Search

Three-hourly net ecosystem exchange (NEE) is estimated at spatial scales of 0.25∘ over the European continent, based on the pre-operational inverse modelling framework “CarboScope Regional” (CSR) for the years 2006 to 2019. To assess the uncertainty originating from the choice of a priori flux models and observational data, ensembles of inversions were produced using three terrestrial ecosystem flux models, two ocean flux models, and three sets of atmospheric stations. We find that the station set ensemble accounts for 61 % of the total spread of the annually aggregated fluxes over the full domain when varying all these elements, while the biosphere and ocean ensembles resulted in much smaller contributions to the spread of 28 % and 11 %, respectively. These percentages differ over the specific regions of Europe, based on the availability of atmospheric data. For example, the spread of the biosphere ensemble is prone to be larger in regions that are less constrained by CO2 measurements. We investigate the impact of unprecedented increase in temperature and simultaneous reduction in soil water content (SWC) observed in 2018 and 2019 on the carbon cycle. We find that NEE estimates during these 2 years suggest an impact of drought occurrences represented by the reduction in net primary productivity (NPP), which in turn leads to less CO2 uptake across Europe in 2018 and 2019, resulting in anomalies of up to 0.13 and 0.07 PgC yr−1 above the climatological mean, respectively. Annual temperature anomalies also exceeded the climatological mean by 0.46 ∘C in 2018 and by 0.56 ∘C in 2019, while Standardised Precipitation–Evaporation Index (SPEI) anomalies declined to −0.20 and −0.05 SPEI units below the climatological mean in both 2018 and 2019, respectively. Therefore, the biogenic fluxes showed a weaker sink of CO2 in both 2018 and 2019 (−0.22 ± 0.05 and −0.28 ± 0.06 PgC yr−1, respectively) in comparison with the mean −0.36 ± 0.07 PgC yr−1 calculated over the full analysed period (i.e. 14 years). These translate into a continental-wide reduction in the annual sink by 39 % and 22 %, respectively, larger than the typical year-to-year standard deviation of 19 % observed over the full period.

The eddy-covariance technique measures carbon, water, and energy fluxes between the land surface and the atmosphere at hundreds of sites globally. Collections of standardised and homogenised flux estimates such as the LaThuile, Fluxnet2015, National Ecological Observatory Network (NEON), Integrated Carbon Observation System (ICOS), AsiaFlux, AmeriFlux, and Terrestrial Ecosystem Research Network (TERN)/OzFlux data sets are invaluable to study land surface processes and vegetation functioning at the ecosystem scale. Space-borne measurements give complementary information on the state of the land surface in the surroundings of the towers. They aid the interpretation of the fluxes and support the benchmarking of terrestrial biosphere models. However, insufficient quality and frequent and/or long gaps are recurrent problems in applying the remotely sensed data and may considerably affect the scientific conclusions. Here, we describe a standardised procedure to extract, quality filter, and gap-fill Earth observation data from the MODIS instruments and the Landsat satellites. The methods consistently process surface reflectance in individual spectral bands, derived vegetation indices, and land surface temperature. A geometrical correction estimates the magnitude of land surface temperature as if seen from nadir or 40∘ off-nadir. Finally, we offer the community living data sets of pre-processed Earth observation data, where version 1.0 features the MCD43A4/A2 and MxD11A1 MODIS products and Landsat Collection 1 Tier 1 and Tier 2 products in a radius of 2 km around 338 flux sites. The data sets we provide can widely facilitate the integration of activities in the eddy-covariance, remote sensing, and modelling fields.

Photosynthesis is a major driver of terrestrial ecosystem dynamics. Unfortunately, gross primary productivity (GPP), or the rate at which solar energy is captured and stored into sugar molecules during photosynthesis, cannot be directly measured from remote sensing (RS) signals. Several RS signals related to vegetation pigments and to canopy structure can, however, serve as proxies for GPP. They can further be combined with different types and degrees of modelling to derive spatio-temporal estimations of GPP. Different strategies exist to do so, which often vary with respect to how much they depend on an in-situ reference for GPP, the gold standard being those derived from eddy covariance (EC) measurements at flux tower sites.Here we investigate several such strategies with a specific goal: to explore the potential contribution of Sentinel satellites to improve GPP estimation. The Sentinel fleet is maintained by the European Union’s Copernicus programme, thereby guaranteeing a certain longevity and enabling the establishment of operational services that do not depend on single satellite missions. The main RS signals we consider are: the OLCI global vegetation index (OGVI) and OLCI terrestrial chlorophyll index (OTCI) from the Sentinel-3 OLCI instrument; daytime and night-time land surface temperature (LST) from Sentinel-3 SLSTR; and sun-induced chlorophyll fluorescence (SIF) from TROPOMI on-board of Sentinel-5-P. We further use time series of Sentinel-2 data to quantify the spatial homogeneity within the observational footprints of these coarser spatial resolution products in order to ensure a proper comparison to flux-tower data. The whole exercise is part of the Sen4GPP project funded by the European Space Agency (ESA).The three strategies we explore to derive GPP are: (1) empirical SIF-based estimation of GPP, including a version involving spatial downscaling to reach a finer resolution of SIF; (2) deterministic modelling based on a quantum yield light use efficiency (LUE) model calibrated on EC flux towers; and (3) purely data-driven machine learning (ML) based on EC measurements at flux towers using dedicated 10-fold cross-validation using the FLUXCOM-X framework. The cross-comparison is done for independent flux tower sites over Europe based on the Warm Winter 2020 database, covering the recent past (2018-2020) when TROPOMI SIF observations are available.The results indicate that the ML approach clearly outperforms the process-based LUE approach, which itself performs better than SIF. However, this order also reflects a decreasing reliance in flux tower data and possibly increasing capacity to extrapolate to situations not present in the learning dataset. The results further indicate that the ML approach using Sentinel data can perform better than a baseline using MODIS data alone, probably due to the inclusion of SIF information. Results also illustrate how ensuring the spatial consistency between grid and tower does improve performance, strengthening the rational for spatially downscaling coarse RS signals such as SIF. Overall, these encouraging results bode well for the potential use of Sentinel data to improve our current capacity to monitor biogeochemical process at global scale.

While eddy covariance (EC) is a standard for measuring total ecosystem evaporation (evapotranspiration, ET), upscaling from the tower to the regional and global scales is still marred with uncertainties. Here, we explore this scale translation via the FLUXCOM-X framework, which links data from EC measurements and remote sensing to machine learning techniques to produce models for generating globally gridded products. In particular, we explore potential sources of uncertainty inherent to this pipeline and how these influence the resulting gridded products: training data quality including site-selection and coverage of in-situ data, and modelling distinct water flux components (i.e., transpiration (T) and abiotic evaporation) individually compared to total evapotranspiration.Overall, changes in the FLUXCOM-X framework compared to previous versions (Jung et al. 2019) results in tangible improvements to the spatial and temporal patterns of the global evapotranspiration products. Furthermore, the predictions of a corresponding transpiration product provide an empirical estimate of plant controlled water fluxes. The resulting global T/ET ratios are consistent with current estimates from isotopic analyses, but with the advantage of high spatio-temporal coverage. Lessons learned from this analysis provide a more targeted line of inquiry into potential avenues for further improvements in global evapotranspiration modeling.Jung, M., Koirala, S., Weber, U. et al. The FLUXCOM ensemble of global land-atmosphere energy fluxes. Sci Data 6, 74 (2019). https://doi.org/10.1038/s41597-019-0076-8

The integration of global land surface remote sensing and in situ measured ecosystem carbon fluxes through machine learning approaches offers a unique data-driven perspective to diagnose the carbon cycle. Earth Observation (EO) data sets from different parts of the electromagnetic spectrum contain specific information on the land surface status, but also on the structural and physiological vegetation conditions. Each EO-derived land surface variable alone has a limited scope, addresses only individual aspects of the complex system, and can be confounded by other factors. Here we use the new generation statistical flux upscaling framework Fluxcom-X to analyse the individual and synergistic contributions of different EO data sets to site-level terrestrial carbon fluxes in tailored cross-validation experiments. Several distinct data streams are explored as predictor variables: land surface temperature (LST) from both polar orbiters and geostationary satellites (MODIS and SEVIRI), far-red SIF from GOME2, multi-spectral vegetation optical depth (VOD) from different sources (ku-band climate archive from Moesinger et al.(2020) and L-band from SMOS), and soil moisture (SM) from ESA CCI. Each predictor variable undergoes a dedicated pre-processing in terms of quality checks and gap-filling. Beyond their overall added value in prediction skill, we are interested in the impacts of the EO predictors on different scales of carbon flux variability (e.g. diurnal, seasonal, seasonal anomalies, inter-annual, and between sites), specifically during situations of unusual water scarcity and surplus. We also compute SHAP values to understand how the machine learning model uses the EO information. Additionally, a second line of analysis addresses the role of acquisition properties for the accuracy of the estimates.The first results for the predictor variable MODIS LST show that the inclusion of MODIS LST improves GPP estimates on all time scales. The model strongly profits from LST as surrogate for moisture availability during dry anomalies, and for light availability during wet anomalies. Regarding the impact of acquisition properties of MODIS, we find that the variability in viewing geometry and overpass time does not affect the accuracy of simulated site-level GPP. However, failing to account for the clear-sky bias in availability of MODIS LST will result in a substantial decrease in accuracy, especially for overcast days.Further experiments will include SEVIRI LST, SIF, VOD, as well as soil moisture, and we will analyse their role in the data-driven simulations of carbon fluxes. The lessons learned from the site-level cross-validation experiments will guide the production of gridded estimates of gross and net carbon fluxes for Europe and the globe.

Quantification of land surface-atmosphere fluxes of carbon dioxide (CO2) fluxes and their trends and uncertainties is essential for monitoring progress of the EU27+UK bloc as it strives to meet ambitious targets determined by both international agreements and internal regulation. This study provides a consolidated synthesis of fossil sources (CO2 fossil) and natural sources and sinks over land (CO2 land) using bottom-up (BU) and top-down (TD) approaches for the European Union and United Kingdom (EU27+UK), updating earlier syntheses (Petrescu et al., 2020, 2021b). Given the wide scope of the work and the variety of approaches involved, this study aims to answer essential questions identified in the previous syntheses and understand the differences between datasets, particularly for poorly characterized fluxes from managed ecosystems. The work integrates updated emission inventory data, process-based model results, data-driven sectoral model results, and inverse modeling estimates, extending the previous period 1990–2018 to the year 2020 to the extent possible. BU and TD products are compared with European National Greenhouse Gas Inventories (NGHGIs) reported by Parties including the year 2019 under the United Nations Framework Convention on Climate Change (UNFCCC). The uncertainties of the EU27+UK NGHGI were evaluated using the standard deviation reported by the EU Member States following the guidelines of the Intergovernmental Panel on Climate Change (IPCC) and harmonized by gap-filling procedures. Variation in estimates produced with other methods, such as atmospheric inversion models (TD) or spatially disaggregated inventory datasets (BU), originate from within-model uncertainty related to parameterization as well as structural differences between models. By comparing NGHGIs with other approaches, key sources of differences between estimates arise primarily in activities. System boundaries and emission categories create differences in CO2 fossil datasets, while different land use definitions for reporting emissions from Land Use, Land Use Change and Forestry (LULUCF) activities result in differences for CO2 land. The latter has important consequences for atmospheric inversions, leading to inversions reporting stronger sinks in vegetation and soils than are reported by the NGHGI. For CO2 fossil emissions, after harmonizing estimates based on common activities and selecting the most recent year available for all datasets, the UNFCCC NGHGI for the EU27+UK accounts for 3392 ± 49 Tg CO2 yr-1 (926 ± 13 Tg C yr-1), while eight other BU sources report a mean value of 3340 [3238,3401] [25th,75th percentile] Tg CO2 yr-1 (948 [937,961] Tg C yr-1). The sole top-down inversion of fossil emissions currently available accounts for 3800 Tg CO2 yr-1 (1038 Tg C yr-1), a value close to that of the NGHGI, but for which uncertainty estimates are not yet available. For the net CO2 land fluxes, during the most recent five-year period including the NGHGI estimates, the NGHGI accounted for -91 ± 32 Tg C yr-1 while six other BU approaches reported a mean sink of -62 [-117,-49] Tg C yr-1 and a 15-member ensemble of dynamic global vegetation models (DGVMs) reported -69 [-152,-5] Tg C yr-1. The five-year mean of three TD regional ensembles combined with one non-ensemble inversion of -73 Tg C yr-1 has a slightly smaller spread (0th–100th percentile of [-135,45] Tg C yr-1), and was calculated after removing land-atmosphere CO2 fluxes caused by lateral transport of carbon (crops, wood trade and inland waters) resulting in increased agreement with the the NGHGI and bottom-up approaches. Results at the sub-sector level (Forestland, Cropland, Grassland) show generally good agreement between the NGHGI and sub-sector-specific models, but results for a DGVM are mixed. Overall, for both CO2 fossil and net CO2 land fluxes, we find current independent approaches are consistent with the NGHGI at the scale of the EU27+UK. We conclude that CO2 emissions from fossil sources have decreased over the past 30 years in the EU27+UK, while large uncertainties on net uptake of CO2 by the land surface prevent trend identification. In addition, a gap on the order of 1000 Tg C yr-1 between CO2 fossil emissions and net CO2 uptake by the land exists regardless of the type of approach (NGHGI, TD, BU), falling well outside all available estimates of uncertainties. However, uncertainties in top-down approaches to estimate CO2 fossil emissions remain uncharacterized and are likely substantial. The data used to plot the figures are available at https://doi.org/10.5281/zenodo.7365863.

Droughts affect terrestrial ecosystems directly and concurrently and can additionally induce lagged effects in subsequent seasons and years. Such legacy effects of drought on vegetation growth and state have been widely studied in tree ring records and satellite-based vegetation greenness, while legacies on ecosystem carbon fluxes are still poorly quantified and understood. Here, we focus on two ecosystem monitoring sites in central Germany with a similar climate but characterised by different species and age structures. Using eddy covariance measurements, we detect legacies on gross primary productivity (GPP) by calculating the difference between random forest model estimates of potential GPP and observed GPP. Our results showed that, at both sites, droughts caused significant legacy effects on GPP at seasonal and annual timescales, which were partly explained by reduced leaf development. The GPP reduction due to drought legacy effects is of comparable magnitude to the concurrent drought effects but differed between two neighbouring forests with divergent species and age structures. The methodology proposed here allows the quantification of the temporal dynamics of legacy effects at the sub-seasonal scale and the separation of legacy effects from model uncertainties. The application of the methodology at a larger range of sites will help us to quantify whether the identified lag effects are general and on which factors they may depend., Biogeosciences, 19 (17), ISSN:1726-4170

Accurate estimates and predictions of the global carbon fluxes are critical for our understanding of the global carbon cycle and climate change. Reducing the uncertainty of the terrestrial carbon sink and closing the budget imbalance between sources and sinks would improve our ability to accurately project future climate change. Net Ecosystem Exchange (NEE), the net flux of biogenic carbon from the land surface to the atmosphere, is only directly measured at a sparse set of globally distributed eddy-covariance measurement sites. To estimate the terrestrial carbon flux at the regional and global scale, a global gridded estimate of NEE must be accurately upscaled from a model trained at the ecosystem level. In this study, the Fluxcom system* is used to train a site-level model on remotely-sensed and meteorological variables derived from site measurements, MODIS and ECMWF ERA5 atmospheric reanalysis data. The non-representative distribution of these site-level data along with missing disturbance histories impart known biases to current upscaling efforts. Observations of atmospheric carbon may provide important additional information, improving the accuracy of the upscaled flux estimate. This study adds an atmospheric observational operator to the model training process that connects the ecosystem-level flux model to top-down observations of atmospheric carbon by adding an additional term to the objective function. The target data are regionally integrated fluxes from an ensemble of atmospheric inversions corrected for fossil-fuel emissions and lateral fluxes. Calculating the regionally integrated flux estimate at each training step is computationally infeasible. Our hypothesis is that the regional flux can be modeled with a limited set of points and that this sparse model preserves sufficient information about the phenomena to act as a constraint for the underlying ecosystem-level model, improving regional and global upscaled products. Experimental results show improvements in the machine learning based regional estimates of NEE while preserving features such as the seasonal variability in the estimated flux. *Jung, Martin, Christopher Schwalm, Mirco Migliavacca, Sophia Walther, Gustau Camps-Valls, Sujan Koirala, Peter Anthoni, et al. 2020. “Scaling Carbon Fluxes from Eddy Covariance Sites to Globe: Synthesis and Evaluation of the FLUXCOM Approach.” Biogeosciences 17 (5): 1343–65.

Global freshwater is becoming an increasingly valuable resource, both due to increased human use as well as due to ecological importance in a changing climate. Understanding the hydrological cycles which govern water availability requires broad scale estimates of terrestrial evaporation, or evapotranspiration, which incorporate the complex signals of plant water use via transpiration. In this regard, evapotranspiration estimated from eddy covariance has proven a valuable resource in understanding ecosystem scale water fluxes at sites around the world, and recent advances in methods for directly estimating transpiration from eddy covariance data provide the opportunity to understand the influence plants have on water cycles. However, linking these ecosystem scale estimates to global scale processes requires a model to act as an intermediary, such as the empirical models used in the FLUXCOM products which train machine learning models on eddy covariance data linked with remote sensing data.Here we look at the next generation of global terrestrial water flux estimates from FLUXCOM, including both the total evapotranspiration and the individual components of transpiration and abiotic evaporation. We benchmark these new estimates against previous FLUXCOM products, as well as compare to the state-of-the-art evapotranspiration estimates from process based models and remote sensing products. The high spatial and temporal scale allows for a close look at how the transpiration to evapotranspiration ratio varies both in space and time. We also outline estimate uncertainties from potential measurement biases to feature selection, and discuss the next steps for high quality empirical water flux estimates.

Much uncertainty remains in measuring the inter-annual and longer-term dynamics of vegetation gross and net primary productivity (GPP, NPP) and the connected land carbon sink. Potential for better GPP estimation lies in newer satellite products representing different processes or vegetation states, but how they capture interannual GPP dynamics remains to be explored. Here, we differentiate shorter- and longer-term vegetation dynamics and their drivers for several Earth-observation-based vegetation proxies and a GPP estimate using time series decomposition. We find that relations between proxies can significantly differ between time scales, along land cover and climate gradients. For GPP estimated at FLUXNET sites, seasonal and multiannual slopes differ by at least 25% for half of the cases investigated, indicating considerable mismatch if multiannual relations were derived from seasonal slopes. Considering time-scale-specific sensitivities between proxies of vegetation productivity may improve estimates of interannual variability in vegetation productivity. publishersversion published

Vegetation indices are the most widely used tool in remote sensing and multispectral imaging applications. This paper introduces a nonlinear generalization of the broad family of vegetation indices based on spectral band differences and ratios. The presented indices exploit all higher-order relations of the involved spectral channels, are easy to derive and use, and give some insight on problem complexity. The framework is illustrated to generalize the widely adopted Normalized Difference Vegetation Index (NDVI). Its nonlinear generalization named, kernel NDVI (kNDVI), largely improves performance over NDVI and the recent NIRv in monitoring key vegetation parameters, showing much higher correlation with independent products, such as the MODIS leaf area index (LAI), flux tower gross primary productivity (GPP), and GOME-2 sun-induced fluorescence. The family of indices constitutes a valuable choice for many applications that require spatially explicit and time-resolved analysis of Earth observation data. Reference: "A Unified Vegetation Index for Quantifying the Terrestrial Biosphere", Gustau Camps-Valls, Manuel Campos-Taberner, Álvaro Moreno-Martı́nez, Sophia Walther, Grégory Duveiller, Alessandro Cescatti, Miguel Mahecha, Jordi Muñoz-Marı́, Francisco Javier Garcı́a-Haro, Luis Guanter, John Gamon, Martin Jung, Markus Reichstein, Steven W. Running. Science Advances, in press, 2021

Terrestrial vegetation couples the global water, energy and carbon exchange between the atmosphere and the land surface. Thereby, vegetation productivity is determined by a multitude of energy- and water-related variables. While the emergent sensitivity of productivity to these variables has been inferred from Earth observations, its temporal evolution during the last decades is unclear, as well as potential changes in response to trends in hydro-climatic conditions. In this study, we analyze the changing sensitivity of global vegetation productivity to hydro-climate conditions by using satellite-observed vegetation indices (i.e. NDVI) at the monthly timescale from 1982–2015. Further, we repeat the analysis with simulated leaf area index and gross primary productivity from the TRENDY vegetation models, and contrast the findings with the observation-based results. We train a random forest model to predict anomalies of productivity from a comprehensive set of hydro-meteorological variables (temperature, solar radiation, vapor pressure deficit, surface and root-zone soil moisture and precipitation), and to infer the sensitivity to each of these variables. By training models from temporal independent subsets of the data we detect the evolution of sensitivity across time. Results based on observations show that vegetation sensitivity to energy- and water-related variables has significantly changed in many regions across the globe. In particular we find decreased (increased) sensitivity to temperature in very warm (cold) regions. Thereby, the magnitude of the sensitivity tends to differ between the early and late growing seasons. Likewise, we find changing sensitivity to root-zone soil moisture with increases predominantly in the early growing season and decreases in the late growing season. For better understanding the mechanisms behind the sensitivity changes, we analyse land-cover changes, hydro-climatic trends, and abrupt disturbances (e.g. drought, heatwave events or fires could result in breaking points of sensitivity evolution in the local interpretation). In summary, this study sheds light on how and where vegetation productivity changes its response to the drivers under climate change, which can help to understand possibly resulting changes in spatial and temporal patterns of land carbon uptake.

[EN] Empirical vegetation indices derived from spectral reflectance data are widely used in remote sensing of the biosphere, as they represent robust proxies for canopy structure, leaf pigment content, and, subsequently, plant photosynthetic potential. Here, we generalize the broad family of commonly used vegetation indices by exploiting all higher-order relations between the spectral channels involved. This results in a higher sensitivity to vegetation biophysical and physiological parameters. The presented nonlinear generalization of the celebrated normalized difference vegetation index (NDVI) consistently improves accuracy in monitoring key parameters, such as leaf area index, gross primary productivity, and sun-induced chlorophyll fluorescence. Results suggest that the statistical approach maximally exploits the spectral information and addresses long-standing problems in satellite Earth Observation of the terrestrial biosphere. The nonlinear NDVI will allow more accurate measures of terrestrial carbon source/sink dynamics and potentials for stabilizing atmospheric CO2 and mitigating global climate change., G.C.-V. was supported by the European Research Council (ERC) under the ERC Consolidator Grant 2014 project SEDAL (647423). M.C.-T. and F.J.G.-H. were supported by the EUMETSAT Satellite Application Facility on Land Surface Analysis (LSA-SAF). SR research was financially supported by the NASA Earth Observing System MODIS project (grant NNX08AG87A). J.A.G. acknowledges the support of NASA ABoVE award number NNX15AT78A. S.W. acknowledges funding from the Emmy Noether Programme (GlobFluo project) of the German Research Foundation (GU 1276/1-1) as well as funding from the European Union's Horizon 2020 research and innovation program under grant agreement 776186 (CHE project) and agreement 776810 (VERIFY project).

The productivity of terrestrial vegetation is determined by multiple land surface and atmospheric drivers. Water availability is critical for vegetation productivity, but the role of vertical variability of soil moisture (SM) is largely unknown. Here, we analyze dominant controls of global vegetation productivity represented by sun���induced fluorescence and spectral vegetation indices at the half���monthly time scale. We apply random forests to predict vegetation productivity from several hydrometeorological variables including multi���layer SM and quantify the variable importance. Dominant hydrometeorological controls generally vary with latitudes: temperature in higher latitudes, solar radiation in lower latitudes, and root���zone SM in between. We find that including vertically resolved SM allows a better understanding of vegetation productivity and reveals extended water���related control. The deep(er) SM control for semi���arid grasses and shrubs illustrates the potential of deep(er) rooting systems to adapt to water limitation. This study highlights the potential to infer sub���surface processes from remote sensing observations., Key Points: Vertically resolved soil moisture (SM) improves the understanding of large���scale vegetation productivity and yields extended water���related controls. Data���driven evidence for the meaningfulness of the long���standing modeling paradigm of vertical soil layer discretization. Comparatively deep SM is most relevant in semi���arid areas and for grasses and shrubs., German Research Foundation, ESA Living Planet Fellowship

Given its simplicity, the light use efficiency (LUE) concept is widely used for the estimation of gross primary productivity (GPP) at the ecosystem scale. In the last three decades, many types of LUE models have been developed to explain the dependencies of GPP to different environmental and meteorological conditions across spatial and temporal scales. Despite the simplicity, LUE models are robust against observations from daily to seasonal scales, though entailing challenges in parameter upscaling. The main differences across LUEs resides in the presence/absence of certain meteorological drivers and in the particular formulations of different response functions to diagnose instantaneous light use efficiency (ε*). Here, we collected different algorithms for describing the meteorological constraints on ε* from literature and performed a factorial experiment by recombining the different response functions between the different models to assess model performance at site and network level. These models were forced and parameterized by remote sensing data, meteorological data and GPP for 177 eddy covariance flux sites from the FLUXNET2015 and LaThuile using a data assimilation approach. The results show that the two selected optimal LUE models had no significant differences in model performances at site-level at daily, weekly, and monthly scales. The Nash-Sutcliffe Model Efficiency Coefficient (NSE) of 50% sites were larger than 0.726 and 0.725 at daily, 0.788 and 0.783 at weekly, 0.836 and 0.834 at monthly and 0.544 and 0.510 at annual scales. Based on the selected models, we further explored the different methods to upscale the optimized parameters: a) site means and medians per plant functional type and climate class, b) random forest regression, using bioclimatic variables and corresponding vegetation index, and c) selection according to the similarity between sites, determined via the NSE in mean seasonal cycle temperature, precipitation, radiation, and vegetation indexes within the same plant functional type. The model efficiencies in cross validation for both models show that using the median parameters per plant functional type had the best performance to upscale parameters from site-level to global-level at daily, weekly and monthly scales. Since the meteorological response functions and the corresponding parameters represent the sensitivity of plant photosynthesis to the meteorological conditions we further explore the relationship between the climate sensitivities and other environmental drivers as well as biophysical plant traits using global retrievals of Sun induced fluorescence. Our results emphasize that novel Earth Observations datasets and transfer learning approaches bridge the LUE formulation tradeoffs between complexity and tractability.

Photosynthesis is one of the most fundamental processes on Earth fuelling life by providing food and energy. Moreover, terrestrial vegetation is a key element in the climate system as it importantly affects exchange processes of carbon, water and energy between the land surface and the atmosphere. In times of a changing climate there is urgent need for detailed knowledge on the factors driving plant activity and for reliable observational systems of the terrestrial vegetation. Satellite remote sensing is the only means to obtain measurements with global coverage, including remote and inaccessible regions, in a spatially and temporally continuous manner. This thesis presents an assess- ment of our current observational capabilities of vegetation dynamics from space. Three complementary approaches of spaceborne ecosystem monitoring are inter-compared: 1) Spectral measurements of the land surface reflectance in the optical range give an indica- tion of the amount of green biomass (as an integrative signal of leaf quantity and quality) and hence of the potential to perform photosynthesis. 2) In the red and far-red spectral regions, satellite instruments register a very small additive signal to the reflected radiance which originates from photosynthetically active chlorophyll pigments, termed sun-induced chlorophyll fluorescence (SIF). 3) Carbon fluxes measured in-situ are upscaled to a global data set of model gross photosynthetic carbon uptake (known as GPP - gross primary production) using empirical relationships with remotely sensed land surface and environ- mental variables. Three case studies focus i) on the spring phenology in boreal forests, ii) on the peak growing season in circumpolar treeless regions, and iii) on phenological changes in ecosystems with varying abundances of trees globally in times of fluctuations in soil moisture availability. The results of all three case studies highlight the intrinsic differences between greenness on the one hand and photosynthetic activity on the other hand. Specifically – for the first time on synoptic scales – a decoupling of photosynthesis (as indicated by SIF and model GPP) and greenness (approximated by various indices derived from spectral reflectance measurements) could be observed in evergreen needleleaf forests during spring recovery. Similarly, a temporal mismatch occurs in northern hemi- sphere forests during the growing season. There, changes in incoming light co-vary with soil moisture and immediately affect photosynthetic performance but barely greenness. Moreover, it has emerged that the timing of peak photosynthesis and peak greenness are asynchronous in tundra areas, which is indicative of differing dynamics. Conversely, there is high consistency between the three approaches regarding the length of growing season in deciduous forests and moisture-related phenological shifts in non-forested ecosystems. The work in this thesis demonstrates that SIF represents an asset for the monitoring of the dynamics of photosynthesis and carbon uptake compared to greenness-based ap- proaches. There are further indications of SIF to track changes in photosynthetic yields. However, despite these promising results for the accurate tracking of photosynthesis from space, further research is required to provide higher resolution data sets with clearer sig- nals. Further, ground-based validation efforts are necessary to improve our mechanistic understanding of physiological and radiative transfer processes controlling the SIF signal.

Mid-to-high latitude forests play an important role in the terrestrial carbon cycle, but the representation of photosynthesis in boreal forests by current modelling and observational methods is still challenging. In particular, the applicability of existing satellite-based proxies of greenness to indicate photosynthetic activity is hindered by small annual changes in green biomass of the often evergreen tree population and by the confounding effects of background materials such as snow. As an alternative, satellite measurements of sun-induced chlorophyll fluorescence (SIF) can be used as a direct proxy of photosynthetic activity. In this study, the start and end of the photosynthetically active season of the main boreal forests are analysed using spaceborne SIF measurements retrieved from the GOME-2 instrument and compared to that of green biomass, proxied by vegetation indices including the Enhanced Vegetation Index (EVI) derived from MODIS data. We find that photosynthesis and greenness show a similar seasonality in deciduous forests. In high-latitude evergreen needleleaf forests, however, the length of the photosynthetically active period indicated by SIF is up to 6 weeks longer than the green biomass changing period proxied by EVI, with SIF showing a start-of-season of approximately 1 month earlier than EVI. On average, the photosynthetic spring recovery as signalled by SIF occurs as soon as air temperatures exceed the freezing point (2-3 °C) and when the snow on the ground has not yet completely melted. These findings are supported by model data of gross primary production and a number of other studies which evaluated in situ observations of CO2 fluxes, meteorology and the physiological state of the needles. Our results demonstrate the sensitivity of space-based SIF measurements to light-use efficiency of boreal forests and their potential for an unbiased detection of photosynthetic activity even under the challenging conditions interposed by evergreen boreal ecosystems.

Two new remote sensing vegetation parameters derived from spaceborne spectrometers and simulated with a three dimensional radiative transfer model have been evaluated in terms of their prospects and drawbacks for the monitoring of dense vegetation canopies: (i) sun-induced chlorophyll fluorescence (SIF), a unique signal emitted by photosynthetically active vegetation and (ii) the canopy scattering coefficient (CSC), a vegetation parameter derived along with the directional area scattering factor (DASF) and expected to be particularly sensitive to leaf optical properties. Here, we present the first global data set of DASF/CSC and examine the potential of CSC and SIF for providing complementary information on the controversially discussed vegetation seasonality in Amazon forests. A comparison between near-infrared SIF derived from the Global Ozone Monitoring Experiment-2 (GOME-2) instrument and the Orbiting Carbon Observatory-2 (OCO-2) (overpass time in the morning and noon, respectively) reveals the response of SIF to instantaneous photosynthetically active radiation (PAR). Large-scale seasonal swings of GOME-2 SIF amount up to 21% (regarding the annual maximum) and peak in October and around February, while OCO-2 SIF peaks in February. However, both time series agree very well if SIF is normalized by overpass time and wavelength. We further examine anistropic reflectance characteristics with the finding that the hot spot effect significantly impacts observed GOME-2 SIF values. On the contrary, our sensitivity analysis suggests that CSC is highly independent of sun-sensor geometry as well as atmospheric effects. The slight annual variability of CSC (3%) shows no clear seasonal cycle, while a relatively high spatial standard deviation points to a high degree of spatial heterogeneity in our study domain within the central Amazon Basin.

High-latitude treeless ecosystems represent spatially highly heterogeneous landscapes with small net carbon fluxes and a short growing season. Reliable observations and process understanding are critical for projections of the carbon balance of the climate-sensitive tundra. Space-borne remote sensing is the only tool to obtain spatially continuous and temporally resolved information on vegetation greenness and activity in remote circumpolar areas. However, confounding effects from persistent clouds, low sun elevation angles, numerous lakes, widespread surface inundation, and the sparseness of the vegetation render it highly challenging. Here, we conduct an extensive analysis of the timing of peak vegetation productivity as shown by satellite observations of complementary indicators of plant greenness and photosynthesis. We choose to focus on productivity during the peak of the growing season, as it importantly affects the total annual carbon uptake. The suite of indicators are as follows: (1) MODIS-based vegetation indices (VIs) as proxies for the fraction of incident photosynthetically active radiation (PAR) that is absorbed (fPAR), (2) VIs combined with estimates of PAR as a proxy of the total absorbed radiation (APAR), (3) sun-induced chlorophyll fluorescence (SIF) serving as a proxy for photosynthesis, (4) vegetation optical depth (VOD), indicative of total water content and (5) empirically upscaled modelled gross primary productivity (GPP). Averaged over the pan-Arctic we find a clear order of the annual peak as APAR ≦ GPP

Stratospheric sudden warmings (SSWs) are extreme events in the polar stratosphere that are both caused by and have effects on the tropospheric flow. This means that SSWs are associated with changes in the angular momentum of the atmosphere, both before and after their onset. Because these angular momentum changes are transferred to the solid Earth, they can be observed in the rate of the Earth's rotation and the wobble of its rotational pole. By comparing observed Earth rotation variations to reanalysis data, we find that an anomaly in the orientation of the Earth's rotational pole, up to 4 times as large as the annual polar wobble, typically precedes SSWs by 20-40 days. The polar motion signal is due to pressure anomalies that are typically seen before SSW events and represents a new type of observable that may aid in the prediction of SSWs. A decline in the length of day is also seen, on average, near the time of the SSW wind reversal and is found to be due to anomalous easterly winds generated in the tropical troposphere around this time, though the structure and timing of this signal seems to vary widely from event to event.

3-hourly Net Ecosystem Exchange (NEE) is estimated at spatial scales of 0.25 degrees over the European continent, based on the pre-operational inverse modelling framework CarboScope Regional (CSR) for the years 2006 to 2019. To assess the uncertainty originating from the choice of a-priori flux models and observational data, ensembles of inversions were produced using three terrestrial ecosystem flux models, two ocean flux models, and three sets of atmospheric stations. We find that the station set ensemble accounts for 61 % of the total spread of the annually aggregated fluxes over the full domain when varying all these elements, while the biosphere and ocean ensembles resulted in much smaller contributions to the spread of 28 % and 11 %, respectively. These percentages differ over the specific regions of Europe, based on the availability of atmospheric data. For example, the spread of the biosphere ensemble is prone to be larger in regions that are less constrained by CO2 measurements. We further investigate the unprecedented increase of temperature and simultaneous reduction of Soil Water Content (SWC) observed in 2018 and 2019. We find that NEE estimates during these two years suggest an impact of drought occurrences represented by the reduction of Net Primary Productivity (NPP), which in turn lead to less CO2 uptake across Europe in 2018 and 2019, resulting in anomalies up to 0.13 and 0.07 PgC yr-1 above the climatological mean, respectively. Annual temperature anomalies also exceeded the climatological mean by 0.46 °C in 2018 and by 0.56 °C in 2019, while standardized-precipitation-evaporation-index (SPEI) anomalies declined to −0.20 and −0.05 SPEI units below the climatological mean in both 2018 and 2019, respectively. Therefore, the biogenic fluxes showed a weaker sink of CO2 in both 2018 and 2019 (−0.22±0.05 and −0.28±0.06 PgC yr-1, respectively) in comparison with the mean −0.36±0.07 PgC yr-1 calculated over the full analysed period (i.e., fourteen years). These translate into a continental-wide reduction of the annual sink by 39 % and 22 %, respectively, larger than the typical year-to-year standard deviation of 19 % observed over the full period.