No abstract

Data on weed species currently found in winter oilseed rape, the extent of their occurrence and regional distinctions were collected in autumn 2005, 2006 and 2007 from a total of 1463 oilseed rape (OSR) fields throughout Germany. Investigations took place in untreated parts of OSR fields at the four to six leaf stage of rape. Weed frequencies and densities were determined by randomly placing a 0.1 m 2 square frame ten times in each plot, identifying and counting all plants within the frame. Results presented here mainly focus the distribution of important weed species and differences in their frequencies and densities between German Federal states.The majority of weed species occurring in winter oilseed rape is not evenly spread over Germany. There are weed species with a limited distribution which can be frequently found in some regions but are absent in others. Sisymbrium officinale, Descurainia sophia, Centaurea cyanus, Anchusa spp., Sonchus spp., Rumex spp. or Euphorbia spp. are some weed species with mainly regional importance. Several ubiquitous weed species show regional differences in their densities. Capsella bursa-pastoris, Chenopodium album, Galium aparine, Lamium spp., Matricaria spp., Stellaria media, Thlaspi arvense, Veronica spp. and Viola arvensis were among the most frequent weed species in OSR in all German federal states though sometimes differing significantly in regional densities. Recent field observations reporting an increased spreading of Geranium spp., Sisymbrium spp. or Anchusa arvensis in intense OSR cropping areas are verified by these results.

Arable weeds always accompany arable farming; however, species and abundances vary in space and change in time. Surveys assess weeds on a large spatial scale. The regional and national importance, the abundances in which weeds occur, is derived from surveys. Environmental and management background data assist to explain the weed vegetation. Arable weed surveys can have an agronomic, a botanical or an ecological purpose-we approach the agronomic viewpoint. Scientific value results from spatial data about the arable weed vegetation for modelling the dependencies and predicting future changes on a large scale. Researchers, botanists, farmers, advisors, herbicide manufacturers and other industry players engaged in weed control can get rationale spatial impressions of trends in the weed vegetation. This review focuses on methodical aspects of weed surveys. Forty-three surveys on farmed fields, mainly but not exclusively undertaken in Europe, done in up to nine crops per survey covering periods of 1 to 67 years and including 59 to 4423 fields were excerpted for suitable methods to conduct and analyse surveys. The methods reflect agronomic and other targets, e.g. from vegetation science. When accessing existing data, the interdependencies of the layout and all targets should be handled with care. When setting up surveys for agronomic purpose, it is worthwhile to include 3 years and apply stratified sampling. We recommend assessing abundances rather than presence only. Weed survey data is most suitable for applying multivariate statistics. We recommend using proven methods, without discouraging to explore others. To collect sound explanatory data demands for robust contacts with land managers. Explanatory variables are important in survey analyses to extract knowledge about the drivers of the weed vegetation. We conclude that extra value can be obtained from existing weed survey data. We propose to setup public-private partnerships as a way to carry on with arable weed surveys.