[R] use of specaccum in routine procedure

Giovanni Bacaro bacaro at unisi.it
Fri Oct 5 17:18:46 CEST 2007


Dear list members, I have a data.frame so shaped:

Sector	Quadrants	Plot	Sic	Time	Species1	Species2	Species3	


..	Species-n
1	1	1	1	5	0	0	1	



.	0
2	1	1	1	12	1	1	1	



.	0
3	1	1	1	34	0	1	0	



.	0
4	1	1	1	23	1	1	0	



.	0
5	2	1	1	22	1	1	1	



.	1
6	2	1	1	10	1	1	1	



.	1
7	2	1	1	2	1	0	0	



.	0
8	2	1	1	2	0	0	1	



.	0
9	3	1	1	12	0	0	0	



.	1



.	


.	


.	


.	


.	


.	



.	



.	



.	



.
16	4	1	1	3	1	1	1	



.	0
1	1	2	1	4	0	1	1	



.	0
2	1	2	1	4	0	0	0	



.	0
3	1	2	1	9	0	0	0	



.	0



..	


..	



	



.	



.	



.	



	



	



.	




3504	876	219	8	12	0	1	0	



.	1


where sector are nested within quadrants, quadrants within plot and plots
within SIC. For each sector time spent to collect
plant assemblage is available.
What I'm trying to perform is an analysis of rarefaction of species
richness considering area and time.
I'm practical with the vegan package but here I need some different analyses.
My first step should be the analysis of species rarefaction with area and
my question is: how many species, on average, would be
collected in an area of one sector? and considering only two? and so on.
The answere could be easy if I don't take care of the spatial organization
of sectors in plots (16 sectors within a plot; for the moment sector in
plot is enaught).
Mainly this is the approach used in Palmer et al. 2002 (Quantitative tools
for perfecting species lists. Environmetrics 13) and called as "extensive
first".
So what I'd like to do is the following procedure:
***Dimension 1 (mean species richness in 1 sector)
Start
->1) select randomly one sector whitin the first plot (plot randomly choosen)
->2) move to the next plot (a different one in respect to the first and
randomly selected, sampling without replacement) and whitin this second
plot, select the second sector
->3) continue until 1 sector is collected for each plot (totally 219 sectors)
->4) use specaccum function ( method = "exact") to calculate
"dim1_sector_curve1" object on the selected 219 sectors.
->5) save in the first column of the data.frame_1 the
"dim1_sector_curve1$richness" values
than (sampling with replacement)
-> repeat point 1),2),3)
->4a) use specaccum function ( method = "exact") to calculate
"dim1_sector_curve2" object on the selected 219 sectors.
->5a) save in the second column of the data.frame_1 the
"dim1_sector_curve2$richness" values
repeat this routine 10000 times.

***Dimension 2 (mean species richness in 2 sector)
Start
->1) select randomly two (different) sectors within the first plot (plot
randomly choosen)
->2) move to the next plot (a different one in respect to the first and
randomly selected, sampling without replacement) and whitin this second
plot, select two more sectors
->3) continue until 2 sectors are collected for each plot (totally 219*2
sectors)
->4) use specaccum function ( method = "exact") to calculate
"dim2_sector_curve1" object on the selected 219*2 sectors.
->5) save in the first column of the data.frame_2 the
"dim2_sector_curve1$richness" values
than (sampling with replacement)
-> repeat point 1),2),3)
->4a) use specaccum function ( method = "exact") to calculate
"dim2_sector_curve2" object on the selected 219*2 sectors.
->5a) save in the second column of the data.frame_2 the
"dim2_sector_curve2$richness" values
repeat this routine 10000 times.

***from dimension 3 to 15, repeat the same procedures with increasing
number of sectors selected (for dimension equal to 16 sectors the obtained
resuld should coincide with the rarefaction curve performed on plots
presence-absence data)

Now, my main questions are: considering the huge amount of randomizations
and calculations, is a normal pc able to run this/these routine/s?
Second, the number of iteraction for each dimension should increase as a
function of the number of sectors aggregated. This mean in my opinion that
the number of randomization for calculation of an approximative mean of
species richness for dimension_15 should be equal to....infinite???
Does someone know if some compiled function in R could provide help for my
problems or where can I look for a solution?

Many thanks
Giovanni Bacaro

-- 
Dr. Giovanni Bacaro
Università degli Studi di Siena
Dipartimento di Scienze Ambientali "G. Sarfatti"
Via P.A. Mattioli 4 53100 Siena
tel. 0577 235415
email: bacaro a unisi.it



More information about the R-help mailing list