Vegetation management in a greasewood-dominated floodplain of the Green River, Colorado

Vegetation structure, the three-dimensional distribution of vegetation in space, is an important indicator of ecological function, especially in shrublands. Moreover, relationships of vegetative manipulations and restoration of understory herbaceous productivity to vegetation structure are not well documented in greasewood (Sarcobatus vermiculatus) systems. Because management decisions at Brown’s Park National Wildlife Refuge, Colorado target a variety of uses, determining these relationships will provide important guidance. We initiated a long-term monitoring project in the summer of 2007 to compare commonly used vegetation structure measurements (e.g. Robel pole) with newly developed methods to describe vegetation in three dimensional space. Thirty-nine plots in six vegetation alliances were randomly placed throughout the floodplain of the Green River on the refuge and baseline monitoring of vegetation characteristics were collected. Of those, twelve 0.78 ha greasewood plots were randomly selected to examine changes associated with the three treatments. We conducted herbicide, mowing and reseeding manipulations and monitored vegetation structure, cover, biomass production, and seedling recruitment on three treatments: mowing coupled with (1) herbicide thinning of greasewood, (2) seeding of native grasses and forbs, or (3) combined seeding and herbicide. All plots were mowed to 15.2 cm in height in March 2008, prior to treatment applications. One-third of each plot was assigned to one of three treatments. Seedings included two grass species alkali sacaton (Sporobolus airoides ), and basin wildrye (Leymus cinereus), and one herbaceous dicot, common gaillardia (Gaillardia aristata). We monitored shrub and basal gap, canopy height, structure, canopy cover, and collected herbaceous biomass along 50 m line transects placed within each third of the circular plots and located at 120 degree azimuth intervals from each other. Shrub density and distribution was documented within a 2 x 50 m belt adjacent to each transect. Vegetation structure decreased following mowing (p < 0.001). Visual obstruction (Robel pole) corresponded to height measurements (p < 0.0001) and was highly variable. Presence of cheatgrass (Bromus tectorum) and litter increased on all plots irrespective of treatment during a wet spring in 2009. Restoration of understory herbaceous cover in dense greasewood stands was most successful in herbicide application only treatments. Densest stands of greasewood had the least understory cover at the beginning of the study but provided greater litter cover following herbicide application. Seeding in absence of cheatgrass control was not successful even though cheatgrass was not recorded in the cover in some plots. Shrub cover did not rebound immediately following treatments even though greasewood is a resprouting species. Although herbicide treatments reduced canopy cover of greasewood, the native perennial understory did not increase by the following spring.