Abstract Detail

Ecological Section

Hackett, Rachel [1], Karbowski, Hillary [2], Monfils, Anna [2].

Site- and landscape-level factors influencing prairie fen plant biodiversity: optimized sampling methodology and preliminary analysis.

Prairie fens are one of the most biodiverse ecosystems in the temperate region. Throughout their limited range in the glaciated Midwest, conservation and restoration efforts are underway to preserve biodiversity and mitigate impacts of anthropogenic disturbances. This project recorded baseline biodiversity measures in prairie fens, addressed optimal sampling techniques, and investigated site- and landscape-level factors influencing plant diversity and quality. Nineteen sites were selected in Michigan’s Jackson Interlobate and Ann Arbor Moraine Ecoregions. In 2012 ten sites were examined twice within one growing season. At each site a random transect-quadrat, area-proportional method was utilized to identify vascular plant species and assign Daubenmire cover classes. Total species richness, exotic species richness, native species richness, Shannon’s Diversity Index, exotic relative abundance, mean coefficient of conservatism, and Floristic Quality Index were calculated as diversity and quality measures. Site-level factors were water quality (pH, nitrates/nitrites, ammonium, and orthophosphate) and soil chemistry (calcium, magnesium, sodium, nitrogen, phosphorus, organic matter, and pH). Landscape-level factors included area, perimeter:area ratio, isolation, and proportion of land cover from the perimeter to 100m, 500m, 1000m, and the watershed boundary. Blocked multi-response permutation procedure indicated that the species composition of the same sites sampled in different seasons were grouped more tightly and independently than sites sampled in the same season, supporting a single season sampling method. Chao rarefaction-based comparison method estimated minimal sample coverage of species richness at 85% for the area-proportional method; this was 5% greater sample coverage than yielded for an equal-area method. Bootstrapped upper and lower limits for sample coverage ranged 81 – 94% for the area-proportional method and 76 - 94% for the equal-area method, supporting an area-proportional method. Nonmetric multidimensional scaling techniques reduced factors to be used in generalized linear models, and then corrected Akaike Information Criterion ranked the models to identify possible explanatory variables for each plant diversity and quality measure. Best-fit models for total and exotic species richness incorporated variables of perimeter:area and either area or isolation. Native species richness models incorporated area and proportion of land cover. Shannon’s Diversity Index models incorporated variables of soil, water, area, and land cover. Exotic relative abundance models incorporated variables of soil, water, perimeter:area, and land cover. Preliminary results indicate the importance of considering both site- and landscape-level factors in prairie fen management. Examination of nine additional sites in 2013 will provide further insight into site- and landscape-level factors and their relationship to plant diversity and quality.

Broader Impacts:

1 - Central Michigan University, Department of Biology, Brooks Hall 217, Mount Pleasant, MI, 48859, USA
2 - Central Michigan University, 180 Brooks Hall, Mount Pleasant, MI, 48859, USA, 989-774-2492

Keywords:
ecology
prairie fen
wetland
sampling
landscape
Diversity
quality.

Presentation Type: Poster:Posters for Sections
Session: P
Location: Grand Salon A - D/Riverside Hilton
Date: Monday, July 29th, 2013
Time: 5:45 PM
Number: PEC025
Abstract ID:697
Candidate for Awards:Ecological Section Best Graduate Student Poster