Hello,
I just recently started working with the different types of home range estimations and have a few questions about this and that. Maybe you could help me out with that?!
My study animals are Sand Lizards (Lacerta agilis), that I tracked with VHF transmitters along a railway track in eastern Germany. I tracked up to 20 animals at the same time, that’s why I have only 3-7 datapoints per animal per day. The transmitters lasted up to 20 days, but most of them were peeled of by the animals earlier. In average I have 33 datapoints per animal. With this information in mind, you can hopefully get an impression of the data quality I am working with. So here are my questions:
1) I read that it is important to report on autocorrelation of the datasets (and also on site fidelity of the animals). The dataset of most of my animals seems to be autocorrelated. This is probably due to the site fidelity of the animals. My question is: How do I interpret this information about autocorrelation and how does that affect my home range estimation?
2) I would like to check if the number of locations was somehow sufficient to calculate proper home range estimates. Therefor I would like to use “area-observation plots”. I am just wondering what to have on the y-axis: if I have this plot for an MCP analysis (for example), do I take the total area of an animals MCP (in m2) or do I use percentages (where my final MCP is 100%)? In the second case, an asymptote would probably have more that 100 % - is that correct? Additionally: Is there a way how to calculate the number of locations randomly from my dataset (for every single animal) or is that usually done just one by one in the same order as my sampling occurred?
3) During the sampling I took notes when I sighted the animal I was tracking. Is there a way to include this information in any home range estimation? Do you think that is a useful information at all?
I would be really glad if you have at least one or another comment on my questions or could recommend some literature on these topics. Thank you very much!!
Alina
Shortly after posting I realized there is also another question here on research gate concering my second topic. And there an R-Skript was recommended, which looks pretty cool (https://github.com/bniebuhr/movecology/tree/master/home_range_accumulation_curves). But anyways, if you have further recommendations on that, that would be great, too!
Hi Alina,
I think it's hard to answer all these questions without knowing the overall questions that you try to answer.
Just some ideas:
ad 1: As the lizards are somehow territorial it is clear that the data points are not completely independent from each other. Maybe it is worth calculating kernel densities or something similar?
ad 2: I would use bootstrapping to calculate the 100% mcp based on different subsets of your dataset. So I would ignore the order.
ad 3: You could add your observations as an additional location when calculating the mcp. However, I would try both - calculate the mcp with and without the additional information so that you see the difference. Be aware that your additional observations are most likely to be biased.
Best,
Christoph
Hi Christoph,
thanks for your quick response! I want to conduct home range analyses to see
- how big the total home range size approximately is (in order to compare that to other findings and to compare it with the size of compensation measures for Sand Lizards, if possible)
- if there are differnces in home ranges between males and females and during differen times of the year: That's why I conducted sampling at three different time slots of the year.
To your third idea: You mean that I add a second location on top of the first to increase weight at this point?
- Badges
- Science topic