Abstract:
My research group is dedicated to tackling the question “how do we find room for biodiversity in an increasingly hot, hungry, and crowded world?” In this talk, I will discuss three projects we have been working on where technological innovations could greatly enhance our ability to conserve wildlife. The first case study centers around migratory birds, which have been disappearing around the world. Determining what these birds require in the way of rest-and-refueling sites during their migrations remains a big challenge, but we are gaining valuable insights using weather radar images, telemetry, and other approaches. However, we could learn a lot more with better image processing and telemetry. The second case study focuses on the movement of people from rural areas to cities. Will this outmigration result in the abandonment of millions of acres of marginal cropland that could then be restored as valuable wildlife habitat? The answer may hinge on our ability to measure landcover at finer scales than is currently feasible. Finally, I will discuss some of our earlier work aimed at understanding how logging affects biodiversity in the tropical forests of Southeast Asia and how advances in sound recording could make this type of research faster, easier, and more reliable.
Bio:
David S. Wilcove is a professor of ecology, evolutionary biology, and public affairs at Princeton University. The primary question driving his research is “How do we find room for biodiversity in an increasingly hot, hungry, and crowded world?” Accordingly, his work focuses on the impacts of farming, logging, hunting, climate change, and other human activities on biodiversity. He and his graduate students and postdocs have studied these issues around the world, from New Zealand to the Amazon Basin and from Siberia to Borneo. Prior to joining Princeton’s faculty in 2001, he worked as a scientist for the Environmental Defense Fund, The Wilderness Society, and The Nature Conservancy. He is the author of two books—No Way Home: The Decline of the World’s Great Animal Migrations (2007) and The Condor’s Shadow: The Loss and Recovery of Wildlife in America (1999)—and many technical and popular articles in the fields of conservation biology, ornithology, and wildlife conservation. Professor Wilcove serves or has served on the boards of directors of the Doris Duke Charitable Foundation, Society for Conservation Biology, Rare, American Bird Conservancy, Natural Areas Association, Galapagos Conservancy, and New Jersey Audubon Society, among other organizations. He has received awards from the Society for Conservation Biology, Defenders of Wildlife, the Pew Foundation, The Nature Conservancy, The Wildlife Society, and the Hauptman-Woodward Medical Research Institute. He received his undergraduate degree from Yale University and his doctorate from Princeton University.
Summary:
Question: how do we find room for biodiversity in an increasingly hot, hungry and crowded world?
Migratory birds
Challenge: we’ve lost many migratory birds over the past several decades
A single bird passes over many countries, states and towns: major jurisdictional challenges
Most management approaches focus on breeding and wintering grounds because that’s where they spend most time
What about the regions they migrate through? Much harder because they spend less time there but also critical to ensure they survive their migrations.
Observation: many songbirds migrate nocturnally
There are so many of them that they can be picked up on weather sensors
Look like clouds to radar
This makes it possible to map migration density
Migration is not very concentrated in space
Birds move in a broad front
Prefer deciduous forest for stopover locations (top 10% of pixels where migrations are observed)
Even in mid-western cornbelt
Very challenging to clean up the data: remove nocturnal bats, insects and non-migratory birds, weather events (e.g. precipitation)
This is currently done manually
If this were automated, it would enable researchers to leverage decades of radar data
For all species: birds, insects and bats
Radar cannot detect bird species but can tell abundances
On-ground humans can detect species but not abundances
Challenge: migration stop-overs are in forests but they are not all the same
The forest has many different levels of the canopy
Different tree types are differently appropriate for birds
LIDAR surveys can help map out the 3D structure of forests most appropriate for birds
Migration via shores
These birds tend to congregate at a few sites
Suspicion: these sites are unusually rich in food
However, these tidal flats are being rapidly developed, are disappearing at 1.2%/year since 1980s (Yellow sea area)
But birds have been declining at a faster rate
Many suspect that the disappearance is caused by decline of tidal flats, at least partially
Birds feeding behavior:
Generalist (can eat in many locations): spend most time closest to the sea wall
Specialist (can eat in fewer locations): spend most time closest to the sea wall
Tide follower: eating area is nearest to water line, moves with tide
Development is mostly focused at the sea wall, which disrupts generalists and specialists
Would be very valuable to have a global map of tidal mudflats
Challenging because they’re exposed for only parts of the day
Need elevation information (measured in meters)
Would enable time series studies to quantify tidal flats losses
Implications for conservation and fisheries (clams and bivalves live in tidal flats)
Global warming and sea level rise will shrink tidal flats
Without development they would move inland
But near-coastal development puts a barrier on the movement of tidal flats
Additional wishlist:
Would love to get photos of birds
Infer their species
Count birds
https://play.google.com/store/apps/details?id=com.labs.merlinbirdid.app&hl=en_GB&gl=US
Rural emigration and crop abandonment
2003-2019
217 Mha cropland expansion
49% created by converting ecosystems
Also, 78.5 Mha cropland abandoned, largely because of urbanization
Abandoned cropland grows through a succession process: grass -> shrubs -> sparse forest -> dense, multi-story forest
Does abandoned cropland stay abandoned?
Have create high-resolution maps of abandonment
Used 30 year 30m resolution satellite images
Observation: abandoned farmland is cultivated within a few decades
14 years before recultivation on average
50% of all abandoned cropland recultivated within 30 years
Need more detailed land cover analysis tools to understand how abandoned cropland evolves over time
Logging tropical forests
Focus on Sundaland: Malaysia/Indonesia
Massive deforestation: 70% of forest converted to cropland, e.g. oil palm plantations
Oil palm plantations are poor habitats (no better than corn fields)
Logging is different: mostly the large trees are removed but the smaller ones are left
Looked at populations of birds and beetles (biodiversity) in logged forests
Unlogged, once-logged (60cm diameter trees removed), twice-logged (40cm diameter trees removed)
Used
Birds: mist nets (capture birds) and point counts (birdwatcher observations)
Dung beetles: baited traps
Discovery: > 75% of species persist in twice logged forests
Good candidates for conservation
Challenge: identifying birds is very hard
Is mostly done by sound because canopy helps them hide
Bird counters need to be extremely skilled
Dataset is incomplete
Would be great to
Collect audio data automatically to get better coverage
Analyze it automatically to get complete coverage from the many audio recordings
Species
Count of individuals
Use drones to get recordings from very remote places
Current market prices for carbon, etc. are far below value of timber