO2 and climate. Here we employ two global databases linking plant functional types to decomposition rates of wood and leaf litter (Cornwell et al., 2008; Weedon et al., 2009) to improve future projections of climate and carbon cycle using an intermediate complexity Earth System model. Implementing separate wood and leaf litter decomposabilities and their temperature sensitivities for a range of plant functional types yielded a more realistic distribution of litter stocks in all present biomes with the exception of boreal forests and projects a strong increase in global litter stocks by 35 Gt C and a concomitant small decrease in atmospheric CO2 by 3 ppm by the end of this century. Despite a relatively strong increase in litter stocks, the modified parameterization results in less elevated wildfire emissions because of a litter redistribution towards more humid regions. [ABSTRACT FROM AUTHOR]

Plant traits - the morphological, anatomical, physiological, biochemical and phenological characteristics of plants and their organs - determine how primary producers respond to environmental factors, affect other trophic levels, influence ecosystem processes and services and provide a link from species richness to ecosystem functional diversity. Trait data thus represent the raw material for a wide range of research from evolutionary biology, community and functional ecology to biogeography. Here we present the global database initiative named TRY, which has united a wide range of the plant trait research community worldwide and gained an unprecedented buy-in of trait data: so far 93 trait databases have been contributed. The data repository currently contains almost three million trait entries for 69 000 out of the world's 300 000 plant species, with a focus on 52 groups of traits characterizing the vegetative and regeneration stages of the plant life cycle, including growth, dispersal, establishment and persistence. A first data analysis shows that most plant traits are approximately log-normally distributed, with widely differing ranges of variation across traits. Most trait variation is between species (interspecific), but significant intraspecific variation is also documented, up to 40% of the overall variation. Plant functional types (PFTs), as commonly used in vegetation models, capture a substantial fraction of the observed variation - but for several traits most variation occurs within PFTs, up to 75% of the overall variation. In the context of vegetation models these traits would better be represented by state variables rather than fixed parameter values. The improved availability of plant trait data in the unified global database is expected to support a paradigm shift from species to trait-based ecology, offer new opportunities for synthetic plant trait research and enable a more realistic and empirically grounded representation of terrestrial vegetation in Earth system models. [ABSTRACT FROM AUTHOR]

Plant litter stocks are critical, regionally for their role in fueling fire regimes and controlling soil fertility, and globally through their feedback to atmospheric CO2 and climate. Here we employ two global databases linking plant functional types to decomposition rates of wood and leaf litter (Cornwell et al., 2008; Weedon et al., 2009) to improve future projections of climate and carbon cycle using an intermediate complexity Earth system model. Implementing separate wood and leaf litter decomposabilities and their temperature sensitivities for a range of plant functional types yielded a more realistic distribution of litter stocks in all present biomes with except of boreal forests and projects a strong increase in global litter stocks and a concomitant small decrease in atmospheric CO2 by the end of this century. Despite a relatively strong increase in litter stocks, the modified parameterization results in less elevated wildfire emissions because of litter redistribution towards more humid regions. [ABSTRACT FROM AUTHOR]

In view of its potential significance for soil organic matter (SOM) cycling, the vertical SOM distribution in the profile should be considered in models. To mechanistically predict the SOM profile, three additional processes should be represented compared to bulk SOM models: (vertically distributed) rhizodeposition, mixing due to bioturbation, and movement with the liquid phase as dissolved organic matter. However, the convolution of these processes complicates parameter estimation based on the vertical SOM distribution alone. Measurements of the atmospherically produced isotope 210Pbex may provide the additional information needed to constrain the processes. Since 210Pbex enters the soil at the surface and bind strongly to organic matter it is an effective tracer for SOM transport. In order to study the importance of root input, bioturbation, and liquid phase transport for SOM profile formation we performed Bayesian parameter estimation of the previously developed mechanistic SOM profile model SOMPROF. 13 parameters, related to decomposition and transport of organic matter, were estimated for the soils of two temperate forests with strongly contrasting SOM profiles: Loobos (the Netherlands) and Hainich (Germany). Measurements of organic carbon stocks and concentrations, decomposition rates, and 210Pbex profiles were used in the optimization. For both sites, 3 optimizations were performed in which stepwise 210Pbex data and prior knowledge were added. The optimizations yielded posterior distributions with several cases (modes) which were characterized by the dominant organic matter (OM) pool: non-leachable slow OM, leachable slow OM, or root litter. For Loobos, the addition of 210Pbex data to the optimization clearly indicated which case was most likely. For Hainich, there is more uncertainty, but the most likely case produced by the optimization agrees well with other measurements. For both sites the most likely case of the final optimization was one where leachable slow OM dominates, suggesting that most organic matter is adsorbed to the mineral phase. Liquid phase transport (advection) of OM was responsible for virtually all organic matter transport for Loobos, while for Hainich bioturbation (diffusion) and liquid phase transport were of comparable magnitude. These results are in good agreement with the differences between the two sites in terms of soil texture and biological activity. [ABSTRACT FROM AUTHOR]

There is a growing consensus that land surface models (LSMs) that simulate terrestrial biosphere exchanges of matter and energy must be better constrained with data to quantify and address their uncertainties. FLUXNET, an international network of sites that measure the land surface exchanges of carbon, water and energy using the eddy covariance technique, is a prime source of data for model improvement. Here we outline a multi-stage process for "fusing" (i.e. linking) LSMs with FLUXNET data to generate better models with quantifiable uncertainty. First, we describe FLUXNET data availability, and its random and systematic biases. We then introduce methods for assessing LSM model runs against FLUXNET observations in temporal and spatial domains. These assessments are a prelude to more formal model-data fusion (MDF). MDF links model to data, based on error weightings. In theory, MDF produces optimal analyses of the modelled system, but there are practical problems. We first discuss how to set model errors and initial conditions. In both cases incorrect assumptions will affect the outcome of the MDF. We then review the problem of equifinality, whereby multiple combinations of parameters can produce similar model output. Fusing multiple independent and orthogonal data provides a means to limit equifinality. We then show how parameter probability density functions (PDFs) from MDF can be used to interpret model validity, and to propagate errors into model outputs. Posterior parameter distributions are a useful way to assess the success of MDF, combined with a determination of whether model residuals are Gaussian. If the MDF scheme provides evidence for temporal variation in parameters, then that is indicative of a critical missing dynamic process. A comparison of parameter PDFs generated with the same model from multiple FLUXNET sites can provide insights into the concept and validity of plant functional types (PFT) — we would expect similar parameter estimates among sites sharing a single PFT. We conclude by identifying five major model-data fusion challenges for the FLUXNET and LSM communities: (1) to determine appropriate use of current data and to explore the information gained in using longer time series; (2) to avoid confounding effects of missing process representation on parameter estimation; (3) to assimilate more data types, including those from earth observation; (4) to fully quantify uncertainties arising from data bias, model structure, and initial conditions problems; and (5) to carefully test current model concepts (e.g. PFTs) and guide development of new concepts. [ABSTRACT FROM AUTHOR]

Eddy covariance data are increasingly used to estimate parameters of ecosystem models. For proper maximum likelihood parameter estimates the error structure in the observed data has to be fully characterized. In this study we propose a method to characterize the random error of the eddy covariance flux data, and analyse error distribution, standard deviation, cross- and autocorrelation of CO2 and H2O flux errors at four different European eddy covariance flux sites. Moreover, we examine how the treatment of those errors and additional systematic errors influence statistical estimates of parameters and their associated uncertainties with three models of increasing complexity - a hyperbolic light response curve, a light response curve coupled to water fluxes and the SVAT scheme BETHY. In agreement with previous studies we find that the error standard deviation scales with the flux magnitude. The previously found strongly leptokurtic error distribution is revealed to be largely due to a superposition of almost Gaussian distributions with standard deviations varying by flux magnitude. The crosscorrelations of CO2 and H2O fluxes were in all cases negligible (R2 below 0.2), while the autocorrelation is usually below 0.6 at a lag of 0.5 h and decays rapidly at larger time lags. This implies that in these cases the weighted least squares criterion yields maximum likelihood estimates. To study the influence of the observation errors on model parameter estimates we used synthetic datasets, based on observations of two different sites. We first fitted the respective models to observations and then added the random error estimates described above and the systematic error, respectively, to the model output. This strategy enables us to compare the estimated parameters with true parameters. We illustrate that the correct implementation of the random error standard deviation scaling with flux magnitude significantly reduces the parameter uncertainty and often yields parameter retrievals that are closer to the true value, than by using ordinary least squares. The systematic error leads to systematically biased parameter estimates, but its impact varies by parameter. The parameter uncertainty slightly increases, but the true parameter is not within the uncertainty range of the estimate. This means that the uncertainty is underestimated with current approaches that neglect selective systematic errors in flux data. Hence, we conclude that potential systematic errors in flux data need to be addressed more thoroughly in data assimilation approaches since otherwise uncertainties will be vastly underestimated. [ABSTRACT FROM AUTHOR]

Abstract: The most widely used method to quantify gross N transformation rates in soils is based on 15N dilution and enrichment principles. To identify rate parameters, 15N-tracing experiments are analysed by models that are linked to algorithms that try to minimize the misfit between modelled and observed data. In currently available 15N-tracing models optimization algorithms are based on the Levenberg–Marquardt method that is suitable for the determination of small number of parameters. Therefore, these models are restricted to a few processes. Methods based on Monte Carlo sampling have the potential to overcome restrictions on parameter numbers but have not been tested for application in 15N-tracing models. Here, for the first time, we use a Markov chain Monte Carlo (MCMC) method with a tracing model to simultaneously determine the probability density functions (PDFs) of the whole set of parameters for a previously published data set [Müller, C., Stevens, R.J., Laughlin, R.J., 2004. A 15N tracing model to analyse N transformations in old grassland soil. Soil Biology & Biochemistry 36, 619–632]. We show that the MCMC method can simultaneously determine PDFs of more than 8 parameters and demonstrate for the first time that it is possible to optimize models where transformations are described by Michaelis–Menten kinetics. Setting the NH4 + oxidation rate to Michaelis–Menten kinetics reduced the misfit by 19%. Together with monitoring diagnostics for parameter convergence, the MCMC method is a very efficient and robust technique to determine PDFs for parameters in 15N-tracing models that contain large number of N transformations and complex process descriptions. [Copyright &y& Elsevier]