Since returning from summer vacation, Darryl has been busy implementing a joint growth and survival model for mountain beech trees in the central South Island, a continuation of the project looking at the effect of a magnitude 6.7 earthquake on tree growth and survival. The purpose of the current analysis is slightly different, with growth and survival being modelled simultaneously once again, and the added complication of using the growth in previous time periods as a predictor variable (or covariate) for future survival. Furthermore, the effect of elevation on survival is also being investigated, and its possible interaction with the growth effect. The intent is to evaluate whether there is evidence that the nature of the relationship between growth and survival changes with elevation, after accounting for the effect of other possible factors that may affect both growth and survival (e.g., competition, available nutrients, etc.). The modelling is being conducted in a Bayesian framework, with each model taking approximately 2.5 days to complete! It's still a work in progress, but a manuscript detailing the results will be prepared and submitted for publication. We'll provide a summary of the results in due course.
A second project Darryl has been working on is analysing data from a rusa deer control operation (using shooting) in Australia, to evaluate the efficiency of the operation, and factors that may impact on various aspects of the efficiency. This is from an animal welfare perspective as it's important that animals do not unduly suffer during such control operations. The models being implemented are very similar to that used previously in a welfare study from a rabbit control operation. Essentially, the models take into account the outcome of a sequence of shots at each animal in terms of whether the animal was hit or not, if it was hit was the animal killed, and if wasn't killed (or hit) did the animal escape. Model parameters are estimated using maximum likelihood approaches, which has been implemented in R. The client intends to publish the results, so once the manuscript has been submitted, we'll share some of the results.
Characterising habitat features of haulout sites for female New Zealand Sea Lions is an analysis that Penelope has been leading (part of a Ministry for Primary Industries project), focusing on areas along the coast of Stewart Island and the south-eastern coast of the South Island. Logistic regression modelling has been used to identify land- and marine-based habitat features that may be potential covariates for predicting the location of haulout sites. The results will be used to identify available locations that have the right mix of characteristics, which female sea lions may begin to use in the future.
Feel free to contact us if you have any questions about these projects!