How We Observe Complexity

Overview

Our goal is to understand how fracking activities interact with sensitive ecological and social systems. To do this, we combine spatial analysis, satellite observation, and machine learning to reveal how infrastructure expansion overlaps with key environmental and human factors — such as water sources, farmland, and communities.

Spatial Risk Indicators – Identifying Environmental Risk Through Proximity

We developed spatial indicators to evaluate the potential risks posed by fracking infrastructure. These indicators measure the distance between wells and nearby resources — such as rivers, irrigation canals, schools, or populated areas.

To build these indicators, we combined peer-reviewed research with local and remote sensing data. We mapped the locations of thousands of fracking wells and calculated how close each one is to rivers, irrigation canals, agricultural areas, schools, and towns. Wells within less than 1 km of these sensitive sites were classified as high-risk, while those 1–2 km away were considered medium-risk. This approach highlights where fracking activities are most likely to threaten water quality, farmland, or community health, helping decision-makers focus monitoring and protective measures where they are needed most.

Monitoring Well Expansion with Satellite Data

Tracking Fracking Infrastructure Over Time. To monitor the expansion of unconventional oil and gas activity, we analyzed high-resolution satellite imagery combined with digital elevation models (DEMs) to map and monitor new well locations.

• Source: PlanetScope satellite imagery
• 3-meter resolution
• 5 spectral bands
• Updated every 5 days

We used these images to detect changes in land use, map fracking infrastructure, and observe how activity spreads across basins.

Machine Learning for Predictive Mapping

Detecting Infrastructure Before It’s Built. We trained machine learning models to detect early signs of land preparation for new extraction wells. This enables predictive monitoring — helping us identify and analyze areas likely to experience future expansion, even before new wells are drilled.

Our model:

• Classifies land cover and use changes
• Flags location preparation patterns
• Enables early alerts for high-risk zones

Mapping Land Use and Ecosystem Pressure

With the same PlanetScope imagery, we produced land use and land cover maps to quantify:

• Cultivated areas
• Natural vegetation
• Changes in ecosystem cover over time

These maps help identify where ecosystems or agricultural lands may be at risk due to infrastructure overlap or expansion.

Why This Matters

This integrated methodology enables us to see what’s often missed — the systems-level interaction between fracking infrastructure, water, agriculture, and human populations.

By combining spatial risk modeling, satellite data, and predictive tools, we offer communities and decision-makers a clearer picture of complex
environmental dynamics — and the tools to act accordingly.