Very low overlap values instead of zero

[James S]

Hello everyone,

I am calculating home ranges for my study species (24 individuals), who have relatively small home ranges situated within a protected area. I am interested in the extent to which their home ranges overlap over a 3 month period and have followed the script from Winner et al 2018/Silva et al 2021. There is a fair bit of variation and the overlap values of home ranges that clearly overlap when plotted look reasonable to me. But I also have some home ranges that look (at least to me) like there is no overlap when plotted. 

However, when I calculate the overlap values, I never get a zero value to indicate no overlap. Instead, I get values that are incredibly low e.g. 10e-17. The attached example below shows two individuals, 19A and 6C, who look like their home ranges are separated by at least 1km (the alldata plot has data from the whole study population, not just 19A and 6C). But the overlap values I get are: 

- 19A-6C low: 4.451392e-79

- 19A-6C est: 1.612460e-63

- 19A-6C high: 1.161637e-49

Does anyone have a suggestion as to why I am getting this? Is there a certain amount of uncertainty in the models and my individuals are all too close together to get certain zeros? My code is below too.

Thanks very much in advance and for having such a useful resource as this group!

James

[Christen Fleming]

Hi James,

10^-49 is, for all intents and purposes, zero.

The reason you get extremely small numbers rather than zero, is because the rasters are truncated and the calculation switches over to a Gaussian tail approximation when they don't touch. This can be useful in some cases, but this level of accuracy is not necessary here.

Best,

Chris

[James S]

Hi Chris,

I figured something like that was the case but I thought I should double check.

Thanks very much for your response! 

James