The Application of Prescribed Fire and Herbicide to Reduce Pennsylvania Sedge (Carex pensylvanica) Cover at the Newaygo Prairies Research Natural Area, Manistee National Forest, Michigan

Sand prairie is part of Michigan’s historic oak-pine barrens ecosystem and once covered approximately 5,000 hectares in northern Lower Michigan in the early to mid-1800s. Dominant plant species include Pennsylvania sedge (Carex pensylvanica), little bluestem (Schizachyrium scoparium), and big bluestem (Andropogon gerardii). Unfortunately, sand prairie has been largely eliminated due to agriculture, reforestation, and fire suppression. Today, approximately 4% of the original extent of sand prairie remains in the state (Kost 2004, Cohen et al. 2015). Similarly, species that depend on these ecosystems such as the federally endangered Karner blue butterfly (Lycaeides melissa samuelis) have also suffered considerable losses (USFWS 1992). Fire suppression has allowed Pennsylvania sedge to dominate the understory of degraded oak-pine barrens and sand prairies and represents a major obstacle for restoration of these areas (Abrams et al. 1985, Nielsen et al. 2003).

Restoration of sand prairie has the potential to help return floral and faunal diversity and ecosystem function to degraded landscapes in the Manistee National Forest, Michigan. The return of prescribed fire may help decrease the cover of Pennsylvania sedge in areas designated for sand prairie restoration. However, little research has been done on the response of Pennsylvania sedge to fire. Previous workers have found that returning fire to fire-suppressed areas may actually result in a modest increase in Pennsylvania sedge cover (Abrams et al. 1985, Nielsen et al. 2003). Although limited, these results indicate that fire alone is at least insufficient at reducing the dominance of this species.

Other common management approaches such as the application of herbicide may be important in decreasing Pennsylvania sedge. Only limited research has been conducted on the response of Carex spp. to herbicide (e.g., Campbell and Burbidge 1993, Ristau 2010), while the response of Pennsylvania sedge to herbicide in particular has not been evaluated. Further, the use of prescribed fire in conjunction with herbicide application has not been evaluated.

Our primary objective was to determine how to effectively reduce the dominance of Pennsylvania sedge on degraded grasslands in the Manistee National Forest, MI through the use of fire and herbicide applied singularly and in combination. We also evaluated the response of the native and non-native plant community to Pennsylvania sedge reduction.

The study area is located within the Newaygo Prairies Research Natural Area and Newaygo Experimental Forest in the Manistee National Forest, Newaygo Co., MI. The experiment is conducted on degraded grassland in three separate management units within the study area. Five treatments of prescribed fire and herbicide application were implemented in each management unit to reduce Pennsylvania sedge cover and biomass: fire followed by herbicide, fire only, herbicide only, herbicide followed by fire, and no fire or herbicide (control). Prescribed fire was applied to the fire and fire herbicide plots on June 4, 2013 by US Forest Service personnel. Glyphosate was applied to herbicide and herbicide fire plots on June 4, 2013 and applied to the fire herbicide plots on June 27, 2013. Herbicide was applied with a backpack sprayer at rate of 1.8 kg a.i./ha (based on control of Carex spp. by Campbell and Burbidge 1993, Ristau 2010). Prescribed fire was applied to the herbicide fire treatment plots on September 4, 2013.

Changes in the post-treatment plant community were determined by evaluating 25 1-m² subplots within each treatment in each management unit for a total of 375 subplots. Each subplot was randomly located within each treatment plot. Species richness and cover were determined for all individual species from each subplot in July 2014 using the point-intercept method. At each 14.3 × 10.0 cm interval within a subplot, we recorded all species that intercepted a 7-mm diameter pin for a total of 54 data points per subplot. Multiple species typically intercepted the pin at each point, so the total number of hits per subplot was normally greater than 54. All species that were found in a subplot, yet did not intercept the pin...