Distribution of the invasive plant species Heracleum sosnowskyi Manden... 71
Distribution of the invasive plant species Heracleum
sosnowskyi Manden. in the Komi Republic (Russia)
Ivan Chadin1, Igor Dalke1, Ilya Zakhozhiy1, Ruslan Malyshev1, Elena Madi1,
OlgaKuzivanova1, Dmitrii Kirillov1, Vladimir Elsakov1
1 Institute of Biology of Komi Scientic Centre of the Ural Branch of the Russian Academy of Sciences, Kom-
munisticheskaya, 28, 167982, Syktyvkar, Komi Republic, Russian Federation
Corresponding author: Ivan Chadin (firstname.lastname@example.org)
Academic editor: Pavel Stoev|Received 14 November 2016|Accepted 27 February 2017|Published 9 March2017
Citation: Chadin I, Dalke I, Zakhozhiy I, Malyshev R, Madi E, Kuzivanova O, Kirillov D, Elsakov V (2017) Distribution
of the invasive plant species Heracleum sosnowskyi Manden. in the Komi Republic (Russia). PhytoKeys 77: 71–80. https://
Resource citation: Chadin I, Dalke I, Zakhozhiy I, Malyshev R, Madi E, Kuzivanova O, Kirillov D (2016) Occurrences
of the invasive plant species Heracleum sosnowskyi Manden. in the Komi Republic (Russia). v. 1.8. Institute of Biology
of Komi Scientic Centre of the Ural Branch, Russian Academy of Sciences. Dataset/Occurrence. http://ib.komisc.
Occurrences of the invasive plant species Heracleum sosnowskyi Manden. in the Komi Republic (northeast-
ern part of European Russia) were recorded and published in the Global Biodiversity Information Facil-
ity (GBIF http://www.gbif.org) using the RIVR information system (http://ib.komisc.ru/add/rivr/en).
RIVR stands for “Rasprostranenie Invasionnyh Vidov Rastenij” [Occurrence of Invasion Plant Species].
is citizen science project aims at collecting occurrence data about invasive plant species with the help
of citizen scientists. Information can be added by any user after a simple registration (concept) process.
However, the data published in GBIF are provided only by professional scientists. e total study area
is approximately 19,000 km2. e GBIF resource contains 10894 H.sosnowskyi occurrence points, each
with their geographical coordinates and photographs of the plants in the locus of growth. e preliminary
results of species distribution modelling on the territory of European North-East Russia presented.
Occurrence, human observation, Heracleum sosnowskyi, hogweed, invasive, geotagged photographs, Komi
Republic, European North-East Russia
PhytoKeys 77: 71–80 (2017)
Copyright Ivan Chadin et al. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0),
which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Launched to accelerate biodiversity research
A peer-reviewed open-access journal
Ivan Chadin et al. / PhytoKeys 77: 71–80 (2017)
“Ecophysiological modelling of invasive plant species distribution. e case of Hera-
cleum sosnowskyi in the taiga zone of the European part of Russia”
e project was supported by a grant of the Russian Foundation for Basic Research and
the Government of Komi Republic (Project No 16-44-110694).
Study area description
e Komi Republic is located in the north-east of the Russian Plain and the western
slopes of the northern Ural Mountains. It is a large and an important biogeographic
boundary that separates the ora and fauna of two continents – Europe and Asia.
On the plain territory of the Komi Republic, a pronounced latitudinal-nature zo-
nation occurs. e extreme north-east is taken by a subzone of the southern tundra.
e forest-tundra is a transition zone between the tundra and taiga. In the Pechora
Province, it has a width of 100–120 km forming the southern periphery of the territory
that has the Bolshezemelskaya tundra. e main type of vegetation in the Republic of
Komi is the boreal (taiga) forest. e taiga zone is divided into the following subzones:
Extreme northern, Northern, Middle, and Southern. e eastern edge of the Republic
in occupied by the Ural Mountains, where altitudinal zonation occurs with distinct
Mountain forest, Alpine tundra, and Cold deserts zones (Gorchakovskij 1975).
A large part of the republic has a climate similar to that of the Atlantic-Arctic
region with a cold temperate (boreal) climate (Brattsev et al. 1997). e territory is a
zone of excessive moisture, widespread marshes, and wetlands. e annual precipitation
exceeds the evaporation and decreases from south to north, from 700 to 550 mm. A
signicant dierence in the climate is observed across the length of the republic from
south to north and from west to east. e duration of the winter in the south of the
republic is 170–180 days and that in the north is 230–250 days. e average tempera-
ture in January (the coldest month) in the south is 15 °C whereas that in the north-east
is –22 °C. Summers are short and warm; the average temperature in July (the warmest
month) is approximately 10°C in the north-east and 17°C in the south. e prevailing
wind directions in winter are south and south-west, and north in summer. e monthly
average wind speed in the taiga zone is 3–4 m/s and that in the tundra area is 6.5 m/s.
Biological diversity of the Komi Republic region includes 929 fungi, 1217 vascu-
lar plants, 653 moss, 1020 lichen, 2,000 algae, more than 3,500 arachnid, more than
6,000 insect, 50 sh, six amphibian, ve reptile, 265 bird, and 57 mammal species.
ere are 237 forest, oristic, meadow, marsh, ichthyological, ornithological, and geo-
Distribution of the invasive plant species Heracleum sosnowskyi Manden... 73
logical reserves and natural monuments on the territory of Komi. e Pechora-Ilych
State Reserve and the Yugyd Va National Park occupy 13.5% of the total territory of
the republic (Ponomarev and Tatarinov 2012).
e project design combines an experimental approach and analysis of results of the
observations. e responses of H.sosnowskyi plants to the changes in the abiotic envi-
ronmental parameters were obtained by instrumental measurements of the morpho-
logical and physiological parameters (including CO2/H2O gas exchange, chlorophyll
uorescence, and heat dissipation) in the plants grown in climatic chambers and ex-
perimental plots. e data of the optimal and critical values of the environmental
factors (heat, light, rainfall, and soil) required for the survival and reproduction of the
plants were used for a joint analysis along with the geographically referenced data of
these factors. e results were arranged in a raster map showing the potential areas of
H.sosnowskyi. e resulting map was veried by a direct comparison with the data of
the eld observations of the habitats of this species and with the correlation simulation
of their geographical distribution.
Data published through
General taxonomic coverage description
e resource contains occurrence data only for one species – H. sosnowskyi Manden.
Species: Heracleum sosnowskyi
Common names: Sosnowsky’s hogweed, plants, vascular plants, owering plants, carrot
Ivan Chadin et al. / PhytoKeys 77: 71–80 (2017)
General spatial coverage
e geographical coverage is essentially limited to the Komi Republic territory located in
the European part of Russia. Currently, all populations of H. sosnowskyi in this area are in-
vasive. is species was introduced into this region in the second half of the 20th century
as a forage crop. Since 2012 varieties of this species are excluded from the register of the
breeding achievements of the Russian Federation (Ocial bulletin 2012; http://gossort.
com/bullets/pdf/bull_176.pdf ). is species is also included in the “specialised catalogue of
weeds” (Information letter 2015; http://antibor.ru/sites/526a0b00d7e1e49744000002/
59°22.48'N and 66°7.12'N Latitude; 48°56.24'E and 60°20.24'E Longitude
28 July 2012 - 23 August 2016
Photographs of plants were taken using consumer cameras. Videos were recorded with
a Car DVR Camera (video 1280×960 pixels at 30 frames/second), mounted on the car
windshield (height from the road surface was 170 cm). e survey was conducted at
speeds of 60–90 km/h. e GPS track was simultaneously recorded with GPS naviga-
tors. e time on the cameras and video recorders were synchronised with the time
displayed on the GPS navigation device.
All the images were geotagged by a GPS track log with “GPS Correlate” soft-
ware (v 1.6.1, https://github.com/freefoote/gpscorrelate) according to the methods
described in the OpenStreetMap Project documentation (Geotagging Source Photos
2016; http://wiki.openstreetmap.org/wiki/Geotagging_Source_Photos). e video
les were broken into frames (one frame per second) and the frames were saved as
“jpeg” les with the program FFmpeg (v 3.1.4 http://www.mpeg.org) followed by
geotagging of these les similar to that of the photographs. e array of images was
hand sorted into two groups: images that contained H.sosnowskyi plants and images
Distribution of the invasive plant species Heracleum sosnowskyi Manden... 75
without these plants. e coordinates of the photographs obtained from a Car DVR
Camera were corrected in the Quantum GIS Geographic Information System (QGIS)
program (v2.16.3 http://www.qgis.org, QGIS Development Team 2016) by shifting
the group of points on the side of the road. All geotagged H. sosnowskyi images were
uploaded to the online database “Occurrence of invasive plant species Heracleum sos-
nowskyi Manden.” (RIVR 2016).
Study extent description
e occurrence data of H.sosnowskyi were collected from an area of approximately
19,000 km2 (Figure 1). Most of the data were collected from the capital area of Komi,
Syktyvkar (61°39.95'N,50°49.53'E) as well as along the roads at a distance of 300km
from Syktyvkar, the directions of which coincide with the ow direction of the major
rivers Vychegda and Sysola belonging to the Northern Dvina basin. A separate clus-
ter of the data was collected from a 664 km (orthodromic) distance in the territory
and suburb of Inta city, located near the Arctic Circle (66° 1.87’N, 60° 8.72’E). A
pronounced sampling bias should be considered before using the data for the species
distribution modelling. Data were collected close to the settlements or the roads con-
necting them, which is a travel time bias (Fourcade et al. 2014). In the case of H.sos-
nowskyi, such a sampling bias may coincide with the actual factors determining the
dispersal of the plants of this species. In most cases, roadsides are the optimal habitats
for this species as they are open and well-lighted with adequate moisture due to the
roadside drainage systems. Moreover, the air ow creates favourable conditions for the
spread of the plants.
e occurrence data consist of the presence data only. Two methods were used for the
creation of occurrence records, which include the data collection along transects (7130
points) and mapping of H. sosnowskyi boundaries that were later converted to regular
points sample (3764 points). e regular points sample coordinates were generated us-
ing the QGIS Desktop software (v 2.16.3). e points were created with a 25 m point
spacing within polygon layers that indicated the H.sosnowskyi population boundaries.
e occurrences were labelled with a tag “Generated Regular Sample” written in the
“occurrence remarks” eld. e “associated media” eld contained the URL of the
locality map showing the generated point pattern with the scale bar and the north end
on top of the map.
Data along transects were collected by recording a video of H.sosnowskyi plants
growing along the roadsides and by taking photographs in the direction perpendicular
to the road at a distance of up to 5 km.
Ivan Chadin et al. / PhytoKeys 77: 71–80 (2017)
Figure 1. Study area. Red points indicate occurrences of Heracleum sosnowskyi described in the data paper.
Quality control description
e published data collected by professional scientists with sustainable skills for the
identication of H.sosnowskyi and its dierences from other similar species in its habi-
tats were published in GBIF whereas that collected by volunteers were accumulated in
the RIVR system. Before publication, data were checked for gross errors in georefer-
encing by visual inspection of the overlay points on the map with the borders of Rus-
sian regions in OpenStreet in the QGIS Desktop.
e presence of duplicate records was checked by running a special SQL script. e
records were counted as duplicated if three elds were the same: the coordinates, the date
of the event, and the le name of the photograph. For many data points (1080 of 10894
points, 10%), the same dates and coordinates were detected; however, they presented a se-
ries of photographs (2 to 13). ese data were saved in the system as they could be of interest
for the assessment of the landscape and the evaluation of plants in the H. sosnowskyi habitat.
Species distribution modelling
e described dataset was used for H. sosnowskyi species distribution modelling (SDM).
e SDM was performed for two plots. Plot 1 was a rectangular, limited by latitudes:
Distribution of the invasive plant species Heracleum sosnowskyi Manden... 77
61.0088°N, 62.1387°N and longitudes: 49.5013°E, 51.5941°E. e area of Plot 1 was
9 180 km2. e Plot 2 was a rectangular, limited by latitudes: 57.0000°N, 70.0000°N,
42.0000°E, 68.0000°E. e area of Plot 2 was 1 857 586 km2. All coordinates were
given in the WGS84 projection (EPSG: 4326).
Two groups of predictors were used. Group 1: the state of the earth’s surface, with
a spatial resolution of 1 second (≈ 30 m) per pixel (data was collected for Plot 1only):
VEG = vegetation cover map derived from classication of satellite images (20 classes);
ROAD = proximity map to the nearest road; AGRO = proximity map to the near-
est borders of agricultural areas. Group 2: bioclimatic variables are derived from the
monthly temperature and rainfall values obtained from WorldClim (Hijmans et al.
2005; http://www.worldclim.org/bioclim) with resolution of 30 second (≈ 1000 m)
per pixel (data was collected for Plot 1 and Plot 2): BIO1 = Annual Mean Temperature;
BIO2 = Mean Diurnal Range; BIO3 = Isothermality; BIO4 = Temperature Seasonal-
ity; BIO5 = Max Temperature of Warmest Month; BIO6 = Min Temperature of Cold-
est Month; BIO7 = Temperature Annual Range (BIO5-BIO6); BIO8 = Mean Tem-
perature of Wettest Quarter; BIO9 = Mean Temperature of Driest Quarter; BIO10
= Mean Temperature of Warmest Quarter; BIO11 = Mean Temperature of Coldest
Quarter; BIO12 = Annual Precipitation; BIO13 = Precipitation of Wettest Month;
BIO14 = Precipitation of Driest Month; BIO15 = Precipitation Seasonality (Coe-
cient of Variation); BIO16 = Precipitation of Wettest Quarter; BIO17 = Precipitation
of Driest Quarter; BIO18 = Precipitation of Warmest Quarter; BIO19 = Precipitation
of Coldest Quarter.
All data were obtained from open sources, either directly or as a result of raw data
processing in geographic information systems. e rights to use the Komi Republic
agriculture area map were acquired under a license agreement with the State Organiza-
tion “Syktyvkar Agrochemical Service Station”.
e presence data of H. sosnowskyi occurrences were obtained as a random sam-
ple of GBIF dataset described in this article. Five hundred randomly chosen presence
points were taken for modelling at Plot 1 and 1000 points for modelling at Plot 2.
Furthermore, 500 (for Plot 1) and 1000 (for Plot 2) randomly distributed points were
used as a background point.
SDM was performed with generalized linear multiple regression model in R (R
Core Team 2014) with dismo package (Hijmans et al. 2017).
Model tting with the predictors VEG, ROAD and AGRO showed statistically
signicant (p < 0.0001) relationship with the dependent variable (H. sosnowskyi pres-
ence in the given point). ROC analysis showed that AUC value for the regression
model was 0.92). ese results were supported by eld observations, invasion history
and ways of H. sosnowskyi seed dispersal. e plant occupies habitats with disturbed
soil cover, spreading rapidly along roads, due to the transfer of seeds by air ow, avoids
shaded and dry habitats (Fig. 2).
Model tting at Plot 1 and Plot 2 with bioclimatic predictors revealed a statis-
tically signicant relationship with eight predictors: BIO2, BIO4, BIO5, BIO6,
BIO7, BIO10, BIO12 and BIO17. e model with all these predictors showed
AUC value 0.99. Prediction with the model obtained within Plot 2 allowed to iden-
Ivan Chadin et al. / PhytoKeys 77: 71–80 (2017)
tify the putative northern H. sosnowskyi range boundary — 67.2000°N, within the
borders of the valley of the Pechora river (Fig. 3). According to the model, the values
of bioclimatic variables in the areas with maximum probability of H. sosnowskyi
presence were as follows (mean and standard deviation): BIO2: 8.3 ± 0.2 °C, BIO4:
112 ± 1 °C, BIO5: 21.2 ± 0.6 °C, BIO6: -21.9 ± 0.3 °C, BIO10: 3.6 ± 0.6 °C,
BIO12: 567 ± 24 mm.
e presence of H. sosnowskyi invasive plants in the northern forest-tundra sub-
zone (66.0000°N) was conrmed by eld observation on the territory of Inta city
(Komi Republic). H. sosnowskyi plants formed monostand and showed high enough
seed productivity (up to 12 000 seeds per plant) in this area.
Figure 2. e prediction map of Heracleum sosnowskyi habitats prepared with the species distribution
model based on vegetation cover map, nearest road proximity map, proximity map to the borders of agri-
cultural areas. e colour scale shows the probability H. sosnowskyi presence.
Distribution of the invasive plant species Heracleum sosnowskyi Manden... 79
Figure 3. e prediction map of Heracleum sosnowskyi habitats prepared with the species distribution
model based on bioclaimatic predictors. e borders of Plot 2 within which the model prediction was
made. e colour scale shows the probability H. sosnowskyi presence.
Object name: Darwin Core Archive Occurrences of the invasive plant species Heracleum
sosnowskyi Manden. in the Komi Republic (European North-East Russia)
Character encoding: UTF-8
Format name: Darwin Core Archive format
Format version: 1.0
Ivan Chadin et al. / PhytoKeys 77: 71–80 (2017)
Publication date of data: 2016-10-19
Licences of use: is work is licensed under a Creative Commons Attribution (CC-BY)
Metadata language: English
Date of metadata creation: 2016-09-07
Hierarchy level: Dataset
Brattsev AA, Brattsev AP, Taskaev AI (1997) Atlas Respubliki Komi po klimatu i gidrologii.
Drofa, 115 pp. [In Russian]
Ponomarev VI, Tatarinov AG (2012) Biologicheskoe raznoobrazie Respubliki Komi. Institut Bi-
ologii Komi NC UrO RAN, 264 pp. [In Russian]
Fourcade Y, Engler JO, Rödder D, Secondi J (2014) Mapping Species Distributions with
MAXENT Using a Geographically Biased Sample of Presence Data: A Performance As-
sessment of Methods for Correcting Sampling Bias. PLoS ONE 9(5): e97122. https://doi.
Gorchakovskij PL (1975) e ora of the Urals highlands. Nauka, Moscow, 283 pp. [In Russian]
Information letter (2015) Information letter of “Rosselhozcentr” from 26.11.2015 No.
1-8/2230. [In Russian]
Ocial bulletin (2012) Ocial bulletin of State Commission of the Russian Federation for
Selection Achievements Test and Protection FGBU “Gossortcommissiya”. MCX 6(176):
QGIS Development Team (2016) QGIS Geographic Information System. Open Source Geo-
spatial Foundation Project. http://www.qgis.org
RIVR (2016) Occurrence of Invasive Plant Species Heracleum Sosnowskyi Manden. http://
Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis (2005) A Very high resolution interpo-
lated climate surfaces for global land areas. International Journal of Climatology 25(15):
Hijmans RJ, Phillips S, Leathwick J, Elith J (2017) dismo: Species Distribution Modeling.
R Core Team (2014) R: A Language and Environment for Statistical Computing. R Foundation
for Statistical Computing, Vienna. https://www.r-project.org