TWEL Nancy Sundell-Turner Thesis

A comparison of landscape based methods for conservation planning
 
Nancy M. Sundell-Turner, MS
Advisor: Amanda Rodewald
 
 
A variety of well-conceived conservation planning approaches have been developed over recent years, all with the critical component of setting conservation priorities for both species protection and land acquisition and management. Ideally, the selection of the sites comprising a conservation portfolio would be based on detailed survey data, but this is not always possible due to time and monetary constraints. When it is not possible or feasible to conduct surveys of even some portion of the species in an area, basic ecological principles can be applied. The overall goal of this thesis is to examine the extent to which commonly used and easily obtained landscape metrics contribute to a conservation planning process that must be completed rapidly and with little financial resources. Specifically, I aimed to determine if landscape metrics can be used to guide the site prioritization process when the relative importance of potential conservation sites to various conservation portfolios is measured using their relative irreplaceability values. This question is examined in the context of a flexible planning process where the scope of a given project (represented by conservation portfolio size) may vary over time. I focused this study on the conservation of riparian forest ecosystems in urbanizing Midwestern landscapes and used bird communities to indicate the ecological value of particular forest tracts.
 
Evaluating the usefulness of landscape metrics in the prioritization process required two steps. First, I compared two easily-applied alternate approaches to using avian species data to indicate the value of a site in a conservation planning context (as measured by relative site irreplaceability). Specifically, I compared the use of avian species richness to the use of a weighting system based on a conservation threat score developed by Partners In Flight. In addition, I compared irreplaceability scores created via two different compilation methods (1) averaged species occurrence data among multiple years versus (2) a cumulative species richness that considers all species recorded on sites over multiple years. These methods were compared using the sign test. Second, I directly examined the relative utility of landscape metrics in predicting the importance (as measured by irreplaceability) of a given site over a range of conservation portfolio sizes. Multinomial logistic regression models were created for 21 different models based on ecological principles and compared using Akaike Information Criterion.
 
My findings suggest that while the method of survey data compilation had little effect on irreplaceability values among sites, use of a weighting scheme that places greater emphasis on vulnerable species can significantly influence site prioritization. My results also confirm that landscape metrics are useful indicators of the value of particular sites within conservation portfolios. In particular, irreplaceability value of a site to a conservation portfolio was most consistently and simply predicted using forest coverage within a 1-km landscape surrounding potential portfolio sites. Use of an additional landscape metric describing human disturbance (e.g. number of buildings, percentage of area covered by roads, pavement, mowed surfaces, or agricultural land) improved model fit substantially (i.e. decreased the log-likelihood score) and is recommended. This combination of forest and disturbance metrics was useful across a range of portfolio sizes (protecting 6-23% of possible sites) and is therefore likely to remain useful even in the context of a planning process whose scope varies over time. These results show that simple landscape metrics can aid land managers and planners that need to make rapid decisions about prioritizing land acquisition, preservation or management activities.