Search

Recent changes in the simulated potential crop yield and biomass production caused by changes in the temperature and global radiation patterns are examined, using the Crop Growth Monitoring System. The investigated crops are winter wheat, spring barley, maize, winter rapeseed, potato, sugar beet, pulses and sunflower. The period considered is 1976-2005. The research was executed at NUTS2 level. Maize and sugar beet were the crops least affected by changing temperature and global radiation patterns. For the other crops the simulated potential yield remained stable in the majority of regions, while decreasing trends in simulated potential yields prevailed in the remaining regions. The changes appear in a geographical pattern. In Italy and southern central Europe, temperature and radiation change effects are more severe than elsewhere, in these areas potential crop yields of more than three crops significantly decreased. In the UK and some regions in northern Europe the yield potential of various crops increased. In a next step the national yield statistics were analyzed. For a large majority of the countries the yield increases of wheat, barley and to a lesser extent rapeseed are leveling off. Several explanations could be given, however, as the simulated yield potential for these crops decreased in various regions, the changing temperature and radiation patterns may also contribute to the diminishing yield increases or to the stagnation. In more than 50% of the investigated countries the maize, potato and sugar beet yields continue to increase. This can be attributed to improving production techniques, new crop varieties, sometimes in combination with an improving climatic potential. In some regions in northern Europe, yields continue to increase. (C) 2010 Elsevier Ltd. All rights reserved.

The Agri-Environmental Zonation (AEnZ) is a biophysical typology based on a recently available detailed database on organic carbon content of the topsoil of Europe, the Environmental Stratification (EnS) and an Agri-mask. The AEnZ is used within the integrated assessment framework of SEAMLESS. The basis for this typology is the Environmental Stratification of Europe (EnS) building mainly on climate and altitude characteristics. The 84 environmental strata were aggregated into 13 environmental zones (EnZs). The environmental zones were then combined with organic carbon topsoil data (OCTOP) to cover the wide range of agri-environmental diversity of Europe. The OCTOP content was selected as soil variable as it explained most of the variation in soils in Europe. The EnZs/OCTOP land units were combined with an Agri-mask representing major obstacles for farming resulting in the final AEnZ typology. The Agri-mask, which is based on CORINE Land Cover, soil, altitude and slope data, divides Europe into three zones with different agricultural potential (suited, unsuited and marginally suited). The AEnZ consists of 238 land types of which 82 classes are referred as suitable for agriculture (75.8% of EU27+). For the SEAMLESS framework two of the three dimensions of the Agri-Environmental Zonation have been used to build the spatial framework to link information on farming and biophysics. The spatial building block of SEAMLESS is thus the Seamzones, that is an overlay of the 13 environmental zones, the seven OCTOP classes and 270 administrative (NUTS2) regions. In total this results in 3,513 Seamzones that are used to structure the biophysical data as well as the data on farming across the EU27+.

Sensitivity analysis is crucial to better understand the behavior of models, both for developers and users. Developers can be supported in avoiding overparameterizations and in focusing their attention only in the processes with a significant impact on the output(s) of interest. Model users can benefit of sensitivity analysis by identifying the most relevant parameters in a particular biophysical context and, therefore, in optimizing the available resources for determining their value, through direct measurements or via calibration. When biophysical, deterministic models are run in a stochastic fashion using weather series, and when other inputs of a model vary substantially, the results of sensitivity analysis may differ, suggesting different, site specific strategies, for operational use. The availability of a generic software component able to be integrated in modeling and simulation environments would hence allow the estimate of differences in the behavior of models in different soil-plant-climate-management scenarios. It is possible to classify the methods for sensitivity analysis developed in the last decades in three groups: the one-factor-at-a-time method, the methods based on regression and the variance-based Monte-Carlo methods. The first group is represented by Morris¿ method, which calculates two metrics: the average (µ) and the standard deviation (s) of the population of the incremental ratios according to an opportune generation of a sample of the possible combination of parameters. The most famous methods belonging to the second group are the Latin Hypercube, the Random and the Quasi-random Lp-Tau. They differ in the method used for generating the sample, while are all based on a linear regression between the differences in the output and those in the values of parameters to calculate sensitivity indices. The third group is based on the decomposition of the total variance in summands of increasing dimensionality and it is able to quantify the effect of the interactions among parameters. The methods based on this principle are Fourier Amplitude Sensitivity Test (FAST), Extended FAST, and Sobol¿s. The last group, and in particular the Sobol¿s method, is considered the most powerful and precise in identifying the output sensitivity to the model parameters. Their drawback is the computational cost since they involve the estimation of k-dimensional integrals. On the other hand, the Morris¿ method is the one requiring the smaller sample for ranking the parameters according to their relevance and it is considered particularly suitable for preliminary screenings of models with several parameters. This paper describes the LUISA (Library User Interface for Sensitivity Analysis) component, based on the SimLab (http://simlab.jrc.ec.europa.eu/) C++ DLL. LUISA has been developed in C# under the .NET platform, with the goal of facilitating implementing sensitivity analysis capabilities on bio-physical model frameworks. As illustrative case studies, spatially distributed sensitivity analysis of two different biophysical models were carried out using the MARS database (http://mars.jrc.ec.europa.eu/), in order to cover the pedo-climatic conditions of Europe. The two models used are the WARM model for rice simulations and the generic crop simulator CropSyst. Results are presented and discussed according to the spatial variability of their relevance and to the identification of clusters based on the parameters ranks., JRC.DG.G.3-Monitoring agricultural resources

Weather generators (WG) can be defined as collections of models to estimate site specific weather data and derived variables. Their use spans from providing inputs to a variety of biophysical models to deriving weather indices. Also, using either global circulation models or local area models inputs, sets of parameters calculated from long term weather series specific to a site can be modified to reproduce via WG synthetic series representing climate change scenarios. Finally, models implemented in WG are used for estimating missing data and to perform quality control on data collected from sensors in weather stations. The models implemented in WG vary from purely empirical to physically based. There are several models to either estimate or to generate each weather variable, with different input requirements. New models are continuously being proposed, and, whether some models to estimate specific variables are commonly accepted as reference methods, the lack of some inputs requires at times using alternate approaches. Currently available WG are applications which implement a predefined set of modelling options, in software implementations which do not allow for independent extensions by third parties. The CLIMA weather generator is a component based application which consist of a set of reusable graphical user interface (GUI) components, and of a set of extensible model components. The latter are subdivided into six namespaces to estimate variables related to air temperature, rainfall, solar radiation, evapotranspiration, wind, and leaf wetness. The time characteristic of the variables estimated varies from a day to ten minutes. Another library allows estimating climatic indices from one year of daily data at the time. The current implementation consists of a total of more than 300 models. Components are usable either via the CLIMA GUI, or via custom developed applications in a client-server architecture. The architecture of components is based on the composite and strategy as keystone design patterns. Models are implemented as single approaches (simple strategies), and as composite models (composite strategies) which are associated to models of finer granularity. Another type of model unit is represented by context strategies, which implement logic to select within associated models. Finally, the GUI allows building composite models which can be saved as libraries, to be reused both within CLIMA for weather series generation, or independently by other applications. The components are implemented as .NET libraries. They implement the test of pre- and post-conditions, and a scalable tracing via .NET listeners. All variables and parameters are documented via a description, units, default, maximum, and minimum values. Components are extensible: new models can be added independently by third parties and detected by the CLIMA application, which can also use them for data generation via building new composite libraries. Each component is made available via a software development kit which includes the code of two sample projects, either to extend or to reuse the component. CLIMA and its model components are freely available for reuse in no-profit applications., JRC.DG.G.3-Monitoring agricultural resources

One of the largest consequences of projected climate change is an increased frequency of extreme weather events or conditions. They impact heavily on crops as they are out of the range of variability which crops are adapted to. As an example, air temperature is a driving variable involved in the simulation of plant development, and it is used for estimating both crop development and photosynthesis limitation at any given level of intercepted solar radiation. Although with different algorithms, these processes are basically included in all crop simulation models. The reduction of leaf area index or the possible death of the plant due to frost are usually not included in crop growth models, as well as the yield losses due to pre-flowering cold shocks inducing spikelet sterility, although these processes are driven by the same variable (temperature). Another example is ozone concentration in the air; it has traditionally been ignored in modelling because taken for constant in a given environment, hence assuming crops being adapted to its concentration. However, if ozone concentration changes, and given its toxicity, its impact on crop growth must be included in models used for scenario analysis. The impact of many extreme events or conditions is usually not included in the simulation models used for the analysis of crop response to climate change scenarios. This is a limitation which needs to be overcome because such conditions are key elements characterizing the weather scenarios as estimated by global circulation models. The .NET C# library AbioticDamage implements models to simulate events the plant is not physiologically prepared to deal with. Models for the simulation of the main damages affecting the production of herbaceous crops due to abiotic factors are implemented in an extensible library, to be linked to crop development and growth models. It implements six categories of damages: frost, cold-induced spikelet sterility, lodging, ozone, salinity, and heat-induced spikelet sterility: The modelling solutions are taken from peer reviewed sources. The component encapsulates the ontology of the variables and parameters defined which is also used to implement optional run-time data quality tests following the design-by-contract approach. The architecture allows for alternate models to simulate the impact of abiotic factors, and it permits extension by third parties. The component is provided together with the documentation of all the equations, the documentation of the code, a set of unit tests, where all the sub-models implemented were tested against the numerical results obtained with the same inputs in a different environment. In this paper we present two case studies where the component AbioticDamage was run linked to the component CropML (Crop Model Library). Gridded weather data with coverage of 27 member states of the European Union (EU-27) for the years 2002 - 2008 were used to simulate abiotic stresses: frost and spikelet sterility on winter wheat and rice grain yields, respectively. Results are discussed in terms of deviation of simulated yields from simulations run without accounting for the impact of stresses. Year 2002 is signalled for severe frost damages affecting France and other important winter wheat EU-27 producing countries. For rice, simulations indicate a roughly constant degree of spikelet sterility (about 5%) affecting the cultivations of Eastern Europe countries. For Mediterranean countries, summer temperature falls may cause severe sterility problems in some years (e.g., 2007). The case studies presented are illustrative of the type of output attainable for dedicated tools for simulation of abiotic damages in cultivated plant species, and their usefulness to interpret the impact of climatic alterations that have occurred in Europe in the last years., JRC.DG.G.3-Monitoring agricultural resources