GIS Used to Study Invasive Weed Abundance in Wisconsin Watershed


Nina Borchowiec and Amanda Little.

Presented at the 36th Natural Areas Conference, “Living on the Edge: Why Natural Areas Matter”, Vancouver, Washington, USA, September 15-18, 2009.

P. arundinacea is a weed that grows invasively across North America. It suppresses native vegetation, ultimately reducing ecological diversity. Knowing how P. arundinacea responds to landscape attributes will help determine how to monitor and manage it. We related P. arundinacea abundance from a statewide data layer created by the Wisconsin Department of Natural Resources. ArcGIS 9.2 was used to calculate the proportions of different soil surface textures, drainage classes, and land-cover types in each watershed to determine NRCS 12-digit watershed characteristics that influenced the abundance of P. arundinacea in wetlands of the Lower Chippewa River Watershed, Wisconsin, USA.

To reduce the number of covarying attributes, we used non-metric multidimensional scaling to create composite variables. We used multiple linear regression to relate these variables to wetland P. arundinacea abundance, as a percentage of wetland land cover dominated by P. arundinacea.

One surface texture and one drainage class variable predicted P. arundinacea abundance (log(y) = 1.23 + 0.467drainvar1 – 0.166surftexvar2, R2 = 29.5%, P < 0.001). Synthetic land cover variables were not significant predictors. Relationships between individual predictors and synthetic variables indicate that P. arundinacea is more abundant in wetland watersheds with more wetland-type muck soils and less abundant with substantial open water. These findings indicate that agriculture may not be a strong driver of P. arundinacea abundance at the watershed level. P. arundinacea is not found in watersheds with somewhat excessively drained fine sandy loam, although it‘s uncertain whether this is a function of the soil properties or associated topographic constraints.