Although several quantitative studies of bird communities had been made inside the urban matrix, very few discuss the problems associated with the differential observation rate created because the visual interference of the buildings. Anybody knows some attempts to standardize the point count or transect count methods in order to obtain quantitative data useful for comparative purposes with ex-urban environments or between cities?
Alvaro,
Urban landscapes can be treated like any other type of habitat. The important thing is to quantify the urban habitat in terms of the co-variates that you believe a priori will have the greatest impact on your detection probabilities. Make measurements of these co-variates in the various types of ex-urban and "natural" habitats you are conducting surveys in. You can classify the types of habitats by their important or defining characteristics (vegetation structure by canopy height, building volume, height, density, cover, space between buildings, road density, automobile traffic, ambient noise levels, human pedestrian volume, relative open sky, relative distance of lib-of-sight observation, etc...). You can quantify aural detectability in different urban habitats experimentally using playback recordings at different distances and heights. Ultimately, you can then determine how the various habitats or habitat parameters influence the detection probabilities - with sufficient data, you may even be able to make these determinations for at least some of your more common target species. You can establish a sampling design that samples these various categories of urban habitat that you designate with sufficient replication in order to compare the bird densities by species and habitat. You can also analyze the data to determine how detection probabilities vary for each of the co-variates, establish models that describe these functions and compare them using AIC. You can then go out and survey again and test the best models form the effort. Point counts can be established in a spatial grid that reduces the amount of space that is not effectively surveyed - to saturate your study areas.
The methodology used to estimate species occurrence, must consider various assumptions on probability of detection, especially concerning the existence and extent of heterogeneities among species detectability in different habitats (Boulinier et al. 1998). A good temporal replication provides most accurate data, needed to resolve the ambiguity between species absence and issues related to non-detection of inconspicuous species or problems of detectability related to high heterogeneus habitat as urbanized environments (Dorazio et al. 2006; Graham et al. 2014).
As demonstrate by other studies using data from large-scale, a powerful tool to calculate for example species richness in heterogeneous habitats, can be the use of capture-recapture models (Jiguet et al. 2005). Anyway, the important always is to "make the error uniform", avoiding the disomogeneous datasets.
Here you find a good example of Methods (the first one) and considerations about detectability in different environments (the second one):
http://onlinelibrary.wiley.com/doi/10.1111/ele.12297/abstract
http://www.tandfonline.com/doi/pdf/10.1080/00063650509461389
Good work!
I can't think of any reason why you would need specific methods to study birds within urban habitats. Buildings can limit your ability to detect animals, but so can dense tree stands or cliffs. An urban habitat can be treated as any other "natural habitat" (if such a thing still exists) and therefore you can use standard methods. Of course the exact method to use will depend on the research question you are trying to answer.
Yes. Please contact Ranjini J. for details as she has completed her PhD on urban bird community.
Thank you very much everyone for the answers. As Pedro says, several previous studies have used the same methods to compare between urban-matrix and ex-urban sites, and the meta-analysis (thanks Faderico) show that it is possible to find good conclusions with that data. But as David and Federico note, I think too that the detectability differences would be keep in mind in the future study designs.
The biggest difference I have found in studying urban vs non urban environments is a very practical one: access. The number of private properties is astronomical, and this presents significant problems in getting access to locations. I found this a problem in randomising sampling procedures. Not an impossible problem to solve, but it needs to be considered, and I found it takes a lot of prior planning to resolve.
Dear Michael,
Yes, the problem for acces to the particular properties is very important. For to obtain enough observations, instead to make a traditional count point, one is obligated to move inside that point, my first question was precisely related to that: it is possible to compare data obtained in urban environments vs. peri-urban if methods differ?
Did you have experience to solve this problem in your researches?
Thanks for your answer!
- Badges
- Science topic
- Similar topics
- Biological Science
- Zoology
- Ornithology