Experimental evaluation of EDRR strategies

Using publically available aerial imagery data from the US Department of Agriculture's Farm Service Agency National Agricultural Inventory Program (NAIP), we identified 5 circular Typha stands at the north end of the wetland, and confirmed their size and invasive Typha dominance with ground-truthing. The stands were isolated from each other but close in proximity (between 20 and 50 m) and were growing in a matrix of native sedge (Carex lasiocarpa and C. aquatilis) and hardstem bulrush (Schoenoplectus acutus) dominated wetland. Stands ranged in size from 2,400–7,800 square meters in 2015 (Figure ​(Figure22).

Plot layout at Sand Island, MI illustrating five isolated Typha stands, completely randomized block design, vegetation sampling transects, vegetation subplots, and the extent of Typha at three points in time (1998, 2008, and 2015).

We used a completely randomized block design that treated individual stands as a block, divided each block into 5 equally sized wedges, and randomly assigned each wedge to one of three treatments: above-water biomass harvest, below-water biomass harvest, or control. Two additional above-water wedges (beyond one per block) were initially assigned in order to allow for repeated treatments to be conducted in year-2. To avoid unbalanced data, one wedge of each treatment per block was randomly selected for all post-treatment comparisons (Figure ​(Figure2).2). We identified the geographic center of each stand by running two 100 m fiberglass measuring tapes along north-south and east-west bearings. To evaluate vegetation responses along a gradient from Typha dominated to native plant dominated wetland, treatment wedges were extended from stand center to 10 m beyond the Typha dominated margin of the stand. We established above-water harvest treatments on August 20–30, 2015 using a low-ground pressure wetland plant harvester (Loglogic Softrak with Cut and Collect system, Devon England, U.K.) and conducted below water treatments using aquatic weed-whackers (Weeders Digest LLC, New Hope, MN, U.S.A.) followed by hand biomass removal.

On August 19–27, 2015, prior to treatment implementation, we established a series of five 1 × 1 m subplots, equally spaced along transects bisecting the middle of each wedge and extending from block center to 10 m beyond the edge of the core of the Typha stand. Thus, each subplot fell along a gradient from the interior to exterior of the Typha stand. We collected vegetation data by assigning aerial cover values (<1–100%) for total vegetative cover, detritus (below water litter), standing dead (above water) litter, and for each plant species. Additionally, we calculated aboveground Typha biomass by measuring the height of each individual stem and using a height-to-dry biomass allometric equation (g = 0.5265e1.751^*height (m), r² = 0.81; Lishawa et al., 2015). One year following treatment implementation, on August 18–20, 2016, we resampled vegetation subplots using the same methods.

We used two high resolution (1 ft pixel), high quality historical aerial photographs from 1998 and 2008 (MDNR, 2001; USGS, 2008) along with field collected data from 2015 to post-dict the invasion history of Typha within the study site following the methods of Lishawa et al. (2013). We determined the extent of Typha in 2015 through field observation; created a shapefile demarcating the stand boundaries; and evaluated the two selected images counter-chronologically and created Typha extent polygons from each. Individual vegetation plots were assigned to one of four stand-age classes representing the first documented date when Typha dominated (subdominant, 2015, 2008, 1998); the subdominant category had scattered Typha stems (<5/m²).