Chapter 7
Spatial Analysis of Ecological Data
7.1 Objectives
Spatial analysis of ecological data is a huge field that could fill several books by
itself. To learn about general approaches in spatial analysis in R, readers may consult
the book by Bivand et al. (2013). The present chapter has a more restricted scope.
After a short general introduction, it deals with several methods that were specifically developed for the analysis of scale-dependent structures of ecological data,
although they can, of course, be applied to other domains. These methods are based
on sets of variables describing spatial structures in various ways, derived from the
coordinates of the sites or from the neighbourhood relationships among sites. These
variables are used to model the spatial structures of ecological data by means of
multiple regression or canonical ordination, and to identify significant spatial structures at all spatial scales that can be perceived by the sampling design. As you will
see, the whole analytical process uses many of the techniques covered in the
previous chapters.
Practically, you will:
• learn how to compute spatial correlation measures and draw spatial correlograms;
• learn how to construct spatial descriptors derived from site coordinates and from
links between sites;
• identify, test and interpret scale-dependent spatial structures;
• combine spatial analysis and variation partitioning;
• assess spatial structures in canonical ordinations by computing variograms of
explained and residual ordination scores.
© Springer International Publishing AG, part of Springer Nature 2018
D. Borcard et al., Numerical Ecology with R, Use R!,
https://doi.org/10.1007/978-3-319-71404-2_7
299
Spatial Analysis of Ecological Data
7.1 Objectives
Spatial analysis of ecological data is a huge field that could fill several books by
itself. To learn about general approaches in spatial analysis in R, readers may consult
the book by Bivand et al. (2013). The present chapter has a more restricted scope.
After a short general introduction, it deals with several methods that were specifically developed for the analysis of scale-dependent structures of ecological data,
although they can, of course, be applied to other domains. These methods are based
on sets of variables describing spatial structures in various ways, derived from the
coordinates of the sites or from the neighbourhood relationships among sites. These
variables are used to model the spatial structures of ecological data by means of
multiple regression or canonical ordination, and to identify significant spatial structures at all spatial scales that can be perceived by the sampling design. As you will
see, the whole analytical process uses many of the techniques covered in the
previous chapters.
Practically, you will:
• learn how to compute spatial correlation measures and draw spatial correlograms;
• learn how to construct spatial descriptors derived from site coordinates and from
links between sites;
• identify, test and interpret scale-dependent spatial structures;
• combine spatial analysis and variation partitioning;
• assess spatial structures in canonical ordinations by computing variograms of
explained and residual ordination scores.
© Springer International Publishing AG, part of Springer Nature 2018
D. Borcard et al., Numerical Ecology with R, Use R!,
https://doi.org/10.1007/978-3-319-71404-2_7
299