I'm working on a conservation project that will soon release ~10 endangered fish that were bred in captivity into a pond. We're hoping this pond serves as a temporal home for a few years, while their natural, polluted habitat is restored. A previous observational study suggests that their spatial distribution within their habitat is aggregated; they stay together in groups of two. This result could impact future relocation programs. But how can I statistically confirm this?

I've found studies that plot the trajectories of individual animals to show how they overlap. Others suggest saying animals were within 1 meter distance from each other in 20 out of 23 readings (made up example). But I'd like a statistic to complement this, I'd appreciate all suggestions.

We will use manual radio telemetry to track individuals and will have a diurnal and nocturnal spatial location (latitude and longitude) for each individual animal on a daily basis during six weeks. We will also analyze how environmental variables affect their distribution, but for this question I'm interested in the social component. Thank you.

  • $\begingroup$ Is the claim that each fish has a single "partner" that it's consistently close to, or merely that the fish stay in groups of two, with the actual composition of those groups not being stable? $\endgroup$ – Kodiologist Mar 20 '17 at 18:43
  • $\begingroup$ Hi, it seems that they have a partner they consistently stay close to. $\endgroup$ – argo Mar 20 '17 at 21:05

This isn't my field, but the first thing that comes to mind is, for each fish, compare its mean distance from its partner to the mean of its mean distances from each of the other fish. The bigger the ratio, the more the fish stays close to its partner in preference to the other fish.

  • $\begingroup$ This is a good idea, thank you. If I also compare the mean distances between a pair with mean random distances and find a difference, would I be able to say they are aggregating? $\endgroup$ – argo Mar 20 '17 at 22:08
  • $\begingroup$ @argo Random distances? Why would you do that? $\endgroup$ – Kodiologist Mar 21 '17 at 1:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.