============================================================================================================
START readme
============================================================================================================

This file focuses more on the details of the data package. 


============================================================================================================
## General
============================================================================================================

+ Author(s): Peter N. Karssemeijer, Luuk Croijmans, Karthick Gajendiran, Rieta Gols, Dirk F. van Apeldoorn, Joop J.A. van Loon, Marcel Dicke, Erik H. Poelman
+ Project: Data underlying the publication: Diverse cropping systems lead to higher larval mortality of the cabbage root fly (Delia radicum)
+ Contact: erik.poelman@wur.nl OR luuk.croijmans@wur.nl


============================================================================================================
## Title
============================================================================================================

Data underlying the publication: Diverse cropping systems lead to higher larval mortality of the cabbage root fly (Delia radicum)

https://doi.org/10.1007/s10340-023-01629-1


============================================================================================================
## Methods
============================================================================================================

# Abstract

We studied how different diversified cropping systems affected the oviposition and abundance of the specialist cabbage root fly Delia radicum, the most important root herbivore in Brassica crops. The cropping systems included a monoculture, pixel cropping, and four variations of strip cropping with varying intra- and interspecific crop diversity, fertilization and spatial configuration. Furthermore, we assessed whether there was a link between D. radicum and other macroinvertebrates associated with the same plants. Cabbage root fly oviposition was higher in strip cropping designs compared to the monoculture and was highest in the most diversified strip cropping design. Despite the large number of eggs, there were no consistent differences in the number of larvae and pupae between the cropping systems, indicative of high mortality of D. radicum eggs and early instars especially in the strip cropping designs. D. radicum larval and pupal abundance positively correlated with soil-dwelling predators and detritivores and negatively correlated with other belowground herbivores. We found no correlations between the presence of aboveground insect herbivores and the number of D. radicum on the roots. Our findings indicate that root herbivore presence is determined by a complex interplay of many factors, spatial configuration of host plants, and other organisms residing near the roots.


# Measurements and data collection

Field setup
This study was carried out in the summer of 2020 as part of a large long-term strip-cropping trial at Droevendaal Experimental Farm in Wageningen, the Netherlands. At the field site, white cabbage (Brassica oleracea var. Capitata), wheat (Triticum aestivum L.), pumpkin (Cucurbita maxima L.), potato (Solanum tuberosum L.), barley (Hordeum vulgare L.), and grass (Lolium multiflorum L.) were grown as main crops. The experimental design consisted of six cropping designs: (1) Reference, i.e. cabbage monoculture of 51 by 45 m (cv. Rivera), (2) Strip, i.e. alternating strips of cabbage (cv. Rivera) and wheat (cv. Lennox), (3) Strip_cultivar; alternating strips of two cultivars of cabbage (cv. Rivera and Christmas Drumhead) and two cultivars of wheat (cv. Lennox and Lavett), (4) Strip_additive, i.e. alternating strips of cabbage (alone) and a combination of wheat with broad bean (Vicia faba L., cv. Pyramid) to add floral and extrafloral nectar to the system for natural enemies of herbivores, (5) Strip_diversity; composite of strips of all six crops with two cultivars per crop and a nectar source in the Poaceae crops (wheat with broad bean, barley with pea, grass with clover), (6) Pixel, adjacent 50 × 50 cm plots with each containing one of the six crops with two varieties per crop and a nectar source in the poaceous crops, in a random design. The same crops and additional nectar sources were planted in the Strip_diversity and Pixel-cropping designs, but with a different spatial  configuration. The setup was an incomplete block design on four adjacent fields. In one field, the reference monoculture (51 by 45 m in size) and a replicate of the strip cropping design were planted next to each other. The other three fields contained replicates of all strip cropping designs (Strip, Strip_additive, Strip_cultivar, Strip_diversity). Every replication of the Strip, Strip_additive and Strip_
cultivar cropping designs consisted of three strips per crop, of which the central two strips of cabbage were included in the measurements to reduce interference between cropping designs. In the Strip_diversity cropping design, a single strip of each crop was planted per replication. In all strip cropping designs, each strip was 3 m wide and 42 or 54 m in length. Two replicates of the Pixel cropping design were planted in the fields, each measuring 12 by 7.5 m consisting of 360 pixels. Planting distance of cabbage was 38 cm within rows and 75 cm between rows, resulting in four rows per strip. In the Strip_cultivar and Strip_diversity cropping designs, the cabbage cultivar “Christmas Drumhead” was planted in the centre two rows of each strip, as every fourth plant. We expected this cultivar to attract shoot herbivores and parasitoids, serving as a trap crop and attracting natural enemies into the strip. This second cultivar was also included in the Pixel-cropping design.
	Fertilization was carried out two weeks prior to planting and varied between the cropping designs. One of the overarching goals of the field trial in which this study was carried out is to study different cropping systems, including differences in fertilization strategies. The Reference, Strip, and Strip_cultivar cropping designs received farm yard manure (35 t/ha), the Strip_additive and Strip_diversity cropping designs were fertilized using organic plant fertilizer (11–0-5 NPK) in a concentration matched to the manure in terms of nitrogen dosage. The Pixel cropping design should have received a similar fertilization as the Strip_additive and Strip_diversity cropping designs; however, due to unforeseen circumstances, this was withheld. Moreover, there were differences in the precrop between the cropping designs, most notably the precrop for the Strip_additive, Strip_diversity and Pixel cropping designs included red clover (Trifolium pratense L.), a nitrogen fixer. The rotation scheme of the long-term field trial ensured that each cropping design was planted on a plot that had a similar cropping design in the previous year, but the six main crops were rotated over a six-year rotation scheme (i.e. a strip containing the Strip_diversity cropping design in one year will be Strip_diversity again in the next year, but with a different main crop). Cabbage plants were planted on the 25th of May and harvested from the 15th of October until the 26th of November. The wheat intercrop was sown in the second half of March and harvested on July 31st, meaning that these rows were empty for two weeks before the last measurements were taken.

Delia radicum oviposition and shoot herbivore monitoring
We assessed the oviposition by cabbage root flies in different cropping systems using felt traps. Strips (4 × 100 cm) of black felt were wrapped around the stem of cabbage plants and secured using a safety pin. Felt traps were collected five days after placing them and the number of eggs in each trap was counted. Three to seven days prior to placing the felt traps, the community of aboveground herbivores on each plant was assessed. Herbivores present on the plant were directly identified to the species level (except thrips and leaf miners). We measured the maximum plant radius (distance between leaf tips furthest apart) as a non-destructive assay of plant size. Per cropping design, 16 cabbage plants were assessed in each cropping design replicate. The first and last 10 m of each strip were not sampled, to account for field edge effects. For the Strip, Strip_additive, Strip_cultivar and Reference, two strips with each eight plants and in the Strip_diversity one strip with 16 plants were sampled. In the Reference monoculture the distance between replicate strips was largely exceeding the distance between replicates of strip cropping treatments, and therefore considered independent measures. The sampled plants were equally spread throughout the strip, and the order of the rows to be sampled was randomized. In the Pixel cropping design, 16 random cabbage plants were chosen among the cabbages that were not on the edges of the Pixel field.

Cabbage root and shoot invertebrate monitoring
To examine how cropping systems affected D. radicum larval and pupal abundance, and how these larvae and pupae correlated with the community of invertebrates associated with cabbage plants, we sampled both the above- and belowground invertebrate communities of cabbage plants. For this purpose, we chose eight cabbage plants (two per plant row) per experimental strip in two rounds: 29–30 June and 17–18 August. As this was a time-consuming sampling effort, timing was important to assure peak D. radicum presence. Therefore, throughout the season, we monitored a small number of cabbage plants in the fields every other week to confirm the presence of D. radicum larvae and pupae. This pilot data, together with the oviposition data, provided the basis for timing the cabbage root and soil collection when D. radicum densities peaked.
	First, we assessed the aboveground community of herbivorous insects by visually examining cabbage plants. Here, all herbivores were identified to the order level (thrips), morphospecies (leaf miners) or species level (all other herbivores) and counted. 
	Secondly, we examined the belowground arthropod community by taking soil samples around the root of the cabbages. For this purpose, we cut the cabbage plant at the base of the stem and we took a soil sample using an auger (20 cm diameter) to a depth of 20 cm with the cabbage tap-root at the centre, one day after the aboveground herbivore monitoring. As such, 16 cabbages were removed from the strips. We expect only minor effects of this removal as the shortest strips already contained about 450 cabbages. Soil samples were placed in plastic bags, secured using a tie-wrap, and stored at 7 °C until further analysis. The Pixel cropping design was not included in this monitoring, as removal of plants would have interfered too much with other measurements in this cropping design.
	Within one week after collection and storage, soil samples were thoroughly searched for macroinvertebrates. The cabbage taproots were carefully opened to find D. radicum larvae feeding within. When found, these cabbage root fly larvae were placed in small containers containing a small piece of rutabaga (Brassica napus var. napobrassica) to rear them to pupation. The collected pupae were stored in glass vials in a climate cabinet (20 ± 1 °C) until eclosion to assess parasitism. While we did not identify Delia spp. to the species level, a previous study indicates that the community in Northern Europe is dominated by D. radicum, and a small fraction of D. floralis. We therefore assume most specimens found in this study to be D. radicum. Empty D. radicum puparia collected from the soil samples were also scored.
	Lastly, as we expected that plant size might also affect D. radicum larval/pupal abundance, we also quantified plant dry biomass. For this purpose, cabbage plant shoots were collected in paper bags and shoot tissue was dried at 70 °C for 2 days and weighed on a “DK-6200-C-M” balance (± 0.1 g).

Pitfall trapping
Furthermore, two times during the season, pitfall traps were placed to quantify the abundance of (egg) predators. Pitfall traps consisted of a plastic cup (8.5 cm diameter) with a layer of water 3 cm high with odourless dish soap placed in the soil up to the rim, which was covered with a plastic roof (12.5 cm diameter) approximately 2 cm above the soil surface. One pitfall trap was placed in a predetermined random position in the one of the central rows of each strip, and traps were left in the field for 5 days. Macroinvertebrates captured in the pitfall traps were preserved in 70% ethanol and identified. Carabid beetles were identified to the species level. Only the activity densities of staphylinid and carabid beetles were statistically analysed.

Statistical analysis
Data were analysed using R, with packages vegan, lme4, emmeans, ggeffects. Individual plants or pitfall traps were the statistical units. Most of the data were counts, which we analysed using Generalized Linear Mixed Models (GLMM) with a negative binomial distribution. Plant quality measurements, i.e. shoot dry biomass and maximum radius, were analysed with a (G)LMM with a gamma or normal distribution. Field (a factor indicating the four fields on which the experiment was replicated) was included as a random factor. We analysed these response variables both using separate datasets for the two rounds, and for both rounds combined. In the case of continuous explanatory variables, we used the ggpredict function to generate conditional predictions of the correlation between the response variable and the explanatory variables whilst keeping other variables and the random factor constant. All models were validated using the DHARMa package, where we tested the residuals of all models. To assess whether the four fields showed similar patterns in oviposition or larval / pupal abundance, we also compared the strip cropping design on all four fields using the same statistical methods as mentioned above. To check if the reference field responded in line with our models from data of all fields combined, we also compared oviposition and larval / pupal abundance between the monoculture reference and strip cropping design on the reference field specifically.
	We performed a principal component analysis (PCA) with auto-scaling using the Hellinger transformation, on the community of aboveground herbivores (oviposition data) and total macroinvertebrate community (soil sample data) separately for each round. Using redundancy analysis (RDA) in which we added the number of cabbage root fly eggs or larvae and pupae as an explanatory variable, we further established whether cabbage root fly oviposition or abundance was correlated with other members of the plant-associated macroinvertebrate community. Finally, we tested for correlations between cabbage root fly oviposition and abundance and other macroinvertebrates using GLMM. For this analysis, species were grouped into explanatory variables based on feeding site and guild into: aboveground chewers, aboveground phloem-feeders, belowground detritivores, belowground predators, and belowground herbivores. No analyses were performed on parasitism rates, as the number of parasitized larvae / pupae found was too low for meaningful comparisons.
	Lastly, we analysed carabid and staphylinid beetle activity densities among the different cropping designs using GLMM with a negative binomial distribution. Here, we used cropping designs, rounds and the interaction between cropping design and round as fixed factors. For these analyses, we used data of both rounds together.


============================================================================================================
## FolderStructure
============================================================================================================

"There is only one parent directory present containing all files."


============================================================================================================
## FolderContents
============================================================================================================

- parent_folder/
Data, scripts, codebooks




============================================================================================================
## Software
============================================================================================================

SoftwareRequired: 

R, Rstudio, spreadsheet software


OtherSoftwareRequirements: 

Excel


============================================================================================================
## FileFormats
============================================================================================================

.csv / .R


============================================================================================================
## CodeBook
============================================================================================================

Please view delia_[dataset]_codebook.csv (where this readme file is also found) for documentation of abbreviations, 
column names, datapoints, etc. The file codebook.csv uses the columns:

+ index = a number used to distinguish the different entries.
+ code = the abbreviation, variable / data / column name used.
+ used = location where the code is used. (filename, foldername, columnname, datapoint, protocol, etc.).
+ meaning = the literal meaning of the code. (e.g., fully written out abbreviation)
+ represents = what the code represents in terms of data or usage. (e.g., units of measurements, 
coding used,more in depth explanation)

Note that this file is ';' delimited. To avoid possible confusion and inconsistencies, sentences within 
cells do not contain reading symbols as comma's or semicolons. When required, separation within sections of
a sentence is made possible using the hashtag symbol (#). Example: The sex of an animal is described as
"m = male pig (boar) # f = female pig (sow)" where the hashtag separates the element in a sentence.


============================================================================================================
## Other
============================================================================================================

[describe any other attention points that will help understandability of your data package; delete this 
explanation]


============================================================================================================
END readme
============================================================================================================
