Title of the dataset:
Data underlying the paper: Multi-species productive grasslands offer advantages over permanent grasslands for arthropod conservation, but preserving semi-natural grasslands remains essential

Creators:
Tim Visser

Contributors:
Tim Visser (project leader Wageningen Environmental Research)
Anthonie Stip (project leader Vlinderstichting)

Related publication:
Multi-species productive grasslands offer advantages over permanent grasslands for arthropod conservation, but preserving semi-natural grasslands remains essential

Description:
This datasets contains the results of both pollinator counts and pyramid catches. Data was collected in the East of the Netherlands in the region 'achterhoek'. All locations had sandy soils. Treatment types were semi-natural grasslands (SNG), multi-species grasslands (MSG), and permanent, conventional grasslands (PG). 

Keywords:
Pyramid catches, pollinator transects, Land Use Intensity (LUI), Arthropods, Grasslands, Netherlands.

Spatial coverage:
Netherlands; region 'achterhoek'. Exact locations can be acquired by contacting the author. 

Temporal coverage:
2021 & 2022. Pollinator counts took place monthly, from april until august. Pyramid traps were placed in the first two weeks of may (period 1), and the last week of July and the first week of august (period 2). 

This dataset contains the following files:
PollinatorData: presence absence data on pollinators.
PyramidCatchesMasterfile: abundance (based on manual counts) and species richness (based on DNA metabarcoding) per grassland, per period, per year. This file also includes the explanatory variables of grassland type, land use intensity (LUI) and plant species richness. 
PyramidCatchesPresenceAbsence: presence absence data on pyramid catches. Results based on DNA metabarcoding. 

Explanation of variables:
See readme sheets in excel files. 

Methods, materials and software:

Pollinators: Pollinator abundance and species richness were assessed using transect counts. In each field, two transects (100m long, 5m wide) were positioned diagonally across the field, staying 10 meters from field edges. Distance between both transects was at least 100 meters. A total of 72 transects were counted once a month from April until August in 2021 and 2022 (720 transect counts in total). Counts were performed in favourable weather conditions following Van Swaay et al. (2018). Bees, butterflies, hoverflies and day-active moths were counted and identified in the field or captured and identified in the lab if necessary. Specimens of Bombus terrestris/lucorum/magnus/cryptarum were grouped into B. terrestris-group. 

Pyramid traps: We used pyramid traps to gain insight in abundance and diversity of arthropods. The base of the pyramid trap covered one square meter. One trap was placed per field, positioned at least 10 meters from field edges. At the top of each trap, a pot was positioned in which arthropods accumulate during a one-week period. This pot was filled with a 250 ml solution of 1 part 100% propylene glycol, 2 parts water, and a drop of soap. 
Sampling took place in 2021 and 2022, in two periods in each year. The first period lasted the first two weeks of May. The second period lasted the last week of July and the first week of August. After the first week we collected the samples, attached a new pot and repositioned the pyramid trap to a new location on the same field, where it would stay until the end of the second week. After collection in the field, samples were rinsed with water and  the number of individuals was counted at order level, except for Diptera and Coleoptera, which were counted on family level. Samples were stored in 96% ethanol at a temperature of –20 °C until subsequent DNA analysis.  

Metabarcoding: For arthropod diversity, we used DNA extraction and amplification following van der Lee et al. (2024), including both DNA from preservative ethanol and from the whole bodies of arthropods using the DNeasy PowerMax Soil DNA Isolation kit (Qiagen), and the COI primers (mlCOIintF, HCO2198; Leray et al. 2013). The bioinformatics and taxonomy assignments were also performed following the protocol described in van der Lee et al. (2024) with the following adjustment: QIIME2 version 2023.2 was used (Bolyen et al. 2019). Any taxon with more reads in a negative control than in any sample was removed. For the run in 2022, a few reads of the positive control were detected in other samples, indicating a small degree of spillover/tag-jumping (Schnell et al. 2015). This resulted in a correction for each OTU where only records consisting of at least 0.002% of the total amount of reads for that OTU were kept. Taxa with identical identifications were then merged. Finally, prior to statistical analyses, the OTU tables were rarefied using the rarefy function of the vegan package for R (Jari Oksanen et al. 2018). Both for manual counting as DNA analysis, only arthropods from the following orders were considered relevant for our study since they are specifically targeted using pyramid traps, Aranaea, Diptera, Coleoptera, Hymenoptera, Hemiptera, Lepidoptera and Thysanoptera. Any arthropods from other orders, as well as fungi, mammals, nematodes, flat worms and gastropods were considered bycatch and removed from the dataset.


This dataset is published under the CC BY (Attribution) license.
This license allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator.

