Skip to main content

The terrestrial Carbon cycle under Climate Variability and Extremes – a Pan-European synthesis

Final Report Summary - CARBO-EXTREME (The terrestrial Carbon cycle under Climate Variability and Extremes - a Pan-European synthesis)

Executive summary:

Climate strongly influences terrestrial ecosystems, their carbon cycle and thus their ability to act as a carbon sink (depending on climate and land-use). Climate change is expected to lead not only to rising mean temperatures across Europe with decreasing precipitation in southern regions and increasing rainfall and snow in northern regions, but also to changes in the magnitude and frequency of extreme weather events. The future fate of European terrestrial biosphere's acting as a net carbon sink is highly uncertain. So far only gradual climate and land-use change (e.g. slow warming) has been seriously considered in predictive carbon cycle studies. Climate variability and extremes will play an important role, but have not been sufficiently accounted for in modelling and experimental studies, thus leading to a critical knowledge gap.

CARBO-EXTREME's overall objective is to obtain a better and more predictive understanding of European terrestrial carbon cycle responses to climate variability and extreme weather events. Therefore we have combined diverse observations of ecosystem responses to climate extremes with computer modeling in a new model-data integration framework to improve our diagnostic and prognostic understanding of climate-carbon interactions on different time-scales. A consistent harmonized multi-source database on the European carbon cycle components for studying climate variability and extreme events was built and a Bayesian model calibration and comparison performed, to improve terrestrial carbon cycle predictions and their uncertainties in scenario analyses.

Project Context and Objectives:

It is beyond scientific doubt that climate change is occurring and that it is being driven by human activity. Climate strongly influences our terrestrial biosphere, the functioning of terrestrial ecosystems and therefore their carbon cycling and carbon sequestration potential. Climate change is expected to lead not only to rising mean temperatures across Europe with decreasing precipitation (i.e. rainfall) in southern regions and increasing precipitation (rainfall and snow) in northern regions, but also to changes in the magnitude and frequency of extreme weather events. The terrestrial carbon cycle causes important feedbacks to the climate and is itself particularly susceptible to extreme climate events. There is general concern that climate change will have fundamental impacts on our ecosystems, and the future fate of European terrestrial biosphere's acting as a net carbon sink is highly uncertain.

Project Results:

2) Description of the main Science and Technology (S&T) results/foregrounds

2.1. WP1 Process study synthesis Lead: UA

Task 1.1 Assessing the role of soil mineralogy in the stabilization of SOC and the sensitivity to temperature and soil moisture extremes Lead: LUH

G.Guggenberger (Task leader), M.O. Gobel [LUH]

Soils were sampled at a permafrost study site at Cherskii with the aim to investigate whether, in addition to the soil environmental conditions (partly permafrost; low redox potential in the active layer and low temperatures in the lower part), the organic matter is also stabilized by mineral-organic associations. We observed that in all horizons 70-80% of the organic matter in these soils was associated with minerals, including cryoturbated horizons and soil from the permafrost itself. The chemical composition of the organic matter indicated that the mineral-bound soil organic matter was quite altered (i.e. formation of carboxylic groups and microbial components on the mineral surfaces).

Task 1.2 Assessing the potential of priming to reactivate sub-surface soil carbon Lead: INRA

S.Fontaine (task leader) [INRA]

80 cm-long intact soil cores were sampled from an upland grassland in central France (Theix, 45°43'N, 03°01'E). Four of these were sown with Festuca arundinacea at a density of 2000 seeds m-2 and four were kept bare for comparison. Plants received a double labelling (13C and 14C) for the whole duration of experiment (511 days). Living plants significantly accelerated the mineralization of SOM as compared to bare soil, and this throughout the duration of the experiment (p-value less than 0.05). The cumulative priming effect almost trippled the SOM mineralization rate in the bare soil. There was a significant interaction between planting and soil depth (p-value less than 0.01).

Task 1.3 Evaluating SOC turnover under different climate variability Lead: UA

I. Janssens (Task leader) and S. Vicca [UA]

The Arrhenius equation is commonly used to describe the relationship of the velocity of a reaction with temperature. This equation predicts an increase in organic matter mineralization with increasing temperature. It also predicts that the temperature response of organic matter mineralization increases with its recalcitrance. These predictions are supported by short term enzyme assays and by soil incubations, which show a positive relationship between temperatures, enzymatic activities and soil respiration up to an optimum temperature. However, soil warming field experiments have shown that the temperature-induced acceleration in soil respiration only occurs for few years, suggesting that mineralization of the large pool of recalcitrant C may not be accelerated by warming. To reconcile these differences, an alternative theory to explain the effect of temperature on velocity of enzymatic reactions and recalcitrant organic matter mineralization was developed by the INRA partner, based on results from focused experiments.

In this alternative theory, we take into account two key processes which are often overlooked while modeling the effect of temperature on enzymatic activities and organic matter mineralization:
1) Energy limitation of microbial decomposers: the availability of fresh energy-rich substrates supplied by autotrophs is low and limiting for decomposer community growth. Moreover, recalcitrant organic matter is too poor in available energy to sustain biological activity. Decomposers degrade recalcitrant organic matter through co-metabolism with fresh energy-rich substrates. This energy limitation of decomposers must be taken into account to predict the effect of temperature on enzymatic reactions.
2) Temperature-dependent inactivation of enzymes: enzymes undergo inactivation along the course of time due to loss of their three dimensional structure induced by Brownian movement. Temperature increases accelerate Brownian movement and thereby also the inactivation of enzymes. Ageing of enzymes is not considered when explaining the response of ecosystems to temperature changes.

2.2. WP2 Network of Ecosystem Manipulation Experiments Lead: ETH

Task 2.1 EME Access and Networking Lead: CNR-IBIMET

F. Miglietta (Task leader) [CNR-IBIMET] and the 12 PIs with associated staff of the individual EMEs

The aim of this task was to coordinate and access a network of 13 research sites with experimental manipulations across Europe. Twelve of the 13 sites had been continuing the prescribed measurements and data collection. One site was destroyed by a wind storm in 2009 and was discontinued (D. Loustau, INRA, France). The remaining sites comprised important land use types (agriculture, forest, heathland, peatland), were run by the respective PIs since quite some years and had been continued, typically with external funding. For each site and for each treatment, time-series of carbon pools and fluxes (soil, vegetation), environmental (meteorology, soil water and temperature) and ecosystem variables (LAI, standing biomass, species composition etc.) had been collected, old as well as new data were submitted to the central database and made available to modellers, particularly for model-data fusion exercises for selected model-site combination. Significant interactions were also developed between the CARBO-EXTREME EME sites and the infrastructure Project EXPEER (7 FP-EU).

Task 2.2 Cross-cutting activity carbon tracing Lead: ETH Zurich

N. Buchmann (Task leader) and S. Burri [ETH Zurich]; A Knohl [U. Göttingen]; M. Bahn and M. Schmitt [UIBK]

In this task, several pulse-labelling experiments with 13CO2 had been carried out at two grassland sites in Switzerland and Austria in 2010 and 2011 to investigate the transfer of newly assimilated carbon from above-ground to below-ground processes in response to summer drought, thus providing detailed info about the short-term responses to drought of the carbon cycle in agroecosystems at two altitudes. In addition, a methodology has been developed to trace carbon flow using mobile laser spectrometry. Allocation speed from the sward to belowground differed among the sites. Drought reduced the incorporation of recently fixed carbon into shoots and increased carbon allocation below-ground relative to total tracer uptake (Burri et al., submitted). Contrary to our hypothesis, we did not find a change in allocation speed in response to drought, although drought clearly reduced soil CO2 efflux rates. 19 days after pulse labeling, only about 60% of total tracer uptake was lost via soil CO2 efflux under drought compared to about 75% under control conditions. Predisposition of grassland by spring drought lead to different responses to summer drought in 2011 compared to 2010, suggesting increased sensitivity of grassland to consecutive drought events as predicted under future climate change (Burri et al., submitted).

Task 2.3 Cross-site meta-analysis from EMEs Lead: UA

I.Janssens (Task leader) and S. Vicca [UA] and the 12 PIs with associated staff of the individual EMEs

In order to reach the objectives of this task, all empirical data from the 12 manipulation experiments had been used. In addition, a large database from 65 additional experimental sites with climate manipulation experiments had been compiled and altogether used for meta-analysis. This data base contains (among others) measurements of biomass production, leaf physiology, vegetation characteristics, foliar and soil chemistry and soil respiration, as well as ancillary data such as soil moisture, soil texture and meteorological data for characterization of the climate. Because all manipulation experiments differed in climatic conditions, species composition, soil properties, et cetera, effects of climate manipulations (e.g. withholding 50% of summer precipitation) could not be analyzed in a conventional way. Instead, all experiments needed to be made comparable before being analyzed.

2.3 WP3 Long term Carbon Measurements and Climate Lead: CEA

Task 3.1 Tree ring data Lead: WSL

D. Frank (Task leader) and F. Babst [WSL]; T. Kun, B. Poulter, P. Ciais [ LSCE]; O.Bouriaud [ICAS]

The goal of this task was to analyze long-term records of annually-resolved radial growth to quantify the variability and sensitivity of key carbon cycle parameters to climate change. Additionally tree-ring data were employed to test, validate and improve carbon cycle models. As the basis for empirical analysis, efforts were directed at both harmonizing existing tree ring data and performing a targeted sampling at selected eddy-covariance sites throughout Europe.

Specific proposed tasks included:
-Extract a dataset of tree ring chronologies (quality, species, harmonization) from existing databases that can be used for carbon applications
-Organize a specific coring programme at eddy-covariance stations to obtain information about recent droughts and ecosystem processes, and relate them to the process monitored on those sites at fine temporal scale
-Derive estimates of stand level and regional woody NPP from tree ring data
-Evaluate how tree ring data compare with eddy-covariance GPP, biometric, remote sensing measurements
-Help to develop modules of tree growth into generic carbon models ; Carbon models evaluation and improvements using tree ring data

Climate sensitivity of forest growth across Europe

Investigations on the climate sensitivity of radial tree growth across Europe were accomplished by compiling and harmonizing a network of existing tree-ring chronologies across the European domain, and subsequently quantifying the climatic constraints on inter-annual growth. In total, datasets from 992 sites representing 36 different tree species across Europe were aggregated and analysed. All tree-ring datasets met certain quality thresholds (e.g. coverage of the 1920-1970 period, a minimum of 5 tree-ring series during this time period, etc.) prior to consistently removing the biological age-trend. Analysis with down-scaled climatology revealed
i) broad biogeographic patterns in the climate controls on growth such as the temperature limitations in Scandinavia and near tree-line zones in Central to Southern Europe,
ii) a high species-specific component in determining radial growth responses to climate variation and
iii) the importance of lagged effects of climate on radial growth (currently not considered in state-of-the-art dynamic global vegetation models).

Tree-ring and eddy-covariance estimates of carbon uptake and allocation

While certain aspects of the of the carbon cycle response to climate change and extremes were able to be quantified with existing data, most site locations were not representative for eddy-covariance stations and also did not allow the absolute quantities of carbon uptake to be estimated due to the lack of appropriate metadata (e.g. stand density, representativeness of sampled trees). Accordingly, CARBO-EXTREME efforts in this task were directed at devising and testing a protocol suitable for biomass estimation and more importantly applying this sampling scheme at eddy-covariance stations to link changes in above ground biomass estimated from tree-rings with alternative measures of ecosystem productivity. Sampling of all trees within plots of fixed size was conducted at 5 eddy-covariance sites throughout Europe balancing species diversity and geographical range with available project resources. For comparison with eddy covariance data and quantification of carbon uptake, tree-ring data were used to reconstruct historical tree diameters which were subsequently up-scaled to site level above ground woody biomass increments using allometric equations and stand measurements.

Testing and improving dynamic global vegetation models with tree-ring data.

The empirical tree-ring data provide insights into ecosystems processes and rates of relevance for current generation dynamic global vegetation models (DGVM’s) and thus future predictions. In task 3.1 we focused on using tree-ring data to test, validate, and improve the LPJ and ORCHIDEE-FM DGVMs representation of NPP sensitivity to temperature and precipitation.

Task 3.2 Tree growth and mortality Lead: UAH

M. A. Zavala (task leader), J. Madrigal-González, P. Ruiz-Benit [UAH];
C. Wirth, S. Rattclife [U Leipzig]; G. Kändler [Forest research Inst. Freiburg];
D. Coomes [U Cambridge]; A. Lehtonen [METLA]

This task has focused on the role of demographic processes -chiefly growth and mortality- on biomass changes at several scales. For this purpose we have compiled and harmonized forest inventory analyses and combine them with forest management records and dendro studies in some localities. We have also use process-based model (ORCHIDEE) for performing experiments to disentangle confounding factors.

Specifically :
1) At an European scale we have investigated how recent climatic changes interact with stand-level processes to modulate decadal biomass changes in the three main biomes (Boreal, Temperate and Mediterranean forests);
2) at a regional scale we have examined drivers of tree mortality along environmental and competition gradients in Iberian forests;
3) at a more local scale we have investigated drivers of forest decay in pine species at the rear edge of their distribution and specifically if carbon fertilization effects can counteract the detrimental effects of increasingly more pronounced droughts.

Recent climatic changes in Boreal, Temperate and Mediterranean forests

European forests have a prominent role in the global carbon cycle and an increase in carbon storage has been consistently reported during the 20th century. Further increase in forest productivity due to increases in temperature and CO2 concentrations, however, could be overwhelmed by increased climatic variability and climatic extremes. Research using plot-level forest inventory data was conducted to identify the relative importance of structural (stand basal area and mean d.b.h.) and environmental factors including mean climate (water availability and minimum temperatures) and recent climatic trends (temperature and precipitation trends in the last 20th century decade) on basal area change (proxy of biomass change). We focused on National Forest Inventories of Spain, Germany and Finland as they are representative of the principal biomes in Europe (e.g. Mediterranean, Temperate, Boreal). Using linear mixed-effects models we observed that stand structure, climate and recent climatic change strongly interact to modulate basal area change.

Patterns and drivers of tree mortality along competition and environmental gradients: implications for modeling changes in forest biomass.

Tree mortality is a key process driving stand dynamics and thus carbon storage dynamics. Hence, understanding major drivers of tree mortality might help us to model more realistically forest carbon dynamics particularly under climate change. Using repeat-measure information from c. 400,000 trees from the Spanish Forest Inventory, we quantified the relative importance of tree size, competition, climate and edaphic conditions on tree mortality of 11 species, and explored the combined effect of climate and competition. Tree mortality was affected by all of these multiple drivers, especially tree size and asymmetric competition, and strong interactions between climate and competition were found. All species showed L-shaped mortality patterns (i.e. showed decreasing mortality with tree size), but pines were more sensitive to asymmetric competition than broadleaved species. Among climatic variables, the negative effect of temperature on tree mortality was much larger than the effect of precipitation. Moreover, the effect of climate (mean annual temperature and annual precipitation) on tree mortality was aggravated at high competition levels for all species, but especially for broadleaved species. The significant interaction between climate and competition on tree mortality indicated that global change in Mediterranean regions, causing hotter and drier conditions and denser stands, could lead to profound effects on forest structure and composition. Therefore, evaluation of potential climatic change on tree mortality ought to consider stand structure, since two systems of similar composition but different structure could radically differ in their response to climate.

Competition and tree age modulated last century pine growth decline in response to high frequency of dry years in a water limited forest ecosystem

Extreme drought over mean temperature or precipitation trends a key factor of climate change affecting forest functions. In a study conducted in inland-dune pinelands in the Iberian Peninsula we examined last century growth trends using decadal direct measurements of wood volume increments recorded in harvested trees belonging to a chronosequence of forest stands. As relevant climatic predictors we considered mean temperature, mean precipitation and three drought indices based on the Standardized Precipitation and Evapotranspiration Index (SPEI) (i.e. mean water balance, frequency of dry years and the most severe drought event in the decade). Results supported that the final outcome of climate change on forest productivity would critically depend more on unusually high frequency of dry years than mean temperature, precipitation or water balance. However this trend can greatly vary with population-level structural factors such as population age distribution and competition, i.e. negative impacts of increasing drought frequency are disproportionally higher in younger trees under high competition.

Disentangling the role of carbon fertilization in water-limited forest ecosystems: confronting long-term observational data and process-based model outputs (ORCHIDEE)

The benefits of living in an enriched CO2 atmosphere can counterbalance the negative impacts of drought and climate change. In particular it is unclear whether the benefits of carbon fertilization in an enriched CO2 atmosphere can offset the costs associated with increasing water stress. Rising concentration of atmospheric carbon might, to a certain extent, reverse the negative impacts of altered water deficits on tree growth by means of enhanced water use efficiency. Long-term observations are scarce in long lived forests species and experiments are hampered by the number of processes involved. In this study we examine the main drivers underlying last century carbon storage in a water-limited Mediterranean ecosystem.

Extreme droughts and forest decay in pine species at the rear edge of their distribution

For Eurosiberian pine species -at the rear edge of their distribution- these detrimental effects can be more patent and extreme droughts can trigger a sequence of chained events leading to forest decay. These processes can be aggravated in high density forest plantations in which thinning processes have not been implemented. We used combined dendro measurements and forest inventory data to examine the relative role of structural and environmental gradients in replanted pine forests in a drought prone area of Southern Spain. As hypothesized, extreme droughts in terms of intensity and length interact with tree density to trigger a forest decay process in these plantations.

Task 3.3 Crop yield variations and climate Lead: UNIABDN

M. Wattenbach, P. Smith [UNIABDN]; X. Wu, M. Mahecha, M. Reichstein [ MPI BGC Jena]

This task focused on yield as records of the impact of climate variability and extremes on cultivated ecosystems. The yield and quality of food crops is central to the well-being of humans and is directly affected by climate and weather. Early studies of climate change on crops mainly focussed on effects of increased carbon dioxide (CO2) level and/or global mean temperature and/or rainfall and nutrition on crop production. However, crops can respond nonlinearly to changes in their growing conditions, exhibit threshold responses and are subject to combinations of stress factors that affect their growth, development and consequently yield. In this context climate variability and changes in the frequency of extreme events are especially important (Porter and Semenov 2008).

Disaggregate national (FAO) and regional (EUROSTAT-NUTS3) yield statistics over the past 30 yrs to produce maps of yield anomalies

In a first step we analysed the response of crops to changes in climatic conditions using a bioclimatic envelope approach to identify potential limitations for a particular crop (Tuck et al., 2006). In the analysis we used the bioclimatic approach in combination with crop yield statistics from Eurostat (EUROSTAT 2010) in order to distinguish between weather- and climate-driven changes in crop yield from management effects. We identified events around 2003, 2006, 1980 and 1959 to be crop yield anomalies that could also be captured by a simple bioclimatic envelope approach identifying them as climatically driven, rather than a result of changes in management caused by, for example, agricultural production subsidies. In contrast, trends in crop yield seem to be mainly driven by improvements in production methods, as the trend could not be captured by the bioclimatic envelope. Identifying the year 2003 as an anomaly for the two central European countries, Germany and France, and not for Spain and Denmark corresponds well with the findings of Ciais et al. (2005). Even though this is only a preliminary analysis, it already identifies other potentially important time periods for a more detailed analysis.

Compare patterns of regional yield anomalies with remote sensing fields

In the follow on activities these data where used to constructed the gridded (0.5 × 0.5 degree) average crop yield data for Barley, Wheat, Grain maize and Potatoes for the period 1975-2009. (CX Eurostat data). Data were compared monthly GIMMS NDVI and up-scaling GPP (0.5 × 0.5 degree) during 1982-2008. Trends in gridded crop yield, NDVI and GPP are subtracted by fitting a cubic smoothing spline.

Test crop models against these data

In this sub-task, we compared the yield and bioclimatic envelope data to the biogeochemical process model simulation such as ORCHIDEE-STICS for the European scale allows us to evaluate if the model is able to capture the dynamics of climate driven yield fluctuations and anomalies. ORCHIDEE-STICS is a coupled (De Noblet-Ducoudre et al., 2004; Gervois et al., 2008) consisting of a dynamic global vegetation model ORCHIDEE (Krinner et al., 2005), and a process-oriented crop model STICS (Brisson et al., 2003). In the conclusion the pattern seen in the wavelets when comparing ORCHIDEE-STICS with yield and bioclimatic envelope in France show few similarities. However, there is a pattern in the period of 1985-1995 in ORCHIDEE and yields that are overlapping. There is no common pattern for between the model and the bioclimatic envelope. The same holds for the comparison of wheat yield, bioclimatic envelope and ORCHIDEE-STICS.

Task 3.4 Analysis of long term ecosystem CO2 fluxes Lead: UNITUS

D. Papale (Task leader) [UNITUS]

The objective of this task consisted in the analysis of ecosystem level carbon fluxes and major pools of European sites being operative for at least 7 years among the CarboEurope eddy covariance sites network. In particular the task addressed
1) the evaluation of the response of CO2 fluxes to climate anomalies and extremes including lag effects and
2) the generalization of the response of fluxes to climate anomalies for different ecosystem types and climate regions which can serve as a ground for modeling evaluation and parameterization.

Evaluation of the response of CO2 fluxes to climate anomalies and extremes including lag effects

Sites being part of the Carbo-Europe IP database were firstly ranked by the length of the available records. The 'long record' sites initiated eddy covariance flux monitoring in 1996 and have been thus continuously operative and delivering data for more than 16 years. A second criterion applied was the existence of ancillary information and finally the representativeness of the sites for different climate and vegetation types. Only sites located in forest ecosystems were selected, since the eddy covariance towers measuring in non-forest ecosystems have been generally installed only in more recent years. The data have been processed according with the standard methodologies and additional tools developed in the context of WP4 in order to make the data comparable and uncertainty estimated. The data have been then analyzed in order to identify effect of climate extremes on fluxes. The first result obtained has been the finding that others processes, in particular ecosystem disturbances, management practices and others confounding factors can have an effect of the fluxes higher than the climate extremes in particular in a relatively short time period like the one analyzed.

Generalization of the response of fluxes to climate anomalies for different ecosystem types and climate regions

Analyzing the role of the main drivers we focused on the impact of air temperature (Tair) and water deficit anomalies on the variability of flux at daily resolution as well as the changes in physiological parameters of photosynthesis and ecosystem respiration (a-quantum yield; ß-max. photosynthetic capacity; rb-basal respiration at 15°C) retrieved by a semi-empirical model parameterized at site level. The variability of daily CO2 flux (NEE, GPP, Reco) has been analyzed in respect with drivers anomalies for different seasons and different Plant Functional Types. In particular the data have been binned into 2.5 percentile classes of the driver anomaly respect to the mean long term average (1989-2010 reconstruction from ECMWF) and regressive analysis performed respect to the anomalies in the three fluxes components. Results have been presented at the final CARBO-EXTREME conference and here below is reported one example. It is possible to see how in general a positive anomaly of the air temperature in March-May has a positive effect on GPP (increase of GPP) at all the sites except for the Mediterranean sites where there is no clear correlation.

Task 3.5 Long term remote sensing analysis Lead: VUA

R. de Jeu (task leader) [VUA]; F. Maignan, F.M. Bréon [LSCE]

Long term soil moisture dataset from passive and active microwave observations

A methodology to develop a long term consistent soil moisture database from multisensor satellite observations has been developed. The data is now free available at (see http://www.esa-soilmoisture-cci.org online). The complete methodology is described by Liu et al., (2011, 2012). The data was released in June 2012 and downloaded by more than 600 different organizations within one year.

Comparison with a global carbon model

The satellite derived soil moisture datasets were compared to soil moisture from the process-based vegetation model ORCHIDEE during the period 2003-2004. We found that the soil moisture products from the satellite and ORCHIDEE correlate well, in particular when considering the root zone soil moisture of ORCHIDEE. However, the root zone soil moisture in ORCHIDEE consistently overestimated the temporal autocorrelation relative to the satellite and in situ measurements. This may be due to the different vertical depth of the two products, to the uncertainty in precipitation forcing in ORCHIDEE, and to the fact that the structure of ORCHIDEE consisting of a single-layer deep soil, does not allow simulation of the proper cascade of time scales that characterize soil drying after each rain event. We concluded that assimilating soil moisture in ORCHIDEE using AMSR-E with the current hydrological model may significantly improve the soil moisture dynamics in ORCHIDEE. The results from this study are published in Rebel et al., 2012.

Soil moisture and vegetation dynamics

The impact of water availability on vegetation was assessed. Two studies were conducted. One analyzed the relationship between the drought index (SPEI) and Net Primary Production (NPP) at a global scale. The other study analyzed the spatiotemporal relationship between the NDVI and satellite soil moisture for Continental Australia.

MODIS and AVHRR NDVI products

We provided netcdf files over Europe, both maps and time series.

MODIS NDVI

We provided directionally corrected NDVI MODIS/Terra daily data, at the Climate Modeling Grid resolution (CMG, 5 km, 0.05°), for the 2000-2008 period. The correction follows the algorithm described in Vermote et al. (2009). Over the European PFTs the noise is estimated to be less than 0.02.

AVHRR NDVI

We used the Land Long Term Data Record LTDR version 3, which includes BRDF effects correction, cloud mask improvement and a better orbital drift correction, for the 1981-2000 period. The data are given on the same CMG grid as MODIS products. We also provided the Quality Assessment field delivered in the LTDR files (see details in http://ltdr.nascom.nasa.gov/ltdr/docs/LTDRDataFormatDescriptions_03_2010_30.pdf ).

NDVI data have been filtered against:
QA bit number 1: Pixel is cloudy
QA bit number 2: Pixel contains cloud shadow
QA bit number 3: Pixel is over water
QA bit number 6: Pixel is at night (high solar zenith angle)

Task 3.6 Soil carbon regional datasets Lead: CU

G. Kirk, P Bellamy [CU]; B. Guenet, F. Moyano, P. Ciais [LSCE]

The main objective of this task was to understand the contribution of climate change and variability vs other drivers on the long term soil C decrease measured by the gridded soil carbon inventory of England and Wales

Modelling the contribution of climate change and variability to the observed long term soil C decrease over England and Wales

Widespread decreases in soil carbon were observed in England and Wales between 1980 and 2000 in the National Soil Inventory (NSI) (Bellamy et al., 2005). Subsequent studies with a simple single pool soil organic matter (SOM) model, which quantified changes in carbon inputs to the soil and rates of decomposition, showed that the changes were largely due to changes in land use or management (Kirk and Bellamy, 2010). In work in Task 3.6 we used the DAYCENT model, initialised in various ways, to study the contribution of climate variation to the observed soil carbon dynamics in England and Wales (Foereid et al., 2012).

Task 3.7. DOC export by streams caused by extreme rainfall events in Scotland and Sweden Lead: CEH

K. Dinsmore, M. Billett [task leader, NERC, CEH Edinburgh], N. Avrosiadi, T. Grabs, K. Bishop [SLU]

Multisite analysis of the role of high flow extremes on aquatic DOC export.

Over the last 12 months we have combined UK and Swedish datasets to carry out a multisite analysis of the role of high flow extremes on aquatic DOC export. This involved 7 UK and 6 Swedish catchments with weekly (UK) or fortnightly (Sweden) DOC concentration data and either continuous (UK) or daily (Sweden) measurements of discharge. Measurement periods range from 5-29 years. Catchments cover a range of runoff regimes and include the presence/absence of spring flood due to snow melt.

Task 3.8 Synthesis of multiple gridded datasets and model results at continental scale Lead: LSCE

T. Kato, P. Peylin, P. Ciais [LSCE]; M. Jung, C. Beer, M. Mahecha [MPI-Jena]

The goal of the synthesis task is to organize and synthesize the 'data mining' and data analysis program of the WP, and to interface with the model developments in WP5 for ensuring an optimal use of all data.

Database status

Data developed from each task are available via the CARBO-EXTREME database directly or via request. The data exchange portal (mainly hosting large spatiotemporal long-term observations and model results) is also online via the link http://www.bgc-jena.mpg.de/geodb/CARBO-EXTREME/Home.php

Synthesis study among models simulated NEE over European continent

We established analysis on consistency between bottom-up and top-down estimates of the CO2 budget of Europe. Estimation of the regional carbon budget has been an important research topic since atmospheric CO2 concentration was started to be measured in 1950s (Keeling, 1960). Consequently, in several recent decades, two different types of model have been developed intensively for estimating regional terrestrial CO2 budgets: Terrestrial Biosphere Model (TBM) and Atmospheric Inversion model (AI), as bottom-up and top-down approaches, respectively.

Synthesis study on regional hot spots / processes of interannual variability of European carbon balance

We aim to quantify year-to-year changes in the dynamics of CO2 fluxes (i.e. net ecosystem exchange, NEE and ancillary variables) contributing to a comprehensive description of the variability of the terrestrial carbon cycle over Europe based on observational evidence. The analysis is based on two methodologically independent upscaling products which represent 22 years of quasi-observational GPP (and NEE) estimates over Europe, and a novel concept for evaluating interannual variability developed in CARBO-EXTREME the so called 'year-to-year differences' (IADs). A manuscript will be submitted.

2.4. WP4 Data infrastructure and model support Lead: UNITUS

D. Papale [UNITUS] and I. Janssens [UA] (Task leaders)]

The database and data processing activity in the project has been focused on the collection, harmonization, processing and distribution of data measured in the context of CARBO-EXTREME but also by groups that although not directly involved in the project wanted to share their measurements and data. This aspect highlights the attractiveness of the scientific activities planned and also the maturity of the scientific community. For this reason the first task for the project database has been to propose, discuss and approve a common data sharing and data use policy that has been accepted by all the participants and contributors and that has been used also as example by others projects and initiatives.

2.5.WP5 Model Data Integration Lead: MPG
Task 5.1: Process-model evaluation and structural development Lead: MPG

task leader, N.Carvalhais [MPG]; M. Kuhnert, P. Smith [UNIABDN]; R. Lardy [INRA]; S. Rolinski [PIK]; M. van Oijen [CEH]

In CARBO-EXTREME, the one-dimensional sectorial models PaSim (grassland model), Daily DayCent (cropland model) and BASFOR (forest model) as well as the one-dimensional generic models JSBACH, ORCHIDEE and LPJmL have been run using forcing data at site level. The comparison to data from ecosystem manipulation experiments and long-term observations as well as more formal model-data fusion exercises (Task 5.2) allowed a detailed evaluation of the models. This exercise was also attractive and interesting for other modelling groups which led to a very successful additional application of the forest model CASTANEA and the Coup model.

Task 5.2 Bayesian model calibration and uncertainty propagation Lead: CEH

M.van Oijen (task leader) [CEH]; N. Carvalhais, C. Beer, M. Mahecha, M. Reichstein, G. Schuermann, S. Zaehle [MPG];
P. Peylin, N. MacBean [LSCE]); T. Thum, [LSCE, now at FMI]; A. Rammig, S. Rolinski [PIK]; D. Papale [UNITUS];
A. Granier, D. Loustau [INRA]; T. Keenan, A. Richardson [(HARVARD]
During the reporting period, extensive model-data fusion exercises have been performed to improve sectorial and generic models. In general, there are two main types of data which have been used successfully: Results from ecosystem manipulation experiments and long-term observations.

Task 5.3 Spatio-temporal multi-scale integration and consistency check Lead: CEA LSCE

P.Peylin (task leader), T. Kato, P.Peylin P.Ciais [LSCE];

The consistency of bottom-up and top-down estimates of the CO2 budget of Europe was assessed using several prognostic and diagnostic terrestrial biosphere models (TBM) as well as atmospheric carbon dioxide inversion schemes (AI). The intercomparison was performed in terms of interannual changes in NEE over Europe. (cf. Deliverable 5.8).

2.6. WP6 Scenario Analysis Lead: UNIABDN

The objective of work package 6 is to provide spatially C flux scenarios based on projections of the European terrestrial carbon cycle and associated uncertainties by applying the improved and calibrated models of WP5 driven by regional climate scenarios, to feed in to the vulnerability synthesis in WP7. Eight different model are used for this approach, the four cross-sectoral models CLM (Community Land Model version 4, Bonan et al. 2003), ORCHIDEE (Krinner et al. 2005), JSBACH (Raddatz et al. 2007), LPJmL (Bondeau et al. 2007) as well as the four sectoral models PaSim (Riedo et al. 1998; Graux et al. 2012), BASFOR (Van Oijen et al. 2005), EPIC (Van der Velde et al. 2012; Balkovic et al., accepted), DAILYDAYCENT (Parton et al. 1998). The sectoral models consider only one of the three land use categories forest (BASFOR), grassland (PaSim) and cropland (EPIC, DAILYDAYCENT). Parameters of most of these models were optimized using experimental data and long-term observations in work package 5.

Task 6.1: Collation of fundamental spatial data sets including regional climate scenarios Lead: ETH Zurich

S.I. Seneviratne (task leader) [ETH Zurich]; C.Beer [MPI-Jena]

For the spatial simulations, the models needed spatial distributed input data that were collected from different projects, which was the aim of Task 6.1. Several climate data sets were collected for the different time periods, as summarized in deliverable 6.2 but finally four data sets (one for 1901-2010 and three projections until 2100) were used for the simulation runs: WATCH forcing data was used for the period 1901-1978, bias-corrected ERA-Interim data was used for the period 1979-2010, and the three projections for the period 2011-2100 assuming the IPCC A1B emission scenario were based on ENSEMBLES outcomes from
1) MPI-REMO,
2) KNMI-RACMO, and
3) DMI-HIRAM.

Task 6.2: Offline model factorial experiments to identify critical thresholds for assessing vulnerability Lead: UNIABDN

P. Smith (task leader) and M. Kuhnert [UNIABDN]; C.Beer [MPI-Jena]

The objective of task 6.2 was to identify critical thresholds for assessing vulnerability of ecosystems. Therefore, the spatial simulations were set up in two experiments to detect impacts of gradual climate change and differences in climate variability on land carbon dynamics. The models were run with the two versions of climate data, CNTL and REDVAR. For the European land surface, ORCHIDEE and LPJmL show increasing stocks until 2100, i.e. the land acts as a C sink. Both models also agree on a stronger sink activity in case of lower variability. The model JSBACH, however, shows a change of dynamics around 2050, suggesting a tipping point after which the land surface stops acting as a C sink and becomes a C source with stronger effects for the lower (REDVAR) variability. Due to the representation of fire activity in LPJmL and dynamic vegetation, this model shows stronger oscillations. Interestingly, this model also predicts a carbon sink in the future. Model differences in 2100 are larger than differences between CNTL and REDVAR forcing or RACMO versus REMO forcing, suggesting less agreement among the models than is seen in the forcing data.

Task 6.3: Offline model ensemble runs with climate scenarios to identify where and when critical thresholds are likely to be reached in the future Lead: UNIABDN

P. Smith (task leader) and M. Kuhnert [UNIABDN]; C. Beer [MPG];

The objective of task 6.3 was to identify where and when critical thresholds are likely to be reached in the future. As mentioned above there are no fixed values that define a threshold of vulnerability, but NEE is a good indicator for climate impacts on the ecosystem. BASFOR is a forest model and considers only spruce in these simulations which suggest that spruce ecosystems will act as a carbon source in future. Accordingly, the negative impacts on spruce may force a change in the species selected for planting on areas currently planted with spruce. The crop model EPIC shows a strong carbon sink in all cases. All sectorial models show a higher sink with lower variability.

Task 6.4: Erosion synthesis modelling to examine the impact of future climate variability / extremes on the vulnerability of soils to erosion Lead: IIASA

M.Obersteiner (task leader) and M.van der Velde [IIASA];

EPIC is the only model that considers erosion in its simulations, and results indicate, that overall, the absolute carbon stored in cropland tends to increase due to higher biomass production with climate change. However, soil organic matter stocks in arable land tend to decrease, due to the higher erosion rates resulting from increased climate variability, and a combination of impacts on biomass production and mineralization rates. Importantly, the regional reduction in soil organic carbon due to climate variability can off-set the gain in carbon stocks resulting from the long term trend in climate alone.

2.7. WP7 Carbon Vulnerability Synthesis Lead: INRA

After publishing (Van Oijen et al. 2013) the probabilistic vulnerability assessment method developed by CarboExtreme (see D7.2) an analysis of the vulnerability of the carbon cycle to climate change and extreme events by sector and ecosystem type had been developed. This has been achieved through projections at the EU scale which combine environment distributions (exposure) and ecosystem response distributions (response) conditional to a given level of environmental exposure. In this way, the direct analysis includes both the effects of variability (in climate) and of uncertainty (in model parameters and structure). An inverse approach has also been developed to address vulnerability and risk in terms of climatic drivers leading to hazardous ecosystem responses. The probabilistic risk assessment has been applied to climate projections (regional climate model forced by the A1B scenario) based on an ensemble of three generic ecosystem models and of 5 sectoral (forests, grasslands and annual crops) models.

Task 7.1 Conditional probabilities of % change in continental scale annual C sink
Lead: UNIABDN

P.Smith (task leader) and M.Kuhnert [UNIABDN]; M. Wattenbach [GFZ Potsdam]

The probability analysis based on the simulation results of work package 6 and considers the results of 8 different ecosystem models that can be divided in four cross-sectoral models (JSBACH, CLM, LPJmL and ORCHIDEE), which consider all land use types, and four sectoral models that only consider forest (BASFOR), grasslands (PaSim) or croplands (EPIC, DAILYDAYCENT). The analysis based on two approaches considering, first, the physical based environmental variable as threshold for a hazardous event (Oijen et al. 2013), second, the extreme impact defined by a biological variable that shows the response of the ecosystem (Rolinski et al., in preparation). In the first approach (from now on called the direct method), the drought impact is defined by the SPEI index (Vicente-Serrano et al., 2010), which is calculated from the difference between rainfall and potential evapotranspiration.

Task 7.2 Continental scale carbon vulnerability indicator Lead: INRA

M. van Oijen (task leader) [NERC-CEH]; JF Soussana [INRA];
A.Rammig S.Rollinski K.Thonicke [PIK]

Two complementary and consistent indicators have been developed. First, the direct probability risk assessment (van Oijen et al. 2013) has provided a first indicator which is the carbon loss by European ecosystems (in TgC) resulting from climate extremes corresponding to droughts (SPEI values below one). Second, the inverse approach has defined threshold values of the SPEI indicator that would turn the European carbon sink into a source. These two indicators are consistent and have been calculated for the European continent and for three SREX regions (Northern Europe, Central Europe and Mediterranean), for natural vegetation and for sectors (arable crop species, grasslands and spruce forest) (see D7.4).

Task 7.3 Recommendation of adaptation and mitigation strategies Lead: PIK

K.Thonicke (task leader), W.Cramer A.Rammig S.Rollinski [PIK]; ]

Two vulnerability approaches, developed under Task 7.2 (direct approach, van Oijen et al. (2013) and the inverse vulnerability approach, Rolinski et al. in prep), were applied to simulation results from large-scale ecosystem models. These models were run for future climate change scenarios using 4 general ecosystem and 3 sectoral models from WP6. Both vulnerability approaches were used to identify the vulnerability and risk to drought under climate change. Different aspects of climate change, building on the design of factorial experiments in WP6, namely the influence of climate variability and CO2 fertilization, were investigated in this task as well to calculate the respective vulnerability on natural vegetation, forests, grasslands and crops. Results were analyzed for the entire European simulation domain and additionally aggregated for 3 SREX regions as well as for entire Europe.

2.8 WP8 Policy Interaction, dissemination Lead : PIK

Task 8.1 Expert elicitation of risks and vulnerabilities Lead: PIK

K.Thonicke (task leader), A.Walz [PIK/Univ. Potsdam], W.Cramer [PIK/IMBE],
D.Frank M.Reichstein [MPG], P.Smith [UNIABDN], JF Soussana [INRA]
The most urgent risks and vulnerability of the Carbon Cycle to extreme meteorological events have been collected, ranked and discussed among senior CARBO-EXTREME researchers and selected external experts invited to a first workshop on April 15/16, 2010. The invited experts included stakeholders and decision-makers in carbon management from seven European countries as well as European Environmental Agency and European Commission.

Task 8.2 Review of project targets in response to stakeholders Lead: PIK

K.Thonicke (task leader), A.Walz [PIK/Univ. Potsdam], W.Cramer [PIK/IMBE],
D. Frank, M. Reichstein [MPG], M. Kuhnert [UNIABDN], D. Frank [WSL], A. Ibrom [DTU]

During the workshop on March 07/08, 2013, we had the chance to present and review the key findings and questions of the CARBO-EXTREME project with you as decision-makers in or knowledge brokers for Carbon Management. The composition of the group covered a wide range of decision-making levels, i.e. from local to global, and a good coverage of Europe, with the exception of Mediterranean regions.

Task 8.3 Dissemination, contributions for workshops, briefings; brochure and flyers, summary for policy makers and web-page Lead: MPG

M.Reichstein (task leader), D.Frank [MPG]; C.Bounama K.Thonicke A.Walz [PIK/Univ. Potsdam]

CARBO-EXTREME communicated obtained scientific knowledge concerning the research topic of the carbon cycle under climate variability and extreme events as well as project activities through:
-over 70 publications (peer-reviewed and non-peer reviewed)
(see http://www.CARBO-EXTREME.eu/index.php/Publications/Project online)
-participation of CARBO-EXTREME researchers in numerous (inter-) national conferences [e.g. the Planet under Pressure Conference London (March 2012), EGU General Assembly Meetings (Vienna, 2011, 2012, 2013), AGU Fall Meeting 2012 (San Francisco), Open Science Conference on Climate Extremes and the Biogeochemical Cycles (Seefeld, April 2013, http://www.bgc-extremes2013.org) and the upcoming 9th International Carbon Dioxide Conference, Beijing, China (June 2013)] - see also 'List of Dissemination Activities'
-via the continuously updated project webpage http://www.carbo-extreme.eu
-project flyer (download via http://www. carbo-extreme.eu)
--the compilation of the final results in a summary for policy makers
-the project brochure
-via several press releases and information about CARBO-EXTREME reported in different media (see http://www.CARBO-EXTREME.eu/index.php/Press/Press online).
These activities were conducted in close collaboration with WP9.

Potential Impact:

Improved understanding of European ecosystems is essential to estimate potential release of carbon from these ecosystems into the atmosphere, and thus their potential further enhancement of climate change. Building on important research on the effect of increasing mean temperatures and changing mean precipitation on ecosystems and their carbon cycle, the effects of an increase in inter-annual variability has much been debated over the past years. CARBO-EXTREME has taken up the discussion and aimed to enhance our understanding of the impact of increasing magnitude and frequency of meteorological extreme events on European ecosystems and their carbon cycle. Bringing together different methodological communities, including experimental ecology, long-term observations and modelling, innovative forms of collaborations could be initiated and new insights obtained over the course of the project.

Participation of CARBO-EXTREME researchers in policy relevant boards:

Philippe Ciais (beneficiary 2 CEA) is co-chair of the Global Carbon Project and the Climate KIC GHG platform. The annual production of the anthropogenic CO2 budget and initial work to produce regional budgets in the RECCAP project are relevant regarding CARBO-EXTREME work because extreme years are correlated with CO2 growth rate anomalies, and the connection between climate and CO2 anomalies is not fully understood.

Nina Buchmann (beneficiary 4 ETH Zurich) is member of the Landwirtschaftlicher Forschungsrat (Agricultural Research Council) of Switzerland (since 2008), member of the Board of Trustees (Kuratorium) of the Öko-Institut e.V. (since 2009) and member of the Steering Committee of the Swiss research programm 'Research for Development (r4d) ', jointly organized by the SNF and the SDC (since 2013).

The 'Open Science Conference on Climate Extremes and the Biogeochemical Cycles' (see http://www.bgc-extremes2013.org online), a joint initiative of CARBO-EXTREME, the US-based network INTERFACE, and the international activity iLEAPS funded by the International Geosphere-Biosphere Program, with over 150 researchers from Europe, Asia and the US, together with the related Nature article 'Wild weather can send greenhouse gases spiralling' (Nature 496,147 doi:10.1038/496147a) from 11th April 2013 contributed substantially to spread the knowledge about climate variability and the carbon cycle and CARBO-EXTREME in the research community.

CARBO-EXTREME researchers participated in numerous inter-)national conferences with talks and posters, e.g. the Planet under Pressure Conference London (March 2012), EGU General Assembly Meetings (Vienna, 2011, 2012, 2013), AGU Fall Meeting 2012 (San Francisco), Open Science Conference on Climate Extremes and the Biogeochemical Cycles (Seefeld, April 2013, http://www.bgc-extremes2013.org) and the upcoming 9th International Carbon Dioxide Conference, Beijing, China (June 2013) [see also 'List of Dissemination Activities'].

Further dissemination of CARBO-EXTREME knowledge were e.g. Open Days at "Long Night of Science" events, the DAAD/BMBF science tour "Understanding Biodiversity and Climate Change" in November 2012, or the use of the project brochure synthesizing project results in Ecology lectures at UAH.

CARBO-EXTREME researchers were also strongly committed to dissemination within the research community. Around 60 peer reviewed scientific publications related to CARBO-EXTREME published until 31.May 2013 by CARBO-EXTREME researchers - 12 of them in high profile journals like Global Change Biology, Nature, Nature Geoscience, New Phytologist, Science (i.e. journals with a Journal Impact Factor 2011 by Thompson Reuters of over 6) - contributed to make the project results highly visible in the research community (see 'Publications' at http://www.CARBO-EXTREME.eu).

Currently (status May 2013), from the 60 peer reviewed CX publications 30% are Open Access (golden way); 8% are published in delayed Open Access journals, the rest is published in closed access journals. As CARBO-EXTREME wants to strengthen Open Access publishing within the project, we will try to make those articles available via the 'Open Access green way': we will contact the respective authors and if legally possible we will create postprints and publish them via the Max-Planck institutional repository linked via the CARBO-EXTREME web page.

In addition a special issue of Biogeosciences on 'Climate extremes and biogeochemical cycles in the terrestrial biosphere: impacts and feedbacks across scales' is planned (see http://www.biogeosciences.net/submission/scheduled_special_issues.html#36 online). The special issue will address the relations between climate extremes and biogeochemical cycling in terrestrial ecosystem, emerging from the global conference on this topic (see http://www.bgc-extremes2013.org online). Its guest editors are amongst others the CARBO-EXTREME members Markus Reichstein and Michael Bahn, and CARBO-EXTREME advisory board member Yiqi Luo.

Contributions to future projects:

CARBO-EXTREME has been contributing substantially to the launch of the AnaEE Project (Infrastructure for Analysis and Experimentation on Ecosystems) which is part of the European Strategy Forum on Research Infrastructures Roadmap (ESFRI) as a new distributed EU infrastructure for ecosystem research. AnaEE entered the Preparatory Phase on November 2012 (3.5 years) and the Construction Phase will commence in 2014.

We see further impact potential as followed:
Exploitation potential 1: The developed concept of probabilistic vulnerability and risk indices was considered a promising communication support by decision- and policy makers. This concept including both a climate as well as a response based hazard definition has great potential to be further exploited and tested by administration of European and national levels. The proposed indices were highly appreciated and well understood, despite their high level of complexity. Decision makers understood immediately the potential of the two complementing approaches of a climate as well as a response based hazard definition, and alternative hazard definitions were discussed in more depth. The one drawback identified for the concept, however, are the high data requirements, which in practical terms allows only using the concepts on modelled data. This, again, is not considered a problem for higher level decision-makers. Future potential to exploit the concept include the discourse-based identification of thresholds for the hazard definition and of a variety of possibly critical response variables. Then the concept could be used to estimate the effect on vulnerability and risk of a variety of measures to protect terrestrial carbon sinks in soils and living biomass under increasing magnitude and frequency of meteorological extreme events.

Exploitation potential 2: CARBO-EXTREME modelling results in conjunction with the developed vulnerability and risk indicators will help to identify hotspots of climate vulnerability against future climate variability for both managed and unmanaged ecosystems in Europe. For agricultural systems, these results indicate not only hotspots of vulnerability, but also priority regions for technical and management based adaptation. Although the range of adaptation options is smaller and turn-over phases much longer for forest ecosystems, the same applies to managed forests. In contrast to the managed systems where priority regions for societal adaptation are indicated, the hotspots of vulnerability in unmanaged ecosystems will reveal areas of high probability of profound alteration within the natural ecosystems with implications on nature conservation concepts.

Exploitation potential 3: Despite some reservations from both sides, the active dialogue between the scientific community and decision-makers has large potential for added-value. Better understanding the physiology of extreme events, namely drought and heat waves, is a major concern to the decision-makers from European administration to local land manager. Based on the experience from our second expert workshop in March 2013, we can - once more - underline the importance of such an exchange for two main reasons, the information exchange itself and the identification of knowledge gaps important to decision-makers. In the case of CARBO-EXTREME, the information generated within the project was of great interest in particular for higher level decision-makers, namely national and EU level where mitigation effects of ecological carbon storage are relevant issues. The second dimension of the same dialogue is the discussion of what both communities do not understand, but what would be of great interest to understand. Contradictory results, possibly even from a variety of data sources and methodological backgrounds, prove here as highly stimulating and enhance the exchange profoundly. This is the honest and unpretentious discussion required to identify common research questions of high relevance for society which can be addressed by innovative scientific settings.

Exploitation potential 4: The added value in better understanding the extreme ecosystem responses, instead of sticking to the extreme meteorological events, is highly appreciated by decision-makers. To identify the circumstances under which extreme ecosystem response occur along the cascade from physiological reaction of single plants to biome shift, should be a priority in ecological research. Here, the collaborative setting of CARBO-EXTREME combining a wide variety of methodological communities is seen as a must to enhance process understanding. So far, short-term extreme responses on the level of physiological reaction and species composition have been investigated on experimental sites. However, a) the results from these sites are contradictory, and b) once experiments had been finished, the ecosystems quickly recovered, so profound ecosystem change has hardly ever been observed within an experimental setting. A coordinated network of long-term experimental sites, possibly funded by diverse national agencies across Europe, is considered an important milestone towards producing necessary process understanding, also amongst decision-makers. In addition to long-term experiments, the need for coordinated efforts is well recognised by scientists as well as decision-makers. Great added-value could be generated in bringing comparable data from different experimental sites and publications together. Standardised methods, measures and metadata on experimental settings are required to enhance comparability and enable such joint analysis of large samples. Such standardisation is not typical for the scientific community, where originality of methodologies and the singularity are important assets. Still, there have been positive examples of similar standardisation efforts recently, such as GEO-BON and EU-BON activities aiming for standard inputs of biodiversity observations, or the highly valuable TRY database where numerous scientists feed data on plant functional traits into following a standard protocol.

Exploitation potential 5: A strong potential is still be exploited in further calibrating modelling approaches with empirical data to increase the prognostic power of scientific ecosystem models. Thanks to the richness and variety of data sources being collected and combined in CARBO-EXTREME, methods could be developed and tested to adapt parameterisation of the existing models to observed data. This data infusion technique has so far been applied to single points in space and has proven successful for grasslands. However, we are still facing big challenges a) for European forests, where management shows a large impact on the carbon cycle; and b) in the calibration across larger areas, if not all European terrestrial ecosystems. Overcoming both these challenges promises a strong rise in the expected accuracy of simulated vulnerability of the terrestrial carbon cycles and potential losses of sequestrated carbon.

List of Websites:
http://www.carbo-extreme.eu

Related documents