Hi all,
I'm attempting to model habitat selection by bandicoots using spatial environmental variables, with the aim to predict habitat use across a landscape.
Using variograms, I've identified spatial auto-correlation within the data. I have subsequently included a correlation factor to the model. Models containing the correlation factor are shown to be a better fit for the data based on AIC model selection.
When confirming whether the spatial auto correlation has been accounted for appropriately within the data - the variogram of the normalized residuals shows a cyclic pattern.
Any advice as to how to remove this spatial dependency, or, the reasons this cyclic pattern is produced would be very much appreciated.
Anthony.
Although it may not be directly related to your question, note that variograms pertain to point data and I suspect your data is not that kind. e.g. does it represent counts for subregions. You might want to look at a series of papers that appear in Ecography 25 (2002) that resulted from a workshop at NCEAS.
Hi Donald,
Thanks for your response. We have converted our tracking data into a spatial count of presences / unit area metric, however, we have since decided to shift our analysis as we thought the complexity of this model wasn't necessary given our core questions.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Software
- R Statistical Package