Phenology and flower species availability define wild bee communities on river embankments
Data files
Jun 30, 2025 version files 1.01 MB
-
BeesOnDykesNetherlands.csv
709.83 KB
-
FlowerCS.csv
15.48 KB
-
FlowerIR.csv
124.47 KB
-
FlowerLS.csv
118.08 KB
-
LocationCoordinates.csv
3.39 KB
-
R_markup_Bees_Dykes_Netherlands.Rmd
36.34 KB
-
README.md
6.40 KB
Abstract
River embankments cover extensive areas in estuaries worldwide, yet their potential for pollinator conservation remains largely unexplored. Even more so than road verges, these embankments may offer widespread habitat for pollinators due to their size and sun-facing warm slopes. Here we present a comprehensive analysis of bee communities from 157 dike grasslands along the lower branches of the Meuse and Rhine rivers in the Netherlands in relation to vegetation composition and landscape characteristics. We encountered nearly half (156) of all Dutch bee species. Almost 10% of all individuals belonged to red-listed species, especially in early spring when we observed rich ground nesting bee communities with their associated kleptoparasites. Wild bee abundance and diversity increased with more flowering plant species present, but saturated at low flower species richness. This early saturation suggests that even modest conservation efforts could effectively alleviate floral resource limitation for pollinators. In contrast to the overall pattern, red-listed species occurrence increased with flower diversity without saturating effects and correlated stronger with landscape level factors. Together, these results highlight the unique conservational value of river embankments for wild bees and illustrate the unused potential at many dike grasslands worldwide.
Dataset DOI: 10.5061/dryad.ffbg79d5b
Description of the data and file structure
Files and variables
File: LocationCoordinates.csv
Description: This file holds all location codes and their respective RD new (EPSG:28992) coordinates.
Variables
- Location: code for the location
- x-coord: x coordinate of the location (EPSG:28992 - Amersfoort / RD new)
- y-coord: y coordinate of the location (EPSG:28992 - Amersfoort / RD new)
File: BeesOnDykesNetherlands.csv
Description: This file holds wild bee survey data and associated variables.
Variables
-
Location: code for the location used (the coordinates are provided in the LocationCoordinates.csv)
-
Year: the year the survey was taken
-
Date: the date of the survey (DD/MM/YYYY)
-
Species: the scientific species name
- Time.start: start time of the survey - in some occasions start time was not noted, and N/A was recorded in the data file. All surveys were always performed between 9 AM and 5 PM.
- Taxon: the taxon the specimen belonged to
- SamplingEffort(mins): minutes spend surveying the location (Minutes)
- Observer: person that collected the data
- UniqueIdentifier: Identifier based on the location + date per location
- StatusRl: TRUE or N/A - TRUE means species is on the national Dutch red list for wild bees
- Genus: genus the species belongs to
- SpeciesName: the species name. N/A when no bee observations during survey
- FlowerSpeciesRichness: number of flowering species present at time of the survey (#Flower Species on Dyke). N/A for few surveys of Ivo where data on flowers was not recorded.
- FlowerAbundance: Flower abundance. N/A for few surveys of Ivo where data on flowers was not recorded.\
IMPORTANT: observers used different, incomparable methods to scale flower abundance. Linde used the Tansley index (here sum of all scores is given), Ivo counted number of floral units along the transect (sum of all FU is given) and Constant counted flowering plants in a 5x5m square (sum of all flowering plants is given). - JulianDate: Sampling date translated to Julian date (1 for the 1st of Januari - 365 for the 31st of December).
- AgricultureArea250m: Area of agriculture (excluding flowering crops) in a 250m radius of the location as extracted from the TOP10NL shapefile. (Area given in square meters)
- GrasslandArea250m: Area of non agricultural (semi-natural) grassland in a 250m radius of the location as extracted from the TOP10NL shapefile. (Area given in square meters)
- ForrestArea250m: Area of forest in a 250m radius of the location as extracted from the TOP10NL shapefile. (Area given in square meters)
- FruitcropArea250m: Area of flowering crop agriculture (e.g. rapeseed, orchards) in a 250m radius of the location as extracted from the TOP10NL shapefile. (Area given in square meters)
- RestArea250m: Area of other elements that bees were not expected to use (e.g. open road, water) in a 250m radius of the location as extracted from the TOP10NL shapefile. (Area given in square meters)
- AreaNNN250m: Area of nature reserve in a 250m radius of the location as extracted from the TOP10NL shapefile. (Area given in square meters)
The next 6 columns give the same information as above but for a 1km radius ([Name]Area1km)
- DistanceToNNN: shortest distance from the location to a nature reserve
- BeeAbundanceNotCorrected: absolute number of individuals counted during the transect, not corrected for sampling effort (#individuals/survey)
- BeeAbundance: Number of individuals counted after correction for sampling effort (#individuals/survey/15mins)
- Windspeed: highest windspeed measured during the sampling day as interpolated from the KNMI database (km/h - also see methods section of DOI 10.1098/rspb.2025.0734)
- Sunshine: percentage of sunshine hours at the sampling day to the total day length as interpolated from the KNMI database (% - also see methods section of DOI 10.1098/rspb.2025.0734)
- Rain: mm of rain (including night) at the sampling day as interpolated from the KNMI database (see methods section)
File: FlowerCS.csv
Description: This file holds the full flower surveys collected by Constant connected to the BeesOnDykesNetherlands.csv data file. Constant counted all flowering plant per species within a quadrant.
- Location: code for the location used (the coordinates are provided in the LocationCoordinates.csv)
- Date: date of sampling (DD/MM/YYYY)
- Species: plant species
- Quantity: number of flowering plants counted
- survey size (m2): the size of the quadrant in which flowering plants were counted (square meters)
File: FlowerIR.csv
Description: This file holds the full flower surveys collected by Ivo connected to the BeesOnDykesNetherlands.csv data file. Ivo counted all floral units per flower species along the transect.
- Location: code for the location used (the coordinates are provided in the LocationCoordinates.csv)
- Date: date of sampling (DD/MM/YYYY)
- Species: plant species
- Number Floral Units: number of floral units counted
File: FlowerLS.csv
Description: This file holds the full flower surveys collected by Linde connected to the BeesOnDykesNetherlands.csv data file. Linde scored flower abundance per species along the transect following the Tansley scale.
- Location: code for the location used (the coordinates are provided in the LocationCoordinates.csv)
- Date: date of sampling (DD/MM/YYYY)
- Transect: Linde estimated Tansley Scale along three 50m transects. The number here gives the transect number at the location. When determining total Tansley score given in the BeesOnDykesNetherlands.csv file Tansley score was first averaged over the three transects.
- Species: plant species
- Tansley scale: flower abundance score on the Tansley scale
File: R_markup_Bees_Dykes_Netherlands.Rmd
Description: this file gives the R code used in the analysis and figure creation of DOI 10.1098/rspb.2025.0734. Subheadings within the code indicate what certain parts of the code are used for. The code was originally written and run in R version 4.2.2. To use the code one should download the BeesOnDykesNetherlands.csv (the other files are supporting information) and indicate directory themselves.