Spatial Capture-Recapture (SCR) methods are relatively new, and are rapidly replacing non-spatial capture-recapture methods for wildlife population abundance estimation. Here is a fairly recent review of these methods.

My introduction to the topic arose from a collaboration with Murray Efford that resulted in this paper estabilishing maximum likelihood SCR methods, and thence Murray's software package ``secr''. My work with PhD students and collaborators has focussed on flexible density surface estimation from SCR surveys (see here), acoustic SCR surveys (see here), continuous-time SCR methods (see here and here), theory unifying Distance Sampling and SCR methods (see here), and most recently methods for dealing with unknown recapture identities (work in progress).

Continuous-time survey models

Most wildlife survey models and abundance estimators treat time as discrete (separate "occasions" for each survey or survey component), but usually surveys happen over some continuous interval, not at discrete points. You loose information when you discard times of events. We can use this information by borrowing ideas from survival analysis and recurrent event analysis.

* Here's a preprint showing how continuous-time models improve conventional Distance Sampling inference.

* Here's a paper developing continuous-time models for Spatial Capture-Recapture

* Here's a paper developing continuous-time models for Distance Sampling with stochastic animal availability.

Surveying with digital detectors

Digital cameras, video recorders and acoustic detectors are increasingly replacing humans as detectors on wildlife surveys. Digital survey data bring new opportunities and new challenges for statisticians. I and colleagues are working on acoustic methods for surveying speicies like frogs, gibbons, birds and cetaceans, and together with HiDef Aerial Survey Ltd, on methods for surveying birds and marine mammals at sea using high-definition video cameras.

Spatial Modelling with Wildlife Survey Data

Knowledge of spatial distribution is important for management and conservation of wildlife populations. By combining models for how animal density depends on habitat and other spatial variables, with models for how we detect animals on surveys, we can use survey data to draw inferences about the things driving animal distribution, as well as about animal abundance. My work in this area is driven by the EPSRC-funded  ESSMod project, which in collaboration with Finn Lindgren at the Unversit of Bath, and Janine Illian and Steve Buckland at St Andrews, is bringing Integrated Nested Laplace Approximation (INLA) methods to bear on Distance Sampling and Spatial Capture-Recapture survey problems.