In this case study, you will learn how to test for existence of spatial structuring of the variation in biotic community composition and how to correctly test for the effects of environment upon community composition when the spatial effects are present in your data, manifested as a spatial correlation among the sampling locations. We also show how to quantify the spatially structured and spatially non-structured fractions of the community variation explained by the environment and the spatial variation in the community composition not related to the environment, and how to visualise these fractions.

You will use a data set first published by Borcard et al. (1992) and describing oribatid mite community in a peat bog. The data set is based on 70 substrate cores, accompanied by measured environmental variables, characterising local conditions: peat density, water contents, substrate type, relief type, and shrub cover. These data are present in the first two data tables of the Oribatida.c5p project: the Oribatids data table with the counts of 35 oribatid taxa and the Environment table with environmental descriptors. The third data table (named Coordinates) stores the X and Y coordinates describing the spatial position of each core within the sampled area, which was 2.5 metres wide and 10 metres long. The additional, Xc and Yc variables represent the centred (i.e. with a zero average) versions of X and Y, and Xc^2 and Yc^2 are the centred second powers of Xc and Yc, respectively. Their use is explained in Section 19.2. Alternatively, you can import the data from the three sheets of the Oribatids.xlsx spreadsheet file. Please note that the data table in the first sheet (Species) must be transposed during the import, as the cores are represented by columns.