Surrounding land-use and management, but not size, predict plant species richness of remnant prairies in Illinois

Background/Question/Methods

Prairies once covered more than 80 percent of the Chicagoland area, but less than 0.2 percent of original, high-quality habitat remains. Remnant prairies are found in a matrix of developed, agricultural, post-agricultural and “natural” land, which may have consequences for species richness. Increased species richness has been linked to increased productivity, invasion resistance, and other ecosystem services. We characterized land use around 36 high-quality prairie remnants in Northeast Illinois to assess effects of surrounding land use on prairie plant diversity, and to determine the scales at which surrounding habitat become important to diversity. We predicted (1) that crop and developed land adjacent to a remnant would reduce species richness and that nearby “natural areas” – (eg: shrubland or forest) would increase richness, and (2) that larger prairies would have higher richness. We characterized land use in one and five km buffers around each site using a 2001 cropscape layer in GIS. Plant species richness data was collected for these sites in 1976 and again in 2001 using 20-30 quarter-meter squared plots along transects. We collected additional data about each site, including community type, successional stage, moisture category, and frequency of prescribed fire.

 

Results/Conclusions

Developed land was the only surrounding land use to significantly impact prairie species richness, and ranged from 9 to 39 percent of area surrounding remnants. Increasing developed area surrounding a prairie remnant had a negative effect on species richness (t = 2.6, R² = 0.13, F_1,33 = 4.85, p = 0.035), while adjacent cropland, shrubland or forest did not significantly impact richness (p>0.05, all analyses using one km buffer). Percent of cropland surrounding the remnant was the best predictor of change in species richness between 1976 and 2001 in regression tree analysis. Remnant size ranged from 0.08 to 32 hectares, but was not a significant predictor of species richness, nor was community type, successional stage or moisture category (p>0.05). Increasing frequency of prescribed fire had a positive effect on richness (R² = 0.20, F_1,27= 6.66,  p = 0.016). In future analyses we will assess the effects of surrounding land use and management on the richness of different plant functional types and the effects of adjacent land-use on richness in restored prairie sites. Our results suggest that land-use and management are more important in predicting plant species richness than remnant size. This is a hopeful finding for a very fragmented habitat, and may have implications for restoration decision-making.