Autocorrelation of residuals from GEE and in comparison to GLM. GEE with correlation structure: fixed performed best for the musdata data set. Spatial autocorrelation is computed as Moran’s I using the acfft function. The figure depicts simulated occurrence data of Mus musculus in response to the degree of pollution and the degree of exposure (for instance, to light, noise or other hypothetical risk factors).

 
  Part of: Carl G, Levin S, Kühn I (2018) spind: an R Package to Account for Spatial Autocorrelation in the Analysis of Lattice Data. Biodiversity Data Journal 6: e20760. https://doi.org/10.3897/BDJ.6.e20760