The package vegtable offers objects and functions for the work with plot observations in vegetation assessments (phytosociological relevés). This object class ensures consistency on data, while it is also allowing flexibility in the degree of completeness and resolution of data. This tutorial is an updated version of a workshop offered in the context of the 16th Workshop of the German Group on Vegetation Databases celebrated in Freiburg in March 2017.
The program of the workshop is accessible here.
This session deals with the management and process of taxonomic lists and retrieving information from vegetation-plot databases. Following software will be required for this session:
Though all required R-packages will be installed in the computer room
previous to this session, here there are instructions, in the case you
try to repeat the sessions in your own computer. Following command line
should be executed in order to install all required packages in
R
:
install.packages(c("devtools", "foreign", "plotKML", "rgdal", "sp", "vegdata",
"vegan", "xlsx"), dependencies=TRUE)
This session will focus on the use of two packages that use GitHub as repository for
sharing and development. The package taxlist
processes taxonomic lists that may be connected to biodiversity
information (e.g. vegetation relevés), while the package vegtable
has
a similar task but handling vegetation-plot records. Since
vegtable
is depending on taxlist
, both
packages should be installed and updated for this session:
library(remotes)
install_github("ropensci/taxlist", build_vignettes=TRUE)
install_github("kamapu/vegtable", build_vignettes=TRUE)
Since both package are been uploaded in the Comprehensive R Archive Network (CRAN), you can also use the common way to get the last released version.
install.packages("vegtable", dependencies=TRUE)
This is the main document of the Workshop and will be distributed
among its participants. Additionally, data sets required for the
examples are provided by the installation of the R-packages
taxlist
and vegtable
.
The package taxlist
define an homonymous S4
class. Among other properties, S4
classes implement a
formal definition of objects, tests for their validation and functions
(methods) applied to them. Such properties are suitable for structure
information of taxonomic lists and for data handling.
This class is composed by four slots, corresponding to
column-oriented tables (class data.frame
in
R). While the slots taxonNames
and
taxonRelations
contain the information on names (labels of
taxon concepts) and taxon concepts, respectively, representing the core
information of taxonomic lists.
most important information regarding taxa (taxon concepts),
information in slots taxonTraits
and
taxonViews
is optional and they can be empty.
Building taxlist
objects can be achieved either through
a step-by-step routine working with small building blocks or by
importing consolidated data sets.
Objects of class taxlist
can be just built from a list
of names formatted as a character vector. To achieve it, the functions
new
and add_concept
are required.
plants <- c("Lactuca sativa","Solanum tuberosum","Triticum sativum")
splist <- new("taxlist")
splist <- add_concept(splist, plants)
summary(splist)
In the case of consolidated species lists, they can be inserted in a
table previously formatted for taxlist
(or
Turboveg). Such tables can include both, accepted names
for plant species as well as their respective synonyms. A formatted
table is installed in the package taxlist
.
File <- file.path(path.package("taxlist"), "cyperus", "names.csv")
Cyperus <- read.csv(File, stringsAsFactors=FALSE)
str(Cyperus)
This table was originally imported form Turboveg and
inherits some of its default columns. Four columns were renamed and are
mandatory for process the table with the function
df2taxlist
:
TaxonUsageID
is the key field of a table of names (slot
taxonNames
).TaxonConceptID
is the link of the name to its
respective concept (key field in slot taxonRelations
.TaxonName
is the name itself as character.AuthorName
is a character value indicating the
author(s) of the respective name.Additionally the functions require a logical vector indicating, which
of those names should be considered as synonyms or accepted names for
its respective concept. This information is provided by the inverse
value of the column SYNONYM
in
Turboveg.
Cyperus <- df2taxlist(Cyperus, !Cyperus$SYNONYM)
summary(Cyperus)
object size: 32.1 Kb
validation of 'taxlist' object: TRUE
number of taxon usage names: 95
number of taxon concepts: 42
trait entries: 0
number of trait variables: 0
taxon views: 0
A direct import from a Turboveg database can be done
using the function tv2taxlist
. The installation of
vegtable
includes data published by Fujiwara et
al. (2014) with vegetation plots collected in forest stands
from Kenya.
library(vegtable)
tv_home <- file.path(path.package("vegtable"), "tv_data")
Fujiwara <- tv2taxlist("Fujiwara_sp", tv_home)
summary(Fujiwara)
For taxlist
objects there are basically two ways for
displaying summaries. The first way is just ot apply the function
summary
to a taxlist
object:
summary(Cyperus)
object size: 32.1 Kb
validation of 'taxlist' object: TRUE
number of taxon usage names: 95
number of taxon concepts: 42
trait entries: 0
number of trait variables: 0
taxon views: 0
Note that this option also executes a validity check of the object,
as done by the function validObject
.
The second way is getting information on a particular concept by adding it concept ID or a taxon usage name as second argument in the function.
summary(Cyperus, "Cyperus dives")
------------------------------
concept ID: 197
view ID: none
level: none
parent: none
# accepted name:
197 Cyperus dives Delile
# synonyms (5):
52000 Cyperus immensus C.B. Clarke
52600 Cyperus exaltatus var. dives (Delile) C.B. Clarke
52601 Cyperus alopecuroides var. dives Boeckeler
52602 Cyperus immensus var. petherickii (C.B. Clarke) Kük.
52603 Cyperus petherickii C.B. Clarke
------------------------------
Specific information for several concepts can be displayed by using a
vector with several concept IDs. Through the option
ConceptID="all"
it is also possible to get a display of the
first concepts, depending on the value of the argument
maxsum
.
summary(Cyperus, "all", maxsum=10)
The installation of taxlist
includes the data
Easplist
, which is formatted as a taxlist
object. This data is a subset of the species list used by the database
SWEA-Dataveg (GIVD ID
AF-006):
The common ways to access to the content of slots in S4
objects are either using the function slot(object, name)
or
the symbol @
(i.e. object@name
). Additional
functions, which are specific for taxlist
objects are
taxon_names
, taxon_relations
,
taxon_traits
and taxon_views
(see the help
documentation).
Additionally, it is possible to use the methods $
and
[
, the first for access to information in the slot
taxonTraits
, while the second can be also used for other
slots in the object.
acropleustophyte chamaephyte climbing_plant facultative_annual
8 25 25 20
obligate_annual phanerophyte pleustohelophyte reed_plant
114 26 8 14
reptant_plant tussock_plant NA's
19 52 3576
Methods for the function subset
are implemented in order
to generate subsets of the content of taxlist
objects. Such
subsets usually apply pattern matching (for character vectors) or
logical operations and are analogous to query building in relational
databases. The subset
method can be apply to any slot by
setting the value of the argument slot
.
Or the very same results:
Similarly, you can look for a specific name.
Objects belonging to the class taxlist
can optionally
content parent-child relationships and taxonomic levels. Such
information is also included in the data Easplist
, as shown
in the summary output.
summary(Easplist)
object size: 761.4 Kb
validation of 'taxlist' object: TRUE
number of taxon usage names: 5393
number of taxon concepts: 3887
trait entries: 311
number of trait variables: 1
taxon views: 3
concepts with parents: 3698
concepts with children: 1343
hierarchical levels: form < variety < subspecies < species < complex < genus < family
number of concepts in level form: 2
number of concepts in level variety: 95
number of concepts in level subspecies: 71
number of concepts in level species: 2521
number of concepts in level complex: 1
number of concepts in level genus: 1011
number of concepts in level family: 186
Note that such information can get lost once applied
subset
, since the respective parents or children from the
original data set are not anymore in the subset. May you like to recover
parents and children, you can use the functions get_paretns
or get_children
, respectively.
summary(Papyrus, "all")
------------------------------
concept ID: 206
view ID: 1
level: species
parent: none
# accepted name:
206 Cyperus papyrus L.
# synonyms (2):
52612 Cyperus papyrus ssp. antiquorum (Willd.) Chiov.
52613 Cyperus papyrus ssp. nyassicus Chiov.
------------------------------
Papyrus <- get_parents(Easplist, Papyrus)
summary(Papyrus, "all")
------------------------------
concept ID: 206
view ID: 1
level: species
parent: 54853 Cyperus L.
# accepted name:
206 Cyperus papyrus L.
# synonyms (2):
52612 Cyperus papyrus ssp. antiquorum (Willd.) Chiov.
52613 Cyperus papyrus ssp. nyassicus Chiov.
------------------------------
concept ID: 54853
view ID: 2
level: genus
parent: 55959 Cyperaceae Juss.
# accepted name:
54855 Cyperus L.
------------------------------
concept ID: 55959
view ID: 3
level: family
parent: none
# accepted name:
55961 Cyperaceae Juss.
------------------------------
Objects of class vegtable
are complex objects attempting
to handle the most important information contained in vegetation-plot
databases. They are defined in the homonymous package, while the species
list will be required as a taxlist
object.
As already mentioned, S4
objects are composed by
slots:
The information contained in the slot is the following:
description
is containing the “metadata” of the
database as a named character vector.samples
is a column oriented table with the
observations of species (as usage IDs) in plots and layers.header
is a table containing information on the
plots.species
is the taxonmic list as taxlist
object.relations
are tables linked to categorical variables in
slot header
(called popups
in
Turboveg).coverconvert
is a list with conversion tables among
cover scales.An empty object (prototype) can be generated by using the function
new
:
Veg <- new("vegtable")
It is also possible to build an object from cross tables, assuming all species included in the table are accepted names:
File <- file.path(path.package("vegtable"), "Fujiwara_2014", "samples.csv")
Fujiwara <- read.csv(File, check.names=FALSE, stringsAsFactors=FALSE)
Fujiwara_veg <-df2vegtable(Fujiwara, 1, 2)
summary(Fujiwara_veg)
The package vegtable
also provides own data sets as
examples. One of those examples is dune_veg
, which contains
the data sets dune
and dune.env
from the
package vegan
.
Other example is Kenya_veg
, which also represent a
subset of the database SWEA-Dataveg (GIVD ID
AF-00-006).
## Metadata
db_name: Sweadataveg
sp_list: Easplist
dictionary: Swea
object size: 9501 Kb
validity: TRUE
## Content
number of plots: 1946
plots with records: 1946
variables in header: 34
number of relations: 3
## Taxonomic List
taxon names: 3164
taxon concepts: 2392
validity: TRUE
As shown in the previous commands, the function summary
displays a general overview of the object’s content, while it also runs
a validity check. An additional display is offered by the function
vegtable_stat
:
vegtable_stat(Kenya_veg)
## Metadata
db_name: Sweadataveg
sp_list: Easplist
dictionary: Swea
object size: 9501 Kb
validity: TRUE
## Content
number of plots: 1946
plots with records: 1946
variables in header: 34
number of relations: 3
## Taxonomic List
taxon names: 3164
taxon concepts: 2392
validity: TRUE
REFERENCES
Primary references: 5
## AREA
Area range (m^2): 150 - 1750
<1 m^2: 0%
1-<10 m^2: 0%
10-<100 m^2: 0%
100-<1000 m^2: 2%
1000-<10000 m^2: 1%
>=10000 m^2: 0%
unknow: 97%
## TIME
oldest: 1983 - youngest: 2014
<=1919: 0%
1920-1929: 0%
1930-1939: 0%
1940-1949: 0%
1950-1959: 0%
1960-1969: 0%
1970-1979: 0%
1980-1989: 36%
1990-1999: 31%
2000-2009: 2%
2010-2019: 1%
unknow: 30%
## DISTRIBUTION
KE: 100%
## PERFORMANCE
01: 83%
02: 17%
03: 0%
04: 0%
05: 0%
06: 0%
07: 0%
08: 0%
09: 0%
10: 0%
11: 0%
12: 0%
Some functions are provided for the access to slots of
vegtable
objects, namely header
and
veg_relation
, taxon_traits
and
taxon_views
(see the help documentation).
For a direct access of the content included in slot
header
, there are the methods $
and
[
:
summary(Kenya_veg$REFERENCE)
veg_relation(Kenya_veg, "REFERENCE")[,1:3]
As a way to generate queries from vegtable
objects, a
method for the function subset
is implemented as well.
Following is the case of the data set dune_veg
generating a
subset with pasture plots.
## Metadata
object size: 22.2 Kb
validity: TRUE
## Content
number of plots: 20
plots with records: 20
variables in header: 6
number of relations: 0
## Taxonomic List
taxon names: 30
taxon concepts: 30
validity: TRUE
summary(pasture)
## Metadata
object size: 19.4 Kb
validity: TRUE
## Content
number of plots: 5
plots with records: 5
variables in header: 6
number of relations: 0
## Taxonomic List
taxon names: 30
taxon concepts: 30
validity: TRUE
Most of the applications dealing with analysis of vegetation tables
(e.g. the package vegan
) will deal with cross tables rather
than with column-oriented database lists. Therefore a function
crosstable
is defined for the conversion of
taxlist
objects to cross tables. Before building a cross
table it will be recommended to transform the cover into percentage
value by using the function transform
.
pasture <- cover_trans(pasture, to="Cover", rule="middle")
Cross <- crosstable(Cover ~ ReleveID + AcceptedName, pasture, max,
na_to_zero=TRUE)
Cross
The first argument in the function is the formula
y ~ x1 + x2 + ... + xn
, where:
y
is the cover value (a column in slot
samples
).x1
is the plot ID or a grouping value from slot
header
which will be used as columns in the cross
table.x2
is the output for species (either
TaxonName
or AcceptedName
).x3
to xn
are additional information for
the rows, such as layer, etc.The second argument is the vegtable
object and the third
is a function, which will be applied in the case of multiple values in a
cell (e.g. multiple records of a species within a plot).
Note that there is also a method dealing with class
data.frame
(see the help documentation).
dune_veg <- cover_trans(dune_veg, to="Cover", rule="middle")
Cross_use <- crosstable(Cover ~ Use + AcceptedName, dune_veg, sum,
na_to_zero=TRUE)
Cross_use
While the previous commands show statistics of groups in a vegetation table, it is also possible to write a presence-absence table:
Cross_pa <- crosstable(Cover ~ ReleveID + AcceptedName, pasture, function(x) 1,
na_to_zero=TRUE)
Cross_pa
The original task of vegtable
was the import of
Turboveg databases in R. Thus one of the oldest
functions in the package is tv2vegtable
.
tv_home <- file.path(path.package("vegtable"), "tv_data")
Fujiwara <- tv2vegtable("Fujiwara_2014", tv_home)
zero values will be replaced by NAs
summary(Fujiwara)
## Metadata
db_name: Fujiwara_2014
sp_list: Fujiwara_sp
dictionary: Swea
object size: 392.8 Kb
validity: TRUE
## Content
number of plots: 56
plots with records: 56
variables in header: 29
number of relations: 2
## Taxonomic List
taxon names: 301
taxon concepts: 254
validity: TRUE
In the case there is an error message produced by inconsistent data structure, there is the possibility of importing the data as a list for further inspection.
Fujiwara_list <- tv2vegtable("Fujiwara_2014", tv_home, output="list")
Export of data for Juice follows a similar procedure as for cross tables.
write_juice(Fujiwara, "RiftValley",
COVER_CODE ~ ReleveID + AcceptedName + LAYER,
header=c("TABLE_NR","NR_IN_TAB","ALTITUDE","INCLINATIO","EXPOSITION",
"COMM_TYPE"),
FUN=paste0)
Besides the vegtable
object, the most important elements
to export are:
formula
indicating how to construct the cross
table.header
to show which variables should be exported for
the head data.coords
with the names of coordinates, if present.FUN
the function used for aggregate multiple
records.The values for coords
should be as decimal degrees and
in the spatial system WGS 84. In Juice such information
will be included in the header data as columns deg_lon
and deg_lat, allowing the display of plots in
Google Earth. The previous command has generated two
files, namely RiftValley_table.txt and
RiftValley_header.txt.
For importing the table, you may follow the steps:
File -> Import -> Table -> From Spreadsheet File (e.g. EXCEL Table)
Open
Braun-Blanquet Codes
and Finish
Now for the head:
File -> Import -> Header Data -> From Comma Delimited File
Open
For geo-referenced plots, you can easily map the location of the
plots, for instance using the package leaflet
.
Then, hopefully you enjoyed the session and did not get too much error messages.
20-06-2022
This workshop was supported by the project GlobE-wetlands.