Graduation date: 1994

A greater than 10-fold increase in Canada goose (Branta canadensis) populations over the past several years has resulted in concerns over grazing impacts on grass seed production in the mid-Willamette Valley, Oregon. This study was designed to develop methods to quantify and statistically analyze goose-grazing impacts on seed yields of tall fescue (Festuca arundinacea Schreb.) and perennial ryegrass (Lolium perenne L.). Yield-mapping-system equipped combines, incorporating global positioning system (GPS) technology, were used to measure and map yields. Image processing of ground-level photography to estimate crop cover and other relevant observations were spatially located via GPS to establish spatial-temporal goose grazing patterns. We sampled each field semi-monthly from mid-winter through spring. Spatially located yield data, soils information, exclosure locations, and grazing patterns were integrated via geographical information system (GIS) technology. To avoid concerns about autocorrelation, a bootstrapping procedure for subsampling spatially contiguous seed yield data was used to organize the data for appropriate use of analysis of variance. The procedure was used to evaluate grazing impacts on seed yield for areas of fields with different soils and with differential timing and intensity of goose grazing activity. We also used a standard paired-plot procedure, involving exclosures and associated plots available for grazing. The combination of spatially explicit photography and yield mapping, integrated with GIS, proved effective in establishing cause-and-effect relationships between goose grazing and seed yield differences. Exclosures were essential for providing nongrazed controls. Both statistical approaches were effective in documenting goose-grazing impacts. Paired-plots were restricted by small size and few numbers and did not capture grazing impacts as effectively as comparison of larger areas to exclosures. Bootstrapping to subsample larger areas of yield for comparison was an effective method of avoiding autocorrelation of data while better representing impacts within a field. Occasional yield increases, ranging from 1 to 5 percent, were recorded following goose grazing. Goose grazing generally resulted in seed yield reductions, ranging up to 20 percent. Later and more intensive grazing tended to increase yield reductions. Newly seeded tall fescue tended to be the most sensitive to grazing. Established perennial ryegrass tended to be more resilient. / Graduation date: 2003