Surveying the arable plant diversity of conventionally managed farmland: a comparison of methods

摘要

Agricultural intensification has led to drastic population declines in Europe's arable plant vegetation, and continuous monitoring is a prerequisite for assessing measures to increase and conserve remnant populations of endangered arable plant species. Unfortunately, strong variation in plot sizes and in-field locations makes comparison of current arable plant monitoring approaches difficult. This study compares different releve approaches in conventionally managed arable fields in Northwest German farmland with respect to plant species detection success and time expenditure. We compared species detection rate and expenditure of time of six different releve types in 45 conventionally managed arable fields (each 15 fields of wheat, maize, and rapeseed): field "Interior" plots (50x2 m); field edge plots: "Edge_30" (30x2 m), "Edge_50" (50x2 m), and "Edge_500" (500x1 m); "Subplots" (four dispersed plots of 5x1 m); and "Corner" plots (50x2 m). To determine species detection rate, the species richness recorded with a survey method was related to the field's total plant species number as estimated from a survey of the entire field edge zone. With a species detection rate of 8.3% (median), interior plots were inadequate for characterizing the field's arable plant vegetation. Edge_500 plots yielded the highest proportion of the field's arable plant species pool (75.6%, including taxa of conservation value), followed by "Corner" plots (45.8%) and "Sublots" (32.6%). Edge_50 and Edge_30 plots detected less than 25% of the field's species pool. The average time needed for a releve was 20 min in Edge500 plots and 5-11 min in the other plot types. We suggest implementing Edge_500 plots as a standard monitoring approach in conventionally managed farmland due to its favorable ratio of detection success to expenditure of time. Our findings should be compared to methodological studies conducted in other regions, in different farmland management systems, and in landscapes of variable complexities.