I'm making changes to a paper and the editor would like some examples of when indices of abundance have been later found to yield misleading results. I guess the tiger track surveys vs photo mark-recapture are one example - any others?
The Guschanski example is a very good one, I have attached a couple of other papers here which might also be helpful, but there are quite a few others that I have read on this topic. Gubili et al. (2009) "Concordance of genetic and fin photo identification in great white sharks" is another one. Stevick et al (2001) "Errors in identification using natural markings" is another. You might also want to read my 2013 paper (see my page) for some references to studies which used both spatial and non-spatial techniques for estimating abundance and density. This paper also shows my own results of using traditional vs. camera trap methods (non-spatial resulted in inflated estimates).
I believe it's quite commonplace in the amphibian and reptile mitigation (capture/translocation) industry in the UK. Standard survey methods/effort levels are dictated by statutory bodies, but the relationship between peak counts and actual population are difficult to pin down. Multipliers are the only realistic tools, based on peak counts from standardised effort. But they don't work for various ecological reasons (habitat quality/extent, spatial patterns of population density, temporal effects, weather effects, and other inherent unpredictability). And to add to this, effort levels are not necessarily scaled appropriately according to the (unknown) population size. For example, if survey is spatially unrepresentative, or ill-timed, it can add very large errors to your predictions. Personally I rely on gut instinct a lot, and am often rewarded by gratifying results. Not very scientific though.
With the New Zealand Critically Endangered grand and Otago skinks (Oligosoma grande and O.otagense) the latter were thought to initially be at much greater risk due to lower observed counts, however, the use of variable detection methods in subsequent years (occupancy and photo-resight) has shown that O.otagense had much lower detection rates - and were overall more widespread and numerous than O.grande and potentially less at risk of extinction. This change was was due to a change from an index (counts) to a estimator approach (variable detection methods). Here is a paper which touches on this:
Article Predator control allows critically endangered lizards to rec...
Dear Hayward,
The indices of abundance will never mislead the results of abundance, it depends on the observers capability to looks into the data set and interpretation of the results. For example tiger track surveys and photo mark recapture have their own difficulties and methodological issues. For example in the track survey, the main factor is the nature of the substrate, but none of the published papers in tigers have never looked in that factor. In addition to that the factors such as breeding season, key resource areas, animals abundance or density, etc. How many papers published in tiger have actually looked into this all factors?. Where as in mark recapture the sampling area, placement of cameras, sampling days, behavioral response, etc will have influence on the estimate. Thus it is reasonable to justify based on resource available for the survey. Proper sampling design is very essential than mere comparison of any methods will not be reliable.
Nathan - thanks heaps for that - exactly the kind of example I was looking for. I'm sure there are thousands that will be found once we compare indices with methods that account for detectability, but NZ is the per-capita epicentre of great science, so not surprising you're leading the way.
Matt, and Bangor is the epicentre of applied scientific training, Brambell Building still standing? Another example that's in prep at the moment is takahe where traditional 'territory surveys' to estimate population size has recently been replaced by known-fate survival modelling (under Nathan Whitmore's guidance) with 80% of the wild population carrying Tx, that has immediately reframed our perspectives on trends in the population both in terms of the point estimate and also the drivers in terms of demographic survival rates.
Yes James - the Brambell Building still stands although the new ECW building stands in its glory beside it!
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping