by murray.efford » Wed Sep 28, 2011 6:12 pm
This is a very wide question! You don't specify the research question, so I assume you just want a population estimate: density and perhaps population size. None of the packages will make up for lack of data. The Density program is probably the place to start because it makes it easy to review the data and summarize it in different ways. It will also fit basic spatially explicit capture-recapture models by the maximum likelihood method. There are more modelling options in the R package 'secr', but there's more to learn before you can use it. Greater modelling flexibility can help with sparse data because it makes it easy to pool some nuisance parameters across areas or seasons (although this can also be done in Density). 'secr' also gives population size estimates from spatially explicit models. Some people claim that Bayesian methods (as in the R package SPACECAP) are better for small datasets, but I've not seen any evidence to support this.
Hope this helps
Murray