Camera-trapping research and the joys of Bayesian statistics

Dear readers,

It’s been a while since we haven’t posted any stories or pictures on the blog and we apologize. I am simply completely overwhelmed with work (this is what I get for wanting to do a PhD 😉 ), Marion was in Europe until very recently and Storme was taking care of foreign students in the Kruger National Park and is now super busy finishing her MSc thesis.

I decided that I would write a bit about what I am doing because you are probably thinking it is taking very long to see the camera-trapping results… Well, it is!

Grey rhebok calf at a natural waterhole

Grey rhebok calf at a natural waterhole

While Storme is writing her MSc thesis and Marion is analyzing what baboons eat thanks to methods involving isotopes, I was busy analyzing and writing three of my data chapters: two on the diet and prey preference and diet overlap of mesopredators and one on camera trapping.

Unfortunately the camera trapping chapter is taking very long because of the data. Yes, it was becoming so complicated that I had to approach a Bayesian statistician to give me some very needed advice. With Justin, we found Allan Clark, at the Department of Statistical Sciences at UCT.

Leopard in Anysberg Nature Reserve

Leopard in Anysberg Nature Reserve

Cape eland in Anysberg Nature Reserve

Cape eland in Anysberg Nature Reserve

The difficulty of my data is that they are very sparse, with lots and lots and lots of zeros. Many data sets have zeros in ecology but they usually have zeros for only a few species or a few sites and then have values (in my case detections) for the other species/sites. In the case of the Karoo, we have many zeros and extremely low detections for most of the species and only a few species have many detections. It makes it really hard to find a model that fits the data properly and allows me to answer my initial questions. I guess it is one of the reasons why not that many researchers work in (semi-)deserts…

Greater kudu in a dry riverbed

Greater kudu in a dry riverbed

Female caracal and kittens drinking at a borehole

Female caracal and kittens drinking at a borehole

But no panic, Allan is here to help me find other ways of analyzing my data and answer my questions. So hopefully we will come out with some good ideas that work well, involving Bayesian statistics. In the meantime, the diet chapters are almost finished and I will present my findings at the 2016 Karoo Parliament in Laingsburg, Central Karoo in September with the rest of the researchers from the Karoo Predator Project. I will post another blog about the conference and our talks titles closer to the date.

Young black-backed jackal on a farm

Young black-backed jackal on a farm

Advertisements
This entry was posted in Uncategorized and tagged , , , , , , , , , , , . Bookmark the permalink.