During my PhD I collected field data of birds populations. Each plot consist in a 1x1km2 square, inside this square I have 9 points counts uniformly distributed. In each point count I recorded presence/absence of birds species. So at the end, for each plot, I had a "relative frequency" from 0-9.
Now I trying a make a bigger data sets, compiling my data with two other data sets. This sets responded more to typical "abundance" measure, number of individual/area or time at each point count (In one case 5 points counts within each plot, the other case I will ask now).
So all the data seems to have quantitative information for a specific area. In my case the area would be 0-9 for 1km2 or 0-1 for each point count (20m radius). While in the other cases will be the radius around each point count (and maybe a mean value of the point counts corresponding to each plot).
So how possible is to compile this sets either as a quantitative measure per plot or a similarity measure between plots?
All this plot are dispersed a land use intensity gradient that goes from natural habitat (forest) to agricultural systems. Therefore maybe we can used a standardized metric of similarity within each data set?
TRY pls to make use of the software called distance. Otherwise try to use any wildlife census technique.
regards
I don't really understand the question but for such presence data, I used 'EstimateS' software to calculate similarity indices and diversity statistics.
One alternative is to take the abundance data and transform it to presence/absence and either perform simple measures of richness or multivariate analyses (eg. Ordination) and do your analyses at the point count, or by pooling the p/a data for each plot. However you'll need to keep in mind that you are using data collected using somewhat different methods, so you'll need to account for this someways, perhaps by estimating 'densities' usingeither the point count area or the pooled area for each plot.
While I recommend the distance software, you have to transform your presence/absence data using your relative scale of 0 - 9 on the basis of very common (9) to absent (0), something similar to Braun Blanquet method of classification of old. The emerging data will then be subjected to multivariate analysis of ordination and complemented with the distance software to establish the spatial character in the bird distribution in your area of research.
Not sure I completely understand the detailed circumstances, but a fundamental question to consider is detection probabilities and how the different study designs and habitat conditions influence detection. Based on my understanding, I'm not optimistic you will be able to do more than come up with rankings of relative abundance; ranks of habitat conditions for each species and ranks of species within a habitat. There are other statistical considerations regarding the point counts.Please see the attached publications.
The cited papers will guide virtually for the research purpose. Thanks Smith.
- Badges
- Science topic
- Similar topics
- Biological Science
- Zoology
- Ornithology