I was recently asked by the New Zealand Forest Owners Association to critique the statistical methods that are commonly used to estimate the volume of woody debris at a site. This work arose out of the need to assess the proportion of woody debris (by volume) that came from forestry operations after Cyclone Gabrielle hit the North Island of New Zealand recently. The work was really interesting, as some of the methods involve probability calculations similar to those described for Buffon’s Needle Problem.

I recently published a paper (which can be found here) with Peter Dillingham, Chuen Yen Hong and others on the use of model averaging when analysing data from a split-plot design. In some split-plot studies the whole-plot variance component may be small relative to that for the sub-plots. When this is the case, we propose using model averaging to combine the results from the usual split-plot analysis with those from an analysis in which the whole-plot variance is set to zero. This leads to confidence intervals for whole-plot treatment means that are narrower than those from the usual split-plot analysis, while still having coverage close to the nominal level.

I now have four online courses running at the Center for Wildlife Studies:

Sampling and Experimental Design This course provides a thorough overview of the principles and techniques that are vital for designing a good research study, both observational and experimental.

Generalised Linear Models This course provides an introduction to a large class of statistical models that are useful in ecology, including models for normal data, proportions and counts. There is emphasis on both understanding the theory and using the methods in RStudio.

Bayesian Statistics in Ecology This course provides an introduction to Bayesian data analysis. There is an emphasis on the flexibility this provides in the types of models that can be used, and plenty of hands-on experience using rjags in RStudio.

Model Averaging in Ecology This course is based on my book and provides an up-to-date overview of the topic, from both the Bayesian and frequentist perspective, including methods such as bagging and stacking. There is emphasis on both understanding the theory and using the methods in RStudio.

My work on estimating maximum population growth rate for New Zealand marine mammal populations has recently been published as part of a report produced by Proteus Wildlife Research Consultants for the New Zealand Ministry of Primary Industries. This report provides an assessment of the risk to marine mammal populations from fisheries in New Zealand waters, and can be found here.

Congratulations to Mark Herse on publishing his paper on population modelling of black swans, with a view to customary management, which can be found here. It was a pleasure to provide Mark with advice on some of the statistical and modelling issues he encountered during this work.

I recently gave a talk at the Swiss Ornithological Institute in Sempach about my work on the population dynamics of tītī (sooty shearwater), which has recently been published here.

The dates have now been set for the two online courses I am running in 2022 for the Center for Wildlife Studies:

Bayesian Statistics in Ecology (6-30 September) This course provides an introduction to Bayesian data analysis. There is an emphasis on the flexibility this provides in the types of models that can be used, and plenty of hands-on experience using rjags in RStudio.

Model Averaging in Ecology (3-30 October) This course is based on my book and provides an up-to-date overview of the topic, from both the Bayesian and frequentist perspective, including methods such as bagging and stacking. There is emphasis on both understanding the theory and using the methods in RStudio.

Tim Jowett and I have just published a paper on single-fit bootstrapping, which provides a simple means of calculating a confidence interval for a non-linear function of model parameters. The idea is not new, but we thought it should be more widely known, as it is simpler and more reliable than a Wald interval based on the delta method. You can find the paper here, and the slides from a talk on this topic is here.

I am part of a research group led by Phil Lyver at Manaaki Whenua (Landcare) that is working with the Rakiura Tītī Islands Administering Body on the population dynamics of tītī (sooty shearwater). Our most recent work has recently appeared in the following papers:

David Fletcher, Jamie Newman, Sam McKechnie, Corey Bragg, Peter Dillingham, Rosemary Clucas, Darren Scott, Sebastian Uhlmann, Phil Lyver, Andrew Gormley, Stewart Bull, Kayne Davis, Renata Davis, Riki Davis, Tane Davis, Lania Edwards, Jane Kitson, Tina Nixon, Michael Skerrett, and Henrik Moller. Projected impacts of climate change, bycatch, harvesting, and predation on the Aotearoa New Zealand tītī (Ardenna grisea) population. Marine Ecology Progress Series 670 (2021): 223-238.

Sam McKechnie, David Fletcher, Jamie Newman, Corey Bragg, Peter Dillingham, Rosemary Clucas, Darren Scott, Sebastian Uhlmann, Phil Lyver, Andrew Gormley, Rakiura Tītī Islands Administering Body, and Henrik Moller. Separating the effects of climate, bycatch, predation and harvesting on tītī (Ardenna grisea) population dynamics in New Zealand: A model-based assessment. PloS one 15.12 (2020): e0243794.

My model averaging course will now be hosted by The Center for Wildlife Studies (CWS). The first offering of this will be in October 2022, and details can be found here. I will also be developing an introductory course on Bayesian Statistics, to be hosted by CWS, that will be offered in September 2022. I’ll post more details on that course nearer the time.