To maintain the equal sampling area in each 100m elevation band, I performed rarefaction and extrapolation of species richness data using iNext package. However, the 95% confidence intervals of the 15 rarefaction curves are overlapped and it is difficult to determine via diagram if the extrapolation gives some real picture of extrapolated species richness in each 100m elevation band. So, is it possible to remove the confidence intervals from the rarefaction and extrapolation curves? Or, is there any other easier package (preferably in R) for the rarefaction of the species richness data?
Any suggestion would be of great help!
I liked EstimateS
http://viceroy.eeb.uconn.edu/estimates/
It might require a bit of processing by hand, but it worked well. It is free, just cite it when publishing your results.
In R: what happens if you put the output into a variable? So something like hist(var) will plot a histogram of the values in "var," but if I write Vbrl
Thanks Timothy A Ebert !
I tried the same for drawing the rarefaction curves without CIs, where I extracted the scores and then tried to plot the scores separately in R but then I missed many things those were earlier displayed in the ggiNext plots, for e.g. the extrapolated portion of the curves, the cut points at each elevation band, e.t.c. May be it is something very easy to do in R, but I don't have that much skill in R.
So, as you have suggested, going to try it in estimateS.
Hi Kuber
I offer this option not knowing the output produced by iNext. As I understand your question it is one of being able to plot the curves for each band without the CI lines which make it difficult to interpret results.
If you are able to get the scores for each band, why do you not plot these in a spreadsheet?
I also tend to use EstimateS but data handling is rather primitive and I still find it better to graph the results in a spreadsheet (I use LibreOffice Calc).
The reason why your CI´s are probably overlapping is either
- because you have too few data per curve
- you have too many curves (some will then always overlap
- or you use the wrong method -> try to use "sample coverage" instead of sample numbers.
For the latter see Chao, A., Chiu, C.-H., and Jost, L. (2016). Statistical challenges of evaluating diversity patterns across environmental gradients in mega-diverse communities. Journal of Vegetation Science 27, 437–438.
If you simply want to remove the Confidence Interval, just type "se = FALSE" in your plotting command...
See ?plot.inext
I think, the easiest program for obtain the rarefaction curves is EstimateS. For this issue I suggest writing to professor Dan Cogalniceanu.
https://www.researchgate.net/profile/Cogalniceanu_Dan
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis