Sunday, 11 March 2012

My research in Dublin - Colour Rings and Social Networks

I've just spent 5 weeks in Dublin and during that time resighted nearly 500 of our colour-ringed brent geese over 3000 times - but a big question is what can we learn from all this information?

My research is focussed on looking at patterns of social interactions in the population, so that we can improve our understanding of what causes them and the impacts that they can have on the individuals themselves. In order to do this I'm using a method known as social network analysis, which I'll attempt to introduce and explain briefly below. It is using this approach that requires such a huge amount of data on where and when we've seen each individual goose.

Part of a flock of brent geese in Kilbarrack containing 2 colour-ringed birds

By resighting colour-ringed individuals in flocks together on many occasions over a short period of time, we can build up a picture of which individuals are found together most often, and whether an individual is always seen with the same other ringed birds or whether it moves between flocks and is seen with lots of other colour-ringed birds. This picture can be represented using a diagram known as a social network. This is a graph where each individual is a point on the graph and lines between the individuals represent whether they have been seen together a set number of times. A simple example is illustrated below.

Part of a Brent Goose social Network. Each red square is a different individual (labelled by its ring code). Lines join individuals that have been seen together 3 times or more

Using social networks you can consider the interactions of an individual within a wider context. You can calculate a number of characteristics of an individual that gives you an idea of the number of associates it has, the strength of these associations, how important an individual is to the structure of the network as a whole. Additionally you look at the overall properties of the network itself, in particular whether individuals occur in little clusters of "friends" or whether all individuals are well connected to each other or not.

All of this can provide us with some very interesting and useful information on the structure of social interactions in the population, and I'll go into more detail on each of these applications in future posts.

No comments:

Post a Comment