No abstract

Filter strips or buffers are areas of grass or other perennial herbaceous vegetation established along waterways to remove contaminants and sediments from agricultural field runoff. In the heavily cultivated regions of the Midwestern United States, these buffer zones established under the Farm Bill provide important habitat for wildlife such as butterflies. The question of how the landscape context of these plantings influences their use has not been adequately researched. We used multiple regression and Akaike's Information Criteria to determine how habitat width and several landscape-level factors (i.e., landscape composition [total herbaceous cover, amount of developed area, and amount of wooded cover] and configuration [herbaceous edge density]) influenced the abundance and diversity of the butterfly community using filter strips in southwestern Minnesota, USA. Habitatsensitive butterfly abundance and all richness and diversity measures were positively correlated with filter-strip width. Butterfly abundance was negatively associated with the amount of developed areas (cities, towns, and roads) within the area of a 1-km radius (3.14 km2) surrounding the sites. Percentage of wooded cover in the landscape was an important variable explaining individual species abundance, although the direction of the relationship varied. Our finding that landscape context influences butterfly use of filter strips highlights the importance of landscape-level approaches to wildlife conservation in agroecosystems.

The value of agricultural conservation practices for wildlife, especially invertebrates, has not been thoroughly explored. One common practice, designed to reduce sedimentation and nutrient loads caused by runoff into streams, is the planting of filter strips. Filter strips are areas of herbaceous vegetation planted between agricultural fields and streams. In 2002 and 2003, we studied the butterfly community in filter strips of a variety of widths and vegetative compositions. We used transect surveys to quantify butterfly abundance and diversity and measured vegetative variables in conjunction with each butterfly survey round.We found overall butterfly diversity (H') and abundance of habitat-sensitive butterflies to be positively correlated with filter strip width. Using stepwise regression, we found that the best models to explain butterfly abundance and the species richness of disturbance-tolerant butterflies include the coverage of forbs and the number of ramets in bloom in the strips, and indicate positive relationships between forbs and the butterfly community (R 2 = 0.33, R 2 = 0.07, respectively). The model that best explained richness of habitat-sensitive butterflies includes vegetation height and density, and is expressed as a positive relationship (R 2 = 0.07). The planting of forbs in filter strips is rare, but may be useful for providing food sources to butterflies. More research on demographic parameters such as reproductive success and mortality is needed to assess whether filter strips represent additional habitat for butterflies, or possible sinks.