Overview of the vegsoup-data repository

A repository of plant species co-occurrence data (vegetation data, phytosociological relevés).

This is our vegetation database (spatial and temporal explicit species co-occurrence data). Essentially, it serves as an incubator for collecting data on vegetation plots. Currently, the repository contains a bunch of unpublished data sets, as well as computerized data taken from the literature. 3,901 genuine relevés were observed by R. Kaiser and/or T. Eberl and 2,587 relevés originate from literature sources and were digitized by R. Kaiser. 1,636 relevés are provided by M. Staudinger. 8,124 relevés are available, involving 2,715 and 1,311 taxa (including bryophytes and lichens). Up to now, R. Kaiser, M. Staudinger, and T. Eberl are the main contributors. You are welcome to join the project!

The bibtex file ./refernces.bib that is contained in each project folder gives intellectual property rights for a particular data set (see also section license). The URL field provides a link to a PDF-file in case of a literature source.

How to access and use the data sets

Data sets containing a file named transcript.txt are ready to be used with the vegsoup R-package. To install this package from within R, type:

# if needed
install.packages("devtools")

library(devtools)

install.packages("vegsoup", repos="http://R-Forge.R-project.org", type = "source")

#   github mirror
install_github("rforge/vegsoup/pkg")

To load a data set into an R session, you may either download a particular *.rda file and attach it to your R-session (first navigate to the respective *.rda file, then right-click on the file, a further click on View Raw will download the file.), or load the data directly from inside an R session:

# if needed
install.packages("RCurl")

library(vegsoup)
library(RCurl)  

URL <- paste0(
	"https://raw.githubusercontent.com/",  # to accesses the raw files
	"kardinal-eros/vegsoup-data/master/",  # this is the repository
	"barmstein%20dta/bs.rda"               # folder and *.rda file
)
load(rawConnection(getBinaryURL(URL)))

A condensed object containing all available data in this repository is contained in the ./mirror folder as mirror.rda.

Licence

This work is licensed under a Creative Commons licence Attribution - NonCommercial - ShareAlike 4.0 International.

We ask to contact the maintainer if you plan to use any data from this repository in a publication that is marked as unpublished. See the @unpublished tag in the references.bib file.

Species designation and taxonomic concepts

All data sets link to a standard list (for further details see here) or use the Austrian setup of the Turboveg data base system.

Notes about computerization of literature data

Any project developed from a published source can be seamlessly built using OCR transcripts. First, we supply PDF files of the (table) sources along with their OCR transcripts. This way, it is possible to scrutinize all steps involved in digitizing the data. All data manipulation steps are documented within a Make file (MakeVegsoup.R). Concerning the interpretation of taxonomic concepts, a table is supplied that translates the taxa in the publication to a reference list (see the translate*.csv files). The URL tag in the references.bib file provides a (stable) link to a full PDF version of the publication – not just the tables and related material used for OCR. A scanned version of the document is stored with the project if no free PDF version is available.

Bibliographic information

Each project folder contains a references.bib file that contains bibliographic information related to a data set. Creating and maintaining bibliographic information for the whole repository is easy using the efficient BibTex format and its readable syntax.

Sampling protocol and data standards

This is a brief summary of the sampling protocol applied by R. Kaiser & T. Eberl and descriptive attributes collected with each sampling unit (plot or relevé) as available in the vegsoup-data repository. Of course, this does not apply to data sets taken from the literature.

Sampling procedure

Different sampling procedures are applied depending on the scope of a particular project.

• Landscape Sampling uses area stratification. That means a survey area is divided into strata (forest, meadow, mire, etc.), and samples within a stratum are replicated depending on the areal extent of the stratum. We always aim to sample all vegetation types in a given survey area that are discernible in the field. The size of the surveyed landscape is typically in the range of 5 to 50 hectares.

• In Type Specific Sampling, we search a possibly large survey area for a specific vegetation type (e.g. ravine forest or rock shrubbery).

• In Species Specific Sampling we sample the vegetation where a particular (rare) species occurs.

• In Rapid Biodiversity Sampling we collect instances of (all) vegetation types in an area as we walk (straight) through the landscape. In this way, we typically do not collect area-dependent replicates.

• Rock Vegetation Sampling is performed by using mountaineering equipment (rope and climbing harness), occasionally free-climbing. Rock vegetation or forest of steep slopes is sampled using a rope along a vertical transect. Usually, 3 to 4 plots are sampled along a 100 m long rope. The first plot is typically sampled just below the rope betray point. The last relevé is taken at the end of the rope. Depending on how long the rappelling (abseiling) needs to be, non-overlapping plots are taken along this line, with gaps between sampling units of at least half the plot size applied.

• Systematic sampling is performed using a rectangular grid or along a (measure tape) transect. We do not give much consideration to homogeneity or other subjective criteria concerning a sample plot, and sampling locations are typically randomly selected. In the case of rock vegetation, the vertical transect's location is constrained by a suitable belay point that can be reached with reasonable effort. Additionally, the danger of falling rocks is significant. Therefore, the vegetation sampled in this context cannot be subjectively selected, and sampling can be considered random.

Estimation scales

We typically apply the 9-point modification of the classical 7-point Braun-Blanquet by Barkman et al. (1964). See Roberts & Peet (2013) for details.

Plot sizes

We use strictly equal-sized plots with edge lengths that are powers of 2 (e.g. 1 × 1, 2 × 2, 4 × 4, 8 × 8, 16 × 16). Grasslands (including alpine vegetation and tundra), marshes, fens and mires are sampled using 16m² area plots (4 m × 4m); scrubs and rock shrubbery use 64 m² plots (8m × 8m) and woodlands are sampled within plots of 256 m² (16 m × 16m). Spring vegetation is sampled using small plots of size 1 m² (1 m × 1 m).

Whenever possible, we use square plots. We record both edge lengths of the plot instead of noting the plot area. The shape of non-square plots can then be estimated by dividing edge max by edge min. In some rare cases (e.g. forests on rock cliffs), it is necessary to switch to elongated plot forms (e.g. 10 × 26 m ≈ 256 m² instead of 16 × 16 m).

Detailed forest surveys use circular plots with a radius of 11m (ca. 380 m²).

Attributes of the sampling units (relevés) that are recorded in the field

Principal standards of the phytosociological relevé follow the notation of Mucina et al. (2000).

Data on field record

plot designation of sample plot (relevé)
date date of sampling (ISO 8601, yyyy-mm-dd)
observer name(s) of the author(s) of the relevé
alliance provisional classification into syntaxon
association optional
plsx edge length of plot, parallel to the hillside
plsy edge length of plot, orthogonal to a hillside
plras radius of a circular plot

Geographic data

location locality, topographic name
expo exposition, aspect
slope slope, inclination

Other fields, such as country, province (district), the nearest village, and altitude, can be obtained with the reverseGeocode function.

Geographic coordinates

latitude coordinate of latitude in decimal degrees (decimal is ».«) based on the WGS 1984 ellipsoid (EPSG:4326), a minus sign or a padding letter »S« and »N« means south and north of the equator, respectively (e.g. 12.345678S or -12.345678).
longitude coordinate of longitude, a minus sign or padding letter »W« and »E» means west and east of the Greenwich zero meridian, respectively (e.g. 87.654321W or -87.654321)
accuracy coordinate uncertainty (precision) in meters

Data on vegetation

We supply the stratum/layer notation as defined in Mucina et al. 2000 (E₀, E₁, E₂, E₃)

Cover of vegetation

cov total cover, total projection of standing vegetation

Cover of tree layer in % (E₃)

htl cover of tree layer (E₃)
t1cov cover of canopy layer (E₃ γ)
t2cov cover of sub-canopy layer (E₃ β)
t3cov cover of lower tree layer layer (E₃ α)

Cover of shrub layer in % (E₂)

scov cover of shrub layer (E₂) s1cov cover of upper shrub layer layer (E₂ β) s2cov cover of lower shrub layer layer (E₂ α)

Cover of herb layer in % (E₁)

hcov cover of herb layer (E₁)

We don't estimate the cover of any sub-strata in the terrestrial herb layer (upper, middle and lower herb layer E₁ α, E₁ β, and E₁ γ, respectively), but we do so for aquatic vegetation:

ncov cover of natant plants (E₁n)
ucov cover of submerged plants (E₁s)

Cover of cryptogam layer in % (E₀)

mcov cover of cryptogam layer (E₀)

Cover of non-vegetated ground in %

litter cover of litter
rock cover of rocks
bare cover of bare soil
soil cover of debris
water cover of free water

Height of vegetation

Height of tree layer in meters, woody plants over 6 m tall or 12 cm in diameter (E₃)

htl height of tree layer (E₃)
htl0 height of emergent layer (E₃ δ), woody plants with crowns towering above the canopy layer (e.g. a tall coniferous tree overtopping the canopy of decidous trees)
htl1 height of canopy layer (E₃ γ)
htl2 height of sub-canopy layer (E₃ β)
htl3 height of lower tree layer (E₃ α)

Height of shrub layer in meters, woody plants ranging 0.5 - 6 m (E₂)

hsl height of shrub layer (E₂), woody plants ranging 0.5 - 6 m. s

We do not distinguish a lower (E₂ α) and upper shrub layer (E₂ β)

Height of herb layer in meters, non-woody phanerogams (E₁)

hhl height of herb layer (E₁)
hhl1 height of upper herb layer (E₁ γ)
hhl2 height of lower herb layer (E₁ α)

We don't distinguish a middle herb layer (E₁ β).
Woody plants that compete with non-woody phanerogams are attributed to the herb layer (juvenile trees).
The height of the cryptogam layer (E₀) is not estimated/measured.

Data on habitat and Notes

comment optional notes
geology type of (bed-)rock
hydrology water regime
soil type and/or texture of the soil

References

M Chytrý & Z Otýpková (2003). Plot sizes used for phytosociological sampling of european vegetation. Journal of Vegetation Science, 14(4):563–570.

L Mucina, J Schaminee, & J Rodwell (2000). Common data standards for recording relevés in field survey for vegetation classification. Journal of Vegetation Science, 11: 769-772.

RK Peet & DW Roberts (2013). Classification of natural and semi-natural vegetation. In Vegetation Ecology (E van der Maarel, & J Franklin, eds.). Wiley-Blackwell.