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

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


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

+ Author(s): Luuk Croijmans, Dirk F. van Apeldoorn, Fabrizio Sanfilippo, Tshelthrim Zangpo, Erik H. Poelman
+ Project: Data underlying the publication: Crop species diversity levels with attract and reward strategies to enhance Pieris brassicae parasitism rate by Cotesia glomerata in strip intercropping
+ Contact: erik.poelman@wur.nl OR luuk.croijmans@wur.nl


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

Data underlying the publication: Crop species diversity levels with attract and reward strategies to enhance Pieris brassicae parasitism rate by Cotesia glomerata in strip intercropping

[ADD DOI Article]


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

# Abstract

We assessed parasitism rates of released large cabbage white (Pieris brassicae) caterpillars by parasitoids on white cabbage (Brassica oleracea) plants in six different cropping systems: 
four different strip cropping designs, a pixel cropping design and a monoculture. These cropping designs differed in the number of crops included, the use of parasitoid attractive 
cultivars in concurrence with a cash cultivar, the use of nectar-providing crops for adult parasitoids, and the spatial arrangement of the crops.
Parasitism rate by the main P. brassicae parasitoid Cotesia glomerata was enhanced by strip cropping of white cabbage with wheat, 
and even further enhanced by the inclusion of four more main crops. Contrastingly, C. glomerata parasitism rate was lower in the most intensive crop mixture, 
i.e., pixel cropping, than in any of the strip cropping designs. The use of attractive cultivars or rewarding floral resources within a strip cropping set-up did not 
significantly further enhance C. glomerata parasitism rate. 



# Measurements and data collection

Study area
This three-year field experiment was performed at two locations in the Netherlands: Wageningen and Lelystad. Both locations are managed organically. The fields in Wageningen are located north of the Wageningen University campus (51°59'24"N, 5°39'36"E, 13 m). The soil type is loamy sand. The surrounding landscape is quite diverse with many ditches, hedgerows, tree lines, relatively small-scale crop fields (< 5 ha), larger grasslands, as well as urban areas especially to the south and east. The fields are directly bordered by small experimental plots from other experiments and the Wageningen Campus to the south, a small scale farm to the west, larger-scale experimental fields to the north and a busy provincial road to the east. The fields in Lelystad are located east of the city of Lelystad and belong to the experimental fields of Wageningen Research Field Crops (52°32'33"N 5°34'23"E, -4 m). The soil type is sandy clay loam, which was reclaimed from the sea in 1957. The surrounding landscape mainly consists of relatively large-scale agricultural fields (10 – 40 ha) with only few non-crop habitats. The main non-crop habitats are ditches with grassy vegetation, occasional flower strips, headlands and some tree lines. The fields are directly bordered by relatively large-scale agricultural experimental fields (5 – 10 ha). 

Experimental design
At Wageningen, three crop pairs were included in the larger experiment: white cabbage – wheat, pumpkin – barley and potato – grass (see table S1 for scientific names). Two exceptions here were that wheat was replaced by oat in 2021 and no pumpkin was grown in 2019 (due to crop failure) leaving these strips without crops during the main growing season. At Lelystad, four crop pairs were included: white cabbage – wheat, sugar beet – barley, potato – grass and carrot – onion. The nine crops that were included in the two locations were chosen based on their importance to Dutch agriculture, as farmers indicated that positive effects on these crops would allow for adoption of crop diversification practices. Differences in crops used per location are based on variation in their performance due to soil type difference of the two experimental sites.
Six distinct cropping designs were included in this study: (1) Monoculture, (2) Strip, (3) Strip_cultivar, (4) Strip_additive, (5) Strip_diversity and (6) Pixel (Figure 1, Table 1). At both locations a large monocultural cabbage field was included as reference. In the systems experiment, the large-scale monoculture was installed for all crop species, limiting the possibility to replicate the cabbage monoculture within the locations because of space and management constraints. Due to its large size (at least 45 x 45m), its width largely exceeded the spatial distance between replicates of strip cropping designs and thus sampling sections in the monoculture covered similar field variation as the strip cropping designs. We installed four strip-cropping designs. In the Strip cropping design we combined strips of cabbage and wheat, to assess whether strip-cropping by itself raises parasitoid pressure. Next, we added an attractive second cultivar in the Strip_cultivar cropping design to see if this would further enhance parasitoid pressure. Simultaneously, in the Strip_additive cropping design we added faba beans (Vicia faba) within the wheat to assess if adding a rewarding nectar source would enhance parasitism rate. Finally, to investigate if further diversification would continue to enhance parasitism rate, we grew all six main crops in a strip-cropping set-up and included both the attract and reward elements of Strip_cultivar and Strip_additive into the Strip_diversity cropping design. We also included a novel cropping design, where we grew crops in a random configuration in a grid of 50 x 50 cm plots called pixels. Here we included all the crop and genetic variation of the Strip_diversity cropping design, but on a smaller spatial scale, to assess how further spatial diversification in an already diverse agricultural system would affect parasitoid pressure. 
All strips were 3 meters wide, and the Strip_diversity and Pixel cropping designs were only included at Wageningen. Furthermore, the Monoculture, Strip and Strip_cultivar cropping designs received animal manure, whereas the Strip_additive, Strip_diversity and Pixel cropping designs only received plant-based fertilizer to maximize the effects of the legume (Tang et al., 2021). In both fertilization regimes, care was taken that all cropping designs would receive similar nitrogen inputs. A full rotation of all main crops was employed at both locations, but the cropping designs were kept in place over the years. When crops were harvested, they were usually replaced by next year’s crop or by a green manure mixture that also adhered in diversity to the cropping designs (see tables S2 – S6 for precise descriptions of cropping practices, planting schemes and seed/plant mixtures). 
The cropping designs at both locations were arranged in an incomplete randomized block design (Figure S1), as including a large-scale monoculture on every field for this multi-year systems experiment was not feasible due to spatial constraints and as it would increase within-block variation. At Wageningen, the cropping designs were divided over four fields. Fields 1, 2 and 3 contained blocks of 3 strips per crop per two-crop cropping designs (Strip, Strip_cultivar and Strip_additive) for each crop pair. Furthermore, each field also contained a set of six strips containing each of the main crops for the strip-cropping design that included six crops (Strip_diversity). The pixel cropping plots were only located on fields 2 and 3, as this cropping design was too labor intensive to be included in each field. Field 4 only contained monocultures of cabbage and wheat, and a set of strips of the two-crop cropping design without further implementations (Strip) to make the comparison with other cropping designs on other fields whilst taking field effects into account. At Lelystad, the cropping designs were divided over three fields. Two fields were arranged in a similar way as fields 1, 2 and 3 at Wageningen, but only included the two-crop cropping designs (Strip, Strip_cultivar and Strip_additive). A third field was arranged as a reference and only contained a monoculture of cabbage and several strips of the two-crop cropping design without further implementations (Strip). 

Organisms
In this agricultural systems experiment, we studied parasitoid pressure by parasitoids in the tri-trophic system of cabbage (Brassica oleracea), the large cabbage white (P. brassicae) and its most prominent parasitoid (Cotesia glomerata) (Poelman, Oduor, et al., 2009). 
We used white cabbage plants (Brassica oleracea var. capitata) of the Rivera cultivar as the cash cultivar. One exception to this occurred at Lelystad in 2021, as Rivera was replaced by a similar cultivar (cultivar Expect), due to a lack of cultivar availability. Within the Strip_cultivar, Strip_diversity and Pixel cropping designs, we replaced one in eight Rivera cabbage plants with the cultivar Christmas Drumhead. This cultivar is known to be more attractive to C. glomerata (Poelman, Oduor, et al., 2009) and to recruit C. glomerata from a further distance compared to other cabbage cultivars (Aartsma, Leroy, et al., 2019). This replacement was always done in the central two plant rows, out of four cabbage plant rows, as this assured that each Christmas Drumhead plant bordered as many Rivera plants as possible. White cabbage was paired with wheat (cultivar Lennox, and Lavett in the cropping designs that included two cultivars) in the strip-cropping designs. When the cabbage was planted, the wheat had already germinated and formed a dense cover. Wheat was harvested after maturation, halfway throughout the growing season of white cabbage, after which a green manure was sown to replace it. 
Pieris brassicae is a butterfly species, whose caterpillars are specialized herbivores of brassicaceous plants. Their immature stages live gregariously and can therefore be voracious pest insects. In nature, important natural enemies of P. brassicae immature life stages are specialized parasitic wasps, such as: Cotesia glomerata and C. rubecula (Poelman, Oduor, et al., 2009). In the Netherlands the gregarious endoparasitoid C. glomerata is the most common parasitoid of P. brassicae, typically being responsible for over 95% of the parasitism incidences of P. brassicae (Poelman, Oduor, et al., 2009). The solitary endoparasitoid C. rubecula has been found in not more than 5% of caterpillars of P. brassicae in our agricultural fields in Wageningen, but can be found to co-occur (multiparasitism) in the same caterpillar with C. glomerata (Poelman et al., 2013; Poelman, Oduor, et al., 2009). Second instar P. brassicae caterpillars were obtained from the rearing maintained at the Laboratory of Entomology, Wageningen University. The original adult butterflies were caught around the Wageningen campus. Caterpillars were reared on Brussels sprout plants (Brassica oleracea var. gemmifera cultivar Cyrus) in a climatized room at 20-22 oC, 50-60% relative humidity and a L16:D8 photoperiod with SON-T light (500 μmol m-2 s-1 in addition to daylight). All parasitoids examined in this study were naturally occurring parasitoids. 

P. brassicae parasitism assessment
To assess field parasitism rates in different cropping systems, we used an existing protocol of release and recapture (Aartsma, Hao, et al., 2020; Aartsma, Pappagallo, et al., 2020; Poelman, Oduor, et al., 2009). Ten caterpillars were released on one (in 2019) or two (in 2020 and 2021) Rivera plants per experimental strip. Here, strips that bordered other cropping designs were excluded to avoid mixing effects of two cropping designs. Within the monoculture, we released caterpillars within six strips of four plant rows, where we always left a full strip in between them. This was done to make the sampling design of the monoculture consistent with the strip-cropping designs and to assure that two sampled plants were at least 6 meters apart. As there was only one Strip_diversity strip per field, we doubled the number of plants on which caterpillars were released per strip for this cropping design (two plants in 2019, four plants in 2020 and 2021). For the Pixel cropping design, we randomly chose four Rivera plants in both Pixel plots. Similarly, we also released ten caterpillars on Christmas Drumhead plants in the Strip_cultivar cropping design, to examine whether indeed this cultivar attracted more parasitoids than the Rivera cultivar in a strip-cropping design. We recaptured all caterpillars that we could find after three days, and stored these in a freezer. Next, we dissected the recaptured caterpillars, identified, and counted any parasitoid eggs or larvae encountered. Eggs and larvae of C. glomerata and C. rubecula were identified by the numbers, size and morphology, as C. rubecula lays one large egg per oviposition event and C. glomerata generally lays 35-100 small eggs per oviposition event (Poelman, Oduor, et al., 2009). The larvae of both parasitoids can be discriminated by the presence of mandibles in first instar C. rubecula larvae that are not distinct in C. glomerata larvae. We sampled at least six rounds per growing season in three years (2019 – 2021) at two locations (Wageningen and Lelystad).

Statistical analysis 
All statistical analyses were done using R, version 4.2.2.
We analyzed four variables that represented parasitism by naturally occurring parasitoids of P. brassicae: (I) “parasitism rate per plant” by C. glomerata, a binary variable defined by either parasitism by C. glomerata did occur in at least one caterpillar on a plant per sampling effort (1) or no caterpillars were parasitized (0), i.e. whether at least one adult C. glomerata visited the plant or not; (II) “parasitism rate per caterpillar” by C. glomerata, a binomial variable defined by the number of parasitized caterpillars out of the number of recaptured caterpillars; (III) the number of C. glomerata eggs per parasitized caterpillar; (IV) parasitism rate per plant by C. rubecula. Parasitism rate per caterpillar (II) showed issues in the dispersal of residuals in all models, because of the high occurrence of either none or all caterpillars being parasitized. Even when corrected for zero-inflation, this issue held as the models could not explain the high number of cases where all caterpillars were parasitized. Due to this statistical issue, we chose to use parasitism rate per sampled plant (I) to explain effects on parasitism rates of C. glomerata, but the results on parasitism rate per caterpillar can still be found in Figure S2 and Table S7. We also analyzed (V) the number of caterpillars that we recaptured from the ten caterpillars previously released, to see if cropping designs affected recapture rate. All five dependent variables were analyzed using generalized linear mixed models (GLMMs), using a binomial distribution for recapture and parasitism rates, and a negative binomial distribution for the number of eggs per parasitized caterpillar. Recapture rate was also corrected for zero-inflation. For this purpose, we used the glmmTMB package.
As the two cropping designs that included all main crops (Strip_diversity and Pixel) were not available at Lelystad, we decided to make a model with data from both locations, but without these two cropping designs. We also made models with the two locations separated, where we did include these two cropping designs for the models made with data from Wageningen. In all models, cropping designs and years were included as fixed factors. Furthermore, to get an indication whether plant size affected recapture rate, we also included a continuous variable for the days since planting for the models on recapture rate. To assess whether the abrupt change in the agro-ecosystem caused by the wheat harvest affected parasitism rates of P. brassicae, we included a variable for whether wheat harvest had happened or not in the models regarding field parasitism. Finally, in all models that included both locations, we also added a fixed variable for locations. To correct for any temporal or spatial variation, we also included round and field numbers as random effects. Furthermore, we included plant number as random effect for the C. glomerata egg abundance per parasitized caterpillar. 
To identify any important interactions among the included fixed variables, we used step-wise backward selection on models that included all potential two-way interactions among fixed factors. Here, only interactions could be removed via model selection, but any fixed factors themselves were retained. The Akaike Information Criterion (AIC) was used in model selection: any model that reduced the AIC whilst dropping an interaction term was preferred. Finally, model validation was performed on the final model after model selection, using the DHARMa package.


============================================================================================================
## 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 / .xlsx / .R


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

Please view Par_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
============================================================================================================
