|
Abstract |
As models become increasingly complex and integrated, uncertainty among model parameters, variables and processes become critical for evaluating model outcomes and predictions. A framework for understanding uncertainty in climate modelling has been developed by the IPCC and EEA which provides a framework for discussion of uncertainty in models in general. Here we report on a review of this framework along with the results of a survey of sources of uncertainty in livestock and grassland models. Along with the identification of key sources of uncertainty in livestock and grassland modelling, the survey highlighted the need for a development of a common typology for uncertainty. When collaborating across traditionally separate research fields, or when communicating with stakeholders, differences in understanding, interpretation or emphasis can cause confusion. Further work in MACSUR should focus on improving model intercomparison methods to better understand model uncertainties, and improve availability of high quality datasets which can reduce model uncertainties. No Label |
|
|
Abstract |
In the MACSUR project, there are several grassland models in use that were designed for and adjusted with data from different climatic regions. To be able to run these modelsfor a wide geographical range, there is a need to validate and calibrate them on the same basis.Therefore, a high-quality dataset is needed, which includes a wide range of climatic conditions, management systems and other variables.Through this search 23 grassland related institutes from eleven countries were found and contacted, where 12 of them responded to the request. Nine institutes from cooler (e.g. Finland) and warmer regions (e.g. Israel) are now willing to provide their experimental data. One contributor is even planning to join the project bringing its own grassland model.These new grassland datasets cover in addition to already available ones (Fig. 1) a wide range of climatic regions for a substantiated calibration and validation of the models. Data supplied by the institutes have been checked for internal consistency and cast into a common format. The data have been passed on to WP L2 (Model intercomparison on climate change in relation to livestock and grassland). No Label |
|
|
Abstract |
Modelling of climate effects on agriculture and food security at the European scale requires a harmonized spatially, explicit database of European land use. It can be used for scaling results of point models to an area. A recent review of land cover maps focused on the global scale (Köchy, 2010). European land use as a subset of global land use is contained in the product GlobCover representing the year 2009 with a resolution of 0.3 km. A European product is the CORINE data set with a resolution of 100 m and a minimum mapping unit of 25 ha representing the year 2006 (version 16, European Environmental Agency, 2012). For scaling the results obtained for individual points to larger regions one needs fine-grained maps using the same categories as represented by the sample points. The CORINE map of pasture cover (Fig. 1) has the advantage of being very fine-grained and the classification being supervised. The visual differences to coarser maps of cover matched to census (Fig. 4), however, indicate, that none of the existing maps is reflecting reality perfectly. Since MACSUR will likely work with official national statistics it may be preferable to use one of the census-calibrated maps. For a better match, official EU spatial reporting schemes may be used at a grain that ensures data privacy of the land owners. No Label |
|
|
Abstract |
Different datasets have been completed and are now available for the analysis of interannual and seasonal variations of productive, reproductive or health data relative to intensively dairy cows and also to establish the relationships between temperature humidity index (THI) and dairy cow performances. Datasets are referred to different European countries (Italy, Belgium, Luxembourg and Slovenia) with different climatic features. All these datasets have data relative to Animal Pedigree (Cow ID, Birth date, Breed, Sire ID and Dam ID), Test-day records (Cow ID, Herd ID, Parity, Calving date, Test date, Milk yield, Milk fat and protein (%), Milk somatic cell score), Reproductive events (Cow ID, Herd ID, Parity, Calving date, AI date, Sire ID, Days Open, NRR-56 day), and Daily meteorological records (Meteo station ID, Zip code of the meteo station, Observation date, Max temperature, Min temperature, Mean temperature, Max relative humidity, Min relative humidity, Mean relative humidity, Solar radiation, Wind speed). The dataset relative to Italy includes also Mortality data (Animal ID, Herd ID, Death date) and Bulk milk quality data (Herd ID, Test date, Fat & protein (%), Somatic cell score, Bacterial count, Herd latitude, Herd longitude, Herd elevation). An additional database is still under construction and will be based on Spanish data from organic dairy farms. No Label |
|