|Luschei, Edward - U. OF WISCONSIN|
Submitted to: Biological Invasions
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: September 21, 2006
Publication Date: July 1, 2007
Repository URL: http://ars.usda.gov/SP2UserFiles/Place/54340000/Publications/BiolEv-Matt.pdf
Citation: Rinella, M.J., Luschei, E.C. 2007. Hierarchial bayesian methods estimate invasive weed impacts at pertinent spatial scales. Biological Invasions 9:545-558. Interpretive Summary: Farmers, ranchers and other rangeland managers need to know how severely weeds are impacting lands so they can make wise weed management decisions. If weeds are having a major negative impact, expensive management (e.g. herbicides, reseeding) is justified, but if impacts are not very severe, less intensive management (e.g. reduced stocking rates, weed containment) may be more sensible. We developed a model that estimates invasive weed (i.e. leafy spurge) impacts on forage production (and related variables) at individual sites and across a 17-state region. Our results indicate that, to accurately estimate leafy spurge impacts on particular sites, managers must measure weed and forage production at the sites. For the 17-state region, we found that leafy spurge reduces cattle carrying capacity by 50 to 217 thousand animals, and reduces grazing land value by 8 to 34 million dollars a year.
Technical Abstract: Without information on the severity of invasive weed impacts, natural resource managers cannot compare the costs and benefits of weed management strategies. The spatial scale of interest to weed managers ranges from very local (e.g. ranchers, park managers) to regional (e.g. policy makers). Our goal was to estimate site-specific and regional impacts of leafy spurge (Euphorbia esula L.) on associated species biomass production (and related variables). Our basic approach was to use an empirical model that characterizes weed densities across the landscape in combination with another empirical model that predicts weed impact from weed density. In developing these models, we gave substantial formal attention to parameter uncertainty and sampling error. Our investigation revealed that, without on-site plant density data, estimates of site-specific leafy spurge impacts are highly imprecise. Supplementing our general predictive model with small quantities of on-site data increased precision considerably. For the 17-state region we considered, 95% Bayesian credibility intervals indicated leafy spurge reduces cattle carrying capacity by 50 to 217 thousand animals, and reduces grazing land value by 8 to 34 million dollars a year. The precision of these estimates would improve substantially if plant density data were collected from randomly selected sites that are occupied by leafy spurge.