Nest survival is one of the most important parameters in the population dynamics of grassland‐nesting ducks (Anas and Aythya spp.) that breed in the Prairie Pothole Region of North America. Grassland habitats used by these species are increasingly threatened by habitat loss and the coincident fragmentation, which may indirectly alter nest survival through effects on predators. Although predators are the dominant cause of nest loss, they are difficult to monitor directly. Thus, indirect analyses of habitat variables are required. Many studies have attempted to address the relationship between fragmentation and nest survival; however, few studies have examined the influence of fragmentation at multiple spatial scales. Understanding how landscape characteristics at multiple spatial scales explain variation in nest survival is important, because no single correct scale is likely to exist for a diversity of landscape metrics. We examined the relationships between habitat variables and duck nest survival (n ≈ 4200 nests) on 18 10.4‐km2 sites selected across a gradient of landscape characteristics in the Missouri Coteau Region of North Dakota. We evaluated both a priori and exploratory competing models of nest survival that considered habitat features measured at nest sites, within nesting patches, and at multiple landscape scales. We used generalized nonlinear mixed‐modeling techniques to model nest survival rates. Information‐theoretic techniques were used to select among competing models. Models that included covariates measured at multiple landscape scales were better than simpler models that included only covariates measured at a single spatial scale. Landscape covariates measured at 10.4 and 41.4 km2 resulted in the best explanation of nest survival. Nest survival was positively related to the amount of grassland habitat, negatively related to the wetland density, and related to the amount of grassland edge in a quadratic manner, with the lowest nest survival at intermediate values of grassland edge. Future research should attempt to determine the causes of these relationships, something we were unable to do with our correlative approach. Conservation efforts focused on maintaining duck populations should seek to maintain landscapes with abundant grassland and to account for the influence of configuration using GIS analyses.

No abstract

Leadership is a critical tool for expanding the influence of conservation science, but recent advances in leadership concepts and practice remain underutilized by conservation scientists. Furthermore, an explicit conceptual foundation and definition of leadership in conservation science are not available in the literature. Here we drew on our diverse leadership experiences, our reading of leadership literature, and discussions with selected conservation science leaders to define conservation-science leadership, summarize an exploratory set of leadership principles that are applicable to conservation science, and recommend actions to expand leadership capacity among conservation scientists and practitioners. We define 2 types of conservation-science leadership: shaping conservation science through path-breaking research, and advancing the integration of conservation science into policy, management, and society at large. We focused on the second, integrative type of leadership because we believe it presents the greatest opportunity for improving conservation effectiveness. We identified 8 leadership principles derived mainly from the "adaptive leadership" literature: recognize the social dimension of the problem; cycle frequently through action and reflection; get and maintain attention; combine strengths of multiple leaders; extend your reach through networks of relationships; strategically time your effort; nurture productive conflict; and cultivate diversity. Conservation scientists and practitioners should strive to develop themselves as leaders, and the Society for Conservation Biology, conservation organizations, and academia should support this effort through professional development, mentoring, teaching, and research.

Native grasslands that support diverse populations of birds are being converted to cropland at an increasing rate in the Prairie Pothole Region of North America. Although limited funding is currently available to mitigate losses, accurate predictions of probability of conversion would increase the efficiency of conservation measures. We studied conversion of native grassland to cropland in the Missouri Coteau region of North and South Dakota (U.S.A.) during 1989-2003. We estimated the probability of conversion of native grassland to cropland with satellite imagery and logistic regression models that predicted risk of conversion and by comparing the overlap between areas of high biological value and areas most vulnerable to conversion. Annualized probability of conversion was 0.004, and 36,540 ha of native grassland were converted to cropland during the period of our study. Our predictive models fit the data and correctly predicted 70% of observed conversions of grassland. Probability of conversion varied spatially and was correlated with landscape features like amount of surrounding grassland, slope, and soil productivity. Tracts of high biological value were not always at high risk of conversion. We concluded the most biologically valuable areas that are most vulnerable to conversion should be prioritized for conservation. This approach can be applied broadly to other systems and offers great utility for implementing conservation in areas with spatially variable biological value and probability of conversion.