Regular issues
Monographs
Special Issues



Southeastern Naturalist
    SENA Home
    Range and Scope
    Board of Editors
    Staff
    Editorial Workflow
    Publication Charges
    Subscriptions

Other EH Journals
    Northeastern Naturalist
    Caribbean Naturalist
    Urban Naturalist
    Eastern Paleontologist
    Eastern Biologist
    Journal of the North Atlantic

EH Natural History Home

Nest-Site Selection and Success of Louisiana Bald Eagles
Nickolas R. Smith, Thomas J. Hess Jr., and Alan D. Afton

Southeastern Naturalist, Volume 16, Issue 3 (2017): 343–361

Full-text pdf (Accessible only to subscribers.To subscribe click here.)

 

Site by Bennett Web & Design Co.
Southeastern Naturalist 343 N.R. Smith, T.J. Hess, Jr., and A.D. Afton 22001177 SOUTHEASTERN NATURALIST 1V6o(3l.) :1364,3 N–3o6. 13 Nest-Site Selection and Success of Louisiana Bald Eagles Nickolas R. Smith1,2,*, Thomas J. Hess Jr.3, and Alan D. Afton4 Abstract - Mating pairs of Haliaeetus leucocephalus (Bald Eagle) nest during winter in Louisiana, and numbers of nests have increased exponentially since the mid-1970s. Active nests have remained relatively concentrated within the south-central and southeastern part of the state, in an area primarily consisting of inland swamps, coastal marshes, and barrier islands, which is referred to as the Basin. However, as the number of nests continues to grow, it is expected that nesting will continue to expand geographically into new habitats. In order to manage an expanding population, it is imperative to first determine parameters that influence nest-site selection. To evaluate site selection and success, we conducted GIS-based analyses to evaluate geographic variables such as proximity to water, landcover, human activity, and other nests. We compared 387 active nests from the 2007–2008 winter nesting season and 1935 random sites, which represented available habitat for site selection. Our results suggest that success of a nest within the Basin was not greatly influenced by the physical characteristics around a site, but sites with the highest probability of being selected for nesting generally had a higher probability of success. Initial selection of a nest site was most influenced by distance to road, number of houses per km2, and landcover within 3 km, but the influence of these variables varied between sites within and outside the Basin. Our results should assist managers in making informed decisions about effects of future developments, conservation activities, and human use on current and future suitable nesting habitat. Introduction Species management frequently relies on knowledge and management of habitats occupied by that species. Researchers often compare characteristics that make up the area where a species is found to the areas that are available in order to better understand factors that may influence site selection (Jones 2001). Habitat selection occurs at hierarchical levels (Johnson 1980), wherein an animal first selects a geographical range, then a home range within, and then selects a place to nest within that home range. Selection may vary at these different levels (Saalfeld and Conway 2010, Thompson and McGarigal 2002), but identifying habitat selection at a landscape level facilitates broad management application, while still providing direction for more-refined management on a local level. Modeling nest-site selection to understand the disproportionate use of particular habitats, especially when related to the success of a site, helps to achieve the ultimate goal of understanding 1School of Renewable Natural Resources, Louisiana State University, Baton Rouge, LA 70803. 2Current address - Ducks Unlimited Inc., One Waterfowl Way, Memphis TN 38120. 3Louisiana Department of Wildlife and Fisheries, Rockefeller Refuge, 5476 Grand Chenier Hwy, Grand Chenier, LA 70743 (deceased). 4US Geological Survey, Louisiana Cooperative Fish and Wildlife Research Unit, Louisiana State University, Baton Rouge, LA 70803. *Corresponding author - nrsmith@ducks.org. Manuscript Editor: Jason Davis Southeastern Naturalist N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 344 the relationship that selection has to the reproductive fitness for the individuals involved (Jones 2001). Nest-site selection on a local level has been extensively studied, indicating that Haliaeetus leucocephalus L. (Bald Eagle) typically select nest sites in large mature trees (Andrew and Mosher 1982, Anthony and Isaacs 1989, Harris et al. 1987, McEwan and Hirth 1979, Saalfeld and Conway 2010, Wood et al. 1989) in close proximity to open water (Andrew and Mosher 1982, Anthony and Isaacs 1989, Harris et al. 1987, McEwan and Hirth 1979, Wood et al. 1989). Other factors that influence nest-site selection include size of water bodies (Anthony and Isaacs 1989, Dzus and Gerrard 1993, Gerrard et al. 1975), prey availability (Gende et al. 1997, Isaacs et al. 1983), human activity/disturbance (Andrew and Mosher 1982, Anthony and Isaacs 1989, Buehler et al. 1991, Guinn 2013, Mundahl et al. 2013, Saalfeld and Conway 2010), and habitat surrounding a site (Andrew and Mosher 1982, Anthony and Isaacs 1989, Buehler 1995, McEwan and Hirth 1979, Mundahl et al. 2013, Saalfeld and Conway 2010, Wood et al. 1989). Harris et al. (1987) examined nest-site characteristics at a local level in southcentral and southeastern Louisiana during 1977–1980, and reported that nests were located primarily in Taxodium distichum (L.) Rich. (Bald Cypress)/ Nyssa aquatica Marshall (Water Tupelo) swamps adjacent to marshes, rivers, canals, bayous, ponds, or lakes. Since then, the number of Bald Eagles nesting in Louisiana has grown rapidly, with no indications of slowing since at least 2008 (Smith et al. 2016). Although Louisiana Bald Eagles still nest primarily in the southcentral and southeastern portion of the state, they are expected to move into other habitats as the population expands (LADWF 2007). Understanding nest-site selection and factors contributing to nest success, especially after major expansion in number of nests and distribution, may allow managers to make informed decisions about potential effects of future developments, conservation activities, and human use. Accordingly, we used nest data collected during the most recent statewide Louisiana nest survey to (1) describe habitats used by nesting Bald Eagles in Louisiana, (2) examine factors influencing landscape-level nest-site selection and success, and (3) identify areas with high potential for future nest sites. Field-site Description Bald Eagles establish territories and nest during winter throughout Louisiana, but their nests are unevenly distributed; the majority of nesting occurs within the south-central and southeastern part of the state (Smith et al. 2016). This area closely matches the boundaries of 2 of Louisiana’s level-IV ecoregions: inland swamps, and the deltaic coastal marshes and barrier islands (Daigle et al. 2006), henceforth collectively referred to as “the Basin” (Fig. 1). The inland swamps are the northern extent of the intertidal basins and comprise a large portion of the Atchafalaya Basin. Their poorly drained soils are dominated by Bald Cypress/Water Tupelo swamps. The deltaic coastal marshes and barrier islands are dominated by brackish and saline marshes. The Basin encompasses 18% of Louisiana and is within the Southeastern Naturalist 345 N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 Mississippi Alluvial Plain, which covers 38% of the state and is one of Louisiana’s 6 level-III ecoregions. Outside the Basin, the Mississippi Alluvial Plain is mostly a flat alluvial area comprised of agricultural land—mainly cotton, corn, soybeans, pasture, and aquaculture— as well as bottomland forests. The South Central Plains cover 40% of the state and are primarily comprised of forested or woodland habitats, with less than 20% of the region in agricultural land. The Western Gulf Coastal Plain comprises 13% of the state and is a relatively flat area with fertile soils; rice and soybean production are the primary land uses in the region. The Southeastern Plains, Mississippi Valley Loess Plains, and Southern Coastal Plain together encompass only 9% of the state and are similarly comprised of a mosaic of cropland, pasture, wetland, forested, and woodland habitats. Methods We used GIS and remote sensing to compile variables that were previously known or hypothesized to influence nest-site selection and succe ss (Table 1). Data Figure 1. Map of Bald Eagle active nest sites (n = 387) from the 2007–2008 winter nesting season in relation to level-III ecoregions and the area referred to as the Basin. Southeastern Naturalist N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 346 for some variables were not available for 2007; to replace these data, we selected data sources from 2006 and 2010 (see Table 1) because we assumed them to be most representative of the values during the 2007–2008 winter nesting season. We compiled information from previous studies on Bald Eagles (Buehler 1995, Harris et al. 1987, McEwan and Hirth 1979, Peterson 1986, Saalfeld and Conway 2010) to guide the development of factors to include in our study. We gave separate consideration to the biological basis of each included variable for both nest-site selection and nest success, as suggested by Burnham and Anderson (2002) and Anderson (2008). Nest-site selection We used nest-location data collected during the 2007–2008 winter nesting season (most recent available statewide survey) to examine landscape-level nest-site selection. The LADWF maintains records of known nest locations as reported by private individuals, state and federal personnel, and the media. These sites were monitored annually, and when new nest sites were detected, they were incorporated into the surveys (Smith et al. 2016). Multiple nests may occur within a nesting Table 1. Summary of data used to model landscape-level nest-site selection and success. Variable Biological indicator Data Source Nest sites Winter 2007–2008 nest LDWFA locations and status Roads Human disturbance TIGER/line roads (2006) US Census Bureau Houses per km2 Human disturbance 2010 population—census US Census Bureau block group 2010 census block group US Census Bureau Nearest nest Nest density Winter 2007–2008 nest LDWF locations Water body Foraging habitat High-resolution NHDB– US Geological Survey linear (1:24,000) High-resolution NHD– US Geological Survey discrete (≥8 ha) NLCDC 2006 (30 m)– MRLC ConsortiumD discrete (≥8 ha) Basin Ecoregion Level III and IV ecoregions US Geological Survey Land cover (0.5 and 3 km) Habitat NLCD 2006 (30 m) MRLC Consortium Open water Woody wetland Emergent herb wetland Developed Agricultural Forest ALouisiana Department of Wildlife and Fisheries. BNational Hydrography Dataset. CNational Land Cover Database. DMulti-Resolution Land Characteristics Consortium. Southeastern Naturalist 347 N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 territory; thus, to reduce pseudoreplication, we excluded from analysis nests not classified as active. Nests were classified as active by the presence of at least 1 of the following: (1) 1 or more adults in or near a nest with signs of nest refurbishment, i.e., presence of fresh nesting material; (2) an adult sitting low in the nest presumably incubating; or (3) the presence of eggs or young. Our analysis of nestsite selection is based on a total of 387 active nest locations. To facilitate comparisons between nest sites and available habitat, we generated 5 corresponding random locations for each nest site (n = 1935). We set selection of each random site to be within 50 km of the corresponding nest as a representation of the relative non-nesting winter home range of Bald Eagles in Louisiana (Smith et al., in press). After looking at the winter home ranges of non-nesting Bald Eagles in Louisiana that were marked with satellite transmitters in a separate part of our research, we assumed that this distance would best reflect the area from which the nesting pair would have first selected their nest site. Within the 50-km radius around each nest, random sites were restricted to areas considered to be suitable habitat (Fig. 2). We based our criteria for determining site suitability on variables associated with nest sites (Andrew and Mosher 1982, Figure 2. Map of areas considered to be suitable habitat in relation to the area referred to as the Basin. Southeastern Naturalist N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 348 Anthony and Isaacs 1989, Dzus and Gerrard 1993, Gerrard et al. 1975, Harris et al. 1987, McEwan and Hirth 1979, Saalfeld and Conway 2010, Wood et al. 1989) and used in the modeling of suitable habitat in other studies (Andrew and Mosher 1982, Grier and Guinn 2003, Saalfeld and Conway 2010, Watts et al. 2008). However, specific values were reflective of characteristics from Bald Eagle nests in Louisiana. Specifically, we classified suitable habitat as: (1) less than 1 km (representative of distance to nearest water body for nest sites) from a water body (discrete water body ≥8 ha or linear water body represented as polygon at 1:24,000 scale), (2) at least 3 km (representative of the observed distance between nests) from another nest, and (3) within suitable landcover. We identified suitable landcover types as: emergent herbaceous wetland, wooded wetland, and forest (Smith 2014) because these types were most likely to have trees that could support a nest. We further restricted the emergent herbaceous wetland landcover type to the area within 1 km of at least 1 other suitable landcover type. We set this condition to exclude large herbaceous wetlands with the lowest probability of containing suitable nest trees, such as coastal marshes. We restricted random sites to these landcover types because 95% of documented nests were located within these landcovers. We used distance to nearest road and number of houses per km2 to index human disturbance around nests and random points. We assumed that sites closer to roads and with more houses per km2 experienced more human disturbance. In Texas, distance to human disturbance was the best predictor of landscape-level nest-site selection (Saalfeld and Conway 2010), but absolute distance to human structures may be misleading because tolerance to human presence may be increased by visual buffers and habituation (Andrew and Mosher 1982, Millsap et al. 2004). We included 2 human-disturbance indices to assess whether either had an effect on nest-site selection by Bald Eagles in Louisiana.We used TIGER/line shapefiles created in 2006 to identify locations of roads, then used a spatial join to calculate the Euclidian distance from a site to the nearest road. We employed 2010 census data to calculate density of houses, then used number of houses and total area within each census-block group to determine number of houses per km2. Landcover type has been included in most studies of Bald Eagle nest-site selection (Andrew and Mosher 1982, Anthony and Isaacs 1989, Buehler 1995, Curnutt and Robertson 1994, Dzus and Gerrard 1993, Gerrard et al. 1975, Guinn 2013, Harris et al. 1987, McEwan and Hirth 1979, Peterson 1986, Saalfeld and Conway 2010, Wood et al. 1989). We hypothesized that the composition of habitat around a site influences nest-site selection and that some landcover types are more influential than others. We followed the 2006 National Land Cover Database (NLCD; Fry et al. 2011) to classify landcover, wherein we combined similar cover types to reduce the number of variables to 6 cover types (open water, wooded wetland, emergent herbaceous wetland, developed, agricultural, and forest; Smith 2014). We determined the proportion of landcover types at 2 spatial scales, 500 m and 3 km, around each site. We set the area within 500 m of a nest to correspond to USFWS (1987) primary management zones. We selected a 3-km scale to represent the observed home-range size of nesting Bald Eagles in Louisiana (Smith et al., in press). We calculated the Southeastern Naturalist 349 N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 proportion of each landcover type within 500 m and 3 km of each site using the ISECTPLYRST tool in Geospatial Modeling Environment (GME; version 7.2.0). Nests were unevenly distributed across the state; the majority of nesting occurred within the south-central and southeastern part of the state (Smith et al. 2016) within the Basin (Fig. 1). We defined the extent of the Basin by combining 2 level-IV ecoregions: (1) inland swamps and (2) deltaic coastal marshes and barrier islands (Daigle et al. 2006), and we examined ecoregion-level variations in nestsite selection within the Basin and outside the Basin. Because our random points were set to be within 50 km of a nest, there were a similar percent of nests and random points within the Basin (81% and 78%, respectively) and outside the Basin (19% and 22%, respectively). Therefore, we expected that the effect of Basin alone would not greatly influence selection, but rather the 2-way interactions of Basin with other variables might show ecoregion-level variations in the other variables. Roads are unevenly distributed across Louisiana, with fewer roads located within the Basin. The area within the Basin is largely comprised of wetland habitats (Daigle et al. 2006), and few roads cross some of the large wetland areas. Outside the Basin, there are very few areas where roads are more than a few kilometers apart. Thus, the observed distribution of roads in Louisiana may not allow for sites to be more than a few kilometers from the nearest road except for within the Basin and, therefore, the importance of distance to the nearest road may vary between areas. For this reason, we included a 1st-order interaction of Basin x distance to road in our analysis. Likewise, the wetland habitat within the Basin is not ideal for human development, and therefore, most of the Basin (81.2%) has ≤5 houses per km2. Outside the Basin, over a third of the area (34.6%) has >5 houses per km2. Thus, we expected that the importance of houses, as an index of human disturbance, would vary between sites within and outside the Basin and we included Basin x houses as a 1storder interaction. General habitat characteristics also varied in relation to whether a site was within or outside the Basin. Woody and emergent herbaceous wetlands are more abundant within the Basin, whereas developed, agricultural, and forested landcover types are more common outside the Basin. Therefore, we included a 1st-order interaction of Basin x landcover type. Aside from these 1st-order interactions with the Basin, we had no biological reason to presume that the influence of nest-site selection would vary for any other interactions. In summary, we considered the following explanatory variables in our analysis of landscape-level nest-site selection: (1) distance to road, (2) houses per km2, (3) land cover, (4) Basin, (5) Basin x distance to road interaction, (6) Basin x houses per km2 interaction, and (7) Basin x landcover type interaction. Nest success within the Basin We used status data for nests active in the survey to classify nests as successful or unsuccessful; we recorded nests as successful if a minimum of 1 young, 8 weeks of age or greater, was observed. We included in our nest success analysis Southeastern Naturalist N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 350 only those nests for which a status could be determined. Of those nests in which a status was determined, there were 39 nests outside the Basin, of which only 7 were unsuccessful. Due to the low sample-size of nests found outside the Basin, we restricted our analysis to nests located within the Basin. Thus, our final analysis of nest success is based on a total of 276 nests of which 234 (85%) were successful and 42 (15%) were unsuccessful. We used the same explanatory variables (distance to road, number of houses per km2, and landcover) in our nest-success analysis as we used in the nest-site selection analysis. We considered hypotheses from the nest-site selection analysis to have similar effects on nest success, i.e., a variable hypothesized to have a negative effect on the probability of a site being selected for nesting would also have a negative effect on the probability of a nest being successful. In addition to the 3 explanatory variables used in the analysis of nest-site selection, we included distance to nearest water body and distance to nearest nest for our analysis of nest success (Table 1). We used distance to nearest water body to index distance to foraging areas. Fish and water-birds comprise the majority of a Bald Eagle’s diet (Buehler 2000, Dugoni et al. 1986); therefore, we hypothesized that successful nest sites would be closer to areas that provide such food sources. Many other studies have reported that nests were close to water (Andrew and Mosher 1982, Anthony and Isaacs 1989, Buehler 2000, Harris et al. 1987, McEwan and Hirth 1979, Wood et al. 1989), and that eagles prefer larger water bodies over smaller ones (Anthony and Isaacs 1989, Dzus and Gerrard 1993, Gerrard et al. 1975), but smaller water bodies may also provide suitable foraging opportunities, especially when located near other water bodies (Peterson 1986). We considered a water body to be any large, discrete body of water ≥8 ha in size (e.g., lakes, ponds, and reservoirs) as well as large linear bodies of water which were represented as polygons rather than lines at a 1:24,000 scale (e.g., rivers, streams, and canals). We used National Hydrography Dataset (NHD) High Resolution Discrete and Linear Waterbody layers and removed all water bodies that did not meet the size requirement. We also removed swamp/marsh water types from the NHD Discrete Waterbody layer because these were represented by landcover types and were not truly representative of unobstructed open water. The NHD had some data gaps, e.g., large water bodies or parts of large rivers, such as parts of the Mississippi River, were not represented; therefore, we supplemented these files with open water from the 2006 NLCD (Fry et al. 2011). We clipped the NLCD raster files open-water landcover class to remove coastal water, used a region group to calculate water body size, and then eliminated areas less than 8 ha in size from the resulting data set. We did not consider distance to nearest water body in models of nest-site selection because random sites were restricted to areas within 1 km of a water body. We used distance to nearest nest, as an index of nest density, to evaluate the hypothesis that nesting density would affect nest success (Dzus and Gerrard 1993, Elliott et al. 2011). we employed the POINTDISTANCE tool within GME to calculate the Euclidean distance to the nearest active nest. Southeastern Naturalist 351 N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 In summary, we considered the following explanatory variables in our analysis of nest success: (1) distance to road, (2) number of houses per km2, (3) landcover, (4) distance to nearest water body, and (5) distance to nearest nest. We did not consider interactions because we had no biological reason to presume that any interaction would influence nest success. Statistical analysis We used stratified logistic regression (PROC LOGISTIC; SAS Institute Inc. 2011) to evaluate the influence of multiple explanatory variables on (1) the probability of a nest site being selected and (2) the probability of a site being successful or unsuccessful. We specified the identification number of the nest and associated random sites in the strata option; the response variable was whether or not a site was used for nesting. We included site characteristics as explanatory variables and the classification of a site (nest/random, successful/unsuccessful) as the response variable. We developed a set of 18 a priori candidate models for the evaluation of landscape-level nest-site selection (Table 2) and a set of 13 a priori candidate models for the analysis of nest success (Table 3). Model selection was based on Akaike’s information criteria, adjusted for small sample size (AICc), where models that best supported the data had the lowest AICc. To be consistent with an AIC approach, we evaluated 1st-order parameters from the top model(s) using an 85% confidence interval (CI) of the parameter estimates (Arnold 2010). Only those parameters that did not overlap zero were considered to be influential in nest-site selection or success. Multicollinearity was inherent due to the nature of our landcover data (Graham 2003). For example, the percent of 1 landcover type present within a buffer influenced the percent of all other landcover types within that same buffer as well as the percent landcover in the smaller/larger buffer, because buffers were inclusive. For this reason, in the nest-site selection and nest-success analyses, all landcover types at a single spatial scale were either included or excluded together from a model and were represented as the variable “landcover” with 6 levels. We separately tested which spatial scale was most influential in nest-site selection and nest success by running the full model from each analysis with landcover within 500 m and then running the full model again with landcover within 3 km. The full model that performed best, as determined by the lowest AICc, was then considered the most influential spatial scale. In these preliminary analyses, landcover within a 3-km radius of sites had a greater influence in nest-site selection and nest success than landcover within a 500-m radius and, therefore, was used for all candidate models subsequently analyzed. To test for multicollinearity in all other variables, we used a correlation matrix (PROC CORR; SAS Institute, Inc. 2011), wherein variables with Pearson correlation coefficients ≥0.7 were considered highly correlated (Dormann et al. 2013); however, none of the variables we considered were found to be highly correlated. Southeastern Naturalist N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 352 Results General relationships Nests within the Basin were located an average of 2.6 km from the nearest road (range = 0.0–21.3 km); whereas, nests outside of the Basin were located an average of 0.6 km from the nearest road (range = 0.0–2.4 km). The distribution of distances from the nearest road and number of houses per km2 for nests and random points were relatively similar except at the extremes. Ninety seven percent of all nests were at least 400 m from the nearest nest, however, nest sites outside the Basin were, on average, 13.7 km from the nearest nest (median = 7.8 km, min–max = 0.7–52.3 km); whereas, they were 2.5 km from the nearest nest within the Basin (median = 1.6, min–max = 0.1–19.3 km). All nests were within 3 km of a large body of water (average distance = 466 ± 26 m). Successful nests were farther from a large body of water, on average, than were unsuccessful nests (473 m and 372 m, respectively). Wooded wetlands made up the largest proportion of landcover types within 3 km of nests (mean = 44.1%); emergent herbaceous wetland was the 2nd-most abundant landcover type (mean = 26.0%) at that scale. Suitable habitat, to which random sites were restricted, encompassed 23,897 km2 in Louisiana. The Basin contains 1/3 of that area defined as suitable and 81% of active nests, despite the Basin only covering 18% of the state (Figs. 1, 2). Nest-site selection The top model accounted for 55% of the Akaike model weight and included distance to road, number of houses per km2, landcover within 3 km, Basin, and all 1storder interactions between Basin and the other 3 variables (Table 2). The top model correctly classified 71.6% of sites. Interactions indicated that the importance of the variables examined (roads, houses, and landcover) were not consistent between sites within and outside the Basin. For example, the importance of emergent herbaceous wetlands was more influential outside the Basin. We did not use an 85% CI to infer the importance of individual variables because the inclusion of all 1st-order interactions in the top model suggested that their importance was not consistent between areas within and outside the Basin. The influence of roads and houses varied between sites within and outside the Basin. Outside the Basin, nest sites and random sites were usually a similar distance from roads (0.55 km and 0.58 km, respectively) and in areas with fewer houses per km2 (13.0 and 21.9, respectively). However, within the Basin, nest sites were usually closer to roads (2.59 km) than random sites (3.50 km) and in areas with more houses per km2 (8.0) than random sites (6.2). Overall, nest sites within the Basin were farther from roads (2.59 km) and in areas with fewer houses per km2 (8.0, than nest sites outside the Basin (0.55 km, 13.0) The areas around nest and random sites were usually comprised of similar percentages for developed, forested, and wooded wetland landcover types. However, outside the Basin, areas around sites contained a larger percentage of developed and forested landcover and less wooded-wetland landcover. The influence of open water and agricultural landcover types was relatively consistent between sites within and Southeastern Naturalist 353 N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 outside the Basin; nests were in areas with more open water and less agricultural land than random sites. However, outside the Basin, areas around sites contained a smaller percentage of open water and a larger percentage of agricultural landcover when compared to sites within the Basin. The influence of emergent herbaceous wetlands had the greatest variation between sites within and outside the Basin. Within the Basin, the percentage of emergent herbaceous wetland landcover around a site was similar for nests and random sites, (28% and 30%, respectively), but outside the Basin, nests were in areas with twice the amount of emergent herbaceous wetland landcover than random sites (16% and 8%, respectively; Fig. 3). Nest success within the Basin The candidate models we considered indicated that nest success within the Basin was best predicted by distance to road, number of houses per km2, and landcover Table 2. Stratified logistic regression models predicting Louisiana Bald Eagle nest sites (n = 387) versus random sites (n = 1935), including number of parameters (K), Akaike’s information criterion adjusted for small sample size (AICc), difference between the AICc of the given model and the model with the lowest AICc (ΔAICC), and Akaike’s model weight (wi). Model K AICC ΔAICC wi RoadA, housesB, landcoverC, basinD, basin x roadE, 15 1199.56 0.00 0.55 basin x housesF, basin x landcoverG Road, landcover, basin, basin x road, basin x landcover 13 1202.01 2.45 0.16 Road, landcover, basin, basin x landcover 12 1202.22 2.66 0.15 Road, houses, landcover, basin, basin x road, basin x landcover 14 1203.50 3.94 0.08 Road, houses, landcover, basin, basin x land cover 13 1203.72 4.16 0.07 Road, landcover 6 1211.50 11.94 0.00 Road, houses, landcover 7 1213.15 13.59 0.00 Houses, landcover, basin, basin x houses, basin x land cover 13 1225.15 25.59 0.00 Land cover, basin, basin x land cover 11 1227.68 28.12 0.00 Houses, land cover, basin, basin x land cover 12 1229.23 29.67 0.00 Landcover 5 1237.29 37.73 0.00 Houses, landcover 6 1238.89 39.33 0.00 Road, basin, basin x road 3 1373.57 174.01 0.00 Road, houses, basin, basin x road, basin x houses 5 1373.99 174.43 0.00 Road 1 1375.93 176.37 0.00 Intercept onlyH 0 1375.93 176.37 0.00 Houses, basin, basin x houses 3 1385.62 186.06 0.00 Houses 1 1388.33 188.77 0.00 ADistance to nearest road (km). BNumber of houses per km2. CProportion of landcover type within 3-km at 5-km levels: open water, developed, forest, agricultural, emergent herbaceous wetland, and reference level set as wooded wetland. DBasin at 2 levels: within Basin and outside Basin; reference level set as outside Basin. E1st- order interaction between outside Basin and distance to nearest road (km). F1st-order interaction between outside Basin and houses per km2. G1st-order interaction between outside Basin and landcover type within 3-km at 5-km levels: open water, developed, forest, agricultural, emergent herbaceous wetland, and reference level set as wooded wetland. HIntercept-only model for benchmark comparison. Southeastern Naturalist N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 354 within 3 km (Table 3). This top model accounted for 68.0% of the Akaike model weight and correctly classified 70.1% of nests as either successful or unsuccessful. Within the top model, distance to roads, number of houses per km2, developed, agricultural, and emergent herbaceous wetlands had parameter estimates with confidence intervals that did not overlap zero, and thus, were considered most influential. The probability of a site being successful decreased as distance to road increased; successful nests were generally closer to roads than unsuccessful nests (2.6 km and 2.9 km, respectively). Number of houses per km2 was marginally influential in predicting the success of a site; successful nests were usually in areas with slightly more houses than unsuccessful nests (7.5 and 5.8, respectively). The probability of a site being successful decreased as the proportion of developed and agricultural land increased, but probability of success increased as proportion of emergent herbaceous wetland landcover increased within 3 km of a nest site. Discussion Nest site selection The interaction between other explanatory variables and the Basin suggest that Bald Eagles use different strategies for nest-site selection depending on whether their home ranges are within the Basin or in another part of Louisiana. However, the Basin interactions that included roads and houses were highly influenced by sites Figure 3. Percent of 6 landcover types within 3 km of nests and random sites within and outside the Basin, including standard error bars. Southeastern Naturalist 355 N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 at the extremes and not necessarily a response to human disturbance. For example, outside the Basin, there were 4 random sites with number of houses per km2 ranging from 147–501, but the maximum number of houses for nests was 124. These few sites at the extreme were likely driving the variable’s inclusion in the model and suggesting that random sites were in areas with more houses than nest sites outside the Basin. Outside the Basin, roads appear to have little effect on nest-site selection because there are very few areas greater than a few kilometers from a road. Within the Basin, nest sites were located in areas closer to roads and in areas with more houses per km2. Within this region, the characteristics that make areas suitable for human development may also provide characteristics which enhance a site’s suitability for nesting. For instance, there is little elevation change within the Basin (Hupp et al. 2008), so areas farther above mean sea level provide stable areas for human development and protection from flooding. Areas with less flooding may be supporting larger Cypress and Tupelo trees (Keim et al. 2013), and thus are better suited for supporting Bald Eagle nests compared to areas that are more frequently inundated. Overall, human disturbance may not affect nest-site selection by Bald Eagles from Louisiana as strongly as in other studies (Andrew and Mosher 1982, Saalfeld and Conway 2010); our results are more consistent with the idea that the influence of human disturbance on Bald Eagle nest-site selection is minimal compared to other factors (McEwan and Hirth 1979, Millsap et al. 2004). Table 3. Logistic regression models predicting successful Louisiana Bald Eagle nest sites (n = 234) versus unsuccessful sites (n = 49), including number of parameters (K), Akaike’s information criterion adjusted for small sample size (AICc), difference between the AICc of the given model and the model with the lowest AICc (ΔAICC), and Akaike’s model weight (wi). Model K AICC ΔAICC wi RoadA, housesB, landcoverC 7 226.17 0.00 0.68 Road, houses, landcover, nearest nestD, nearest waterE 9 229.03 2.86 0.16 Houses, landcover, nearest water 7 229.86 3.70 0.11 Landcover, nearest water 6 233.99 7.82 0.01 Land cover 5 234.60 8.44 0.01 Nearest nest 1 235.13 8.97 0.01 Land cover, nearest nest 6 235.16 8.99 0.01 Nearest nest, nearest water 2 235.84 9.68 0.01 Intercept onlyF 0 237.42 11.26 0.00 Nearest water 1 238.15 11.98 0.00 Houses 1 238.94 12.78 0.00 Road 1 239.14 12.98 0.00 Road, houses 2 240.82 14.66 0.00 ADistance to nearest road (km). BNumber of houses per km2. CProportion of landcover type within 3 km at 5 levels: open water, developed, forest, agricultural, emergent herbaceous wetland, and reference-level set as wooded wetland. DDistance to nearest nest (km). EDistance to nearest water body (km). FIntercept only model for benchmark comparison. Southeastern Naturalist N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 356 The model results being driven by a few sites at the extremes may have been an artifact of random sites being restricted to areas considered to be suitable of supporting Bald Eagle nesting. This restricted designation of random sites may have created relative uniformity for many aspects, whereas if the parameters we used to locate random sites were not restricted, we may have expected greater distinction between nest and random sites. However, our results from this greater distinction would likely have been less informative than our more restrictive analysis because they would likely have suggested that nest-site selection is driven by factors already confirmed, such as that Bald Eagles nest near water and in areas with suitable nest trees. Thus, we argue that our model should be used to determine the probability of a site being used for nesting within areas deemed to be suitable for nesting; i.e., within 1 km of a large body of water, and in a forested, wooded wetland, or emergent herbaceous wetland landcover type with trees capable of supporting a nest. Landcover type around a site was an influential variable in our top model. Landcover type within 500 m provided less predictive power than did landcover within 3 km for both the nestsite selection models and the nest-success models. This outcome may be explained by the fact that Bald Eagle nest-site selection at a local level is relatively homogenous; most nests are in large trees and the area immediately surrounding the nest is comprised of a wooded landcover type (Andrew and Mosher 1982; Anthony and Isaacs 1989; Buehler 1995, 2000; Harris et al. 1987; Saalfeld and Conway 2010; Wood et al. 1989). Nest sites had more open-water than did random sites, which probably provided more foraging habitat. Large bodies of open water are the primary foraging habitats for Bald Eagles (Buehler 2000), and many studies have found that their presence has been influential in nest-site selection (Andrew and Mosher 1982, Anthony and Isaacs 1989, Buehler 2000, McEwan and Hirth 1979). However, a congregation of smaller water bodies may also provide suitable foraging opportunities (Peterson 1986), which may be the case for the nests located farthest from a large body of water. Our analysis accounted for Bald Eagles that selected for areas with multiple smaller foraging areas within reasonably close proximity because we used proportion of open water around a site. Thus, our results should better account for differing foraging strategies in relation to water than did previous studies which used only a linear distance to nearest large water body. As expected, wooded and emergent herbaceous wetlands made up the greatest proportion of the area around sites within the Basin because these 2 landcover types were the most abundant in this region. Likewise, wooded wetlands were also the most abundant outside the Basin; however, we found that nests were in areas with twice as much emergent herbaceous wetland than were random sites outside the Basin. Like open water, emergent herbaceous wetlands provide abundant foraging opportunities for Bald Eagles in Louisiana (Harris et al. 1987). In addition to the fish resource, emergent herbaceous wetlands also support large numbers of waterfowl that winter in Louisiana (Baldassarre 2014, Michot 1996) and which make up a large portion of the prey consumed during the nesting period (Dugoni et al. 1986). Southeastern Naturalist 357 N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 Thus, emergent herbaceous wetland landcover has a discernible influence on nestsite selection, especially when it is limited within the landscape. Bald Eagles selected against agricultural land; however, it was still the 3rd-most abundant landcover type around nests outside the Basin, which may be due to a landscape-level change in habitat characteristics between the different ecoregions. Bald Eagles may be first selecting nest sites within the Basin, where there is more woody and emergent herbaceous wetland landcover, but as these areas become occupied and nesting expands outside the Basin, the habitats shift to more forested and agricultural landcover (Daigle et al. 2006). Although these habitats are more abundant outside the Basin, their use may not be proportional with other habitat types. Examination of resource utilization by nesting individual Bald Eagles may provide better insight into the relationship that various habitat types play in nestsite selection. In general, nest-site selection outside the Basin appears to be mainly driven by habitats with more open water, less agricultural land, and more emergent herbaceous wetlands. The Basin may provide ideal habitat for Bald Eagle nesting, and thus, may be driving the dissimilarity in nest numbers between the Basin and the rest of the state. As of 2008, the number of nesting Bald Eagles in Louisiana has increased exponentially (Smith et al. 2016). If the number continues to increase, we would expect that more nests will be found outside the Basin because the area within the Basin may be reaching a carrying capacity. If this is the case, our findings may be used to prioritize conservation of areas wherein the probability of nesting would be greatest. Our findings highlight the importance of the Basin and its habitats for the stability of Louisiana’s nesting Bald Eagles. A lack of sediment input, subsidence, and sea-level rise threaten the region (Daigle et al. 2006); Louisiana loses almost a football field-sized amount of of coastline every hour (Couvillion et al. 2011). Besides land loss, the ecosystem can be drastically altered by changes in salinity and hydrological regimes, creating changes in the characteristics of the wetland habitats and the species that occupy those areas (Boesch et al. 1994). These threats may be exacerbated by projected increases in the frequency and severity of hurricanes in the area (Knutson et al. 2010). Bald Cypress/Water Tupelo swamps are relatively resistant to the effects of hurricanes (Shaffer et al. 2009); however, fresh and brackish marshes can be severely affected (Barras 2006). Past hurricanes directly destroyed a large proportion of nests in Louisiana, and although they showed no short-term effect on the population because many of the nests were rebuilt (Hess et al. 1994), the long-term effects are unknown. Together, these threats should be considered in the future conservation and management of Bald Eagles in Louisiana. Nest success within the Basin Nest success within the Basin followed similar patterns as observed for nestsite selection. Thus, areas with the highest probability of being selected for nesting generally had a higher probability of success. However, our top model Southeastern Naturalist N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 358 provided relatively weak evidence that landscape-level variables could be used to accurately predict nest success. Our top model correctly classified 70.1% of nests; however, there was an uneven number of successful versus unsuccessful nests. Thus, predicting that all nests were successful would have provided 84.8% correct classification. Bald Eagles generally have high rates of nest success (Buehler 2000, Driscoll et al. 1999, Saalfeld et al. 2009, Watts et al. 2008), and Louisiana has one of the highest documented rates (Smith et al. 2016). Future research looking at the long-term trend of success for nest sites may provide a better understanding of the landscapelevel characteristics that influence success, and then sites could be grouped by above average, average, or below average success. Analysis of this long-term trend of site success should also better account for nest failures from factors that cannot be characterized through remote sensing. We only considered landscape-level variables that could be characterized using remote sensing, whereas nest success can be affected by other things such as weather, prey availability, disease, and/or the age or skill level of the nesting pair (Buehler 2000, Elliott et al. 2011, Forslund and Pärt 1995, Gende et al. 1997, Millsap et al. 2004). However, of the variables we considered, distance to road, number of houses per km2, and the amount of developed, agricultural, and emergent herbaceous wetland landcover within 3 km appear to be the most influential in predicting the success of a nest within the Basin. Nest success may not be greatly impacted by the physical characteristics around a site, but the initial selection of a site appears to be influenced by at least some landscape-level factors, as shown by our models. Overall, factors associated with habitat degradation and the ability of the Bald Eagle to adapt to a changing environment may be the driving force behind a healthy and expanding nesting population in Louisiana. With these results, managers may be able to focus efforts on the protection of current and future suitable habitat, emphasizing areas with the highest probability of nesting. Acknowledgments We dedicate this manuscript to the memory of Thomas Hess Jr., our co-author who devoted much of his career to protecting the Bald Eagles of Louisiana. Financial support for this study and its publication were provided by the Louisiana Department of Wildlife and Fisheries and the US Fish and Wildlife Service, Division of Federal Aid, through Louisiana State Wildlife Grant T-98, the Rockefeller Wildlife Refuge Trust, the US Geological Survey- Louisiana Cooperative Fish and Wildlife Research Unit, and the School of Renewable Natural Resources at Louisiana State University. We acknowledge the work of W. Dubuc, R. Aycock, G. Melancon, J. Linscombe, and all the individuals from US Fish and Wildlife Service and Louisiana Department of Wildlife and Fisheries who have assisted with Louisiana’s nest-monitoring program since 1975. We also thank the landowners who contributed information on nesting eagles. We thank W. Selman, D. Blouin, and L. Wang for providing critical comments on this manuscript. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the US Government. Southeastern Naturalist 359 N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 Literature Cited Anderson, D.R. 2008. Model-based Inference in the Life Sciences: A Primer on Evidence. Springer, New York, NY. 184 pp. Andrew, J.M., and J.A. Mosher. 1982. Bald Eagle nest-site selection and nesting habitat in Maryland. Journal of Wildlife Management 46(2):382–390. Anthony, R.G., and F.B. Isaacs. 1989. Characteristics of Bald Eagle nest sites in Oregon. Journal of Wildlife Management 53(1):148–159. Arnold, T.W. 2010. Uninformative parameters and model selection using Akaike's information criterion. Journal of Wildlife Management 74(6):1175–1178. Baldassarre, G.A. 2014. Ducks, Geese, and Swans of North America, Second Edition. Johns Hopkins University Press, Baltimore, MD. 1088 pp. Barras, J.A. 2006. Land-area changes in coastal Louisiana after the 2005 hurricanes: A series of three maps. US Geological Survey Open–File Report 06-1274. US Geological Survey, Reston, VA. Boesch, D.F., M.N. Josselyn, A.J. Mehta, J.T. Morris, W.K. Nuttle, C.A. Simenstad, and D.J.P. Swift. 1994. Scientific assessment of coastal wetland loss, restoration, and management in Louisiana. Journal of Coastal Research, Special Issue No. 20. 103 pp. Buehler, D.A. 1995. A geographic information system to identify potential Bald Eagle breeding habitat for southeastern United States rivers and reservoirs. Proceedings of the Annual Conference of Southeastern Association of Fish and Wildlife Agencies 49(1):293–303. Buehler, D.A. 2000. Bald Eagle (Haliaeetus leucocephalus). No. 506, In A. Poole and F. Gill (Eds.). The Birds of North America. The Birds of North America, Inc., Philadelphia, PA. 40 pp. Buehler, D.A., T.J. Mersmann, J.D. Fraser, and J.K.D. Seegar. 1991. Effects of human activity on Bald Eagle distribution on the Northern Chesapeake Bay. Journal of Wildlife Management 55(2):282–290. Burnham, K.P., and D.R. Anderson. 2002. Model Selection and Multimodel Inference: A Practical Information-theoretical Approach, Second Edition. Springer, New York, NY. 488 pp. Couvillion, B.R., J.A. Barras, G.D. Steyer, W. Sleavin, M. Fischer, H. Beck, N. Trahan, B. Griffin, and D. Heckman. 2011. Land-area change in coastal Louisiana from 1932 to 2010: US Geological Survey Scientific Investigations Map 3164. US Geological Survey, Reston, VA. 12 pp. Curnutt, J.L., and W.B. Robertson. 1994. Bald Eagle nest-site characteristics in south Florida. Journal of Wildlife Management 58(2):218–221. Daigle, J.J., G.E. Griffith, J.M. Omernik, P.L. Faulkner, R.P. McCulloh, L.R. Handley, L.M. Smith, and S.S. Chapman. 2006. Ecoregions of Louisiana (color poster with map [map scale 1:1,000,000]), descriptive text, summary tables, and photographs). US Geological Survey, Reston, VA. Dormann, C.F., J. Elith, S. Bacher, C. Buchmann, G. Carl, G. Carré, J.R.G. Marquéz, B. Gruber, B. Lafourcade, P.J. Leitão, T. Münkemüller, C. McClean, P.E. Osborne, B. Reineking, B. Schröder, A.K. Skidmore, D. Zurell, and S. Lautenbach. 2013. Collinearity: A review of methods to deal with it and a simulation study evaluating their performance. Ecography 36(1):27–46. Driscoll, D.E., R.E. Jackman, W.G. Hunt, G.L. Beatty, J.T. Driscoll, R.L. Glinski, T.A. Gatz, and R.I. Mesta. 1999. Status of nesting Bald Eagles in Arizona. Journal of Raptor Research 33(3):218–226. Southeastern Naturalist N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 360 Dugoni, J.A., P.J. Zwank, and G.C. Furman. 1986. Foods of nesting Bald Eagles in Louisiana. Journal of Raptor Research 20(3/4):124–127. Dzus, E.H., and J.M. Gerrard. 1993. Factors influencing Bald Eagle densities in Northcentral Saskatchewan. Journal of Wildlife Management 57(4):771–778. Elliott, K.H., J.E. Elliott, L.K. Wilson, I. Jones, and K. Stenerson. 2011. Density-dependence in the survival and reproduction of Bald Eagles: Linkages to Chum Salmon. Journal of Wildlife Management 75(8):1688–1699. Forslund, P., and T. Pärt. 1995. Age and reproduction in birds: Hypotheses and tests. Trends in Ecology and Evolution 10(9):374–378. Fry, J.A., G. Xian, S. Jin, J.A. Dewitz, C.G. Homer, L. Yang, C.A. Barnes, N.D. Herold, and J.D. Wickham. 2011. Completion of the 2006 National Land Cover Database for the conterminous United States. Photogrammetric Engineering and Remote Sensing 77(9):858–864. Gende, S.M., M.F. Wilson, and M. Jacobsen. 1997. Reproductive success of Bald Eagles (Haliaeetus leucocephalus) and its association with habitat or landscape features and weather in southeast Alaska. Canadian Journal of Zoology 75(10):1595–1604. Gerrard, J.M., P. Gerrard, W.J. Maher, and D.W.A. Whitfield. 1975. Factors influencing nest-site selection of Bald Eagles in northern Saskatchewan and Manitoba. Blue Jay 33(3):169–176. Graham, M.H. 2003. Confronting multicollinearity in ecological multiple regression. Ecology 84(11):2809–2815. Grier, J.W., and J.E. Guinn. 2003. Bald Eagle habitats and responses to human disturbances in Minnesota. Final Report. Natural Heritage and Nongame Research Program, Minnesota Department of Natural Resources, St. Paul, MN. 44 pp. Guinn, J.E. 2013. Generational habituation and current Bald Eagle populations. Human– Wildlife Interactions 7(1):69–76. Harris, J.O., P.J. Zwank, and J.A. Dugoni. 1987. Habitat selection and behavior of nesting Bald Eagles in Louisiana. Journal of Raptor Research 21(1):27–31. Hess, T.J., W.G. Perry, and W. Dubuc. 1994. Effects of Hurricane Andrew on Louisiana’s nesting Bald Eagles. Proceedings of the Annual Conference Southeastern Association of Fish and Wildlife Agencies 48(1):395–400. Hupp, C.R., C.R. Demas, D.E. Kroes, R.H. Day, and T.W. Doyle. 2008. Recent sedimentation patterns within the central Atchafalaya Basin, Louisiana. Wetlands 28(1):125–140. Isaacs, F.B., R.G. Anthony, and R.J. Anderson. 1983. Distribution and productivity of nesting Bald Eagles in Oregon, 1978–1982. Murrelet 64(2):33–38. Johnson, D.H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61(1):65–71. Jones, J. 2001. Habitat selection studies in avian ecology: A critical review. Auk 118(2):557–562. Keim, R.F., T.J. Dean, and J.L. Chambers. 2013. Flooding Effects on Stand Development in Cypress-Tupelo. Pp. 431–437, In J.M. Guldin (Ed.). Proceedings of the 15th Biennial Southern Silvicultural Research Conference, Hot Springs, AR. US Department of Agriculture, Forest Service, Technical Report SRS-175. Southern Research Station, Asheville, NC. 568 pp. Knutson, T.R., J.L. McBride, J. Chan, K. Emanuel, G. Holland, C. Landsea, I. Held, J.P. Kossin, A.K. Srivastava, and M. Sugi. 2010. Tropical cyclones and climate change. Nature Geoscience 3(3):157–163. Southeastern Naturalist 361 N.R. Smith, T.J. Hess, Jr., and A.D. Afton 2017 Vol. 16, No. 3 Louisiana Department of Wildlife and Fisheries (LADWF). 2007. Louisiana Bald Eagle program status report for 2006–2007. Status Report 5. Louisiana Department of Wildlife and Fisheries, Baton Rouge, LA. McEwan, L.C., and D.H. Hirth. 1979. Southern Bald Eagle productivity and nest-site selection. Journal of Wildlife Management 43(3):585–594. Michot, T.C. 1996. Marsh loss in coastal Louisiana: Implications for management of North American Anatidae. Gibier Faune Sauvage Game and Wildlife 13(2):941–957. Millsap, B., T. Breen, E. McConnell, T. Steffer, L. Phillips, N. Douglass, and S. Taylor. 2004. Comparative fecundity and survival of Bald Eagles fledged from suburban and rural natal areas in Florida. Journal of Wildlife Management 68(4):1018–1031. Mundahl, N.D., A.G. Bilyeu, and L. Maas. 2013. Bald Eagle nesting habitats in the Upper Mississippi River National Wildlife and Fish Refuge. Journal of Fish and Wildlife Management 4(2):362–376. Peterson, A. 1986. Habitat-suitability index models: Bald Eagle (breeding season). Biological Report 82(10.126):23. US Fish and Wildlife Service, Washington, DC. 25 pp. Saalfeld, S.T., and W.C. Conway. 2010. Local- and landscape-habitat selection of nesting Bald Eagles in East Texas. Southeastern Naturalist 9(4):731–742. Saalfeld, S.T., W.C. Conway, R. Maxey, C. Gregory, and B. Ortego. 2009. Recovery of nesting Bald Eagles in Texas. Southeastern Naturalist 8(1):83–92. SAS Institute Inc. 2011. SAS/STAT User’s Guide, 9.3 Edition. Cary, NC. Shaffer, G.P., W.B. Wood, S.S. Hoeppner, T.E. Perkins, J. Zoller, and D. Kandalepas. 2009. Degradation of Bald Cypress–Water Tupelo swamp to marsh and open water in Southeastern Louisiana, USA: An irreversible trajectory? Journal of Coastal Research Special Issue 54:152–165. Smith, N.R., T.J. Hess Jr., and A.D. Afton. 2016. History and nesting population of Bald Eagles in Louisiana. Southeastern Naturalist 15(1):12–25. Smith, N.R., A.D. Afton, and T.J. Hess Jr. In press. Winter breeding and summer nonbreeding home ranges of Bald Eagles from Louisiana. American Midland Naturalist. Thompson, C.M., and K. McGarigal. 2002. The influence of research scale on Bald Eagle habitat selection along the lower Hudson River, New York (USA). Landscape Ecology 17(6):569–586. US Fish and Wildlife Service (USFWS). 1987. Habitat-management guidelines for the Bald Eagle in the southeast region. US Fish and Wildlife Service, Atlanta, GA. 25 pp. Watts, B.D., G.D. Therres, and M.A. Byrd. 2008. Recovery of the Chesapeake Bay Bald Eagle nesting population. Journal of Wildlife Management 72(1):152–158. Wood, P.B., T.C. Edwards, and M.W. Collopy. 1989. Characteristics of Bald Eagle nesting habitat in Florida. Journal of Wildlife Management 53(2):441–449.