Selecting High-dispersal Species for Precision Prairie Reconstruction on the Northwestern Great Plains

Selecting High-dispersal Species for Precision Prairie Reconstruction on the Northwestern Great Plains

Peter Lesica (corresponding author: Conservation Biology Research, 929 Locust, Missoula, MT 59802, lesica.peter{at}gmail.com) and Stephen V. Cooper (Conservation Biology Research, Missoula, MT)

Millions of hectares of the Northern Great Plains were planted with the introduced crested wheatgrass (Agropyron cristatum) on abandoned croplands to prevent soil erosion (Lesica and Deluca 1996). These fields are generally poor in native grasses and forbs and provide little habitat for many species of native vertebrates (Reynolds and Trost 1980, Urness 1986, Lloyd and Martin 2005). Tilling or herbicide application followed by broadcast or drill seeding has been used to restore these fields. However, the large amount of prairie dominated by crested wheatgrass makes restoration an expensive proposition. One way to reduce the cost of restoring native diversity to fields dominated by a few exotic grasses is to restore islands of native species within the fields. These islands allow natural dispersal of native species to the remainder of the field. This method, called Precision Prairie Reconstruction (PPR), enhances native species richness and density in old fields dominated by a few non-native grass species (Grygiel et al. 2009, 2018). The PPR technique, in which ca. 25% of a field is restored in island-like patches, costs approximately one-third as much as traditional herbicide application/drill-seeding or rototill/broadcast seeding methods and obtains equal or better increases in native diversity.

Success of the PPR method in enhancing native plant diversity across the entire field relies on dispersal and establishment of the species planted in the restored islands. However, dispersal ability varies among species (Simberloff 2009), so planting species with good dispersal ability will help improve PPR restoration success. We conducted a study that employed a novel sampling scheme to examine factors influencing the invasion of native species into crested wheatgrass fields in northeast Montana with an experimental design inadvertently begun ca. 80 years ago when crested wheatgrass fields were planted in native steppe. We ranked species for their ability to disperse throughout crested wheatgrass fields. Our results will help managers choose which native species to plant in PPR islands in order to obtain a more diverse native composition across the entire field.

The semi-arid grasslands of eastern Montana were homesteaded in the early 1900s, and millions of hectares of native prairie were plowed and planted with wheat. During the great drought of 1928–1937 there was a mass exodus from the region, and most of the farm fields were planted with crested wheatgrass to help stabilize soil and prevent erosion. Crested wheatgrass has been planted on 6 to 10 million ha in the northern Great Plains (Lesica and DeLuca 1996). Our study sites were all located on lands planted with crested wheatgrass ca. 80 years ago. Our 24 study sites were located on the Great Plains in central Montana, east of the Rocky Mountains, between the towns of Fort Benton on the northwest and Roundup on the southeast (see Lesica and Cooper 2019 for additional information).

Our study was conducted in June of 2010 and 2011, during the height of the growing season when species were above ground and identifiable. We located two circular, 0.8 ha macroplots (50 m radius) at each site. The Edge macroplot was in the crested wheatgrass field within 50 m of native vegetation. The Center macroplot was either in the center of the crested wheatgrass field or as far from native vegetation as possible. With two exceptions, centers of Edge and Center macroplots were greater than 90 m apart (mean = 184 m). Density is the number of plants of each species counted in 12 stratified-random microplots per macroplot. Individual stems or clumps of stems were counted for rhizomatous species (see Lesica and Cooper 2019 for further details).

We can estimate a plant’s relative dispersal ability by comparing its density in the Center macroplot to that in the Edge macroplot. We expected propagule pressure of native species to be greater for Edge macroplots than for Center plots because Edge macroplots are adjacent to, and Center macroplots are distant from, native vegetation. The difference in species abundance between Center and Edge macroplots is a measure of dispersal ability because biotic/abiotic interactions were approximately constant within and across the two macroplots. Only species that occurred in density microplots of at least five macroplots were used in this analysis. We quantified dispersal ability by calculating a “Dispersal Ratio,” the mean density in the Center macroplot divided by the mean density in the Edge macroplot (Mean DensityCenter/Mean DensityEdge). We characterized dispersal ratio values less than ca. 0.5 as dispersal limitation. The change in density between Edge and Center macroplots was assessed with a nonparametric Mann-Whitney test for each species. Plant nomenclature follows Lesica (2012).

Dispersal ability varied greatly among species. Seventeen of the 24 native species recorded in microplots of at least five macroplots declined in density from Edge to Center macroplots (Dispersal Ratio < 1.0, Table 1). All shrubs were less common in Center macroplots compared to Edge macroplots. Artemisia frigida and Opuntia polyacantha were significantly less dense in Center macroplots compared to Edge macroplots, and the density of big sagebrush (Artemisia tridentata) was marginally less in Center macroplots. The common grasses except Stipa comata had Dispersal Ratios near or greater than 1.0, showing no indication of dispersal limitation. Density of Achillea millefolium in Edge microplots was greater than in Center macroplots (Table 1). Astragalus missouriensis and Symphyothrichum falcatum also had low Dispersal Ratios, but the differences in density were not significant (p > 0.19).

Dispersal (propagule movement) is only one part of successful colonization. Propagules must also develop into established seedlings which involves several factors (Harper 1977). However, the canopy cover of crested wheatgrass did not differ between Edge and Center macroplots across all sites, and there was little difference in life-form composition (Lesica and Cooper 2019). Consequently, our data support the hypothesis that native-species abundance differed between Edge and Center macroplots due to differential dispersal.

Many plant species have propagules with special structures to aid dispersal, and it is generally believed that, all else being equal, these plants should have greater dispersal compared to species lacking such structures (van der Pijl 1982). Support for this hypothesis from our study was equivocal. Most members of the Family Asteraceae have a pappus that is thought to be an adaptation for dispersal (Cousens et al. 2008). Seven species in our study were Asteraceae; four of them have pappus-bearing seeds, and three do not. The three species that lack pappus (Artemisia frigida, A. tridentata and Achillea millefolium) were among the most poorly dispersed species with the lowest Dispersal Ratios (Table 1). However, two of the four pappus-bearing Asteraceae species (Gutierrezia sarothrae, Symphyotrichum falcatum) also had low dispersal ratios.

Diacon-Bolli et al. (2013) found that grasses dispersed better than forbs overall in Swiss calcareous grasslands and concluded that the grass superiority was due to greater release height. This may have been a factor in our study system as well; both Schedonnardus paniculatus and Bouteloua gracilis have short stature compared to the majority of the graminoids that dispersed well. However, two of the better dispersers, Poa secunda and Carex eleocharis, also have short stature. Many grasses have awns or bristles that become embedded in animal fur and promote dispersal (Rabinowitz and Rapp 1981, Fischer et al. 1996). However, in our study all three of the poorly dispersed grasses (Schedonnardus paniculatus, Stipa comata, Bouteloua gracilis) have awns, while the only grass species without awns (Poa secunda) was one of the best dispersers.

Our study allows managers attempting to restore native composition to crested wheatgrass fields in northeast Montana and adjacent North Dakota and Saskatchewan using the PPR method to prioritize species with a high Dispersal Ratio. These species are more likely to move from the PPR islands into the unplanted 75% of the crested wheatgrass fields. Dispersal limitation can also be an issue with whole-field restoration seedings. Often some species will establish in one portion of a field but not in another (P. Lesica observations). Species with greater ability to disperse and establish are better for restoration seed mixes because they will promote faster and more complete establishment of native vegetation before crested wheatgrass is able to attain significant recruitment. Our methods can be used to help select restoration species for PPR where crested wheatgrass or other non-native species, such as Bromus inermis (smooth brome), have been widely planted.