Title: | Relationship Between Species and Groups of Sites |
---|---|
Description: | Functions to assess the strength and statistical significance of the relationship between species occurrence/abundance and groups of sites [De Caceres & Legendre (2009) <doi:10.1890/08-1823.1>]. Also includes functions to measure species niche breadth using resource categories [De Caceres et al. (2011) <doi:10.1111/J.1600-0706.2011.19679.x>]. |
Authors: | Miquel De Cáceres [aut, cre] |
Maintainer: | Miquel De Cáceres <[email protected]> |
License: | GPL (>= 2) |
Version: | 1.7.15 |
Built: | 2024-08-21 10:26:25 UTC |
Source: | https://github.com/emf-creaf/indicspecies |
The resource use data for a set of 10 hypothetical bird species during two different seasons. There are five resource classes, and their relationships are given in matrix resourceD
.
data(birds)
data(birds)
birdsbreed - a data frame containing the resource use of 10 birds for a set of five resources during the breeding season.
birdswinter - a data frame containing the resource use of 10 birds for a set of five resources during the winter season.
resourceD - a distance matrix (object of class dist
) containing the relationships between the five resources.
data(birds)
data(birds)
Creates species combinations to be used in indicator value analyses.
combinespecies( X, min.order = 1, max.order = 3, min.occ = 1, FUN = min, verbose = FALSE, add.names = TRUE, ... )
combinespecies( X, min.order = 1, max.order = 3, min.occ = 1, FUN = min, verbose = FALSE, add.names = TRUE, ... )
X |
A community data table with sites in rows and species in columns. This table can contain either presence-absence or abundance data. |
min.order |
Minimum number of species conforming species combinations. |
max.order |
Maximum number of species conforming species combinations. |
min.occ |
Threshold for minimum occurrence. Combinations with lower values are not kept. |
FUN |
Function to be calculated for all species forming the species combination. Accepted values are |
verbose |
If TRUE, prints the results of each step. |
add.names |
If TRUE, adds the names of the species combinations to the output matrix. Species combination names are lists of species concatenated using character '+'. |
... |
Additional arguments for function |
This function allows creating a data table where rows are sites and columns are combinations of species. Values for a given column of this matrix are derived from the abundance values of the species forming the corresponding combination. In particular, the abundance value for a given combination in a given site is equal to the value returned by function 'FUN' (normally the minimum) among the site values of all species forming the combination. The matrix 'XC' returned by this function can be used in functions strassoc
and signassoc
. Alternatively, indicators
and related functions provide a more elaborated way to explore the indicator value of the simultaneous occurrence of sets of species (i.e. species combinations).
An list with:
XC - A matrix containing the abundance/occurrence of each species combination.
C - A binary matrix describing the set of species forming each combination ('0' means that the species is not included, and '1' means that the species is included).
Miquel De Cáceres Ainsa, EMF-CREAF
De Cáceres, M., Legendre, P., Wiser, S.K. and Brotons, L. 2012. Using species combinations in indicator analyses. Methods in Ecology and Evolution 3(6): 973-982.
indicators
, strassoc
, signassoc
library(stats) ## Loads species data data(wetland) ## Create species combinations Y <- combinespecies(X=wetland, max.order=3, min.occ=5, verbose=TRUE) ## Creates three site groups using kmeans wetkm <- kmeans(wetland, centers=3) ## Calculate indicator value of species combinations for each of the three site groups strassoc(Y$XC, cluster=wetkm$cluster,func="IndVal.g") ## Calculate point biserial correlation value of species combinations ## for each of the three site groups strassoc(Y$XC, cluster=wetkm$cluster,func="r.g")
library(stats) ## Loads species data data(wetland) ## Create species combinations Y <- combinespecies(X=wetland, max.order=3, min.occ=5, verbose=TRUE) ## Creates three site groups using kmeans wetkm <- kmeans(wetland, centers=3) ## Calculate indicator value of species combinations for each of the three site groups strassoc(Y$XC, cluster=wetkm$cluster,func="IndVal.g") ## Calculate point biserial correlation value of species combinations ## for each of the three site groups strassoc(Y$XC, cluster=wetkm$cluster,func="r.g")
Function coverage
calculates the proportion of sites of the target site group where one or another indicator (a species or a species combination) is found. Parameters are used to select valid indicators before calculating coverage. Function plotcoverage
plots the coverage against the positive predictive value threshold used to select indicators.
coverage( x, y = NULL, selection = NULL, minstat = NULL, At = NULL, Bt = NULL, type = "stat", alpha = NULL ) plotcoverage( x, y = NULL, by = 0.05, type = "stat", max.order = NULL, group = NULL, alpha = NULL, add = FALSE, xlab = expression(A[t]), ... )
coverage( x, y = NULL, selection = NULL, minstat = NULL, At = NULL, Bt = NULL, type = "stat", alpha = NULL ) plotcoverage( x, y = NULL, by = 0.05, type = "stat", max.order = NULL, group = NULL, alpha = NULL, add = FALSE, xlab = expression(A[t]), ... )
x |
An object of class ' |
y |
An object of class ' |
selection |
A logical vector restricting the set of indicators used to calculate the coverage. |
minstat |
Minimum value of the statistic for selecting indicators. |
At |
Minimum value of positive predictive value (A) for selecting indicators. |
Bt |
Minimum value for sensitivity (B) for selecting indicators. |
type |
Specifies how to select indicators: (1) using the value of the statistic ( |
alpha |
Significance level for selecting indicators. |
by |
Rate of increase in the predictive value threshold (At). |
max.order |
The maximum number of species conforming species combinations (allows examining the effects of increasing the order of combinations).This parameter is only used when the function is called using objects of class ' |
group |
Either an integer or a character string indicating the site group or site group combination for which plot is desired. This parameter is only used when the function is called using objects of class ' |
add |
Flag indicating whether the function should draw on the existing plot. |
xlab |
Label for the x-axis. |
... |
Additional plotting parameters that are passed to the |
The coverage
of a set of indicators was defined in De Cáceres et al. (2012) as the proportion of sites in a given site group where one or several indicators are found. This value allows assessing how often the site group will be able to be determined. If all indicators of a site group are rare, then the indication system will not be useful, regardless of how much restricted to the site group the indicators are. The coverage value is a generalization of quantity B of IndVal, that applies to a group of indicators instead of a single one. Function plotcoverage
plots the coverage against the positive predictive value threshold (At) used to select indicators, as in De Cáceres et al. (2012). Functions coverage
and plotcoverage
can be executed using either an object of class 'indicators
', or an object of class 'multipatt
'. However, the parameters that apply to each case are slightly different. When using coverage
and plotcoverage
on objects of class 'multipatt
' one is expected to calculate the coverage for those indicators that are significant (see alpha
parameter), although other constraints to select valid indicators can be used. When using coverage
and plotcoverage
on objects of class 'indicators
' one is expected to calculate the coverage for indicators that have values of A larger than a specified threshold (see At
parameter). In this latter case, it may be advisable to use stat="lowerCI"
, so that indicators with broad confidence intervals are not included in the selection.
When used with an object of class 'indicators
', function coverage
returns the proportion of sites of the target site group where one or another indicator (species combination) is found. When used with an object of class 'indicators
', function coverage
returns a vector containing the coverage value for each site group or site group combination.
Miquel De Cáceres Ainsa, EMF-CREAF
De Cáceres, M., Legendre, P., Wiser, S.K. and Brotons, L. 2012. Using species combinations in indicator analyses. Methods in Ecology and Evolution 3(6): 973-982.
indicators
, multipatt
, pruneindicators
library(stats) data(wetland) ## Loads species data ## Creates three clusters using kmeans wetkm <- kmeans(wetland, centers=3) ## Run indicator analysis with species combinations for the first group sc <- indicators(X=wetland, cluster=wetkm$cluster, group=1, verbose=TRUE, At=0.5, Bt=0.2) ## Determine the coverage of the selected set of indicators coverage(sc) ## Plot the coverage against the threshold At plotcoverage(sc) plotcoverage(sc, max.order=2, add=TRUE, lty=2) ## Runs the combination analysis using IndVal.g as statistic wetpt <- multipatt(wetland, wetkm$cluster, control = how(nperm=999)) ## Determines the coverage for each site group combination coverage(wetland, wetpt, alpha = 0.05)
library(stats) data(wetland) ## Loads species data ## Creates three clusters using kmeans wetkm <- kmeans(wetland, centers=3) ## Run indicator analysis with species combinations for the first group sc <- indicators(X=wetland, cluster=wetkm$cluster, group=1, verbose=TRUE, At=0.5, Bt=0.2) ## Determine the coverage of the selected set of indicators coverage(sc) ## Plot the coverage against the threshold At plotcoverage(sc) plotcoverage(sc, max.order=2, add=TRUE, lty=2) ## Runs the combination analysis using IndVal.g as statistic wetpt <- multipatt(wetland, wetkm$cluster, control = how(nperm=999)) ## Determines the coverage for each site group combination coverage(wetland, wetpt, alpha = 0.05)
Determines the indicator value of species combinations.
indicators( X, cluster, group, func = "IndVal", min.order = 1, max.order = 5, max.indicators = NULL, At = 0, Bt = 0, sqrtIVt = 0, control = how(), permutations = NULL, print.perm = FALSE, nboot.ci = NULL, alpha.ci = 0.05, XC = TRUE, enableFixed = FALSE, verbose = FALSE ) ## S3 method for class 'indicators' print( x, At = 0, Bt = 0, sqrtIVt = 0, alpha = 1, selection = NULL, confint = FALSE, ... ) ## S3 method for class 'indicators' summary(object, ...) ## S3 method for class 'indicators' plot(x, type = "sqrtIV", maxline = TRUE, ...)
indicators( X, cluster, group, func = "IndVal", min.order = 1, max.order = 5, max.indicators = NULL, At = 0, Bt = 0, sqrtIVt = 0, control = how(), permutations = NULL, print.perm = FALSE, nboot.ci = NULL, alpha.ci = 0.05, XC = TRUE, enableFixed = FALSE, verbose = FALSE ) ## S3 method for class 'indicators' print( x, At = 0, Bt = 0, sqrtIVt = 0, alpha = 1, selection = NULL, confint = FALSE, ... ) ## S3 method for class 'indicators' summary(object, ...) ## S3 method for class 'indicators' plot(x, type = "sqrtIV", maxline = TRUE, ...)
X |
A community data table with sites in rows and species in columns. This table can contain either presence-absence or abundance data. |
cluster |
A vector containing the classification of sites into site groups. |
group |
The label corresponding to the target site group. |
func |
The indicator value variant to be used, either "IndVal" (non-equalized) or "IndVal.g" (group-equalized). |
min.order |
Minimum number of species conforming species combinations. |
max.order |
Maximum number of species conforming species combinations. |
max.indicators |
Maximum number of valid indicators to be kept. If |
At |
Threshold for positive predictive value used to select valid indicators. Combinations with lower values are not kept. |
Bt |
Threshold for sensitivity used to select valid indicators. Combinations with lower values are not kept. |
sqrtIVt |
Threshold for (square root of) indicator value. Combinations with lower values are not kept. |
control |
A list of control values describing properties of the permutation test design, as returned by a call to |
permutations |
A custom matrix of permutations, to be used if |
print.perm |
If TRUE, prints permutation numbers after each set of 100 permutations. |
nboot.ci |
Number of bootstrap samples for confidence intervals. If |
alpha.ci |
Error in confidence intervals. |
XC |
If TRUE, outputs the abundance/occurrence matrix of species combinations. |
enableFixed |
If TRUE, uses species that occur in all sites as fixed elements and creates combinations with the remaining ones. |
verbose |
If TRUE, prints the results of each step. |
x |
An object of class 'indicators'. |
alpha |
Threshold for statistical significance of indicator value. Combinations with higher p-values are not kept. |
selection |
A logical vector used to restrict, a priori, the species combinations to be printed. |
confint |
Flag to indicate that confidence interval bounds are desired when printing. |
... |
Additional arguments for functions |
object |
An object of class 'indicators'. |
type |
Statistic to plot. Accepted values are "IV" (indicator value), "sqrtIV" (square root of indicator value), "A", "LA", "UA", (positive predictive value and confidence limits), "B", "LB", "UB" (sensitivity and confidence limits). |
maxline |
Flag to indicate whether a line has to be drawn joining the maximum values for each order of combinations. |
Function indicators
creates explores the indicator value of the simultaneous occurrence of sets of species (i.e. species combinations). The method is described in De Cáceres et al. (2012) and is a generalization of the Indicator Value method of Dufrêne & Legendre (1997). The minimum and maximum number of species conforming the species combination can be controlled using min.order
or max.order
. For each combination of species it determines its positive predictive value (A), sensitivity (B) and the square root of indicator value (sqrtIV). Statistical significance of indicators for the target site group is determined by internal calls to function signassoc
. Additionally, if nboot.ci
is not null then bootstrap confidence intervals are determined with the specified alpha
level, as explained in De Cáceres & Legendre (2009). The combinations to be kept can be restricted to those whose positive predictive value, sensitivity and/or indicator value are equal or greater than input thresholds. Function print
allows printing the results in a nice table, whereas summary
provides information about candidate species, combinations and coverage of the set of indicators. Function plot
draws the statistics against the order (i.e. the number of species) of the combination.
An object of class indicators
with:
candidates |
The vector of initial candidate species. |
finalsplist |
The vector of species finally selected for combinations. |
C |
A matrix describing all the combinations studied. |
XC |
A matrix containing the abundance/occurrence of each species combination. |
A |
Positive predictive power of species combinations. If |
B |
Sensitivity of species combinations. If |
sqrtIV |
Square root of indicator value of species combinations. If |
sign |
P-value of the permutation test of statistical significance. |
group.vec |
A logical vector indicating the membership to the target group. |
Miquel De Cáceres Ainsa, EMF-CREAF
De Cáceres, M., Legendre, P., Wiser, S.K. and Brotons, L. 2012. Using species combinations in indicator analyses. Methods in Ecology and Evolution 3(6): 973-982.
De Cáceres, M. and Legendre, P. 2009. Associations between species and groups of sites: indices and statistical inference. Ecology 90(12): 3566-3574.
Dufrêne, M. and P. Legendre. 1997. Species assemblages and indicator species: The need for a flexible asymetrical approach. Ecological Monographs 67:345-366.
predict.indicators
,pruneindicators
, coverage
, multipatt
, strassoc
, signassoc
library(stats) data(wetland) ## Loads species data ## Creates three clusters using kmeans wetkm <- kmeans(wetland, centers=3) ## Number of sites in each group table(wetkm$cluster) ## Run indicator analysis with species combinations for the first group sc <- indicators(X=wetland, cluster=wetkm$cluster, group=1, verbose=TRUE, At=0.5, Bt=0.2) #Prints the results print(sc) ## Plots positive predictive power and sensitivity against the order of ## combinations plot(sc, type="A") plot(sc, type="B") ## Run indicator analysis with species combinations for the first group, ## but forcing 'Orysp' to be in all combinations sc2 <- indicators(X=wetland, cluster=wetkm$cluster, group=1, verbose=TRUE, At=0.5, Bt=0.2, enableFixed=TRUE)
library(stats) data(wetland) ## Loads species data ## Creates three clusters using kmeans wetkm <- kmeans(wetland, centers=3) ## Number of sites in each group table(wetkm$cluster) ## Run indicator analysis with species combinations for the first group sc <- indicators(X=wetland, cluster=wetkm$cluster, group=1, verbose=TRUE, At=0.5, Bt=0.2) #Prints the results print(sc) ## Plots positive predictive power and sensitivity against the order of ## combinations plot(sc, type="A") plot(sc, type="B") ## Run indicator analysis with species combinations for the first group, ## but forcing 'Orysp' to be in all combinations sc2 <- indicators(X=wetland, cluster=wetkm$cluster, group=1, verbose=TRUE, At=0.5, Bt=0.2, enableFixed=TRUE)
This function studies the association between species patterns and combinations of groups of sites.
multipatt( x, cluster, func = "IndVal.g", duleg = FALSE, restcomb = NULL, min.order = 1, max.order = NULL, control = how(), permutations = NULL, print.perm = FALSE )
multipatt( x, cluster, func = "IndVal.g", duleg = FALSE, restcomb = NULL, min.order = 1, max.order = NULL, control = how(), permutations = NULL, print.perm = FALSE )
x |
Community data table |
cluster |
A vector representing a partition of sites |
func |
Species-site group association function. Four values are accepted |
duleg |
If TRUE, site group combinations are not considered, only the original site groups, like in Dufrêne & Legendre (1997). Internally, |
restcomb |
A vector of integer values used to restrict the combinations of site groups to those with ecological sense according to the analyst. The default |
min.order |
An integer indicating the minimum order of site group combinations (by default |
max.order |
An integer indicating the maximum order of site group combinations to be considered: |
control |
A list of control values describing properties of the permutation design, as returned by a call to |
permutations |
A custom matrix of permutations, to be used if |
print.perm |
If TRUE, prints permutation numbers after each set of 100 permutations. |
This function creates combinations of the input clusters and compares each combination with the species in the input matrix x. For each species it chooses the combination with a highest association value. Best matching patterns are tested for statistical significance of the associations. Four association indices are possible (some less than for strassoc
): "IndVal", "IndVal.g", "r" and "r.g". Indicator value indices will return the pattern that better matches the species observed pattern, whereas correlation indices will return the pattern that creates a highest inside/outside difference. Details are given in De Cáceres et al. (2010). The user can restrict the combinations in three ways: (1) by using duleg=TRUE
, which leads to consider single site-groups only; (2) by setting the minimum and maximum order of combinations using min.order
and max.order
; or (3) by using restcomb
to restrict combinations at will. In order to carry out the third way, values in restcomb
must be the indices of combinations that appear in the column index
of object sign
(see below).
Complex permutation designs are allowed through the function how
from package "permute". If those are not enough, the user can set control = NULL
and specify a custom matrix of permutations to test with parameter permutations
.
An object of class multipatt
with:
func |
The name of the function used. |
comb |
A matrix describing all the combinations studied. |
str |
A matrix the association strength for all combinations studied. |
A |
If |
B |
If |
sign |
Data table with results of the best matching pattern, the association value and the degree of statistical significance of the association (i.e. p-values from permutation test). Note that p-values are not corrected for multiple testing. |
This function gives the same results as function indval
in package "labdsv" when used setting func="IndVal.g"
and duleg=TRUE
, excepting the fact that the square root IndVal values is returned instead of the original IndVal.
Miquel De Cáceres Ainsa, EMF-CREAF
Florian Jansen, Institute of Botany and Landscape Ecology, Ernst-Moritz-Arndt-University
De Cáceres, M. and Legendre, P. 2009. Associations between species and groups of sites: indices and statistical inference. Ecology 90(12): 3566-3574.
De Cáceres, M., Legendre, P., Moretti, M. 2010. Improving indicator species analysis by combining groups of sites. Oikos 119(10): 1674-1684.
Dufrêne, M. and P. Legendre. 1997. Species assemblages and indicator species: The need for a flexible asymetrical approach. Ecological Monographs 67:345-366.
summary.multipatt
, strassoc
, signassoc
, how
library(stats) data(wetland) ## Loads species data wetkm <- kmeans(wetland, centers=3) ## Creates three clusters using kmeans ## Runs the combination analysis using IndVal.g as statistic wetpt <- multipatt(wetland, wetkm$cluster, control = how(nperm=999)) ## Lists those species with significant association to one combination summary(wetpt) ## Lists those species with significant association to one combination, ## including indval components. summary(wetpt, indvalcomp=TRUE)
library(stats) data(wetland) ## Loads species data wetkm <- kmeans(wetland, centers=3) ## Creates three clusters using kmeans ## Runs the combination analysis using IndVal.g as statistic wetpt <- multipatt(wetland, wetkm$cluster, control = how(nperm=999)) ## Lists those species with significant association to one combination summary(wetpt) ## Lists those species with significant association to one combination, ## including indval components. summary(wetpt, indvalcomp=TRUE)
Function nichepref
computes the species resource preference from a the
species resource use (and resource availability when given). Function nichecentroid
computes the centroid on the resource space for a set of species. Function nichevar
computes the multivariate resource variance for a set of species (i.e. niche breadth).
In all functions resources are given in distance matrix D
, the species resource use is given in P
and the availability of resources, if present, are given by vector q
.
nichecentroid( P, D = NULL, q = NULL, mode = "multiple", Np = NULL, Nq = NULL, nboot = 1000, alpha = 0.05 ) nichepref( P, D = NULL, q = NULL, mode = "multiple", Np = NULL, Nq = NULL, nboot = 1000, alpha = 0.05 ) nichevar( P, D = NULL, q = NULL, mode = "multiple", Np = NULL, Nq = NULL, nboot = 1000, alpha = 0.05 )
nichecentroid( P, D = NULL, q = NULL, mode = "multiple", Np = NULL, Nq = NULL, nboot = 1000, alpha = 0.05 ) nichepref( P, D = NULL, q = NULL, mode = "multiple", Np = NULL, Nq = NULL, nboot = 1000, alpha = 0.05 ) nichevar( P, D = NULL, q = NULL, mode = "multiple", Np = NULL, Nq = NULL, nboot = 1000, alpha = 0.05 )
P |
Data frame containing the relative or absolute usage that a set of species (in rows) make of a set of resources (in columns) |
D |
Object of type |
q |
Vector with the availability of each resource |
mode |
Either |
Np |
Vector with the number of observations per species from which the values in |
Nq |
The number of observations per species from which the values in |
nboot |
Number of boostrap samples used to compute bias-corrected percentile confidence intervals |
alpha |
Used to set the confidence level (i.e. |
The method is described in De Caceres et al. (2010). If the distance matrix is not specified
(i.e. if D=NULL
) the function assumes that all resources are at a maximum distance (d=1
).
If the resource availability vector q
is given then the values in P
are taken as
assessments of resource use and the species preference is calculated taking into account resource
availability. Otherwise resource use is equated to resource preference. Moreover, most functions can
compute bootstrap confidence intervals following the bias-corrected percentile method (Manly 2007).
If mode = "multiple"
and Np != NULL
, bootstrap samples for a given species are generated
assuming a multinomial distribution with the proportions calculated from the corresponding row values
in P
, and the number of observations comes from the corresponding element in Np
.
If mode = "single"
then the bootstrapped units are the rows of matrix P
. In both cases,
if Nq
is indicated the availability of resources is also bootstrapped.
The bias-corrected percentile method was described for overlap niche measures
in Mueller and Altenberg (1985) and is extended here for all niche metrics.
Function nichepref
returns a matrix of species relative preference.
Function nichevar
returns a vector with the variance of the resources
used for each species in P
. Function nichecentroid
returns a matrix niche centroid in the resource space for each species in df
. If bootstrap confidence intervals are asked then the three functions also compute two extra data containing respectively the lower and upper bounds of the confidence intervals obtained following the bias-corrected percentile method. Function nichearea
returns the area of the convex hull occupied by the resources used for each species in P
.
Miquel De Caceres Ainsa, EMF-CREAF
Mueller, L.D. and L. Altenberg. 1985. Statistical Inference on Measures of Niche Overlap. Ecology 66:1204-1210.
Manly, B.F.J. 2007. Randomization, bootstrap and Monte Carlo methods in biology. Chapman and Hall texts in statistical science series. 2nd edition.
De Caceres, M., Sol, D., Lapiedra, O. and P. Legendre. (2011) A framework for estimating niche metrics using the resemblance between qualitative resources. Oikos 120: 1341-1350.
See nicheoverlap
for descriptors comparing two niches.
# Loads example data data(birds) # The niche metrics using distances among resources and assuming equal availability of resources nichepref(birdsbreed, D = resourceD) nichevar(birdsbreed, D = resourceD) nichecentroid(birdsbreed, D = resourceD) # The niche metrics using distances among resources and computes # 95 percent confidence intervals nichepref(birdsbreed, D = resourceD, mode="multiple", Np = rowSums(birdsbreed), Nq = 100) nichevar(birdsbreed, D = resourceD, mode="multiple", Np = rowSums(birdsbreed), Nq = 100) nichecentroid(birdsbreed, D = resourceD, mode="multiple", Np = rowSums(birdsbreed), Nq = 100) # Same computations with different resource availability nichepref(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple") nichevar(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple") nichecentroid(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple") # The niche metrics using distances among resources and # computes 95 percent confidence intervals nichepref(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple", Np = rowSums(birdsbreed), Nq = 100) nichevar(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple", Np = rowSums(birdsbreed), Nq = 100) nichecentroid(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple", Np = rowSums(birdsbreed), Nq = 100)
# Loads example data data(birds) # The niche metrics using distances among resources and assuming equal availability of resources nichepref(birdsbreed, D = resourceD) nichevar(birdsbreed, D = resourceD) nichecentroid(birdsbreed, D = resourceD) # The niche metrics using distances among resources and computes # 95 percent confidence intervals nichepref(birdsbreed, D = resourceD, mode="multiple", Np = rowSums(birdsbreed), Nq = 100) nichevar(birdsbreed, D = resourceD, mode="multiple", Np = rowSums(birdsbreed), Nq = 100) nichecentroid(birdsbreed, D = resourceD, mode="multiple", Np = rowSums(birdsbreed), Nq = 100) # Same computations with different resource availability nichepref(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple") nichevar(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple") nichecentroid(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple") # The niche metrics using distances among resources and # computes 95 percent confidence intervals nichepref(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple", Np = rowSums(birdsbreed), Nq = 100) nichevar(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple", Np = rowSums(birdsbreed), Nq = 100) nichecentroid(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple", Np = rowSums(birdsbreed), Nq = 100)
Functions nicheoverlap
and nichedispl
compute the overlap and centroid distance between pairs of resource distributions. In both cases resource relationships are given in the distance matrix D
and the resource use data are given in data frame P1
(and in some modes also P2
).
nichedispl( P1, P2 = NULL, D = NULL, q1 = NULL, q2 = NULL, mode = "multiple", Np1 = NULL, Np2 = NULL, Nq1 = NULL, Nq2 = NULL, nboot = 1000, alpha = 0.05 ) nicheoverlap( P1, P2 = NULL, D = NULL, q1 = NULL, q2 = NULL, mode = "multiple", Np1 = NULL, Np2 = NULL, Nq1 = NULL, Nq2 = NULL, nboot = 1000, alpha = 0.05 )
nichedispl( P1, P2 = NULL, D = NULL, q1 = NULL, q2 = NULL, mode = "multiple", Np1 = NULL, Np2 = NULL, Nq1 = NULL, Nq2 = NULL, nboot = 1000, alpha = 0.05 ) nicheoverlap( P1, P2 = NULL, D = NULL, q1 = NULL, q2 = NULL, mode = "multiple", Np1 = NULL, Np2 = NULL, Nq1 = NULL, Nq2 = NULL, nboot = 1000, alpha = 0.05 )
P1 |
Data frame containing the amount of usage that a set of species (in rows) make of a first set of resources (in columns) |
P2 |
Data frame containing the amount of usage that a set of species (in rows) make of a second set of resources (in columns). Not used if |
D |
Object of type |
q1 |
Vector with the availability of each resource corresponding to P1 |
q2 |
Vector with the availability of each resource corresponding to P2 |
mode |
Either |
Np1 |
Vector with the number of observations per species from which the values in |
Np2 |
Vector with the number of observations per species from which the values in |
Nq1 |
The number of observations from which the values in |
Nq2 |
The number of observations from which the values in |
nboot |
Number of boostrap samples used to compute bias-corrected percentile confidence intervals. |
alpha |
Used to set the confidence level (i.e. |
The method is described in De Caceres et al. (2011). If the distance matrix is not specified (i.e. if D=NULL
) the function assumes that all resources are at a maximum distance (d=1
). If the resource availability vector q1
(and q2
if supplied) is specified, then the values in P1
(and P2
if supplied) are taken as assessments of resource use and the species preference is calculated taking into account resource availability. Otherwise, resource use is equated to resource preference (i.e. all resources are considered equally available). The functions can compute bootstrap confidence intervals following the bias-corrected percentile method (Manly 2007). If mode = "multiple"
and Np1
and Np2
are not null, bootstrap samples for a given niche are generated assuming a multinomial distribution with the proportions calculated from the corresponding row values in P1
(resp. P2
), and the number of observations comes from the corresponding element in Np1
(resp. Np2
). Similarly, if mode = "pairwise"
and Np1
is not null, bootstrap samples for each niche are generated assuming a multinomial distribution with the proportions calculated from the corresponding row values in P1
, and the number of observations comes from the corresponding element in Np1
. Finally, if mode = "single"
then the bootstrapped units are the rows of matrices P1
and P2
. In both cases, if Nq1
(and Nq2
) is indicated, the availability of resources is also bootstrapped. The bias-corrected percentile method is described for overlap niche measures in Mueller and Altenberg (1985).
Function nicheoverlap
(resp. nichedispl
) returns the overlap (resp. the distance between centroids) between the each pair of rows in P1
and P2
. If mode = "multiple"
or mode = "single"
the values are returned as a data frame. If mode = "pairwise"
a matrix of values is returned instead. If bootstrap confidence intervals are asked then the functions also compute the lower and upper bounds of a confidence interval obtained following the bias-corrected percentile method. Upper and lower bounds are returned as additional columns of the data frame in mode = "multiple"
or mode = "single"
or as additional matrices of a list in mode = "pairwise"
.
Miquel De Caceres Ainsa, EMF-CREAF
Mueller, L.D. and L. Altenberg. 1985. Statistical Inference on Measures of Niche Overlap. Ecology 66:1204-1210.
Manly, B.F.J. 2007. Randomization, bootstrap and Monte Carlo methods in biology. Chapman and Hall texts in statistical science series. 2nd edition.
De Caceres, M., Sol, D., Lapiedra, O. and P. Legendre. (2011) A framework for estimating niche metrics using the resemblance between qualitative resources. Oikos 120: 1341-1350.
See nichevar
for descriptors of single niches.
# Loads example data data(birds) # The overlap and displacement metrics using distances among # resources and assuming equal availability of resources nicheoverlap(birdsbreed, birdswinter, D = resourceD, mode="multiple") nichedispl(birdsbreed, birdswinter, D = resourceD, mode="multiple") # The overlap and displacement metrics using distances among resources # and computes 95 percent confidence intervals nicheoverlap(birdsbreed, birdswinter, D = resourceD, mode="multiple", Np1 = rowSums(birdsbreed), Np2 = rowSums(birdswinter), Nq1 = 100, Nq2 = 100) nichedispl(birdsbreed, birdswinter, D = resourceD, mode="multiple", Np1 = rowSums(birdsbreed), Np2 = rowSums(birdswinter), Nq1 = 100, Nq2 = 100) # Same computations with different resource availability q <- c(0.18, 0.24, 0.22, 0.21, 0.15) nicheoverlap(birdsbreed, birdswinter, D = resourceD, q1 = q, q2 = q, mode="multiple") nichedispl(birdsbreed, birdswinter, D = resourceD, q1 = q, q2 = q, mode="multiple") nicheoverlap(birdsbreed, birdswinter, D = resourceD, q1 = q, q2 = q, mode="multiple", Np1 = rowSums(birdsbreed), Np2 = rowSums(birdswinter), Nq1 = 100, Nq2 = 100) nichedispl(birdsbreed, birdswinter, D = resourceD, q1 = q, q2 = q, mode="multiple", Np1 = rowSums(birdsbreed), Np2 = rowSums(birdswinter), Nq1 = 100, Nq2 = 100) # The overlap metrics using distances among rows of 'birdsbreed' nicheoverlap(birdsbreed, D = resourceD, mode="pairwise")
# Loads example data data(birds) # The overlap and displacement metrics using distances among # resources and assuming equal availability of resources nicheoverlap(birdsbreed, birdswinter, D = resourceD, mode="multiple") nichedispl(birdsbreed, birdswinter, D = resourceD, mode="multiple") # The overlap and displacement metrics using distances among resources # and computes 95 percent confidence intervals nicheoverlap(birdsbreed, birdswinter, D = resourceD, mode="multiple", Np1 = rowSums(birdsbreed), Np2 = rowSums(birdswinter), Nq1 = 100, Nq2 = 100) nichedispl(birdsbreed, birdswinter, D = resourceD, mode="multiple", Np1 = rowSums(birdsbreed), Np2 = rowSums(birdswinter), Nq1 = 100, Nq2 = 100) # Same computations with different resource availability q <- c(0.18, 0.24, 0.22, 0.21, 0.15) nicheoverlap(birdsbreed, birdswinter, D = resourceD, q1 = q, q2 = q, mode="multiple") nichedispl(birdsbreed, birdswinter, D = resourceD, q1 = q, q2 = q, mode="multiple") nicheoverlap(birdsbreed, birdswinter, D = resourceD, q1 = q, q2 = q, mode="multiple", Np1 = rowSums(birdsbreed), Np2 = rowSums(birdswinter), Nq1 = 100, Nq2 = 100) nichedispl(birdsbreed, birdswinter, D = resourceD, q1 = q, q2 = q, mode="multiple", Np1 = rowSums(birdsbreed), Np2 = rowSums(birdswinter), Nq1 = 100, Nq2 = 100) # The overlap metrics using distances among rows of 'birdsbreed' nicheoverlap(birdsbreed, D = resourceD, mode="pairwise")
Data set used for the dietary analysis of two feral pigeon (Columba livia) populations. There are six seed types, and their distance relationships are given in matrix dfood
.
data(pigeons)
data(pigeons)
diet.barcelona - a data frame containing the seed consumption (in grams) of pigeons taken from the population in Barcelona.
diet.moia - a data frame containing the seed consumption (in grams) of pigeons taken from the population in Moia.
dfood - a distance matrix (object of class dist
) containing the relationships between the six seed types. Distances were calculated from data describing the nutritional content and dimensions of each seed type.
De Caceres, M., Sol, D., Lapiedra, O. and P. Legendre. (2011) A framework for estimating niche metrics using the resemblance between qualitative resources. Oikos 120: 1341-1350.
data(pigeons)
data(pigeons)
Function plotniche
draws a resource niche on the resource space computed by means of principal coordinates analysis. Resource relationships are given in distance matrix D
, the species resource use is given in P
and the availability of resources is given by vector q
.
plotniche( P, D = NULL, q = NULL, mode = "multiple", Np = NULL, Nq = NULL, nboot = 1000, alpha = 0.05, species = NULL, axes = c(1, 2), chull = TRUE, bubbles = TRUE, writeName = FALSE, add = FALSE, col = "black", lty = 1, ... )
plotniche( P, D = NULL, q = NULL, mode = "multiple", Np = NULL, Nq = NULL, nboot = 1000, alpha = 0.05, species = NULL, axes = c(1, 2), chull = TRUE, bubbles = TRUE, writeName = FALSE, add = FALSE, col = "black", lty = 1, ... )
P |
Data frame containing the relative or absolute usage that a set of species (in rows) make of a set of resources (in columns). |
D |
Object of type |
q |
Vector with the availability of each resource. |
mode |
Either |
Np |
Vector with the number of observations per species from which the values in |
Nq |
The number of observations per species from which the values in |
nboot |
Number of boostrap samples used to compute bias-corrected percentile confidence intervals. |
alpha |
Used to set the confidence level (i.e. |
species |
Specifies which species niche is to be plot. This parameter is mandatory and can be either an numeric index or a string for a species name. |
axes |
PCoA axes used for plotting. |
chull |
Logical flag indicating whether or not convex hulls should be drawn (only in |
bubbles |
Logical flag to draws bubbles proportional to resource preference data. |
writeName |
Logical flag indicating whether or not the name of the species should be drawn beside the centroid. |
add |
If |
col |
Color of the centroid and confidence interval arrows. |
lty |
Line type of the confidence interval arrows. |
... |
Additional graphical parameters. |
The method is described in De Caceres et al. (in prep). If the distance matrix is not specified (i.e. if D=NULL
) the function assumes that all resources are at a maximum distance (d=1
). If the resource availability vector q
is given then the values in P
are taken as assessments of resource use and the species preference is calculated taking into account resource availability. Otherwise resource use is equated to resource preference. The function can also plot bootstrap confidence intervals following the bias-corrected percentile method (Manly 2007). If If mode = "multiple"
and Np != NULL
, bootstrap samples for a given species are generated assuming a multinomial distribution with the proportions calculated from the corresponding row values in P
, and the number of observations comes from the corresponding element in Np
. If mode = "single"
then the bootstrapped units are the rows of matrix P
. In both cases, if Nq
is indicated, the availability of resources is also bootstrapped. The bias-corrected percentile method was described for overlap niche measures in Mueller and Altenberg (1985) and is extended here for all niche metrics except nichearea
.
Miquel De Caceres Ainsa, EMF-CREAF
Mueller, L.D. and L. Altenberg. 1985. Statistical Inference on Measures of Niche Overlap. Ecology 66:1204-1210.
Manly, B.F.J. 2007. Randomization, bootstrap and Monte Carlo methods in biology. Chapman and Hall texts in statistical science series. 2nd edition.
De Caceres, M., Sol, D., Lapiedra, O. and P. Legendre. (2011) A framework for estimating niche metrics using the resemblance between qualitative resources. Oikos 120: 1341-1350.
See nichevar
and nicheoverlap
to obtain a resource niche metrics.
# Loads example data data(birds) plotniche(birdsbreed, D = resourceD, mode="multiple", species=10) plotniche(birdsbreed, D = resourceD, mode="multiple", Np = rowSums(birdsbreed), Nq = 100, species=10) plotniche(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple", species=10) plotniche(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple", Np = rowSums(birdsbreed), Nq = 100, species=10) #draw two species plotniche(birdsbreed, D = resourceD, mode="multiple", Np = rowSums(birdsbreed), Nq = 100, species=10, writeName=TRUE, bubbles=FALSE, chull=FALSE) plotniche(birdsbreed, D = resourceD, mode="multiple", Np = rowSums(birdsbreed), Nq = 100, species=1, writeName=TRUE, bubbles=FALSE, chull=FALSE, add=TRUE, col="red", lty=2)
# Loads example data data(birds) plotniche(birdsbreed, D = resourceD, mode="multiple", species=10) plotniche(birdsbreed, D = resourceD, mode="multiple", Np = rowSums(birdsbreed), Nq = 100, species=10) plotniche(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple", species=10) plotniche(birdsbreed, D = resourceD, q = c(0.18, 0.24, 0.22, 0.21, 0.15), mode="multiple", Np = rowSums(birdsbreed), Nq = 100, species=10) #draw two species plotniche(birdsbreed, D = resourceD, mode="multiple", Np = rowSums(birdsbreed), Nq = 100, species=10, writeName=TRUE, bubbles=FALSE, chull=FALSE) plotniche(birdsbreed, D = resourceD, mode="multiple", Np = rowSums(birdsbreed), Nq = 100, species=1, writeName=TRUE, bubbles=FALSE, chull=FALSE, add=TRUE, col="red", lty=2)
Function predict.indicators
takes an object of class indicators
and determines the probability of the indicated site group given a community data set. If no new data set is provided, the function can calculate the probabilities corresponding to the original sites used to build the indicators
object.
## S3 method for class 'indicators' predict(object, newdata = NULL, cv = FALSE, ...)
## S3 method for class 'indicators' predict(object, newdata = NULL, cv = FALSE, ...)
object |
An object of class 'indicators'. |
newdata |
A community data table (with sites in rows and species in columns) for which predictions are needed. This table can contain either presence-absence or abundance data, but only presence-absence information is used for the prediction. If |
cv |
A boolean flag to indicate that probabilities should be calculated using leave-one-out cross validation (i.e recalculating positive predictive value of indicators after excluding the target site). |
... |
In function |
Function indicators
explores the indicator value of the simultaneous occurrence of sets of species (i.e. species combinations). The method is described in De Cáceres et al. (2012) and is a generalization of the Indicator Value method of Dufrêne & Legendre (1997). The current function predict.indicators
is used to predict the indicated site group from the composition of a new set of observations. For communities where one or more of the indicator species combinations are found, the function returns the probability associated to the indicator that has the highest positive predictive value (if confidence intervals are available, the maximum value is calculated across the lower bounds of the confidence interval). For communities where none of the indicator species combinations is found, the function returns zeroes.
If newdata = NULL
, the function can be used to evaluate the predictive power of a set of indicators in a cross-validated fashion. For each site in the data set, recalculates the predictive value of indicators after excluding the information of the site, and then evaluates the probability of the site group.
If confidence intervals are available in x
, function predict.indicators
returns a matrix where communities are in rows and there are three columns, correspoinding to the probability of the indicated site group along with the confidence interval. If confidence intervals are not available in x
, or if cv = TRUE
, then predict.indicators
returns a single vector with the probability of the indicated site group for each community.
Miquel De Cáceres Ainsa, EMF-CREAF
De Cáceres, M., Legendre, P., Wiser, S.K. and Brotons, L. 2012. Using species combinations in indicator analyses. Methods in Ecology and Evolution 3(6): 973-982.
Dufrêne, M. and P. Legendre. 1997. Species assemblages and indicator species: The need for a flexible asymetrical approach. Ecological Monographs 67:345-366.
indicators
, pruneindicators
coverage
, multipatt
, strassoc
, signassoc
library(stats) data(wetland) ## Loads species data ## Creates three clusters using kmeans wetkm <- kmeans(wetland, centers=3) ## Run indicator analysis with species combinations for the first group sc <- indicators(X=wetland, cluster=wetkm$cluster, group=1, verbose=TRUE, At=0.5, Bt=0.2) ## Use the indicators to make predictions of the probability of group #1 ## Normally an independent data set should be used, because 'wetland' was used to derive ## indicators. The same would be obtained calling 'predict(sc)' without further arguments. p <- predict(sc, wetland) ## Calculate cross-validated probabilities (recalculates 'A' statistics once for each site ## after excluding it, and then calls predict.indicators for that site) pcv <- predict(sc, cv = TRUE) ## Show original membership to group 1 along with (resubstitution) predicted probabilities ## and cross-validated probabilities. Cross-validated probabilities can be lower for sites ## originally belonging to the target site group and higher for other sites. data.frame(Group1 = as.numeric(wetkm$cluster==1), Prob = p, Prob_CV = pcv)
library(stats) data(wetland) ## Loads species data ## Creates three clusters using kmeans wetkm <- kmeans(wetland, centers=3) ## Run indicator analysis with species combinations for the first group sc <- indicators(X=wetland, cluster=wetkm$cluster, group=1, verbose=TRUE, At=0.5, Bt=0.2) ## Use the indicators to make predictions of the probability of group #1 ## Normally an independent data set should be used, because 'wetland' was used to derive ## indicators. The same would be obtained calling 'predict(sc)' without further arguments. p <- predict(sc, wetland) ## Calculate cross-validated probabilities (recalculates 'A' statistics once for each site ## after excluding it, and then calls predict.indicators for that site) pcv <- predict(sc, cv = TRUE) ## Show original membership to group 1 along with (resubstitution) predicted probabilities ## and cross-validated probabilities. Cross-validated probabilities can be lower for sites ## originally belonging to the target site group and higher for other sites. data.frame(Group1 = as.numeric(wetkm$cluster==1), Prob = p, Prob_CV = pcv)
This function allows reducing drastically the number of species combinations to be retained for a given target site group.
pruneindicators( x, At = 0, Bt = 0, sqrtIVt = 0, alpha = 1, max.indicators = 4, verbose = FALSE )
pruneindicators( x, At = 0, Bt = 0, sqrtIVt = 0, alpha = 1, max.indicators = 4, verbose = FALSE )
x |
An object of class ' |
At |
Threshold for positive predictive value. Combinations with lower values are not kept. |
Bt |
Threshold for sensitivity. Combinations with lower values are not kept. |
sqrtIVt |
Threshold for (square root of) indicator value. Combinations with lower values are not kept. |
alpha |
Threshold for statistical significance of indicator value. Combinations with higher p-values are not kept. |
max.indicators |
Maximum number of species combinations to be kept. If |
verbose |
If TRUE, prints the results of each step. |
First, the function selects those indicators (species or species combinations) with valid positive predictive value, sensitivity and indicator value, according to the input thresholds. If the object 'speciescomb
' contains confidence intervals, then the lower bounds are used to select the valid indicators. Second, the function discards those valid indicators whose occurrence pattern is nested within other valid indicators. Third, the function evaluates the coverage
of the remaining set of indicators and explores subsets of increasing number of indicators, until the same coverage is attained and the set of indicators is returned. If the maximum allowed members is attained (max.indicators
) then the set of indicators with maximum coverage is returned.
An object of class 'indicators
' with only the species combinations selected.
Miquel De Cáceres Ainsa, EMF-CREAF
De Cáceres, M., Legendre, P., Wiser, S.K. and Brotons, L. 2012. Using species combinations in indicator analyses. Methods in Ecology and Evolution 3(6): 973-982.
De Cáceres, M. and Legendre, P. 2009. Associations between species and groups of sites: indices and statistical inference. Ecology 90(12): 3566-3574.
library(stats) data(wetland) ## Loads species data ## Creates three clusters using kmeans wetkm <- kmeans(wetland, centers=3) ## Run indicator analysis with species combinations for the first group sc <- indicators(X=wetland, cluster=wetkm$cluster, group=1, verbose=TRUE, At=0.5, Bt=0.2) ## Finds the 'best' subset of indicators sc2 <- pruneindicators(sc, At=0.5, Bt=0.2, verbose=TRUE) print(sc2)
library(stats) data(wetland) ## Loads species data ## Creates three clusters using kmeans wetkm <- kmeans(wetland, centers=3) ## Run indicator analysis with species combinations for the first group sc <- indicators(X=wetland, cluster=wetkm$cluster, group=1, verbose=TRUE, At=0.5, Bt=0.2) ## Finds the 'best' subset of indicators sc2 <- pruneindicators(sc, At=0.5, Bt=0.2, verbose=TRUE) print(sc2)
This function computes the permutation p-value of the association between a species vector and a vector of memberships to a site-group. Four different tests of hypotheses arise from considering either presence/absence or quantitative species data, and from using a non-equalized or group-equalized approach.
signassoc( X, U = NULL, cluster = NULL, mode = 1, alternative = "greater", control = how(), permutations = NULL, print.perm = FALSE )
signassoc( X, U = NULL, cluster = NULL, mode = 1, alternative = "greater", control = how(), permutations = NULL, print.perm = FALSE )
X |
Community data table (rows are sites, columns are species). |
U |
A matrix of group memberships. Each column corresponds to one site-group. Membership values can be crisp or fuzzy. If this matrix is not provided, vector gmv will be used. |
cluster |
A vector of numeric group labels for sites. |
mode |
Association mode, 0 (site-based) or 1 (group-based). |
alternative |
Alternative statistical hypothesis: "greater" (positive association), "less" (negative association), "two.sided" (either positive or negative). |
control |
A list of control values describing properties of the permutation design, as returned by a call to |
permutations |
A custom matrix of permutations, to be used if |
print.perm |
If TRUE, prints permutation numbers after each set of 100 permutations. |
Input data for this function is the species matrix X and either a matrix of group memberships (U) or a vector of numeric group labels (cluster). This R function works for both presence/absence and quantitative species data, depending on the values of the input matrix X.
If mode = 0
, the null ecological hypothesis is that the frequency (or abundance) of the species of interest in sites belonging to the site group is not higher than the frequency (or abundance) in sites not belonging to it. If mode = 1
, the null ecological hypothesis is that the relative frequency (average abundance) of the species of interest is not higher in the target site group than in other groups. See De Cáceres and Legendre for more details.
Complex permutation designs are allowed through the function how
from package "permute". If those are not enough, the user can set control = NULL
and specify a custom matrix of permutations to test with parameter permutations
.
Returns a matrix of p-values, where species are in rows and groups are in columns. Two additional columns indicate the group with lowest p-value and the p-value for this group after Sidak's correction for multiple testing.
Users should be aware that the significance test in signassoc
is not exactly the same as the one in indval
from labdsv
package. The signassoc
function is using the preference for the target group (either non-equalized or group-equalized) as test statistic. After every permutation the preference for the target group is recalculated. The function is therefore testing the null hypothesis stating that the preference of the species for a given site group is due to chance only (as in Bakker 2008). The test is repeated for every group, and this is the reason why there are as many p-values as groups. In contrast, the indval
function from labdsv
package uses the maximum preference value as test statistic, and the maximum preference value is recalculated after each permutation. The maximum preference may correspond to other groups than the one chosen for the unpermuted data. indval
function from labdsv
package is therefore testing the null hypothesis saying that the group with observed maximum preference is not such, because the maximum preference was in that group due to chance only. In order to get the consistent results compared to the indval
function, users should use the function multipatt
, along with the option duleg=TRUE
.
Miquel De Cáceres Ainsa, EMF-CREAF
Bakker, J. 2008. Increasing the utility of Indicator Species Analysis. Journal of Applied Ecology 45: 1829-1835.
De Cáceres, M. and Legendre, P. 2009. Associations between species and groups of sites: indices and statistical inference. Ecology 90(12): 3566-3574.
library(stats) data(wetland) ## Load species data wetkm <- kmeans(wetland, centers=3)## Create three clusters using ## Look for species whose abundance is significantly higher in one of the three groups signassoc(wetland, cluster=wetkm$cluster, mode=1, control = how(nperm=999)) ## Look for species whose abundance is significantly higher in sites belonging ## to one group as opposed to sites not belonging to it. signassoc(wetland, cluster=wetkm$cluster, mode=0, control = how(nperm=999))
library(stats) data(wetland) ## Load species data wetkm <- kmeans(wetland, centers=3)## Create three clusters using ## Look for species whose abundance is significantly higher in one of the three groups signassoc(wetland, cluster=wetkm$cluster, mode=1, control = how(nperm=999)) ## Look for species whose abundance is significantly higher in sites belonging ## to one group as opposed to sites not belonging to it. signassoc(wetland, cluster=wetkm$cluster, mode=0, control = how(nperm=999))
This function computes the strength of the association between a species vector and a vector of memberships to the site-group. Several statistics are possible, following De Cáceres and Legendre (2009). The choice of the most appropriate statistic depends on the application. If required, bootstrap confidence interval bounds are also computed.
strassoc( X, cluster, func = "r", group = NULL, nboot.ci = NULL, alpha.ci = 0.05, c = 1 )
strassoc( X, cluster, func = "r", group = NULL, nboot.ci = NULL, alpha.ci = 0.05, c = 1 )
X |
Community data table (rows are sites, columns are species). |
cluster |
A vector of numeric group labels for sites. |
func |
The association index to be used. Accepted function names: "r", "r.g", "IndVal", "IndVal.g", "A", "A.g", "B", "cos", "cos.g", "r.ind", "r.ind.g", "s.ind", "s.ind.g" (lowercase values are also accepted). |
group |
Group for which association values are to be computed. If group=NULL, association values will be computed for all groups. |
nboot.ci |
Number of bootstrap samples. If |
alpha.ci |
Error in confidence intervals. |
c |
Vector of total abundance per site (used in functions "r.ind", "r.ind.g", "s.ind", "s.ind.g" only). |
This R function is applicable to both presence-absence and quantitative species data, depending on the values in the input matrix X. Indices "r","r.g","cos" and "cos.g" correspond to the general correlation generalization ("r" is the point-biserial correlation coefficient), whereas "r.ind", "r.ind.g", "s.ind" and "s.ind.g" correspond to the individual-based generalization. Both approaches give the same results for binary (presence-absence) species data. Indicator value indices "IndVal" and "IndVal.g" are partial generalizations of the indices of the presence-absence indices. The "IndVal" index of Dufrêne & Legendre (1997) is called "IndVal.g" here and strassoc
returns the square root of the original index. Indices "A","A.g" and "B" are the asymmetric quantities into which indicator values "IndVal" and "IndVal.g" can be decomposed. See De Cáceres and Legendre for details on the relationships between these indices and their usage. Bootstrap confidence intervals are computed using the simple percentile method (Manly 1997).
Returns a matrix of association values, where species are in rows and groups are in columns. If nboot.ci
is not null, then a list is returned, where 'stat' contains the matrix of association values, and the lower and upper confidence limits are given in supplementary matrices named 'lowerCI' and 'upperCI' respectively.
This function gives the same association values as function indval
in package "labdsv" when used setting func="IndVal.g"
, excepting the fact that the square root IndVal values is returned instead of the original IndVal.
Miquel De Cáceres Ainsa, EMF-CREAF
De Cáceres, M. and Legendre, P. 2009. Associations between species and groups of sites: indices and statistical inference. Ecology 90(12): 3566-3574.
Dufrêne, M. and P. Legendre. 1997. Species assemblages and indicator species: The need for a flexible asymetrical approach. Ecological Monographs 67:345-366.
Manly, B. F. J. 1997. Randomization, bootstrap and Monte Carlo methods in biology. Chapman and Hall Texts in Statistical Science Series.
## Load species data data(wetland) ## Create three clusters using wetkm <- kmeans(wetland, centers=3) ## Compute Dufrene and Legendre's IndVal strassoc(wetland, wetkm$cluster, func="IndVal.g") ## Compute correlation, with bootstrap 95 percent confidence intervals strassoc(wetland, wetkm$cluster, func="r.g", nboot.ci =100)
## Load species data data(wetland) ## Create three clusters using wetkm <- kmeans(wetland, centers=3) ## Compute Dufrene and Legendre's IndVal strassoc(wetland, wetkm$cluster, func="IndVal.g") ## Compute correlation, with bootstrap 95 percent confidence intervals strassoc(wetland, wetkm$cluster, func="r.g", nboot.ci =100)
This is a summary
function for multipatt
objects.
## S3 method for class 'multipatt' summary( object, alpha = 0.05, minstat = NULL, At = NULL, Bt = NULL, indvalcomp = FALSE, ... )
## S3 method for class 'multipatt' summary( object, alpha = 0.05, minstat = NULL, At = NULL, Bt = NULL, indvalcomp = FALSE, ... )
object |
An object returned by function |
alpha |
Significance level for selecting species in the summary. |
minstat |
Minimum value of the statistic for selecting species in the summary. |
At |
Minimum value of positive predictive value (A) for selecting indicators. This argument is effective only if |
Bt |
Minimum value for sensitivity (B) for selecting indicators. This argument is effective only if |
indvalcomp |
Logical flag used to report components "A" (or "A.g") and "B" of indicator values. This argument is effective only if |
... |
Additional arguments of the |
The summary function for multipatt
objects presents the results of the analysis by grouping species following their associated site group combination. Only significantly associated species are listed for each combination. The user can limit the species shown by setting a significance level and the minimum value for the strength of the association.
Miquel De Cáceres Ainsa, EMF-CREAF
strassoc
, signassoc
, multipatt
library(stats) ## Loads species data data(wetland) ## Creates three clusters using kmeans wetkm <- kmeans(wetland, centers=3) ## Runs the combination analysis using IndVal.g as statistic wetpt <- multipatt(wetland, wetkm$cluster) ## Lists those species with significant association to one combination summary(wetpt) ## Lists those species with significant association to one combination, ## including indval components summary(wetpt, indvalcomp=TRUE)
library(stats) ## Loads species data data(wetland) ## Creates three clusters using kmeans wetkm <- kmeans(wetland, centers=3) ## Runs the combination analysis using IndVal.g as statistic wetpt <- multipatt(wetland, wetkm$cluster) ## Lists those species with significant association to one combination summary(wetpt) ## Lists those species with significant association to one combination, ## including indval components summary(wetpt, indvalcomp=TRUE)
Vegetation of the Adelaide river alluvial plain (Australia). This data set was published by Bowman & Wilson (1987) and used in Dale (1988) to compare fuzzy classification approaches.
data(wetland)
data(wetland)
A data frame with 41 sites (rows) and 33 species (columns). Abundance values are represented in abundance classes.
Bowman, D. M. J. S. and B. A. Wilson. 1986. Wetland vegetation pattern on the Adelaide River flood plain, Northern Territory, Australia. Proceedings of the Royal Society of Queensland 97:69-77.
Dale, M. B. 1988. Some fuzzy approaches to phytosociology. Ideals and instances. Folia geobotanica et phytotaxonomica 23:239-274.
data(wetland)
data(wetland)