This week I’ve been working on two main things. Neither have anything to do with species distribution modelling as such, though the second might in the future.
Firstly, getting the R scripts for the analysis of some nematode physiology data finished. This dataset is a bit messy. Basically looking at survival as measured through LT50 (Lethal temperature that kills half of the individuals in a sample). The nematodes get taken to a set temperature, held there for an hour and then are scored alive/dead after 24 hours. Its pretty straightforward really, and you can just apply binomial glms with logit link function to model the relationship between temperature and survival. So just at looking at this trait measured across two different species, and seeing different curves its kind of nice:
Each colour is a species, for the upper and lower lethal temperatures. Its clear from this that they have different temperature tolerance profiles. Where it gets messy is in the repeat of the experiments. For the second type of experiments I looked at how pre-acclimation of each of the species might affect this LT50 trait value. So, I kept the nematodes at 24 hours above and below their optimum temperature (25 degrees Celsius – induced acclimation.
The blue and red here represent the first batch of experiments (A) and second batch (B). Everything was setup the same way, except I didn’t do the scoring on the second batch, someone else did for me. Especially for Species A LLT there is a huge difference in survivorship. Its proving a bit of a headache to analyse, but I think I’m going to consider each cohort of data separately and see if the acclimation*set temperature is preserved. Differences could be due to a number of things, such as nematode age, difference in scoring alive/dead (observer bias is a big deal in physiology).
Bah, never expect clean datasets. Some work do to there!
Secondly, I’ve started working on a paper that’s come from the workshop on Insect Invasions. Last year I started collecting trait and other potential parameter information from the literature for a group of insects that I think could be a potentially good one to illustrate some arguments about climate change. With the coauthors now before we start mapping that one out.
p.s. the shared legend script for ggplot2 is available from Hadley Wickham’s git