Spatial Analysis: A Guide For Ecologists by Mark R. T. Dale & Marie-Josée Fortin

Author:Mark R. T. Dale & Marie-Josée Fortin [Dale, Mark R. T.]
Language: eng
Format: epub, azw3, pdf
Publisher: Cambridge University Press
Published: 2014-08-30T21:00:00+00:00

8.4.7 Example of the use of non-spatial and spatial regressions

To illustrate how non-spatial and spatial regressions behave when applied to spatially autocorrelated data, we investigated how the relationship we can detect between bird distributions and landcover types varies according to the regression method used. Specifically, we studied how the local variability in data on bird occurrence is related to the local variability in land-cover types over a study area of 920 000 km2 in southern Ontario (Canada), based on 920 sampling units of 10 × 10 km each. The local variability was assessed by computing the rate of change among adjacent sampling units using the lattice-wombling algorithm (see Chapter 9 for mathematical details). Rates of change are the first partial derivatives of a quantitative variable in both the x and y directions. The rates of change for the bird data were based on the occurrence of 60 species (mostly passerines; Figure 8.16a) as given in the Ontario breeding bird atlas (OBBA 2001–2005; Cadman et al. 2007). Of the 60 species, 27 species are associated with woods and forest habitat; 11 with grassland, agricultural and open habitat; 10 with shrub and early successional habitat; seven with urban and suburban habitat; and five with wetlands (see Polakowska et al. 2012). The Ontario Land Cover data (OMNR 1998) were used to quantify the rates of change in the amounts of landcover types per sampling unit (agriculture, mixed forest, deciduous forest, coniferous forest, water bodies, urban; see Figure 8.16b).

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.