Considerable uncertainties exist on how increased biofuel cropping affects biodiversity. Regarding oilseed rape, the most common biofuel crop in the EU, positive responses of flower-visiting insects to plentiful nectar and pollen seem apparent. However, previous investigations on this insect guild revealed conflicting results, potentially because they focused on different taxonomic groups representing a narrow range of ecological traits and considered only short time periods. Here, using trap nests in landscapes with independent gradients in area of oilseed rape and seminatural habitats, we assessed the whole community of cavity-nesting bees and wasps, including early- and late-emerging species. Our study's temporal resolution allowed determination of flowering and postflowering effects of oilseed rape on these species' richness, abundance, and mortality. Species richness of cavity-nesting bees and wasps significantly increased with oilseed rape, although nesting activity was considerably higher after mass flowering. In addition to increasing richness independently of oilseed rape, the amount of seminatural habitat in the landscape was the sole positive driver of insect abundance once the community's dominant species was accounted for as a covariate. Thus, growth of the co-occurring species' community is not stimulated by the resource pulse provided by oilseed rape early in the year, but by persistent resources provided by seminatural habitats after mass flowering. Early individuals of bivoltine species' first generations accumulated in seminatural habitats when these habitats were scarce, but became increasingly diluted when habitat availability increased. Once established, later foraging females generally benefited from the resource availability of seminatural habitats when initializing the second generation. We conclude that mass-flowering crops, despite covering only a short interval of the community's main activity phase, benefit bee and wasp species richness. However, seminatural habitats are crucial in maintaining viable communities of flower-visiting insects at the landscape scale, mitigating potential negative effects of high land-use intensities in modern agro-ecosystems.
The rapid expansion of biomass and biofuel production in agricultural systems will result in major land-use changes at large spatial scales (Koh, 2007). In the EU, oilseed rape is the most common oleaginous crop for biofuel production (FAO, 2008) and the production area has more than doubled within the past 20 years (Fig. 1). However, significant uncertainties exist about the effects of this extensive increase in biomass and biofuel cropping on biodiversity, especially at the regional scale (Eggers et al., 2009; Dauber et al., 2010). Here, to reduce these uncertainties, we studied the diversity effects of the mass flowering of oilseed rape on the community of trap-nesting bees and wasps in modern agro-ecosystems.
Considering the bounty of nectar and pollen that mass-flowering crops supply, it seems plausible to expect positive responses of flower-visiting insects to oilseed rape at the landscape scale. Increased abundances of short-tongued social bumblebees with increasing area of oilseed rape early in the year seem to confirm this assumption. This numerical increase in worker bees, however, fails to translate into improved sexual reproduction later in the year when food resources are usually scarce in modern agro-ecosystems (Westphal et al., 2003, 2009). Moreover, long-tongued bumblebees were negatively affected by an increasing area of oilseed rape. Once mass flowering had ceased, short-tongued bumblebees, in contrast to long-tongued ones directly utilizing this resource early in the year, increased nectar robbing on long-tubed plants, the legitimate resource of long-tongued flower visitors (Diekötter et al., 2010).
In contrast to social bumblebees producing only worker bees early in the year, the direct production of propagable females and males in solitary wild bees argues for a positive effect of oilseed rape on the reproductive success. Such an effect has recently been shown for the polylectic and phenologically early mason bee Osmia rufa (Jauker et al., 2012). Considering this trait specificity in pollinator responses, here, we were interested in the effect of mass-flowering oilseed rape at the community level of cavity-nesting bees and wasps as potential shifts in the structure of this community may not only be of direct conservation concern but also constrain the buffering of ecosystem services (i.e., pollination or biological control) from environmental changes (cf. Brittain et al., 2013).
According to analyses with high temporal resolution, the overall positive effect of mass-flowering oilseed rape on O. rufa's reproductive success resulted from early reproductive benefits that outweighed postflowering disadvantages of increased spillover of parasites and parasitoids in close proximity to oilseed rape fields (Jauker et al., 2012). Such negative postflowering effects may be mitigated by flower-rich and continuous seminatural habitats via diluting antagonist spillover or providing habitat for hyperparasitoids (cf. Jauker et al., 2012). Additional benefits of seminatural habitats might include the provision of above-ground cavities or specific soil microhabitats for nesting (Cane et al., 2007; Steffan-Dewenter & Schiele, 2008). However, seminatural habitats have become increasingly scarce in modern agricultural landscapes (Potts et al., 2010). This shortage might become more severe as the area of energy crops, including mass-flowering crops increases with an increasing demand for biofuel (Koh, 2007; FAO, 2008).
While the effects of seminatural habitats on cavity-nesting bees and wasps in agro-ecosystems have been investigated before (e.g., Holzschuh et al., 2010), we are not aware of any study providing information on the response in the structure of this complex community to the increasing amount of mass-flowering crops in comparison to seminatural habitats. Specifically, we tested whether (i) species richness; and (ii) abundance of cavity-nesting bees and wasps were associated with the area of mass-flowering oilseed rape and seminatural habitats. By separating the flowering and postflowering phase of oilseed rape in our study, we were able to attribute total community responses across the year specifically to species with an early or late phenology. Because some species with an early phenology, i.e., the first generations of bivoltine species, hatch from their nests during the year, we also analyzed the relationship between (iii) nests vacated in the field and mass-flowering crops or seminatural habitats. Finally, we analyzed patterns of (iv) mortality in relation to availability of oilseed rape and seminatural habitats.
Materials and methods
The study was carried out in the Nidda catchment in Central Germany, an area dominated by farmland (~50%) and woodland (~30%) interspersed with settlements and seminatural habitats (center: Echzell, 50°23′0″N, 8°53′0″E). In this area, twelve spatially separate (minimum distance 5 km) habitat elements that represent typical nesting sites for cavity-nesting pollinators (i.e., shrubs, hedges, and forest edges) were selected as study sites for trap nest location. Percentage area of oilseed rape and percentage area of seminatural habitats were quantified around the study sites for eight radii (250, 500, 750, 1000, 1250, 1500, 1750, 2000 m; Table 1) based on an updated ground-truthed digital map derived from high resolution color-infrared aerial photographs (0.5 × 0.5 m) from 2005 using ArcMap 10 (ESRI, Redlands, CA, USA). No other mass-flowering crops were detected during the updating procedure. Coverages of oilseed rape and seminatural habitats were uncorrelated at all scales (P ≥ 0.369). Seminatural habitats included fallows, orchards, field margins, tree rows, hedges, shrubs, and forest edges (width 10 m), which may provide nesting or foraging resources for the community of cavity-nesting bees and wasps. In addition, the area of the study sites (320–1 367 282 m²) and the distance between their trap nests and the nearest oilseed rape field margin (0–480 m) were measured. There were no significant correlations between these additional variables and the percentages of oilseed rape or seminatural habitats (P ≥ 0.107), except between the distance to the nearest oilseed rape field and the percentage of oilseed rape at the two smallest scales for which, however, no variables were retained in any model.
Table 1. Percentage areas of oilseed rape and seminatural habitats in the landscape surrounding study sites for different radii (scale: 250–2000 m). Given are mean values for eleven study sites and minimum (min) and maximum (max) values
Oilseed rape (%)
Seminatural habitats (%)
Two trap nests fixed on a wooden post 150 cm above the ground were set up at the edge of each selected study site in March 2008, 1 week before the flowering of oilseed rape. Each trap nest consisted of two plastic tubes (length: 30 cm; radius: 5 cm) filled with internodes of common reed Phragmites australis Cav. (diameter: 2–10 mm). At the end of May 2008, after the mass flowering had ceased, internodes with nests were partly extracted from the plastic tube, individually marked with a red permanent marker on the outside of the internode, and repositioned to enable discrimination between nests built during and after mass flowering. The trap nests were collected in October 2008 and stored in a climate chamber (4 °C) until March 2009. All solitary bees and wasps emerging from trap nests were counted (‘abundance’) and determined to species level (‘species richness’). Nests of bivoltine species' early first generations were vacated before the trap nests were returned to the laboratory and thus counted without further species determination or abundance record (‘vacated nests’). Nests where eggs, larvae, or pupae died before hatching were counted without further species determination and abundance record. This nest number in relation to the number of hatched nests per site constituted the ‘mortality rate’.
Trap nests from one site were excluded from statistical analyses due to contamination by road and tire wear from a nearby uphill motorway that prohibited proper colonization of their reed internodes. For the remaining sites, data from both trap nests were pooled per site. Because bee and wasp individuals emerging from different brood cells of the same nest lack independence, we used the number of nests as sampling unit when testing for sampling effects on species richness (t = 1.7999, df = 9, P = 0.105) and evaluating completeness of sampling. Species-accumulation curves produced with R-function specaccum (method ‘random’) in the vegan package (Oksanen et al., 2012) indicated that for various sites sampling was incomplete (see Supporting Information). Consequently, we estimated species richness using the nonparametric estimator Chao 1 (Chao, 1984) with R-function estimateS in vegan (Oksanen et al., 2012):
where Sobs = the number of species per site; F1 = the number of observed species represented by a single nest per site (singletons); and F2 = the number of observed species represented by two nests per site (doubletons; Magurran, 2004).
As Osmia rufa exceeded accumulated abundances of all other trap-nesting pollinators by 6.2 times on average, it was excluded from all dependent variables, except (i) estimated species richness. This allowed for a detailed analysis of otherwise blurred diversity responses in the co-occurring species. Instead, the number of nests of O. rufa entered all models as an additional explanatory variable. As this number of nests strongly correlated with O. rufa's abundance (t = 19.15, df = 9, P < 0.001), i.e., the number of provisioned cells in these nests, we thus accounted for potential competition of this dominant species with the remaining pollinator community not only for nesting but also for food resources. In addition, we correlated O. rufa abundances across the year as well as during and after mass flowering with those of all other bees and wasps in the community using R-function cor.test. Following the exclusion of O. rufa nests for calculation of dependent variables, we then determined the number of (ii) emerging bee and wasp individuals for each site; (iii) vacated nests of bivoltine species' first generation; and (iv) mortality rates.
Prior to testing for landscape effects, each untransformed response variable was correlated with the scale-dependent landscape variables percentage area of oilseed rape and percentage area of seminatural habitat across all scales with Spearman Rank Correlations using R-function cor.test (cf. Steffan-Dewenter et al., 2002). For each response variable, landscape variables were selected at the scale that yielded the highest r-value for entering the full model.
Landscape variables at these selected scales were then used for transforming dependent variables, applying R-function boxcox in package MASS (Venables & Ripley, 2002) to meet normality and homoscedasticity assumptions. In addition to these explanatory variables of main interest, the model specified in this transformation procedure also contained the three scale-independent explanatory variables (1) distance to nearest oilseed rape field, (2) area of study site, and (3) the number of nests of O. rufa. Mortality rate was transformed by taking the arcsin of the square root of the value. Each of these transformed dependent variables was then again correlated with the percentage areas of oilseed rape and seminatural habitats at all spatial scales to determine the decisive spatial scale to enter full models together with the three scale-independent explanatory variables already used in the transformation procedure. Analyses were conducted using generalized linear models (glm) and the corrected Akaike information criterion (stepAICc) for stepwise (direction = both) model selection in R 2.14.2 (R Development Core Team, 2012). The model yielding the lowest AICc according to this procedure always had the lowest number of predictor variables and was selected as the final model. We used R-function lm on final models to retrieve adjusted R2 values.
Altogether, 1540 trap-nesting bees and wasps from 16 species were recorded. O. rufa was represented by 1208 individuals (78%). The remaining nine bee species contributed 247 individuals; six wasp species contributed 85 individuals. Neither across the year (t = 0.6034, df = 8, P = 0.563) nor for the separate periods during (t = −0.4999, df = 8, P = 0.631) nor after (t = 1.0756, df = 8, P = 0.314) the mass flowering of oilseed rape were abundances of O. rufa and the total of all other trap-nesting bees and wasps significantly correlated. On average, 66% ± 21 (standard deviation) of all individuals per site hatched from nests built during the mass flowering of oilseed rape. However, early nests were dominated by O. rufa. Once this dominant species was excluded, 78% ± 31 individuals hatched from nests built after the mass flowering of oilseed rape (332 individuals, Fig. 2).
From a total of 516 nests, hatching bees and wasps were determined for 133 nests, 220 nests of bivoltine species' first generations were vacated during exposure in the field without record, and bees or wasps died before hatching in 163 nests. Based on the records from 133 nests with species identification, estimated species richness of trap-nesting bees and wasps across the year was significantly positively related to the percentage area of oilseed rape at the 1750 m scale and seminatural habitat at the 1500 m scale (Fig. 2a–b; Table 2). In absolute numbers, estimated species richness increased twofold from approximately three to six species with an increase in oilseed rape from 4 to 15% and from 2.5 to 5 species with an increase in seminatural habitats from 2 to 14%. While both landscape variables were not significantly related to estimated species richness during the mass flowering of oilseed rape, after the mass flowering the percentage areas of oilseed rape and seminatural habitats were significantly positively related to estimated species richness at the 2000 m and 1000 m scale, respectively (Fig. 2c–d; Table 2). In absolute numbers, observed species richness after the mass flowering of oilseed rape increased from approximately two to five species with an increase from 4 to 15% in oilseed rape and from 2 to 17% in seminatural habitats.
Table 2. Generalized linear model results for four dependent variables. Explanatory variables, the corresponding spatial scale (radius) and the adjusted R² are given for the most parsimonious model. Models were reduced using a stepwise (direction = both) selection with AICc as a criterion to omit terms. For mortality rate, there was no significant final model
Adjusted R2 (whole model)
Estimated species richness (Chao1)
Seminatural habitats were also positively related to the abundance of trap-nesting pollinators across the year at the 1250 m scale (Fig. 2e; Table 2) and after the mass flowering of oilseed rape at the 1000 m scale (Fig. 2f; Table 2).
The number of nests from which individuals of bivoltine first generations hatched in the field was significantly negatively related to the percentage area of seminatural habitats at the 1750 m scale across the year and during the mass flowering of oilseed rape (Fig. 2g–h; Table 2).
There were no significant effects of any of our explanatory variables on the mortality rate (i.e., percentage of nests where individuals died before hatching).
Here, we showed that species richness of the community of cavity-nesting bees and wasps comprising early- and late-emerging uni- and bivoltine species increased with the area of oilseed rape. Unlike the dominant wild bee species O. rufa, the abundance of the community's subdominant species was not related to oilseed rape. Congruent with previous studies, richness of bees and wasps also increased with the area of seminatural habitats, as did abundance. This overall abundance effect of seminatural habitats was dominated by the postflowering period, when most nests were built. This indicates that mainly late univoltine bees and wasps and early bivoltine species' second generations benefited. The decreasing number of vacated nests with a decreasing proportion of seminatural habitats, in contrast, suggests a dilution effect for early univoltine species and bivoltine species' first generations.
In contrast to the positive effect on O. rufa (Jauker et al., 2012), a polylectic and univoltine solitary wild bee species with an early phenology utilizing oilseed rape (Holzschuh et al., in press), the total abundance of the remaining community of cavity-nesting bees and wasps was not related to mass-flowering oilseed rape. Considering that the majority (78%) of individuals of these remaining species in the community emerged from nests that were built after the mass flowering of oilseed rape had ceased, this seems less surprising. Given this relatively later phenology of most recorded species, however, the observed increase in species richness of cavity-nesting pollinators with oilseed rape is unexpected. An explanation may be that, similar to the early but univoltine O. rufa (Jauker et al., 2012), the bounty of pollen and nectar resources provided by mass-flowering oilseed rape benefited the first generation of bivoltine species like O. caerulescens. These reproductive benefits in landscapes with high proportions of oilseed rape might then translate into enhanced abundances in this species' second generation, which may render species usually difficult to detect more likely to occur in trap nests. Yet, these increased abundances in subordinate species may be too low to be detected statistically. Alternatively, the increase may be compensated by abundance declines in subdominant species so that – besides on the dominant species O. rufa (cf. Jauker et al., 2012) – an abundance effect of oilseed rape on the remainder of the community of cavity-nesting bees and wasps does not become apparent (cf. Arnan et al., 2011).
The possible mechanism causing the observed positive effect of oilseed rape on species richness may be a competition cascade. Similar to ant communities (Arnan et al., 2011), benefits of the dominant species from mass-flowering oilseed rape may result in increased competition (e.g., nesting resources, cf. Steffan-Dewenter & Schiele, 2008) with the subdominant species. This in turn may lead to a competitive release for the species-rich subordinate compartment of the pollinator community. Because reductions in species number and/or abundance of subdominant species are compensated by an increase in the subordinate compartment, such a cascading effect would also explain the lack of correlative patterns between abundances of O. rufa and the total of the remaining cavity-nesting bee and wasp species (subdominant and subordinate species combined) at any time. Considering previous results on O. rufa (Jauker et al., 2012), however, adverse effects on bee and wasp diversity could have also been expected from parasitoids benefiting from mass-flowering oilseed. Yet, due to relatively small sample sizes, mortality of cells, which was neither related to oilseed rape nor seminatural habitats, could not be differentiated into different processes, such as abortion or parasitation. Future research is required before a concluding evaluation of mass-flowering oilseed rape modulating population dynamics of cavity-nesting pollinators will be possible.
As a consequence of the predominantly later phenology of the species recorded in our trap nests, species richness of cavity-nesting bees and wasps was positively associated with the proportion of seminatural habitats that provide a diverse set of alternative floral resources once the mass flowering has ceased (Mandelik et al., 2012). By providing diverse and more permanent foraging as well as nesting and overwintering resources, seminatural habitats are known to facilitate not only species richness (Steffan-Dewenter et al., 2002; Le Féon et al., 2010) but also bee abundances (Steffan-Dewenter et al., 2002; Jauker et al., 2012). Congruently, in the present study, the abundance of cavity-nesting bees and wasps was positively related to seminatural habitats. Analyses at higher temporal resolution revealed that this relationship was only apparent after the mass flowering of oilseed rape. Thus, being masked in the whole-year analysis, the lack of such an abundance effect during mass flowering indicates the importance of higher temporal resolution in understanding the modulation of annual community dynamics by pulsed and permanent resource availability.
In addition to univoltine (e.g., Heriades truncorum, Hylaeus difformis) or partially bivoltine (e.g., Megachile centuncularis, Discoelius zonalis) species with a late phenology, especially the second generation of early bivoltine species, such as Ancistrocerus nigricornis, A. trifasciatus, or Osmia caerulescens may have benefited in their abundance from resources offered by increasing amounts of seminatural habitats after the mass flowering. As suggested by coarse taxon identification based on brood cell residues (i.e., clay cell walls and cocoon fragments), these latter species' first generations might be responsible for the observed negative relationship between seminatural habitats and early vacated nests. The first nest-building individuals of the year might distribute themselves across the increasingly available seminatural habitats, while they are concentrated to few locations when these elements providing valuable nesting sites are scarce. A similar dilution of pollinators was reported in grasslands with increasing foraging resources (i.e., mass-flowering crops; Holzschuh et al., 2011). Once the first generation is established, this trend seems reversed, as foraging females initializing the second generation may benefit from increasing availability of resources in seminatural habitats. Potentially, this population growth of the second generation exceeds the dilution effect of the first generation, thus resulting in the positive response of recorded late abundances to seminatural habitats. To substantiate this assumption, however, a higher temporal resolution is needed in the assessment of species-specific nesting activity throughout the year.
Our results may contribute to a more complete evaluation of how increased biofuel cropping may affect the diversity of flower-visiting insects and potentially the services (i.e., pollination and biocontrol) they provide in modern agro-ecosystems. The increase in species richness of cavity-nesting bees and wasps with oilseed rape observed in this study and positive responses of pollinator abundance and reproductive success previously shown suggest that overall the community of flower-visiting insects benefit from mass-flowering oilseed rape. Yet, insect phenology seems to be a very important trait in affecting the response to oilseed rape. Disadvantages or transient benefits of this mass-flowering crop previously shown for long- or short-tongued bumblebees, respectively, further confirm the trait specificity in this response. With regard to the most important general outcome of the invariable value of seminatural habitats for bee and wasp diversity, we strongly recommend securing high proportions of seminatural habitats enabling early positive effects of mass-flowering biofuel crops on pollinators and biocontrol agents to be carried on throughout the year and temporarily dynamic disservices to be buffered against in intensively used modern agro-ecosystems.
We thank farmers for their cooperation and Christoph Scherber for providing the stepAICc R-code.