Glass Building Bird Collision Risk Maps

ISEF Category: Earth and Environmental Sciences

Subcategory: Environmental Effects on Ecosystems  ·  Difficulty: Advanced  ·  Setup: Home Setup  ·  Time: Full Year

The Hook

Glass can turn a city tower into a deadly mirror for birds. A bird sees sky, trees, or open space, then flies straight into a wall it cannot read. You can study where that risk is highest without owning a field site. Public collision reports and satellite data can help you build a real risk map.

What Is It?

An ecological trap is a place that looks safe or useful to an animal, but ends up harming it. For migratory birds, glass buildings can act like that trap because reflections hide the real barrier. A bird may read a mirrored tree line or open sky as a path forward, then collide with the glass.

Your project asks a geographic question, not a lab chemistry question. You combine reported bird collision locations with satellite-based measures of urban form, then ask whether certain areas line up with higher risk. Think of it like predicting potholes on a road map, except the road is bird flight and the hazard is transparent glass.

Why This Is a Good Topic

This makes a strong science fair topic because you can test a clear pattern with real data, not just opinion. It connects to bird conservation, urban planning, and window safety in a way people can act on. You can learn geospatial analysis, data cleaning, and risk mapping, all with public datasets and free software.

Research Questions

• How does the density of reported bird collisions change with urban glass-facade indices?
• What is the effect of nearby green space on collision counts near glass buildings?
• Does building height change the relationship between glass exposure and bird collision risk?
• To what extent do collision hotspots overlap with migratory flyway corridors?
• Which satellite-derived land-cover features best predict reported bird collisions?
• How does collision risk differ between downtown cores and mixed-use neighborhoods?
• What is the effect of seasonal migration timing on the spatial clustering of collision reports?

Basic Materials

• Laptop with internet access.
• Spreadsheet software such as Google Sheets or Excel.
• Free GIS software such as QGIS.
• Public bird collision reports from city agencies, nonprofits, or open datasets.
• Sentinel-2 imagery from NASA or the Copernicus Data Space ecosystem.
• Land cover and vegetation layers from NOAA, USGS, or local open-data portals.
• Digital notebook for tracking data sources, code, and map versions.

Advanced Materials

• Laptop or desktop with strong memory for GIS layers.
• QGIS or ArcGIS Pro if your school provides access.
• Python with GeoPandas, Pandas, NumPy, Matplotlib, and Rasterio.
• ImageJ for measuring facade reflectance proxies from image sets when appropriate.
• High-resolution aerial imagery from city open-data portals or USGS EarthExplorer.
• Building footprint datasets from municipal GIS portals or OpenStreetMap exports.
• Bird migration timing data from eBird, USGS, or NOAA-linked sources.
• Statistical software such as R for spatial regression and model diagnostics.

Software & Tools

• QGIS: Builds maps, overlays collision points, and compares urban features across neighborhoods.
• Google Earth Engine: Screens large satellite image sets and derives land-cover proxies at scale.
• Python: Cleans datasets, joins spatial layers, and runs reproducible analysis.
• R: Fits spatial statistics and checks whether your risk model holds up.
• ImageJ: Measures brightness, contrast, and image-based facade proxies when you use imagery.

Experiment Steps

1. Define the collision outcome you will measure, such as total reports, hotspot density, or collision rate per area.
2. Choose the spatial unit you will compare, such as blocks, census tracts, or grid cells.
3. Build your predictor set from satellite, land-cover, and building-related layers that can act as glass exposure proxies.
4. Clean the public collision data so duplicated reports, missing coordinates, and time mismatches do not distort your map.
5. Test which environmental features separate high-risk areas from low-risk areas, then check for spatial clustering.
6. Translate the results into a risk map and explain where your model is strong, where it is uncertain, and why.

Common Pitfalls

• Mixing report counts with true collision rates, which makes dense neighborhoods look riskier just because more people filed reports.
• Using satellite proxies that track general urban density but do not actually capture glass exposure.
• Failing to align collision dates with migration season, which blurs the effect you want to measure.
• Ignoring spatial autocorrelation, which can make nearby areas look like independent evidence when they are not.
• Treating public collision reports as complete records, which can hide reporting bias near schools, parks, or wealthy districts.

What Makes This Competitive

A strong project does more than draw a map. It tests whether your glass proxy really predicts bird collisions after you control for land cover, building density, and reporting bias. You can also compare several modeling approaches and show which one holds up best on unseen locations. If you add uncertainty maps, seasonal effects, or a validation step against a second city, your work looks much closer to real environmental research.

Project Variations

• Compare collision risk near office towers, residential towers, and mixed-use buildings to see whether building type matters.
• Swap reported collisions for citizen-science observations of bird strikes and test whether the pattern changes.
• Focus on one migratory season or one city district, then build a finer-scale hotspot map with stronger local detail.

Learn More

• USGS Bird Collision Studies: Search the USGS website for bird-window collision research, mapping methods, and conservation summaries.
• NOAA Climate Data and Mapping Resources: Search NOAA for environmental layers that help explain seasonal bird movement and habitat patterns.
• NASA Earthdata: Find Sentinel-2 and related remote sensing data through NASA Earthdata Search.
• PubMed: Search for review articles on bird-window collisions, urban ecology, and spatial risk modeling.
• Urban Ecology journal: Read peer-reviewed studies on how cities affect wildlife, usually through your school library or journal search tools.
• MIT OpenCourseWare, Intro to GIS: Use free course materials to learn mapping, spatial layers, and geographic analysis basics.