Woody Encroachment in North Dakota Grasslands: Using Predictive Modeling to Examine Key Players and Management Implications

Date of Award


Document Type


Degree Name

Master of Science (MS)


Earth System Science & Policy


Woody shrub encroachment is one of the major threats to grasslands, closely connected to other issues such as fire regimes, herbivore grazing patterns, and climate change, and is currently a global trend. Consequences of shrub encroachment can significantly impact many of the ecological characteristics of grasslands, including grassland biogeochemistry and resource availability and biodiversity. Studies conducted on grassland ecosystems have pointed to grassland management, in the form of prescribed burning, grazing, or some combination of the two, as a way to prevent woody encroachment. Prevention of woody encroachment may be particularly important to preserving grasslands, as management may not be effective in reducing shrub coverage once shrubs have become established.

Knowledge of the names and potential g> ographic distributions of woody shrub species that are possible encroachment threats ;s lacking; such knowledge is necessary to identify likely problem species and indicate vdiere they could occur. Predictive niche modeling provides a way to assess whether certain shrub species could potentially pose encroachment problems in selected locations. This study aimed to create a species list of the woody species that inhabit the Northern Great Plains and predict the spatial distributions of these species using ecological niche modeling in order to assess the occurrence risk of species in North Dakota. Additionally, managers of grassland sites throughout North Dakota were interviewed to leain about management practices and the status of encroachment at each.

The final species list, compiled by consulting multiple online databases, personal communications, and scientific literature, consisted of 385 woody species including both native and naturalized introduced species. Locality data were colleted for each species from the Global Biodiversity Information Facility. Maxent 3.0.4 Bela was used to create predictive niche models for woody species using a suite of fifteen continental-scale environmental data layers. There were 293 species predicted to have a mean probability of presence of 1.0% or higher, the threshold used as an indication for significant likelihood of presence; management efforts should account for these species. Grassland managers of 19 sites provided names of 26 species recognized to be involved in encroachment; all of these species except for one had average probabilities of presence above the 1.0% threshold for North Dakota. Important biases may exist due to factors such as quality of input data, environmental layers used, and spatial scale.

Site conditions and management practices varied widely among the sites, but 7 sites were found to actively manage for woody encroachment. There were 14 sites using fire and 13 sites using grazing as management in some capacity. Only 5 sites using fire also were experiencing encroachment by woody species, and only 1 site with burning frequencies of 2-5 years or less were said to be experiencing encroachment. Overall, sites with prescribed burns at least every 2-5 years seemed to have a lower rate of encroachment, and grazing was found to reduce encroachment rates in the absence of fire on two sites. The interview portion is limited by sample size, variability in site characteristics, and possible participant bias.

This document is currently not available here.