• Effects of Grazing Intensity, Precipitation, and Temperature on Forage Production

      Patton, Bob D.; Dong, Xuejun; Nyren, Paul E.; Nyren, Anne (Society for Range Management, 2007-11-01)
      Questions have been raised about whether herbaceous productivity declines linearly with grazing or whether low levels of grazing can increase productivity. This paper reports the response of forage production to cattle grazing on prairie dominated by Kentucky bluegrass (Poa pratensis L.) in south-central North Dakota through the growing season at 5 grazing intensities: no grazing, light grazing (1.3 +/- 0.7 animal unit months [AUM] ha-1), moderate grazing (2.7 +/- 1.0 AUM ha-1), heavy grazing (4.4 +/- 1.2 AUM ha-1), and extreme grazing (6.9 +/- 2.1 AUM ha-1; mean +/- SD). Annual herbage production data were collected on silty and overflow range sites from 1989 to 2005. Precipitation and sod temperature were used as covariates in the analysis. On silty range sites, the light treatment produced the most herbage (3 410 kg ha-1), and production was reduced as the grazing intensity increased. Average total production for the season was 545 kg ha-1 less on the ungrazed treatment and 909 kg ha-1 less on the extreme treatment than on the light treatment. On overflow range sites, there were no significant differences between the light (4 131 kg ha-1), moderate (4 360 kg ha-1), and heavy treatments (4 362 kg ha-1; P > 0.05). Total production on overflow range sites interacted with precipitation, and production on the grazed treatments was greater than on the ungrazed treatment when precipitation (from the end of the growing season in the previous year to the end of the grazing season in the current year) was greater than 267.0, 248.4, 262.4, or 531.5 mm on the light, moderate, heavy, and extreme treatments, respectively. However, production on the extreme treatment was less than on the ungrazed treatment if precipitation was less than 315.2 mm. We conclude that low to moderate levels of grazing can increase production over no grazing, but that the level of grazing that maximizes production depends upon the growing conditions of the current year. 
    • Spring Precipitation as a Predictor for Peak Standing Crop of Mixed-Grass Prairie

      Wiles, L. J.; Dunn, Gale; Printz, Jeff; Patton, Bob; Nyren, Anne (Society for Range Management, 2011-03-01)
      Ranchers and range managers need a decision support tool that provides a reasonably accurate prediction of forage growth potential early in the season to help users make destocking decisions. Erroneous stocking rate decisions can have dire economic and environmental consequences, particularly when forage production is low. Predictions must be based on information that is easily obtained and relevant to the particular range. Our goal was to evaluate monthly precipitation in spring months as a potential predictor of forage production compared to annual and growing-season precipitation. We analyzed the relationships between grazed and ungrazed peak standing crop (PSC) and precipitation using nonlinear regression and a plateau model, Akaike’s information criterion for model selection, and data from three locations: Streeter, North Dakota; Miles City, Montana; and Cheyenne, Wyoming. The plateau model included a linear segment, representing precipitation limiting production, and a plateau, an estimate of average production when precipitation is no longer the limiting factor. Both the response and predictor variables were rescaled so variability in production from average production was related to variability in precipitation from the long-term average. We found that grazing did not affect the relationship between PSC and precipitation, nor were annual or growing-season precipitation good predictor variables. The best predictor variable was total precipitation in April and May for Montana, May and June for North Dakota, and April, May, and June for Wyoming, with r2 ranging from 0.74 to 0.79 for precipitation less than long-term average. These results indicate that spring precipitation provides useful information for destocking decisions and can potentially be used to develop a decision support tool, and the results will guide our choice of possible predictor models for the tool.