goeveg (version 0.7.9)

synsort: Sorting functions for synoptic tables

Description

This function sorts synoptic tables from syntable function output. Sorting criteria can be either numerical values in synoptic tables, such as group-wise frequencies or fidelity measures, as well as combined criteria that also take into account differential character (according to the criteria defined by Tsiripidis et al., 2009).

The algorithm aims to sort species to blocked structure considering the defined criteria and input tables, with the best characterizing species on the top of the block, followed by species with descending importance for plant community description.

Usage

synsort(
  syn1,
  syn2 = syn1,
  matrix = NULL,
  groups,
  group_order = NULL,
  method = "allspec",
  min1 = 0,
  min2 = 0,
  samples = FALSE,
  manual_order = NULL
)

Value

Returns an (invisible) list composed of:

• $output Sorting method description

• $species Information to species included in the output table

• $samplesize Sample sizes in groups

• $syntable Sorted synoptic table, with the numeric values of syn1 in the left-side columns and differential character of species on the right-side of the output table. See Tsiripidis et al. (2009) for details and criteria for the assignment of a differential species as p = positive, n = negative, pn = positive/negative.

• $others Species that are omitted in Synoptic table due to their failing reaching the given threshold values min1 and min2. Sorted alphabetically.

• $samples Sorted original species-sample matrix, with original Plot-IDs (as column names) and the group identity (Group_No as first row of output samples table) (only when `samples = TRUE`)

• $omitted_groups Names of groups removed because they were not listed in `group_order`

Arguments

syn1

Input synoptic table 1, a data frame with numerical data format, usually from syntable function output. See Details for input table format. The values of this table will be displayed in the final output table.

syn2

Optional second input table with additional numeric or differential character sorting criteria.

matrix

Optional species-sample matrix, already used for syntable function input; used only when calculating the sorted species-sample matrix (`samples = TRUE`). Site names are imported from row names.

groups

Group identities for samples (identical to syntable function input). A vector/factor of length nrow(matrix) (one group per sample/row). Ensure matching order of group identity and samples in matrix for correct allocation of group numbers to samples.

group_order

Optional vector giving the desired order of group levels. If provided, only groups listed here are included in the output.

method

Sorting algorithm and synoptic table output options (method = c("allspec", "alldiff", "totalfreq", "manual")).

min1

Group-wise threshold minimum value for species shown in the final sorted synoptic table. Species below that minimum will be listed in the output ($others section).

min2

Threshold minimum value for considering species values of a numerical second input table syn2. Species below that minimum will not be displayed in final synoptic table, but will be listed in the output ($others section).

samples

Logical; if TRUE, a sorted species-sample matrix is returned. Defaults to FALSE.

manual_order

Optional character vector of species names (matching row names of syn1) to impose a manual row order when method = "manual". Species not listed are appended and ordered by their overall frequency (row sums of syn1, descending).

Author

Jenny Schellenberg, Friedemann von Lampe

Details

Different types of sorted synoptic tables can be created with this function. Methods "allspec" and "alldiff" are based upon the calculation of group-wise frequencies. They basically build species blocks left-to-right, for one group at a time (1..k): For group i = 1, all species whose row maximum equals the value in column i are taken and sorted in descending order within that block. This is repeated for all subsequent groups, appending species step-by-step. For groups 2..k, duplicates that were already bound to a previous group are dropped.

• method = "allspec" (default): Computes a sorted synoptic table based on one or two numeric input tables, e.g. percentage or absolute frequencies, or phi fidelity values. When specifying a second table with syn2, the threshold minimum value defined in min2 will be applied for considering species values of syn1 for sorting.

• method = "alldiff": Includes differential speciescharacter as sorting criteria. Table syn1 must be numeric (e.g. percentage frequency) and syn2 must contain information on differential character (output from syntable function with defined type = "diffspec"). The result table shows ALL diagnostic and non-diagnostic species, as long as they match the min1 and min2 thresholds. The algorithm detects highest group values of species calculated from syn1 as base for sorting, but will consider differential character criterion from syn2 as well. Species with high values in syn1 AND positive differential character will then be listed on the top of a species block. Within such a block, the differentiating and high-abundant species are sorted in a way favoring species that are positive in only one or at least few groups.

• method = "totalfreq": Sorts species by their overall frequency in syn1 (row sums) in descending order. Only species with at least one group value \(\ge\) min1 are kept in the table; the rest are listed in $others.

• method = "manual": The rows are ordered by manual_order (species not present are ignored). Any remaining species are appended, sorted by their overall frequency (row sums of syn1, descending). Species must still pass min1 (max across groups \(\ge\) min1); the rest go to $others.

References

Bruelheide, H. (2000): A new measure of fidelity and its application to defining species groups. Journal of Vegetation Science 11: 167-178. tools:::Rd_expr_doi("https://doi.org/10.2307/3236796")

Chytry, M., Tichy, L., Holt, J., Botta-Dukat, Z. (2002): Determination of diagnostic species with statistical fidelity measures. Journal of Vegetation Science 13: 79-90. tools:::Rd_expr_doi("https://doi.org/10.1111/j.1654-1103.2002.tb02025.x")

Sokal, R.R. & Rohlf, F.J. (1995): Biometry. 3rd edition Freemann, New York.

Tsiripidis, I., Bergmeier, E., Fotiadis, G. & Dimopoulos, P. (2009): A new algorithm for the determination of differential taxa. Journal of Vegetation Science 20: 233-240. tools:::Rd_expr_doi("https://doi.org/10.1111/j.1654-1103.2009.05273.x")

See Also

syntable

Examples

### Synoptic table of Scheden vegetation data using syntable()-function:
# classification to create a vector of group identity
library(cluster)
pam1 <- pam(schedenveg, 4)


### One input table for sorting:
## Synoptic table with percentage frequency of species in clusters, all species
unordered <- syntable(schedenveg, pam1$clustering, abund = "percentage",
                      type = "percfreq")   # Unordered synoptic percentage frequency table
sorted <- synsort(syn1 = unordered$syntable, matrix = schedenveg, 
                  groups = pam1$clustering, method = "allspec", min1 = 0,
                  samples = TRUE)
sorted             # view results
if (FALSE) {
# Export sorted synoptic table
write.csv(sorted$syntab, "syntab.csv")
# Export sorted species-sample matrix with original releve data for postprocessing
write.csv(sorted$samples, "output_species_sample.csv")}

## Synoptic table with only phi values
phi <- syntable(schedenveg, pam1$clustering, abund = "percentage",
                     type = "phi")         # calculates cluster-wise phi for each species
phi_table <- synsort(syn1 = phi$syntable, matrix = schedenveg, groups = pam1$clustering,
                     method = "allspec", min1 = 0.3, samples = TRUE)
phi_table     # view results

## Synoptic table with total frequency (global ranking)
total <- synsort(syn1 = unordered$syntable,
               groups = pam1$clustering,
               method = "totalfreq",
               min1 = 5)
total         # view results

### Two numerical tables for sorting:
## Synoptic table showing percentage frequencies, but only for species with minimum phi-value
## of 0.3 AND exclude species with less than 25% percentage frequency

unordered <- syntable(schedenveg, pam1$clustering, abund = "percentage",
                      type = "percfreq")   # Unordered synoptic percentage frequency table
phitable <- syntable(schedenveg, pam1$clustering, abund = "percentage",
                     type = "phi")         # calculates cluster-wise phi for each species
# now sorting and arranging
phi_complete <- synsort(syn1 = unordered$syntable, syn2 = phitable$syntable,
                       matrix = schedenveg, groups = pam1$clustering, method = "allspec",
                       min1 = 25, min2 = 0.3, samples = TRUE)
phi_complete      # view results

### Differential species analysis
differential <- syntable(schedenveg, pam1$clustering, abund = "percentage",
                         type = "diffspec")

## Synoptic table with percentage frequency (only species >25%) and
## differential character.
complete <- synsort(syn1 = unordered$syntable, syn2 = differential$syntable,
                    matrix = schedenveg, groups = pam1$clustering,
                    method = "alldiff", min1 = 25, samples = TRUE)
complete            # view result table
differential$differentials  # list differential species for groups