Figures and data
Effect of crop configuration (monoculture versus strip cropping) on ground beetle biodiversity.
(a) Location of experimental sites in the Netherlands. (b) Field set-up of the two crop configurations: monoculture and strip cropping. At Lelystad and Wageningen, strip cropping consisted of 3-m wide crop strips of two crops (pairs), and multiple crop pairs were assessed. At Almere and Valthermond, strip cropping consisted of 6-m wide crop strips of eight crops combined (Fig. S7). (c) Sample-based species accumulation curves of all year series from monocultures (brown) and strip cropping (green), in Almere from 2021 and 2022. This was the only location where an equal number of samples were taken in the monocultures and in strip cropping on a similar area. Ground beetle species include Poecilus cupreus (left) and Pterostichus melanarius (right). Photo credit: Ortwin Bleich, retrieved from: www.eurocarabidae.de. (d-e) Overall relative change in field-level ground beetle (d) taxonomic richness and (e) activity density. Positive values indicate higher richness or activity density in strip cropping, negative values in monocultures. (f-j) Overall relative effect of crop configuration on ground beetle (f) taxonomic richness, (g) activity density, (h) absolute evenness, (i) inverse Simpson index, and (j) Shannon entropy. (k) Effect of crop configuration on ground beetle taxonomic richness for each combination of location, year and crop. Barley-mixture consists of a mixture of barley-bean (2020) or barley-pea (2021). Squares indicate estimated means, the bar indicates the 95% confidence interval. Asterisks indicate significant differences among the crop configurations. Empty panels indicate combinations of years and locations that were not sampled. When no estimated mean and confidence interval are shown, crops were not grown or sampled in that year. Open circles indicate individual year series to visualize sample size (Table S7).
Effect of crop configuration on ground beetle community composition.
Results from permanova analysis using Hellinger’s transformation for data from the three locations with species level data (see Table S4 for analyses per location). Crops were a nested variable within years, as these differed among years. Years were nested in locations, as the years that were studied differed among locations. Bold letters indicate significant effects (α = 0.05).
Ground beetle species associated with crop configuration (monoculture (a) versus strip cropping (b)).
Results obtained by indicator species analyses with the four locations analysed separately (Table S6). Locations are indicated between brackets (Al=Almere; Le = Lelystad; Va = Valthermond; Wa = Wageningen). Data from Almere in 2020 was excluded from the analysis as ground beetles were identified up to genus level only. An asterisk indicates the significance of the association between ground beetle specie and crop configuration (*: α < 0.05; **: α < 0.01; ***: α < 0.001). Rare species indicated by red “^” (Turin, 2000). Photo credit: Ortwin Bleich, retrieved from: www.eurocarabidae.de.
Effect of crop configuration on ground beetle activity density and absolute evenness.
Effect of crop configuration on activity density (a) and absolute evenness (d) per combination of location (columns), year (rows) and crop (x-axis); and the effect of crop configuration on activity density (b) and absolute evenness (c) in the case-study in Wageningen, consisting of two early spring sampling rounds in 2022. Activity density here is the total number of ground beetles captured in year series, which might differ in sampling frequency (Table S1). Absolute evenness is calculated from the total number of ground beetles captured in year series. Empty panels indicate combinations of years and locations that were not sampled. Squares indicate estimated means, the bar indicates the 95% confidence interval. When no estimated mean and confidence interval are shown, then those crops were not grown or sampled in that year. Asterisks indicate significant differences among the crop configurations (α = 0.05). Open circles indicate individual year series (Table S1).
Effect of crop configuration on ground beetle inverse Simpson index and Shannon entropy.
Effect of crop configuration on inverse Simpson index (a) and Shannon entropy (d) per combination of location (columns), year (rows) and crop (x-axis); and the effect of crop configuration on inverse Simpson index (b) and Shannon entropy (c) in the case-study in Wageningen, consisting of two early spring sampling rounds in 2022. Inverse Simpson index and Shannon entropy were calculated from the total number of ground beetles captured in year series. Empty panels indicate combinations of years and locations that were not sampled. Squares indicate estimated means, the bar indicates the 95% confidence interval. When no estimated mean and confidence interval are shown, then those crops were not grown or sampled in that year. Asterisks indicate significant differences among the crop configurations (α = 0.05). Open circles indicate individual year series (Table S1).
The effect of crop configuration on ground beetle community composition.
Data from all datasets that included species level data (Almere (○), Lelystad (▽), Wageningen (▴)) are used for visualizing the first RDA and PC axis. Each axis shows the percentage explained variation. Colour of the dots indicates crop configuration (brown = monoculture, green = strip cropping), and the direction of each crop configuration on the RDA axis (x-axis) is mentioned. Red arrows indicate species placement and name codes are given close to the tip of the arrow, meaning of the name codes can be found in Table S3.
The effect of crop configuration on ground beetle community composition in Almere.
The first RDA and PC axis are visualized. Each axis shows the percentage explained variation. Shape of the dots indicates year (▽ = 2021, ▴ = 2022), colour of the dots indicates crop configuration (brown = monoculture, green = strip cropping), and the direction of each crop configuration on the RDA axis (x-axis) is mentioned. Red arrows indicate species placement and name codes are given close to the tip of the arrow, meaning of the name codes can be found in Table S3.
The effect of crop configuration on ground beetle community composition in Lelystad (left) and Wageningen (right).
The first RDA and PC axis are visualized. Each axis shows the percentage explained variation. Shape of the dots indicates year (○ = 2019, ● = 2020, ▽ = 2021, ▴ = 2022), colour of the dots indicates crop configuration (brown = monoculture, green = strip cropping), and the direction of each crop configuration on the RDA axis (x-axis) is mentioned. Red arrows indicate species placement and name codes are given close to the tip of the arrow, meaning of the name codes can be found in Table S3.
Effect of crop configuration and crop on ground beetle community composition in Wageningen.
Here, we use data from Wageningen in 2021 (left column; panels a, c, e) and 2022 (right column; panels b, d, f) from three fields including three crop pairs being cabbage and oat (panel a and f), barley and pumpkin (panel b and c), and grass and potato (panel d and e). The first and second RDA axis are visualized. Each axis shows the percentage explained variation. Colour of the dots indicates crop configuration (brown = monoculture, green = strip cropping), shapes indicates crop (□ = barley, ▪ = pumpkin, ○ = grass, ● = potato, △= oat, ▴ = cabbage). The red lines and abbreviated names indicate how specific ground beetles species correlate with the RDA axes, species names are given in Table S3.
Field maps of experimental lay-out per year and location.
Field maps are shown in chronological order (oldest to newest) per location: Almere (a), Lelystad (b), Valthermond (c) and Wageningen (d). Colors indicate distinct crops, whereas diagonal lines indicate a crop mixture (only for Valthermond). Brighter colours with black lining around the field indicate areas in which samples included in this study were taken.
All ground beetle species found among the four locations, and their species codes as used in several figures.
Total number of ground beetles caught per species (or genus), per location. For some locations, ground beetles were identified up to genus level, these are underlined. “N/A” indicates that this taxa was identified to a different taxonomic level for the specific location. Locations are indicated with abbreviations (Al=Almere; Le = Lelystad; Va = Valthermond; Wa = Wageningen).
Effect of crop configuration on ground beetle community composition.
Results from permanova analyses using Hellinger’s transformation for data from the three locations with species level data. “Crop species” is a nested variable within years, as these differed among years. Years were nested in locations, as the years that were studied differed among locations. P-values in bold typeset indicate significant effects (α = 0.05).
Effect of crop configuration and crop species on ground beetle community composition.
Results from pairwise permanova analyses for crop pairs pumkin-barley, cabbage-oat, and potato-grass in 2021 (a) and 2022 (b) in Wageningen. Values show F-values for the comparison between the crop configurations and crops in crossing rows and columns. Bold numbers indicate significant differences between combinations of crop configurations and crops (α = 0.05).
Effect of crop configuration on ground beetle indicator species.
Results from indicator species analyses with the four locations being analysed separately. Data from Almere in 2020 data was excluded from the analysis as ground beetles were identified up to genus level in 2020. Specificity refers to the predictive value of the species as indicator of the specific crop configuration whereas sensitivity refers to the percentage of samples of the respective crop configuration that included the species. “Stat” refers to the indicator value, which is the square root of the product of specificity and sensitivity. Bold letters indicate a significant association between species and crop configurations per location (α = 0.05).
Effect of crop configuration on crop yield.
Yield results were retrieved from published and unpublished studies on effects of strip cropping on crop yield in similar locations, years, and crops as this study. Mean crop yield is presented in ton per hectare (t/ha). When known, standard deviations of mean crop yield are given (± SD). When a crop is indicated with NC (not collected) the crop yield was not collected due to an inconsistent sampling method (potato, 2020, Almere), crop failure (broccoli, 2020, Almere; celeriac, 2021, Almere), unavailable machine-harvest data (grass, Almere, 2020, 2021, 2022) and undocumented reasons (barley/beans, 2020, Valthermond). Unavailable data include cabbage (2019) and potato (2020) in Lelystad; and pumpkin (2020, 2021, 2022), barley (2020, 2021, 2022), oat (2021, 2022), potato (2021, 2022), grass (2021, 2022), and cabbage (2022) in Wageningen.
Sampling effort.
The total numbers of pitfall traps placed per location, year and crop. Rounds indicate the number of times pitfall traps were placed and were pooled within year series.
