Northern Prairie Wildlife Research Center
The protection afforded stands after the creation of TRNP and the absence of fires in stands we sampled would have allowed ash plants alive in the early 1950's to continue growth and maturation. Green ash and choke cherry trees are relatively short-lived, and, in stands we sampled, many of the larger plants of both species were nearing the ends of their lifespans. As old plants die, sunlight at ground level increases which stimulates plant growth in the ground stratum until new green ash or choke cherry plants mature. This process presumably occurs on a patch by patch basis in uneven aged stands. In even aged stands, whether the evenness was created by grazing, fire, or tree harvest, these normal stand dynamics may result in waves of maturation and decay.
Snowberry, which is likely to increase under heavy grazing by livestock (Girard et al. 1987), did not increase significantly between the 1985 and 1996 in green ash stands while climax graminoids did. Invasive forbs and exotic grasses also increased but did not appear to be replacing native species in the interiors of green ash stands. Leafy spurge (Euphorbia esula) occurred in only two stands we measured and was not recorded in any stands in 1994 and 1996. Exotic plants, especially leafy spurge, increased in grass communities and riparian forest communities in TRNP during 1985 to 1996 despite an aggressive biological and chemical control program (Leafy Spurge Scientific Advisory Panel 1994, Trammel and Butler 1995, Irby and unpubl. data).
Deer and elk were the only ungulates that made any use of the interiors of green ash stands that we sampled. Signs of browsing by deer were observed in all of the stands, but the intensity of use of stand interiors was low. Forb and climax graminoid canopy did increase as deer density declined, but densities of palatable shrubs decreased or remained relatively stable. Neither the increases nor decreases were likely to be attributable to ungulate browsing. Over all transects and years, only 8% of shrub stems counted in 1-m2 plots showed evidence of more than 10% removal of current year's growth. Median use for most shrub species counted in 1-m2 plots was zero in all years (Irby, Unpubl. data.).
Elk have been identified as a factor in degradation of deciduous tree communities in the Rocky Mountains (Singer et al.1998, White et al. 1998) and probably had negative impacts on deciduous tree stands in Wind Cave National Park (R. Klukas, National Park Service, Omaha, pers. comm.). When we examined vegetation in green ash stands with zero to 12 years of elk use in TRNP, we did not find any consistent patterns of negative changes in vegetation categories expected to be susceptible to overuse by elk.
The lack of detectable elk effects on vegetation in TRNP can be explained by four factors. First, elk numbers have been regulated in TRNP. The herd is reduced to approximately 200 when it exceeds 400 individuals (based on aerial counts). Linear models developed by Westfall et al. (1993) predicted possible deleterious effects on vegetation when elk numbers exceed 400 for long time periods in a system with approximately 400 bison, 150 horses, and 800 deer. TRNP staff have used this numerical guideline in managing populations of all four species. Second, initial ungulate impacts on mature green ash stands are likely to occur in stand edges rather than stand interiors. All of our sample sites were placed in the middle of draws to avoid edge effects in vegetation measurement. Third, change takes time under any stocking rates (Laurenroth et al 1994, Hobbs 1996, Biondini et al. 1998). We measured changes in stands over only approximately 10% of the lifespan of the trees dominating the canopy in mature stands during one regeneration cycle. Ultimate effects of ungulate herbivory could take longer to be expressed or could vary significantly with the successional status of a stand. Fourth, ungulates are only one of several environmental factors that affect plant species. An increase in ungulate use of plants could have been effectively masked at any point over the sample period by an increase in ground moisture or seasonal variation in temperature (Olson et al. 1985, Hart et al. 1988, Biondini and Manske 1996, Biondini et al. 1998).
Results of logistic regression and correlation analysis suggested that precipitation during the growing season one year before vegetation measurements were taken was an important determinant of change in vegetation. Two other factors, the month in which vegetation measurements were taken and the sum of growing season precipitation one year before vegetation measurement and in the year of measurement, were identified as important in correlation analysis. Vegetation measurements taken after mid-August were likely to show lower canopy coverage in ground stratum vegetation than vegetation measurements taken in mid-July in most years as a result of late summer desiccation. Shrub counts also tended to decline over the growing season as young shrubs died due to browsing, crowding, or desiccation. The importance of precipitation a year before vegetation measurement may be an artifact of the pattern of wet and dry years we sampled or it may reflect the two-year interval in vegetation measurement. In all of our analyses of short term effects, plants had two growing seasons in which to change. Above average rainfall during the growing season of years in which we did not sample stands could stimulate increased growth in the non-measurement year, and movement of water retained in upland soils to adjacent drainages could contribute to plant growth in the drainages at the beginning of the growing season in which the next round of vegetation measurements were made. A wet season in the year of vegetation measurement, would reinforce positive changes that occurred in the previous non-measurement year. Drought would have an inverse effect and reinforce negative changes.
Our results indicate that researchers planning tests of grazing systems designed to decrease negative effects of livestock on green ash communities need not exclude native ungulates (as long as the ungulates occur in low to moderate densities) but should concentrate on sample timing and should add covariates to account for precipitation differences between sampling periods. Researchers should attempt to sample in a relatively narrow time period and to adjust this period for phenological variability associated with variability in precipitation during the growing season. We attempted to concentrate our sampling in a four-week period in July and August with samples collected early in the sampling window in dry years and later in the window in wet years. This was not precise enough to avoid differences in growth stage in perennial species and presence of annual species that may affect interpretation of experimental results. The standard deviations we calculated (16% to greater than 200% of associated means) demonstrate the high variability among green ash stands for vegetation categories that researchers would be likely to measure in tests of grazing systems. The lowest variability occurred in the tree stratum and the highest in the ground stratum.