Submitted to: Society for Range Management Meeting Abstracts
Publication Type: Abstract Only
Publication Acceptance Date: 1/5/2004
Publication Date: 1/16/2004
Citation: Rinella, M.J., Sheley, R.L., Goodman, D. 2004. Toward a decision support system for leafy spurge-infested plant communities. Society for Range Management Meeting Abstracts #8. On CD.
Technical Abstract: Technology estimating the probability that various species or functional group compositions will result from different management options, would enable invasive plant managers to make well-informed decisions. In ongoing efforts to develop that technology we used competition experiments to develop/select deterministic time series models that predict grass and leafy spurge response to management. The predictive capability of these models were evaluated using data from herbicide, revegetation and selective plant removal experiments. Predicted lines were centered on observed values for most accuracy assessments indicating that there are no consistent differences between the system used to develop the models and the system used to assess them. Bayesian parameter estimation techniques were used to develop stochastic versions of the deterministic models. Prior probability distributions on competition parameters were developed from herbicide, revegetation and selective plant removal experiments. These prior distributions and non-informative prior probability distributions describing intrinsic rates of population increase, carrying capacities and standard deviations of error terms were updated with competition experiment data using likelihood functions. Because the response variables from the stochastic models are probability distributions, probabilities that particular shifts in leafy spurge and grass production will result from management actions are estimated.