We are sampling tadpoles in wetlands to determine mortality as a function of a few different biotic and abiotic variables. We know (roughly) how many tadpoles are in the wetlands because we count either tadpoles or egg masses. We use a 0.5 m^2 pipe sampler to collect data and estimate abundance using the methods for plot sampling from Borchers et al. 2002 on estimating animal abundance.
I am concerned we are underestimating abundance in large wetlands because we can only sample a set amount which means the probability of sampling tadpoles is lower in larger wetlands. I believe the probability of detection for the area in that we sample is 1 or very close to it (as required for the assumptions of plot sampling) but as wetland size increases, the likelihood of actually sampling tadpoles decreases.
Does anyone have any suggestions on how to deal with what the influence of differences in wetland size (via different detection probabilities) might be on our abundance estimates?
Cheers,
Paul
” the probability of sampling tadpoles is lower in larger wetlands”
that's true! in most of the cases we find more tadpoles in small, temporary pools than in larger, even if permanent, waterbodies
just as example, here is a link to a video-clip showing the amount of tadpoles in a pool not deeper than 15 - 20 cm...
https://www.youtube.com/watch?v=0rSzzP92DNE
Hi Paul,
I think that the best to do to take into account detection probabilities would be to use hierarchical model, so that you can model the detection process using a simple binomial process. You can also include some spatial variation too.
You can refer to:
- Dodd, C. K., and Dorazio, R. M. (2004), “Using Counts to Simultaneously Estimate Abundance and Detection Probabilities in a Salamander Community,” Herpetologica, 60, 468–478.
- Royle, J.A. & Dorazio, R.M. (2008) Hierarchical modeling and inference in ecology: the analysis of data from populations, metapopulations and communities. Academic Press. (especially chapters 1 & 2).
If you have troubles to find these chapters, do not hesitate to ask me
Cheers,
Guillaume
Article Hierarchical models of animal abundance and occurrence
I would be wary about making inferences about lower detection rates in larger water bodies being due to sampling methods. I seem to recall that there is some pretty good evidence out there that medium sized habitats will have higher abundances due to complex predator prey interactions and ecosystem size. I.e. lower relative abundance may actually be representative true abundance not a sampling effect.
Thanks for the warning Josh, those are exactly the things I am trying to discriminate between. I tried to do the initial sampling immediately after hatching for that very reason, to minimize the reduction in abundance due to predation and to examine the difference in detectability across wetland sizes. Cheers.
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India