During fall migration, hundreds of millions of birds fly over North America on their way south. For many songbirds that nest in northern forests, choosing which patches of habitat to land in during migration – to rest and refuel – can have a critical influence on their health and energetics. Migration is incredibly energy-intensive, and stopover sites should ideally contain enough food resources to sustain migratory birds on their journeys.
However, decades of forest change have resulted in certain areas of forest having fewer arthropod resources than other areas. For birds making habitat selection decisions from the sky, it may not be possible to distinguish high-quality forest habitat from forest habitat that is lacking food resources.
Research
We were curious how songbirds were distributed across the forests of the southwestern Baraboo Hills during two life stages: the post-fledging phase when young birds have left the nest, and fall migration. We hypothesized that there would be more birds in sites with higher food resource availability.
To measure bird distribution, we used bioacoustic recorders to capture 24-hour soundscapes in three common forest types: upland oak woodlands (which are managed with prescribed fire and tree thinning), unmanaged uplands, and stream gorges. Additionally, we used malaise traps to measure aerial insect biomass available in the upland habitats.
We examined the bioacoustic data in two few ways, first by calculating an acoustic index or a mathematical summary of a soundscape. The index we chose is called Acoustic Complexity Index (ACI) and it was designed to capture bird activity. ACI patterns across our study area revealed that bird activity was clustered into three seasonal peaks that aligned with: post-fledging, early fall migration, and late fall migration. Additionally, bird activity was higher in managed upland woodlands early in the season, and higher in stream gorges late in the season.
Next, we investigated whether ACI was correlated with aerial insect biomass, which we measured in three upland sites across a gradient of management. We found that arthropod biomass and ACI were both higher in managed uplands than in unmanaged uplands.
This finding aligns with our research conducted during songbird nesting season – you can read about how managing for oak woodland ecosystems can enhance food resources for insectivorous birds, and result in more robust and complex soundscapes.
Finally, we went beyond examining an acoustic index, and used machine learning to identify the calls of individual bird species in our bioacoustic data. After validating these identifications, we modeled the probability of detecting common bird species in each forest habitat type across the season.
This figure tells a more complex story. The bird species we examined followed seasonally-dynamic habitat use patterns: managed uplands were used frequently during post-fledging while unmanaged uplands were used frequently during fall migration, and stream gorges were not frequently used. This underscores the importance of maintaining diverse and interconnected forest habitats, where a mosaic of managed and unmanaged areas supports the needs of forest birds throughout the annual-cycle.
The full research paper is available here, and was published in the Passenger Pigeon, the Wisconsin Society for Ornithology’s journal.
Baraboo Hills Ecology Research Collective 