The concept of CC-TAME is to model explicit land use on farm/forest management practice level taking into account the emerging technological changes in the land-use sector and its associated industries. CC-TAME will combine regional climate models with biophysical ecosystem models, which are rich in technology representation, with state of the art bottom-up type economic sector models embedded in the theory of modern welfare economics. A technologically explicit bottom-up approach on the farm/forest management practice level to full fledged sector analysis allows the CC-TAME consortium to assess “The efficiency of current and future land-use adaptation and mitigation processes” on various levels.
In state of the art climate models, maps of land use are incorporated as time-invariant lower boundary conditions. This means, calculations of present and future climate assume that land use is not changing with time. To assess the climate change impact on land use as well as the feedback of land use changes on the simulated climate, the following two-step approach will be applied.
The working group will apply multiple bio-physical process models (EPIC, DNDC, Sundial/RothC, and SWAT) to analyse production and environmental impacts of agricultural systems (e.g. crop residue systems, crop rotations, agro-forestry and grazing systems, precision farming), bio-energy systems, and other non-forest land-use systems (e.g., wetlands) at plot, watershed, and EU27 scales. Particularly, EPIC, DNDC, and Sundial/RothC have been already made operational at EU scale in separate EU 6th framework projects (e.g., INSEA, CAPRI-DYNASPAT). A common and harmonized delineation process of homogenous response units will allow joint application of these models and increase validity and comparability of model outputs and consistency in integration of bio-physical impact trajectories in economic land use optimization models. Simulation experiments will be carried out with computed extreme weather events to assess bio-physical vulnerability thresholds for major EU agricultural production regions as well as for the interface of terrestrial and aquatic ecosystem functioning. Agricultural and bio-energy systems are assessed with respect to multifunctional agriculture exhibiting trade-offs in externalities of mitigation and adaptation strategies. This will allow assessment of sustainable development pathways in European agriculture that show increased resilience to climate change and simultaneously considering the multi-policy framework (Cross Compliance, Water Framework, Programme for Rural Development).
To achieve its objectives the following tasks are introduced:
Currently and potentially available agricultural and bio-energy production systems will be specified with respect to their agronomic and bio-geographical characteristics in EU27. These agricultural systems will encompass production options in arable and pasture land utilization and include: alternative crop residue systems i.e., conventional, reduced or minimum tillage as well as straw harvesting regimes, food and non-food crop rotations of annual and perennial crops, inter-cropping systems, agro-forestry systems, grazing and other forage production systems, precision farming, fertilization and irrigation regimes, manure handling systems, cover crop management systems. The agronomic characterization of major food, non-food and forage production systems basically includes planting and harvesting dates, plant growth control (timing and amount of fertilization, irrigation, pesticide application), tillage operation schedules, and grazing and cutting intensities using available European information systems (NewCronos, FADN, etc.) and other project outputs (INSEA, ENFA, CAPRI-DYNASPAT, etc.).
Other non-forest land-use systems (e.g., wetlands, nature conservation areas, land abandonment) that primarily provide environmental and amenity benefits to the society or are severely exposed to ecological vulnerabilities will be also included in this analysis. Modelling these non-forest land-use systems particularly justifies the multi-bio-physical modelling approach to complement the weaknesses of each model. The level of management in these land use systems will range from ‘land abandonment’ to several extensification levels.
The INSEA developed data infrastructure of Homogenous Response Unit (HRU) will be used, after further validation efforts and climate change scenario integration, to establish a common input data portal for EPIC, DNDC, Sundial/RothC, and SWAT analyses at watershed and EU27 scales. The high quality data will play a major integrative role between the models and regular workshops will assure optimal and correct use of the data as well as enhance mutual and adaptive learning between and within data providers and model users/developers. Such data and model infrastructure will allow uncertainty quantification across models on a set of model output indicators by using the same input data. In addition, field experimental data from NitroEurope and CarboEurope will be used to exercise Bayesian calibration on the plot-scale models (EPIC, DNDC, and Sundial/RothC) to allow uncertainty to be better quantified than previously done. It will also allow validation of European data sources and model outputs as well as an assessment of model and output comparability at EU27 scale. In addition, this infrastructure will also allow integration of plot-scale models (EPIC, DNDC, and Sundial/RothC) with the watershed-scale model SWAT. SWAT will be applied in selected EU watersheds to account for up- and down-stream effects in the water, carbon, and nutrient cycles across complex landscapes.
The Sundial/RothC model (Bradbury et al., 1993; Smith et al., 2005) will simulate C and N transformations in the soil and GHG emissions under a large variety of European cropping systems. Impacts on soil C (Smith et al., 2005; Schröter et al., 2005), and GHG fluxes (Flynn et al., 2005) as well as on indicators of water and soil quality (Smith et al., 2007d) will be examined. Results will be compared to outputs from the other biophysical models i.e., DNDC and EPIC. Sundial allows for multiple year runs, which will provide a useful comparison to models performing a series of single year runs (e.g., DNDC).
The Denitrification-Decomposition (DNDC) model (Li, 2000; Li et al., 1992) is a process-oriented computer simulation model of soil carbon and nitrogen biogeochemistry. It is a developed for the use at regional scale. DNDC is a multi-ecosystem model designed for assessing the emissions of N2O, CH4, and NH3 from the soil into the atmosphere and the stock changes of organic carbon. The model has been tested against numerous field data sets of nitrous oxide (N2O) emissions and soil carbon dynamics (Li et al., 2005). DNDC has also been widely used also for regional modelling studies. DNDC will incorporate management information from Sundial/RothC and EPIC. On the other hand DNDC will inform the other models on the speciation of gaseous emissions. The model is already operational on European scales.
The Environmental Policy Integrated Climate (EPIC) integrates components from CREAMS (Knisel, 1980), SWRRB (Williams et al., 1985), GLEAMS (Leonard et al., 1987), and is continuously expanded and refined to allow simulation of many processes important in agricultural and forest land use management (Sharpley and Williams, 1990; Williams, 1995 and 1996). A major carbon cycling routine was performed by Izaurralde et al., (2006) based on the approach used in CENTURY (Parton et al., 1994). Efforts will focusing on model algorithm to address greenhouse gas emissions (e.g., N2O, CH4). The drainage area considered by EPIC is generally a field-size area – up to 100 hectare – where weather, soil, topography, and management systems are assumed to be homogeneous. The major components in EPIC are weather simulation, hydrology, erosion-sedimentation, nutrient and carbon cycling, pesticide fate, plant growth and competition, soil temperature and moisture, tillage, cost accounting, and plant environment control. EPIC operates on a daily time step, and is capable to simulate hundreds of years if necessary. The optional Green and Ampt infiltration equation simulates rainfall excess rates at shorter time intervals (0.1 h). Different options for simulating several processes with different algorithm – five PET equations, six erosion/sediment yield equations, two peak runoff rate equations, etc., which allow reasonable model applications in very distinct natural areas will be performed. EPIC will be used to compare management systems and their effects on water, nitrogen, phosphorus, pesticides, organic carbon, and sediment transport, on organic carbon sequestration, and eventually on green house gas emissions. The management components that will be analysed include crop rotations, crop/grass mixes, tillage operations, irrigation scheduling, drainage, furrow diking, liming, grazing, burning operations (e.g., on prairies), tree pruning, thinning and harvest, manure handling (e.g., lagoons), and fertilizer and pesticide application rates and timing. The Epic model is already operational on EU25.
The Soil Water Assessment Tool (SWAT) is a river basin or watershed scale model developed by Arnold et al. (1995) for the USDA Agricultural Research Service (ARS). SWAT was developed to predict the impact of land management practices on water, sediment and agricultural chemical yields in large complex watershed with varying soils, land-use and management conditions over long periods of time. SWAT incorporates features of several ARS models and is a direct outgrowth of the SWRRB model (Simulator for Water Resources in Rural Basins) (Williams et al., 1985; Arnold et al., 1990). Specific models that contributed significantly to the development of SWAT were CREAMS (Chemicals, Runoff, and Erosion from Agricultural Management Systems) (Knisel, 1980), GLEAMS (Groundwater Loading Effects on Agricultural Management Systems) (Leonard et al., 1987), and EPIC (Erosion-Productivity Impact Calculator) (Williams et al., 1984). Interfaces for the model have been developed in Windows (Visual Basic), GRASS, and ArcView. SWAT has also undergone extensive validation. The model has been already applied for several river basins (e.g. Danube), and watersheds in the European Union. The SWAT Model will be applied to a number of vulnerable watersheds. The impact of adaptation strategies will be assessed with respect to climate change risks and pressures from land-use change.
A set of bio-physical vulnerability thresholds will be identified with respect to food and non-food production systems and ecological functioning (e.g., wetlands) in the terrestrial and aquatic interface. This effort will be jointly done with the economic modelling group to produce spatially explicit bio-physical and economic vulnerability matrices. Simulation experiments will be carried out with computed extreme weather events using REMO to establish a functional relationship between large-scale agricultural production outages and the specific agricultural system. A catalogue of efficient and effective mitigation and adaptation options will be established in assessing the bio-physical potentials of alternative agricultural systems in a geographical representation. The geographical representation will include sub-national and EU27-national as well as bio-geographical and watershed scales.
The INSEA developed 2-step approach of HRU delineation will be used to consistently integrate bio-physical impact trajectories in economic land use optimization models. Bio-physical impact vectors (e.g., crop yields, soil carbon sequestration, GHG-emissions, nitrate emission, sediment transport, etc.) will be spatially and temporally indexed based on common activity units (i.e., wheat production in ha) and input in the economic land use models for joint economic and environmental analyses at farm, regional, and sectoral scale. All bio-physical process models will run in the HRU framework and will therefore be able to supply production and environmental indicators to the economic models. Modelling spatially explicit production (emission) possibility curves using the array of alternative agricultural systems and other non-forest land-use systems allows a fuller representation and assessment of land use change under climate and global change. In addition, capturing the bio-physical dynamics in water, N & C cycles of alternative land use and management systems will allow identifying optimal management plans and policy incentives necessary for adapting European agriculture towards a sustainable development pathway.
At the level of EU 27 the CC-TAME modelling framework will locate vulnerable conditions with regard to climate change mitigation and subsequently identify efficient management options for climate change adaptation and mitigation. A multi-model approach will be utilized to bridge various temporal and spatial scales and provide a consistent modelling framework to cover responses to a changing climate as well as to alternative management practices. Mature, intensively tested plot/stand level process-based forest models from different biogeographical regions (boreal, mediterranean, temperate and alpine) will be used for calibration of a continental scale model with sufficient flexibility while providing computational efficiency and low input data requirements which is a necessity at such scale. Natural disturbance factors (fire, storm, bark beetles) will be scaled and integrated in the continental modelling approaches. For selected regions in major biomes plot-level and continental scale forest models (i.e., NUTS-2) are run side by side to check for consistent upscaling results under various climate and management scenarios and to quantify uncertainties from modelling approaches as well as from data availability. Based on relevant EU policies and other external drivers continental scale modelling approaches will generate various management strategies spatially explicit for EU27 and link to economic modelling and optimization. Landuse strategies optimized at continental scale (e.g., spatial resolution at level NUTS-2) will be disaggregated for selected regions in different biomes to identify consequences of optimised management strategies and their externalities at the operational scale. Biophysical vulnerability thresholds will be identified with regard to mitigation and sustainable supply of timber to bio-based industries. This novel approach will allow for a detailed analysis of climate and management effects as well as for a sound assessment of trade-offs and externalities of mitigation and adaptation strategies on other forest functions. Moreover, the impacts of extreme events and natural disturbances will be addressed in such analysis.
In order to fully exploit the potential of the multi-model approach of CC-TAME the coordinated application of the tools at all stages of the project has to be secured. This includes a harmonized understanding in the regionally stratified calibration and validation of the continental scale models with mature, fully operable stand and plot level models as well as concerted definition of management strategies with regional focus including harmonized wording, definitions and required data sets for European scale modelling.
The continental scale models are calibrated utilizing inter alia the regional modelling design and data represented by the project consortium. Utilizing regionally validated models for various biomes allows for a thorough parameterisation and serves also as a basis for verification exercises with regard to simulated response to climate change and management of the continental scale models.
Regional modelling serves two major tasks within the project: (a) to evaluate continental scale modelling approaches against aggregated plot-level simulations for selected regions, and (b) to disaggregate management pathways optimised by EUFASOM to the level of operational management decision making. Regional modelling is conducted for main biogeographical regions.
Output of plot-level models will be aggregated and compared with continental models to evaluate the scaling assumptions in (a) calibration procedures, (b) implementation of management practices, and (c) disturbance modelling as well as data limitations at continental scale. Results of this task will be used in an overall assessment of modelling approaches.
Hierarchical process-based stand growth model PipeQual (including stand, tree, whorl and branch submodels) will be used for simulations of boreal stands (Mäkelä et al., 1997, Mäkelä 2003). PipeQual will be applied to a test region in Finland utilizing detailed tree level information of permanent sample plots of forest inventory. (For a small subset of plots we also have repeated measurements of soil organic layer). Model output will be combined with Yasso soil model (Peltoniemi et al., 2004, Liski et al., 2005) which allows us to simultaneously analyse dynamics of vegetation and soil in terms of carbon sequestration potential. The applied process-based stand model is earlier applied, e.g., for economical optimization of stand management (Hyytiäinen et al., 2004).
In CC-TAME PICUS is applied to a test region in Austria utilizing detailed, plot level forest inventory information as well as regionalized climate data. In the selection of the test region a particular focus lies on capturing the spatial heterogeneity of Central European landscape ranging from sub-continental low lands to high alpine terrain. This test region is used to both calibrate and test the assumptions of the continental scale models applied (cf. but also forms the basis for the detailed assessment of macro-economically optimised management strategies with regard to their operational implications and externalities at the regional scale. In the latter, the model is applied to quantify trade-offs between forest services such as timber production, C sequestration and bioenergy production and assess the potential distribution and pressures from large-scale socio-economic drivers. Moreover, interactions and implications of disturbances and extremes under current and changed climate will be addressed.
Analogue to the other two biomes a sample region representing mediteranean forest conditions will be utilized in the project. Drawing on a large scale inventory of Mediterranean forests in Spain the model GOTILWA+ will be utilized to assess particularly regional impacts of mitigation and adaptation strategies. The models high physiological level of detail is prime to a reliable assessment of responses and adaptation possibilities of trees and forest ecosystems to climatic changes. Moreover, the extensive database of the Spanish sample region serves as basis for the regional parameterisation and validation of the continental scale simulation tools under Mediterranean conditions.
CC-TAME also accounts for detailed assessment of effects of socio-economic drivers on forest services by dis-aggregating optimized FASOM results for selected regions employing regional forest models. This feedback to regional scale grants that (i) detailed implications of strategic pathways can be scrutinized at the level of the operational management decision maker; (ii) externalities of adaptation and mitigation measures can be quantified at subnational level.
In CC-TAME continental scale modelling relies on several models in order to be able to capture all relevant effects of alternative land use policies and management strategies on the forest sector and reveal potential uncertainties. This modelling framework essentially provides the biophysical backbone for the optimisation of management strategies with regard to their adaptation and mitigation potentials (EUFASOM) but also addresses explicitly natural disturbances at a larger scale and tackles interactions between the forest sector and other land uses.
The OSKAR model will, spatially explicit on NUTS2 level for EU27, generate the biophysical quantification of a large number of adaptation and mitigation strategies under climate change which are then integratively optimised by the FASOM model. Model-related limitations and uncertainties at continental scale are analysed inter alia by employing a second simulation approach. In extension to the cross-model comparison we aim at highlighting sensitive parameters which control different processes in different parts of the forestry sector C balance by applying the FORMICA model (UBA et al., 2006, Böttcher et al., forthcoming).
In CC-TAME the continental scale assessment is extended to global scale by applying the model GFM (Kindermann et al., 2006). Development of biomass and C stocks at global scale are assessed in the model which serves also as main interface for the assessment of land-use change induced feedbacks to mitigation objectives. In integrating biophysical and socio-economic drivers GFM is a main linkage between biophysical and economic modelling in the CC-TAME framework. Analysis for Europe will be conduced on a 1km resolution using the European species map, the European biomass map (in cooperation with EFI) and imputs from the plot level models.
With regard to biophysical forest modelling disturbances and extreme events are addressed in CC-TAME at various scales. At plot to regional level detailed impacts of disturbances can be studied for the chosen sample regions expanding methodologies from already existing work in the consortium (e.g., Seidl et al., 2007a,c). Options to integrate disturbances and extreme events (e.g., fire, storm, snow, bark beetles) into the continental scale modelling framework will be developed. This is of particular importance since disturbances influence not only the recent carbon balance of forest regions but have also implications for the age class structure of forest landscapes which affect carbon sinks and sources of that landscape many decades in the future. Within the CC-TAME consortium such ventures can again make extensively use of the nested modelling approach, facilitating the scaling of complex ecosystem processes such as natural disturbances.
In order to integrate the results of the biophysical models to the economic framework, the economic models used in the project need to apply compatible definitions with regard to land use options and land (agricultural and forestry land) management policies considered and with regard to resource and product scopes. When the harmonisation of a definition is not applicable, mapping rules for exchange of information between different models need to be developed. The harmonization criteria will be elaborated. All the researchers working with the different economic models (see below) will contribute to this task.
This task will be to integrate the detailed farm-level model FAMOS with the regional-level model PASMA using farm and regional spatially explicit land use management options (e.g. organic and conventional crop rotations, bio-energy crops, grazing, meadow and forestry activities) as well as regional price vectors to mimic regional land-rent markets which are iteratively adjusted to some equilibrium state. Both models will be linked with EUFASOM using its endogenously estimated commodity price vectors to integrate international market responses at farm and regional levels. FAMOS and PASMA will be also used to validate the outputs from the EU models (AROPAj, EUFASOM), because very detailed farm level data are available in Austria.
AROPAj will be extended from EU15 coverage to EU-25 with the help of the newly available FADN data from the new member states. The FADN data will be checked for its quality to be used for policy analysis. AROPAj will develop its own baseline scenarios and these will be compared to CAPRI baselines to be harmonized at later stages. The biomass sector will be included. AROPAj will deliver cost and policy information to EU FASOM and will receive price vectors and other market information from policy scenarios EU FASOM. AROPAj results for farm-types will further be downscaled to grids.
The CAPRI model will be used for baseline scenario computations for the Post-Kyoto time horizon. CAPRI and FASOM will establish a joint market and activity database according to the harmonization criteria. CAPRI will incorporate information from the biophysical models to gain higher flexibility with respect to agricultural management changes and become more comparable to the other models. CAPRI will implement forest response curves from EUFASOM to intgrate possible land exchanges between agriculture and forestry. These response curves will be EU country and policy specific and include cost, carbon, and land data. Exogenous resource and market data will be compared and harmonized.
The regional forest model will be a stand alone global forest sector model based on the forest sector part of EU FASOM and the SF-GTM with additional input data about technology costs and demand side information which cannot be shared publically due to data confidentiality issues. However, linking functions will be established to harmonize with the publicly available FASOM model and data. The regional forest sector model will also be fully linked to all continential scale biophysical forestry models with special focus on forest sink and bioenergy analysis.
EUFASOM and GLOBALFASOM will implement econometrically estimated trends for agricultural crop yields, demand functions, and resource endowments from CAPRI. Common Agricultural Policy data and restrictions from CAPRI will be crosschecked and harmonized. Available species choices (animals, crops and trees), management regimes, and processing alternatives will be compared and harmonized with PASMA, AROPAJ, GTM, and CAPRI.
BEWHERE will inform AROPAj, FAMOS, PASMA, CAPRI and FASOM about detailed cost structures of bioenergy supply chains (incl. Pulp and paper mills) with special attention to economies of scale, transportation costs, and marginal cost reduction due to polyproduction (e.g. heat, power, fuel) based on its geographically explicit analysis linked to results. Conversely, BEWHERE will use market price vectors, land rent information and demand information form the other models.
GFM will be used link carbon sequestration, timber and bioenergy supply and demand information to Global FASOM for long-term analysis. GFM will use dynamic price vectors for “down-scaling” Global FASOM runs to the grid level. Spatial resolution for the European version of GFM will be 1 km grid for plotting changes in vegetation cover, carbon stock changes and harvests.
In order to allow consistent policy analysis, we will establish model interfaces, which serve the transmission of input and output data between the models working in different scales and sectors. In particular, the main macro-economic models used in the analysis, the EUFASOM, needs to be linked consistently with the more detailed, site-specific models, and with the models simulating the sectors exogenous to the EUFASOM.
There is great uncertainty over the future climate and over the response of the crop and tree growths to climatic changes. We develop models, mostly rooted in financial theory that can be used to assess what are the optimal land management strategies of different homogenous farms/forest enterprises, when accounting for these uncertainties and those arising from socio-economic development and policies. Portraying the uncertainty with a scenario tree or Adaptive Monte Carlo Optimization techniques, we formulate the problems for the enterprises optimizing risk metrics such as the conditional value at risk (CvaR). We compare the outcome with the strategies suggested by the deterministic models and consider what the implications would be at the aggregate market level.