You are on page 1of 9


Basic and Applied Ecology 9 (2008) 682–690

Insect pests in winter oilseed rape affected by field and landscape characteristics
Johann G. ZallerÃ, Dietmar Moser, Thomas Drapela, Claudia Schmo ¨ ger, Thomas Frank
Department of Integrative Biology and Biodiversity Research, Institute of Zoology, University of Natural Resources and Applied Life Sciences, Gregor Mendel Strasse 33, A-1180 Vienna, Austria Received 24 January 2007; accepted 18 October 2007

Winter oilseed rape (OSR, Brassica napus) cropping is often associated with an intensive use of pesticides. The transformation of structurally rich landscapes into more monotonous landscapes may be partly responsible for this, because non-crop habitats believed to benefit natural enemies have been eliminated. We examined the influence of field (soil quality, nitrogen fertilization) and landscape characteristics (OSR area and isolation, non-crop area, landscape diversity, proportions of grassy fallows and woody areas) on three major European OSR pest groups: pollen beetles, stem weevils, and brassica pod midges. Twenty-nine landscape sectors ranging from structurally poor to complex were studied at eight spatial scales (radii 250–2000 m) centered in the studied OSR fields. Abundances of pollen beetles and stem weevils were significantly positively correlated with soil quality and negatively related to OSR area in the surroundings. Generally, abundances of all groups were positively related to woody areas, but not related to grassy fallow area. Pod midges and stem weevils tended to respond primarily to landscape variables at small (250–500 m) and medium (1000–1250 m) scales, while pollen beetles responded at medium to large (1000–2000 m) scales. The results are discussed in relation to differences in overwintering strategies and mobility of pest insects. Strategies at the field and landscape level, aiming to reduce pest pressure in OSR fields, are also discussed. ¨ kologie. Published by Elsevier GmbH. All rights reserved. r 2007 Gesellschaft fu ¨rO

Der Anbau von Winterraps (OSR, Brassica napus) ist oft mit einem intensiven Pestizideinsatz verbunden. Zumindest teilweise fu ¨ r diese Entwicklung verantwortlich gemacht wird die Umgestaltung von strukturreichen zu monotonen Landschaften, da dabei naturnahe Habitate verloren gehen, die fu ¨ dlinge ¨ r die natu ¨ rlichen Gegenspieler der Scha fo ¨ rderlich sind. Wir untersuchten den Einfluss von Feld- (Bodenqualita ¨ t, Stickstoffdu ¨ ngung) und Landschaftscharakteristika (Anteil und Isolation der Rapsfla ¨ chen, naturnahe Fla ¨ chen, Landschaftsdiversita ¨ t, Anteil an Grasbrachen und Geho ¨ lzfla ¨ chen) auf die Abundanzen von drei in Europa wichtigen Gruppen von Rapsscha ¨ dlingen: Rapsglanzka ¨ fer, Gefleckter Kohltriebru ¨ ngelru ¨ ssler und Großer Rapssta ¨ ssler und Kohlschotenmu ¨ cke. An 29, von strukturarm bis komplex reichenden Landschaftssektoren wurden auf acht konzentrischen Skalen um das Untersuchungsfeld (Radius 250–2000 m), die Beziehungen zwischen Scha ¨ dlingsabundanzen und Feld-/Landschaftvariablen untersucht. Die Abundanzen der Rapsglanzka ¨ fer und der Sta ¨ ngelru ¨ ssler korrelierten signifikant positiv mit der Bodenqualita ¨ t und signifikant negativ mit der Rapsfla ¨ che in der Umgebung. Generell waren die Abundanzen aller
ÃCorresponding author. Tel.: +43 1 47654 3205; fax: +43 1 47654 3203.

E-mail address: (J.G. Zaller). ¨ kologie. Published by Elsevier GmbH. All rights reserved. 1439-1791/$ - see front matter r 2007 Gesellschaft fu ¨rO doi:10.1016/j.baae.2007.10.004

J.G. Zaller et al. / Basic and Applied Ecology 9 (2008) 682–690 683

Scha ¨ dlingsarten positiv beeinflusst von der umgebenden Waldfla ¨ che aber unbeinflusst vom Anteil an grasigen Brachen. Kohlschotenmu ¨ ngelru ¨ heren (250–500 m) und ¨ cken und Sta ¨ ssler reagierten vorwiegend auf Landschaftsvariablen im na mittleren Umfeld (1000–1250 m), wa ¨ hrend Rapsglanzka ¨ fer auf mittleren bis großen Skalen reagierten (1000–2000 m). ¨ berwinterungsstrategien und der Mobilita Die Resultate werden in Bezug zu den unterschiedlichen U ¨ t der Scha ¨ dlinge diskutiert. ¨ kologie. Published by Elsevier GmbH. All rights reserved. r 2007 Gesellschaft fu ¨rO
Keywords: Canola; Ceutorrhynchus napi; Ceutorrhynchus pallidactylus; Dasineura brassicae; Insect pests; Landscape ecology; Meligethes aeneus; Oilseed crop; Pest–crop interactions; Spatial scales

In agroecological research it has been appreciated that plant–insect interactions and other ecological processes depend on scales much larger than a single habitat (Levin, 1992; Schneider, 1994; Wiens, 1989). Moreover, studies using a landscape approach where area proportions, spatial arrangement and connectivity of habitats were considered have shown that the influence of landscape complexity and structure tends to be species-specific and communities are made up of species with different spatial strategies (Kareiva & Wennergren, 1995; Pickett & Cadenasso, 1995; Thies & Tscharntke, 1999; Wiens, Schooley, & Weeks, 1997; With, Pavuk, Worchuck, Rhonda, & Fisher, 2002). Crop–pest interactions have mainly been studied on single-pest species by focusing either on the impact of field parameters or on landscape structure but only rarely included both ¨ stman, Ekbom, & Bengtsson, 2001a). We factors (O investigated how the abundances of three major functional groups of insect pests in winter oilseed rape (OSR, Brassica napus L., canola) responded to field parameters and landscape characteristics at various spatial scales. We are not aware of any other studies that have examined insect pest–field/landscape interactions on more than one group of insect pests. The importance of OSR as a source for industrial and nutritional oil has been increasing worldwide during the last decades as reflected by an increase of its acreage from 8.2 million ha to 27.0 million ha from 1970 till 2005 ( In many regions this is accompanied by a dramatic disproportionate increase of pesticide application (Gianessi & Marcelli, 2000). Prevalent pest problems are often attributed to the transformations of formerly heterogeneous landscapes with high proportions of seminatural non-crop habitats to more monotonous landscapes mainly dominated by arable land and its local decrease of biodiversity (Kareiva, 1990; Pickett & Cadenasso, 1995) and detrimental consequences for interactions between pests and beneficial organisms (Menalled, ¨ stman, Ekbom, Marino, Gage, & Landis, 1999; O Bengtsson, & Weibull, 2001b; Polis, Anderson, & Holt, 1997; Roland & Taylor, 1997; Thies & Tscharntke, 1999).

Pest species considered in the current study include (i) stem weevils (Ceutorhynchus pallidactylus Marsh. and Ceutorhynchus napi Gyll., Coleoptera, Curculionidae) that lay eggs in leaf petioles or midribs of OSR plants while the larvae tunnel in the stems; (ii) pollen beetles (Meligethes aeneus Fabr. and M. viridescens Fabr., Coleoptera, Nitidulidae) that feed on pollen and destroy flower buds and (iii) brassica pod midge (Dasineura brassicae Winn, Diptera, Cecidomyiidae) that lay eggs into OSR pods where the hatched larvae consume the seeds as well as tissue of the pod walls and cause the pods to split prematurely (Alford, Nilsson, & Ulber, 2003). Stem weevils either pupate in the soil of the OSR field or in woodland and grassy field boundaries and are not very mobile; the pod migdes have more than one generation a year and pupate in the OSR field from which the weakly flying adults invade new OSR fields of the next year; the very mobile pollen beetles hibernate in woods and other sheltered sites (see Alford et al., 2003 and literature therein). With the exception of pollen beetles these pest species have never been studied in a landscape context. Because pest species must migrate to colonize the crop the spatial patterns of OSR fields and potential overwintering sites were assumed to affect their abundances. The specific objectives of this study were to determine (i) whether field and landscape characteristics influence the major OSR pests in different ways, (ii) at which spatial scales the influence of landscape variables are effective and (iii) possible consequences for field and landscape management. Generally, we hypothesized that the abundances of these specialized pests should be more related to variables describing OSR in the landscape than to non-crop variables and that less mobile D. brassicae would respond to landscape characteristics at smaller scales than the more mobile Ceutorhynchus and Meligethes species.

Materials and methods
Study area and landscape structure
The study region (about 240 km2) is located in the relatively flat region of Lower Austria, about 40 km east

684 J.G. Zaller et al. / Basic and Applied Ecology 9 (2008) 682–690

of Vienna (altitudes of arable land ranging from 130–250 m a.s.l.; coordinates of the central area: 161570 E, 481040 N). During the study period from April to July 2005 the average temperature was 16.9 1C; the rainfall during this period amounted to 336 mm (average temperature and rainfall in these months between 1990 and 2004 was 16.0 1C and 230 mm, respectively; weather data were provided from the meteorological station Bruck/Leitha located about 16 km to the west of the study region). The region mainly consists of farmland on chernozem soils cropped according to integrated pest management guidelines (main crops: wheat, maize, barley, OSR, sunflower, sugar beet, poppy seed) interspersed with semi-natural non-crop areas such as fallows, hedges and forest remnants. The 29 OSR study fields were randomly selected from this region that covered a landscape complexity gradient ranging from structurally poor to complex (average minimum distance between study fields: 1198 m). Scale effects were investigated by calculating landscape variables at eight circular sectors ranging from 250 to 2000 m radius around our study plots using the software packages ArcGis 9.1 and ArcView GIS 3.3 (ESRI Redlands, CA, USA). Intensive field surveys using real-color orthophotos (minimum resolution 0.25 m) were conducted in 2005 to assess the current landscape composition. Landscape variables calculated were proportions of OSR fields, non-crop area, area of grassy fallows and woody areas (see Appendix A: Table 1). To characterize the landscape diversity of the study area, the Shannon–Wiener index (Krebs, 1994) was calculated based on 14 habitat types (arable land, grassy fallows, woody fallows, copse, dry grassland, dry shrubland, dry forests, riparian forest, woods along rivers, hedges, settlements, aquatic areas, roads, vineyards). Isolation of OSR fields in the landscape was calculated by using a negative exponential weighting function, where isolation was calculated based on the distances of neighbouring OSR fields to our study field and OSR area of a given spatial scale using the formula: Isolation OSRi ¼ ÀS(eÀdistance  OSR areaj)/ SeÀdistance (Kruess, 2003). Information on field variables was obtained by a questionnaire among the participating farmers (see Appendix A: Table 1). The soil index specifies the natural yield capacity of a field in relation to the highest yielding area of the country (0ysoil with lowest yield capacity, 100ysoil with highest yield capacity) and is an integrative measure of soil quality that encompasses additionally the soil type, humus content, soil depth, texture, density, structure, lime ¨ BG, 2001). content, gleying and soil congregation (O

at seeding rates between 3.1 and 5.2 kg haÀ1. OSR fields were fertilized and treated with herbicides, fungicides and insecticides following common agricultural practice until December 2004. Starting January 2005, 1 ha of each of the OSR fields was excluded from pesticide applications and used for sampling insect pests and crop plants for this study. A buffering zone of at least 10 m to the next sprayed OSR was kept. The preceding crop was winter barley for 18 fields, winter wheat for 10 fields and poppy seed for one field (there was no relation between the preceding crop and total pest abundances: r2 ¼ 0.032, P ¼ 0.354; logistic regression analysis). Stem weevil larvae were sampled in late April 2005 by removing 25 randomly chosen OSR plants along a 50 m transect located in the central area of each study field, stems were dissected subsequently and weevil larvae therein counted. Two stem weevils (C. pallidactylus and C. napi) are major pests in the region with C. pallidactylus comprising more than 80% of the stem weevil abundance. Adult pollen beetles were sampled during OSR flowering in late April 2005 at sunshine and low wind velocity between 10 a.m. and 2 p.m. on 25 randomly chosen OSR plants along a 50 m transect located in central area of each field. A plastic bag was put over the top raceme, the raceme was cut off, and pollen beetles were counted in the laboratory. Pollen beetle communities comprise several species of the genus Meligethes. In our study region, M. aeneus is the predominant one, M. viridescens can be observed only sporadically (Kraus & Kromp, 2002). Brassica pod midge larvae (D. brassicae W) were assessed in late May 2005 in 100 randomly chosen pods from the top racemes of OSR plants along a 50 m transect located in the central area of each field after dissecting the pods in the laboratory. For better comparability of fields, all pest abundance data were expressed on a 1-m2 basis by using OSR stand density data (average number of stems assessed in two 1m2 frames). OSR plants in the study plots were cut within two 1-m2 frames per field in the third week of June 2005 and yields were assessed after threshing of pods and wind sifting.

Statistical analyses and calculations
Effects of field and landscape variables on abundance of OSR pest species were analysed using an ordinary least-squares regression (OLS) model. Response variables were tested for normality using Shapiro–Wilk W-statistic and log-transformed when necessary to meet criteria for regression analyses (Zar, 1996). To analyse scale effects, we compared the predictive power of landscape variables measured at the eight spatial scales (250, 500, 750, 1000, 1250, 1500 and 2000 m radius centered in our study fields) in a univariate regression

Crop management and pest sampling
OSR (cv. Californium) was sown by the farmers on the 29 fields between 20 August and 14 September 2004

J.G. Zaller et al. / Basic and Applied Ecology 9 (2008) 682–690 685

approach. Prior to multivariate analyses we tested the predictor variables for intercorrelations (see Appendix A: Table 2). Of the intercorrelating variables we excluded those from the subsequent analyses that did not characterize specific non-crop elements (N fertilization level, landscape diversity, non-crop area, OSR isolation). Significant predictor variables were selected by performing a forward stepwise selection followed by a backward elimination procedure (Yee & Mitchell, 1991). Possible non-linear effects were accounted for by using restricted quadratic and cubic spline functions (Harrel, 2001). Linearity of the predictor variables were tested by comparing linear and non-linear terms (Wald tests, F-statistics; Po0.05). At each step of the variable selection, the significance of partial effects was tested by dropping each predictor term from the model (Wald tests, F-statistics; Po0.05). Because variables arising from such nested measurement designs can be correlated, we allowed only the most significant of a group of correlated variables (e.g., one landscape variable at different radii and different variables at the same radius) to be included into the next step of variable selection. To test for possible spatial autocorrelation we used Moran’s I correlograms of residuals of the landscape regression model (Legendre & Legendre, 1998; Lichstein, Simons, Shriner, & Franzreb, 2002).

The investigated landscape sectors varied considerably in their structure and complexity, and OSR fields differed in soil index and nitrogen fertilization (see Appendix A: Table 1). Overall, we found great variation in pest abundances among the studied OSR fields with mean abundances (7SE) of brassica pod midge 126716 mÀ2, pollen beetle adults 346777 mÀ2 and stem weevil larvae 3074 mÀ2. Abundance of pollen beetle adults was significantly positively correlated with stem weevil larvae (r ¼ 0.650, Po0.001) and pod midge larvae (r ¼ 0.440, P ¼ 0.015); abundance of pod midge larvae was not correlated with stem weevil larvae (r ¼ 0.318, P ¼ 0.086). Univariate regression analyses showed that only abundances of pollen beetles (r2 ¼ 0.359, P ¼ 0.001) and stem weevils (r2 ¼ 0.291, P ¼ 0.002) but not of pod midges (r2 ¼ 0.015, P ¼ 0.492) were positively related to soil index; nitrogen fertilization levels were not significantly related to pest abundances. Pollen beetles showed a negative relationship to OSR area (Fig. 1B) and positive relations with the isolation of OSR in the landscape (Fig. 1E), the proportion of non-crop (Fig. 1K, woody areas (Fig. 1N) and landscape diversity (Fig. 1H. Stem weevil larvae were negatively related to OSR area (Fig. 1C), but positively related to OSR

isolation (Fig. 1F) and the proportion of woody areas (Fig. 1O). Abundance of pod midge larvae was positively related to landscape diversity (Fig. 1G) and the proportion of woody areas (Fig. 1M). For most pest species responses to landscape variables showed clear maxima at certain scales (Fig. 1). Abundance of pod midge larvae was explained by proportion of woody areas up to 500 m only but no relation to woody areas at greater radii (Fig. 1M). Additionally, pod midge abundance was significantly associated with landscape diversity at 1000 and 1250 m radius but was unrelated at other scales (Fig. 1G). Pollen beetle abundance was related to OSR area across all tested scales (Fig. 1B) and related to OSR isolation at a low (250 m) and larger radii (X1250 m; Fig. 1E). Additionally, pollen beetle abundance was significantly related to landscape diversity at 1500 m (Fig. 1H, noncrop area between 1250–1750 m radius (Fig. 1K and woody areas above a radius of 250 m (Fig. 1N). Abundance of stem weevil larvae was significantly explained by OSR area between 500 and 1000 m (Fig. 1C), by the isolation of OSR at 750 and 2000 m radius (Figs. 1F) and by woody areas between 250 and 1000 m radius and at 2000 m radius (Fig. 1O). Abundances of tested pest species showed generally no relation to the area of grassy fallows (data not shown). Multivariate OLS analysis for pod midge abundances revealed that no other variable except the proportion of woody areas could significantly enhance the predictive power of the model (univariate model result: r2 ¼ 0.279, P ¼ 0.009). The multivariate model for the pollen beetles (r2 ¼ 0.770, Po0.001) contains two variables, the proportion of OSR at 1000 m radius (partial r2 ¼ 0.368, Po0.001) and soil index (partial r2 ¼ 0.102, P ¼ 0.0137) with a distinct negative response to OSR area and a curvilinear relationship to soil index. The plot of the partial effects indicated that pollen beetle abundance was affected by the soil index mainly below 4% OSR area in the landscape (Fig. 2A). Above 4% OSR area soil index played only a marginal role in modulating the primary correlation of pollen beetle abundance to OSR area. Maximum pollen beetle abundance is predicted by this model at lowest levels of OSR area and average soil index values (Fig. 2A). The final model for the stem weevil data (r2 ¼ 0.550, P ¼ 0.019) also consisted of two variables: soil index (partial r2 ¼ 0.335, P ¼ 0.001) and proportion of woody areas at a radius of 250 m (partial r2 ¼ 0.116, P ¼ 0.017). The plot of the partial effects shows that abundance was highest when soil index values were slightly above average and proportion of woody areas is high (Fig. 2B). While abundance showed a linear response to woody area, the relation to soil index is curvilinear and resembles a saturation curve (Fig. 2B). The Moran’s I correlograms gave no indication for spatial patterns in the residuals of the OLS model.

686 J.G. Zaller et al. / Basic and Applied Ecology 9 (2008) 682–690

Fig. 1. Relationships between abundance of pod midge larvae (D. brassicae), adult pollen beetles (M. aeneus+M. viridescens) and stem weevil larvae (C. pallidactylus+C. napi) and landscape variables derived from univariate ordinary least-square regression analyses. Scale-dependent predictors are displayed for all investigated radii. Asterisks denote statistical significance: *Po0.05, **Po0.01. With the exception of graphs B+C all relationships are positive.

OSR yield was significantly negatively correlated with pollen beetle abundance (Fig. 3B), but unrelated to the abundance of pod midge larvae (Fig. 3A) or stem weevils (Fig. 3C).

Pest responses to field characteristics
Abundances of pollen beetle adults and stem weevil larvae showed strong relations to soil quality but remained unrelated to nitrogen fertilization levels; pod midges showed no relation to field characteristics tested. Because pollen beetles and stem weevils do not overwinter or pupate in current OSR fields, their soil quality is unlikely to have any direct effect on their current

abundance. However, it is most likely that bottom-up effects via the nutritional quality or the canopy microclimate of OSR could have influenced the searching efficiency of pollen beetles and stem weevils and thus altered their abundances (Walters, Young, Kromp, & Cox, 2003). In fact, glucosinolates and their catabolites in OSR have been found to be important cues for host selection by pests of crucifers, aiding both orientation to and recognition of the host plant (Bartlet, 1996). In the current study we used a double-zero cultivar that was reduced in its glucosinolate and erucic acid content, however other secondary compounds may have been affected by soil quality (Nicholsorians, 1991). The curvilinear shape of the relationship between pollen beetle and stem weevil abundance and soil index with a maximum at average levels indicates that at low soil index, OSR quality was not suitable for pests. On the other hand at a higher soil quality the crop could have

J.G. Zaller et al. / Basic and Applied Ecology 9 (2008) 682–690 687

exhibited a better ability to protect itself from herbivores via the production of secondary compounds (Cipollini & Bergelson, 2002). It remains to be tested experimentally, whether soil quality could also have affected the colour and odour of OSR plants and thereby altered their attractiveness to pest species.

Pest responses to landscape characteristics
Generally, we hypothesized our specialist pest species to be more responsive to the proportion and isolation of OSR than to non-OSR elements in the landscape (Holt,

Fig. 2. Partial effects of field and landscape variables on the abundance of (A) pollen beetles adults (M. aeneus+M. viridescens) and (B) stem weevil larvae (C. pallidactylus+C. napi) derived from multivariate ordinary least-square regression analyses.

Lawton, Polis, & Martinez, 1999). This was confirmed for pollen beetles, partly confirmed for stem weevils but not for pod midges. Pollen beetles and stem weevils were more abundant in landscapes with less OSR area and less abundant when more OSR area was available. This is in contrast to findings by Thies, Steffan-Dewenter, and Tscharntke (2003), however could be explained by the different methodological approaches. The colonization by pollen beetle of a few potted OSR plants located in old field margins in the former study may have differed considerably from that of field-grown plants in the current study. Generally, the pest distribution patterns could be expected to mirror areas of OSR and overwintering sites of the preceding year that enabled the buildup of a certain landscape pest pool that was then dispersed among available OSR area in the current year (Hokkanen, 2000). However, since in our region OSR cropping history and non-crop areas did not change considerably during the last years because agri-environmental programmes provided subsidies to promote both OSR cropping and extensification in the region (BMLFUW, 2006), we assume that pest migration processes are most likely responsible for these findings. Pollen beetle abundance and non-crop area were positively related suggesting that more complex landscapes support a greater variety of alternative host plants for pollen beetles, and thereby complex landscapes may have enhanced pest populations (Frenzel & Brandl, 1998). However, this finding appears to be in contrast to that of others (Thies et al., 2003; Thies & Tscharntke, 1999) who showed a negative relationship between pollen beetle damage (i.e., number of podless stalks) and non-crop area. Because pollen beetle abundance and damage caused are well correlated (Zaller, Moser, Drapela, Schmo ¨ ger, & Frank, 2008) several explanations could be considered for this discrepancy. First, the parameter non-crop area per se does not adequately describe landscapes because noncrop usually encompasses various habitats (e.g., fallows, woody areas, meadows) that can exert contrasting effects on pest abundances. For instance, in our study

Fig. 3. Relationship between OSR yield and abundance of pod midge larvae (D. brassicae), pollen beetles (M. aeneus+M. viridescens) and stem weevil larvae (C. pallidactylus+C. napi). Formulae, r2 and P-values derived from linear regression fits (n ¼ 29).

688 J.G. Zaller et al. / Basic and Applied Ecology 9 (2008) 682–690

non-crop was mainly comprised of grassy fallows and woody areas, but only the latter showed a significant effect on pest abundances. Moreover, in our study grassland elements were mainly composed of grassy fallows while in the former studies pastures and meadows formed the grassland elements (Thies et al., 2003; Thies & Tscharntke, 1999). Second, contrasting findings might again result from the fundamentally different methodological approaches used. The positive relations between pollen beetles and non-crop in our study also suggest that natural enemies promoted by non-crop areas (e.g. hymenopterous parasitoids – Thies ¨ stman et al., & Tscharntke, 1999; carabid beetles – O 2001b; spiders – Schmidt, Roschewitz, Thies, & Tscharntke, 2005; Drapela et al., unpublished data) had no detrimental influence on pests in our region. Indeed we observed that parasitoids of pollen beetles seemed to play only a minor role in our landscapes/ region (Zaller et al., unpublished data) and there is evidence that besides beneficial organisms also pests are supported by non-crop structures (Thies, Roschewitz, & Tscharntke, 2005). Pod midge abundance was unrelated to OSR variables but positively related to landscape diversity and woody areas. This finding seems somewhat surprising because pod midge adults are reported to emerge from the OSR field of the last year and in the following spring migrate to the current OSR fields (Alford et al., 2003). However, it also has been observed that second and third generations of these polyvoltine pests are overwintering in stands of other Brassicaceae (Paul, 2003). Thus, our data suggest that for this species a more diverse landscape with more woody areas increases the likelihood of finding alternative host plants. However, admittedly our understanding of landscape effects on pod midges is still too scarce to be able to further interpret our findings. Two general trends for the three species groups could be seen in our data: (i) pest abundances were consistently positively related to woody areas, and (ii) an absence of any relation between pest abundances and grassy fallows. For pollen beetles, the positive effect of woody areas can be explained by their ability to serve as overwintering sites. For stem weevils and pod midge that are thought to overwinter in non-woody areas it is possible that woody areas in the vicinity of the actual overwintering habitats could have climatically influenced these habitats (e.g., reduced wind exposition or increased relative humidity; Foggo, Ozanne, Speight, & Hambler, 2001). No detrimental effects of grassy fallows on populations of OSR pests could be identified although these sites have frequently been shown to be important overwintering habitats for natural enemies ¨ stman et al., 2001b; Thies of OSR pest species (O et al., 2003). The lack of a relation between pests and grassy fallows also suggests that a widely used assump-

tion that fallows are refuges for insect pests (Lethmayer, Nentwig, & Frank, 1997) could not be substantiated by our data.

Spatial aspects of pest responses
Our data showed that the scale at which each insect group is influenced is related to each groups dispersal capacity. Generally, pod midges seemed to respond to landscape factors at small (250–500 m; woody areas) and medium scales (1000–1250 m; landscape diversity), pollen beetles showed effects at medium to large scales (1000–2000 m; all variables except grassy fallows) while stem weevils were related to landscape factors mainly at small to medium scales (500–1000 m; OSR area, isolation OSR, woody areas). A limitation of available OSR food at certain scales seemed unlikely to be responsible for these relationships, however dispersal limitations might be the reason for these patterns (Kruess & Tscharntke, 1994, 2000; Tscharntke & Kruess, 1999; With, Cadaret, & Davis, 1999). Pod midges are known to be weak fliers with females usually dispersing no more than a few hundred meters from their emergence sites. This might explain their response to woody areas at the very smallest scales, however our data also indicate that responses at medium scales are possible. In contrast, pollen beetles can easily cross distances greater than that considered in the current study (Fritzsche, 1957) and are therefore affected by landscape characteristics on medium to large scales. Information on the mobility of stem weevils is generally scarce but they are reported to preferably make short flights (Schmutterer, 1956). Based on the current results, however, this did not seem to hinder them to react to landscape effects across all scales. Overall, this study is a first attempt to understand how abundances of different pest species might be affected by landscape characteristics and at what spatial scale this occurs, however, much more research is necessary to be able to elucidate the underlying processes particularly regarding pest migration patterns and dispersion between metapopulations of pests in a landscape. From a farmer’s perspective pollen beetles appear to be the most important species because they showed the most pronounced negative impact on yields, however, data also showed that the studied pest species were generally co-occurring in our fields. Our results suggest that crop rotation schemes commonly in place in order to reduce the carry-over of OSR pest and disease problems should be expanded by a landscape perspective. However, more multi-scale assessments conducted in different regions are necessary to adequately pinpoint how habitat features and the spatial configuration of crop and non-crop elements can affect the abundance of pest populations.

J.G. Zaller et al. / Basic and Applied Ecology 9 (2008) 682–690 689

We are grateful to the farmers for participating in this research project, the staff of the University Research Farm Groß-Enzersdorf for providing post-harvest equipment and to the Austrian Science Fund for supporting J.G.Z., D.M. and T.D. (Grant no. P16972). The regional governments of Lower Austria and Burgenland provided maps and aerial photographs of the project area. Help by Norbert Schuller and Erhard Tesarik in the field and laboratory is gratefully acknowledged.

Appendix A. Supplementary materials
Supplementary data associated with this article can be found in the online version at doi:10.1016/j.baae. 2007.10.004.

Alford, D. V., Nilsson, C., & Ulber, B. (2003). Insect pests of oilseed rape crops. In D. V. Alford (Ed.), Biocontrol of oilseed rape pests (pp. 9–42). Oxford, UK: Blackwell. Bartlet, E. (1996). Chemical cues to host–plant selection by insect pests of oilseed rape. Agricultural Zoology Reviews, 7, 89–116. BMLFUW (Ed.) (2006). Gru ¨ner Bericht 2006. Vienna, Austria: Bundesministerium fu ¨ r Land- und Forstwirtschaft, Umwelt und Wasserwirtschaft. Cipollini, D. F., & Bergelson, J. (2002). Interspecific competition affects growth and herbivore damage of Brassica napus in the field. Plant Ecology, 162, 227–231. Foggo, A., Ozanne, C. M. P., Speight, M. R., & Hambler, C. (2001). Edge effects and tropical forest canopy invertebrates. Plant Ecology, 153, 347–359. Frenzel, M., & Brandl, R. (1998). Diversity and composition of phytophagous insect guilds on Brassicaceae. Oecologia, 113, 391–399. ¨ kologie der Rapsscha Fritzsche, R. (1957). Zur Biologie und O ¨dlinge aus der Gattung Meligethes. Journal of Applied Entomology, 40, 222–280. Gianessi, L. P., & Marcelli, M. B. (2000). Pesticide use in US crop production: 1997. Washington, DC, USA: National Center for Food and Agricultural Policy. Harrel, F. E. (2001). Regression modeling strategies. New York, USA: Springer. Hokkanen, H. M. T. (2000). The making of a pest: Recruitment of Meligethes aeneus onto oilseed Brassicas. Entomologia Experimentalis et Applicata, 95, 141–149. Holt, R. D., Lawton, J. H., Polis, G. A., & Martinez, N. D. (1999). Trophic rank and the species–area relationship. Ecology, 80, 1495–1504. Kareiva, P. (1990). Population dynamics in spatially complex environments: Theory and data. Philosophical Transactions of the Royal Society of London Series B – Biological Sciences, 330, 175–190.

Kareiva, P., & Wennergren, U. (1995). Connecting landscape pattern to ecosystem and population processes. Nature, 373, 299–302. Kraus, P., & Kromp, B. (2002). Parasitization rates of the oilseed rape pests Ceutorhynchus napi, Ceutorhynchus pallidactylus (Coleoptera, Curculionidae) and Meligethes aeneus (Coleoptera, Nitidulidae) by Ichneumonids in several localities of eastern Austria. IOBC/WPRS Bulletin, 25, 117–122. Krebs, C. J. (1994). Ecology: The experimental analysis of distribution and abundance (4th ed.). New York, USA: Harper Collins College Publishers. Kruess, A. (2003). Effects of landscape structure and habitat type on a plant–herbivore–parasitoid community. Ecography, 26, 283–290. Kruess, A., & Tscharntke, T. (1994). Habitat fragmentation, species loss, and biological control. Science, 264, 1581–1584. Kruess, A., & Tscharntke, T. (2000). Effects of habitat fragmentation on plant–insect communities. In B. Ekbom, M. E. Irwin, & Y. Robert (Eds.), Interchanges of insects between agricultural and surrounding landscapes (pp. 53–70). Doordrecht, The Netherlands: Kluwer Academic Publishers. Legendre, P., & Legendre, L. (1998). Numerical ecology (2nd ed.). Amsterdam, The Netherlands: Elsevier Science. Lethmayer, C., Nentwig, W., & Frank, T. (1997). Effects of weed strips on the occurrence of noxious coleopteran species (Nitidulidae, Chrysomelidae, Curculionidae). Journal of Plant Diseases and Protection, 104, 75–92. Levin, S. A. (1992). The problem of pattern and scale in ecology. Ecology, 73, 1943–1967. Lichstein, J. W., Simons, T. R., Shriner, S. A., & Franzreb, K. E. (2002). Spatial autocorrelation and autoregressive models in ecology. Ecological Monographs, 72, 445–463. Menalled, F. D., Marino, P. C., Gage, S. H., & Landis, D. A. (1999). Does agricultural landscape structure affect parasitism and parasitoid diversity? Ecological Applications, 9, 634–641. Nicholsorians, C. M. (1991). Environmentally induced differences in plant traits – consequences for susceptibility to a leaf-cutter ant. Ecology, 72, 1609–1623. ¨ BG. (2001). Bodenaufnahmesysteme in O ¨ sterreich. MitteiO lungen der o ¨sterreichischen bodenkundlichen Gesellschaft, 62, 219pp. ¨ stman, O ¨ ., Ekbom, B., & Bengtsson, J. (2001a). Landscape O heterogeneity and farming practice influence biological control. Basic and Applied Ecology, 2, 365–371. ¨ stman, O ¨ ., Ekbom, B., Bengtsson, J., & Weibull, A.-C. O (2001b). Landscape complexity and farming practice influence the condition of polyphagous carabid beetles. Ecological Applications, 11, 480–488. Paul, V. H. (2003). Raps. Krankheiten-Scha ¨dlinge-Schadpflanzen. Gelsenkirchen-Buer, Germany: Verlag Th. Mann. Pickett, S. T. A., & Cadenasso, M. L. (1995). Landscape ecology: Spatial heterogeneity in ecological systems. Science, 269, 331–334. Polis, G. A., Anderson, W. B., & Holt, R. Y. (1997). Toward an integration of landscape and food web ecology: The dynamics of spatially subsidized food webs. Annual Reviews in Ecology and Systematics, 28, 289–316.

690 J.G. Zaller et al. / Basic and Applied Ecology 9 (2008) 682–690

Roland, J., & Taylor, P. D. (1997). Insect parasitoid species respond to forest structure at different spatial scales. Nature, 386, 710–713. Schmidt, M. H., Roschewitz, I., Thies, C., & Tscharntke, T. (2005). Differential effects of landscape and management on diversity and density of ground-dwelling farmland spiders. Journal of Applied Ecology, 42, 281–287. Schmutterer, H. (1956). Zur Lebensweise und Beka ¨ mpfung des Großen Rapsstengelru ¨ sslers Ceutorhynchus napi (Gyll). Zeitschrift fu ¨r angewandte Entomologie, 39, 302–315. Schneider, D. C. (1994). Quantitative ecology: Spatial and temporal scaling. New York, USA: Academic Press. Thies, C., Roschewitz, I., & Tscharntke, T. (2005). The landscape context of cereal aphid–parasitoid interactions. Proceedings of the Royal Society London B, 272, 203–210. Thies, C., Steffan-Dewenter, I., & Tscharntke, T. (2003). Effects of landscape context on herbivory and parasitism at different spatial scales. Oikos, 101, 18–25. Thies, C., & Tscharntke, T. (1999). Landscape structure and biological control in agroecosystems. Science, 285, 893–895. Tscharntke, T., & Kruess, A. (1999). Habitat fragmentation and biological control. In B. A. Hawkins, & H. V. Cornell (Eds.), Theoretical approaches to biological control (pp. 190–205). Cambridge: Cambridge University Press.

Walters, K. F. A., Young, J. E. B., Kromp, B., & Cox, P. D. (2003). Management of oilseed rape pests. In D. Alford (Ed.), Biocontrol of oilseed rape insect pests (pp. 43–71). Oxford, UK: Blackwell. Wiens, J. A. (1989). Spatial scaling in ecology. Functional Ecology, 3, 385–397. Wiens, J. A., Schooley, R. L., & Weeks, R. D. (1997). Patchy landscapes and animal movements: Do beetles percolate? Oikos, 78, 257–264. With, K. A., Cadaret, S. J., & Davis, C. (1999). Movement responses to patch structure in experimental fractal landscapes. Ecology, 80, 1340–1353. With, K. A., Pavuk, D. M., Worchuck, J. L., Rhonda, K. O., & Fisher, J. L. (2002). Threshold effects of landscape structure on biological control in agroecosystems. Ecological Applications, 12, 52–65. Yee, T. W., & Mitchell, N. D. (1991). Generalized additive models in plant ecology. Journal of Vegetation Science, 2, 587–602. Zaller, J. G., Moser, D., Drapela, T., Schmo ¨ ger, C., & Frank, T. (2008). Effect of within-field and landscape factors on insect damage in winter oilseed rape. Agriculture, Ecosystems and Environment, 123, 233–238. Zar, J. H. (1996). Biostatistical Analysis (3rd ed.). Englewood Cliffs, USA: Prentice-Hall.