The decline of Nearctic-Neotropical migratory birds over the last two-decades remains one of the most pressing issues in conservation biology. The Department of Defense (DoD) manages 25 million acres of land and has the joint responsibility of managing these high-quality forest, desert, prairie, and other lands while concurrently engaging in military readiness activities that can alter the structure and function of those habitats. Given that forest loss and fragmentation threaten terrestrial biodiversity worldwide, it is critical for DoD to understand what species are affected by these processes.

In this project, researchers describe a multi-institutional collaboration (Smithsonian, Institute for Bird Populations, Oregon State University and US Army Engineer Research and Development Center) with two primary aims: 1) to validate and test the performance of three avian sampling approaches (point-counts, Monitoring Avian Productivity and Survivorship (MAPS) and intensive demography) to characterize source-sink dynamics, and 2) to determine which local and landscape factors influence distribution, demography, and overall population persistence on both DoD installations as well as in surrounding landscapes. 

Technical Approach

This research was carried out at Jefferson Proving Ground (now Big Oaks National Wildlife refuge), NSA Crane, and within surrounding Indiana state forests and parks. Given the limitations of collecting intensive demographic data for multiple species, this work was primarily focused on Wood thrush (Hylocichla mustelina), a DoD species of concern. Where possible, researchers used the larger spatial scale data from point-count sampling and MAPS to assess the factors that influence demographic performance and population persistence for specific species of concern and the broader community of birds.

Characterizing how populations are spatially structured by demographic processes across heterogeneous habitats (source-sink dynamics) remains an important monitoring tool. High quality, habitats typically yield a demographic surplus where births exceed deaths, and emigration exceeds immigration, whereas low quality, habitats generate demographic deficits where deaths exceed births and immigration exceeds emigration. To date, a variety of approaches that vary greatly in effort and data resolution have been used to determine population growth trajectories and persistence probabilities. This hierarchy of methodological approaches includes: 1) point-counts that provide coarse demographic data on distribution, abundance, and richness and can be implemented over large spatial-scales with low unit effort, 2) Monitoring Avian Productivity and Survivorship that provide higher resolution demographic data on some vital rates (e.g., survival), but are limited in scale largely because of unit effort and, 3) intensive demographic monitoring that can provide high resolution demographic data on most vital rates, but is also limited in spatial monitoring extent and requires the greatest unit effort. 


The report is organized into four chapters that address the projects broad objectives. Chapter one provides detailed results on the point-count and occupancy component of the research. This chapter first assesses the impact of temporary emigration on occupancy models’ ability to characterize population dynamics, and second, applies occupancy models to quantify the independent effects of habitat amount, fragmentation, and their interaction on dynamic avian distribution patterns. These results will provide the DoD useful information about how to sample and model dynamic distributions of unmarked animal communities, identify local and landscape features that have positive and negative effects on forest bird communities, and provide information about occurrence “hotspots” on and around military bases for avian species of concern. Chapter two provides detailed results for MAPS constant effort mist-net monitoring. This chapter characterizes two key vital rates (survival and fecundity) for modeling population growth, compares those rates to those derived from intensive demographic monitoring, and assesses the local and landscape predictors of Wood thrush productivity. These results provide essential information on the accuracy of previously collected vital rate data on DoD installations and identify the landscape features that drive regional Wood thrush productivity. These results also provide essential information about the spatial scale of inference for MAPS measures of productivity and how they should be applied to management decisions. Chapter three reports on the intensive demographic monitoring component of the research. This chapter details a full-annual cycle integrated population model that provides key vital rates (survival, fecundity, immigration, etc.) for both multiple age classes and across the breeding and non-breeding season. This chapter specifically addresses the role of immigration, and to a lesser extent emigration, in source-sink dynamics. Model results are used to understand the relative contribution that different vital rates make to population growth. In addition, the chapter also details the key drivers of habitat specific demography and how these variable demographic rates differentially contribute to population dynamics. These results provide the DoD an essential understanding of which vital rates drive population growth, how those processes vary with habitat quality, and details the role of breeding vs. non-breeding season population limitation. Chapter 4 is a synthesis of results from the first three chapters. This chapter specifically compares population growth estimates from the hierarchy of approaches (low resolution and effort counts to high resolution and effort intensive demography) to assess the accuracy and precision as well as the scale at which each approach can inform population dynamics. In addition, the researchers discuss the pros and cons of each methodological approach and their utility for assessing changes in population growth. Given that self-recruitment rate is the best indicator of population persistence, this chapter reviews which local and landscape features best predict the contribution of each local Wood thrush population. Finally, the researchers compare the results from the occupancy modeling and demographic monitoring to broadly discuss the relationships between distributional patterns and population dynamics for a suite of passerine species on DoD installations.


Using dynamic occupancy and species distribution models to understand species sensitivity to landscape configuration

Researchers developed species distribution models (SDMs) for 30 species. SDMs were constructed by modeling observed point count data as a function of explanatory variables extracted from Landsat 5 TM imagery using boosted regression trees (BRTs). The BRTs for 18 of these species performed well when tested against held-out validation data (AUC ≥ 0.7), and thus they used them to produce region-wide, spatially-explicit maps of occurrence probability. These maps will be useful to DoD for identifying biodiversity hotspots, or high probability use areas for species of interest in southern Indiana.  In the future, researchers intend to use these spatially-explicit models to refine the estimates of the effect of forest loss and fragmentation on bird distributions and dynamics.

Estimating vital rates of Wood thrush at Monitoring Avian Productivity and Survivorship (MAPS) Stations

The results of the MAPS monitoring provide important insight into Wood thrush population dynamics at both the local and regional scale. MAPS data as presented in the Final Report were able to provide vital demographic information about 65 species of forest birds many of which have shown sharp declines over the last decade. Although vital rates (productivity and survival) are only highlighted for the focal species of the grant, Wood thrush, future work will examine comparable analyses for a host of species to examine similarities/ differences in demographic performance on DoD installations.

Demography across the annual cycle: spatial and temporal variation in vital rates drives population dynamics in a migratory bird

These results underscore the importance of understanding how spatial and temporal variation in vital rates contribute to population dynamics. Previous empirical work largely identified sources and sinks based on single site studies that failed to account for how vital rates change across the annual cycle. The results highlight that migration may be a critical phase of the annual cycle for Wood thrush population dynamics.

Synthesizing what different methodological approaches tell us about avian demography on Department of Defense installations

The results definitively show that the point count and intensive demographic approaches the researchers demonstrated provide very similar estimates of λ. The MAPS protocol, on the other hand, produced estimates that were not significantly, and in some cases not positively, correlated with either the point count or demographic estimates. As mentioned, the researchers hypothesize that these results are indicative of similarities in spatial areas sampled by point counts and demographic monitoring, while the MAPS protocol samples an area that is geographically much larger. Validation of this hypothesis is beyond the scope of this study, but future work should investigate the source of this discrepancy.

As an addendum to the report, the researchers are providing DoD land managers with a user’s guide to begin thinking strategically about how to structure monitoring programs to meet specific management objectives.

  • Bird,

  • Habitat,