PR - Recreating Location From Non-spatial Data – Sample Size Requirements To Reproduce The Locations Of Farms In The European Farm Accountancy Data Network
Individual farm accountancy data sources such as the European Farm Accountancy Data Network (FADN) include no specific information on the spatial location of farms. However, spatial characteristics and site conditions determine the farms’ production potential and its influence on the surrounding environment. Spatially explicit models that make use of the FADN data need to be able to recreate a landscape including the location of the farms in a plausible way. This paper investigates the minimum sample size of farm locations required to insure the ability to reproduce a reliable map of a given region. This is done by analysing relative locations between all the 1871 farms present in the Danish river Gudenå watershed. As we have detailed information about each of the farms we can categorize the farms in groups in a way similar to what one would be able to do with farms from a FADN sample. By utilising the rich information that the FADN sample contains to create a multidimensional spatial set of requirements that the farms on average have to meet it is possible to reduce the number of available locations to a minimum. This investigation is divided into the following two-step procedure: First the variability of an individual farms spatial relationship is investigated with regard to variation in sample size and composition. Secondly is the average values investigated with regard to variation in sample size and composition.
Keywords: FADN, spatial location, methodology.
Organization(s): IAMO (Leibniz-Institute of Agricultural Development in Central and Eastern Europe) (1)