Friday, 12 October 2012

More news from the virtual brent world

As more and more brents begin to appear in Dublin its beginning to get frustrating being stuck in an office in Cornwall still. There was even a single light-bellied brent (almost certainly from "our" population) just down the road in Falmouth the other day but it didn't seem to stick around for long unfortunately...

I'm still working with virtual geese, hoping to learn a few things about social interactions and social networks that will contribute to us being able to understand and analyse all of the data that 3 winters of fieldwork will provide. A fortnight ago I wrote about a virtual population of geese, and how we could sample flocks from this and look at social networks in a similar way we would in the field, using social groups as a proxy for "interactions". We have used simulations of this population to answer the first of the two main questions we hoped to use them for, and are now busy adapting the code we used so we are in a position to answer the second of the two.

In the meantime I have focussed more on what happens in a single group social group of geese. When we construct a social network for the whole population we tend to assume that all individuals in a group interact (the Gambit of the Group assumption). However, this seems unlikely to be the case as a goose can only "interact" with geese immediately around it. By looking at how movement of geese within groups influences how many other geese in the flock they interact with we can see how accurate this Gambit of the Group assumption is.

Our simulations create a flock of geese and instructs them to move around randomly within their flock (see the video of a program running below). We vary how much each goose can move and look to see how this effects the properties of the social network for the group.


video

This is already producing some interesting results, and I hope that as we continue to work on the "rules" that govern the movement of each goose we can begin to really get to grips with what within flock networks of social interactions really look like in brent geese. For one thing, I'm particularly excited to see how adding in dominance hierarchies might effect which individuals are important in forming connections in the network of interactions.

Hopefully more to come from me soon!

No comments:

Post a Comment