Introduction

Mining companies' reported production volumes are sometimes unreliable. Thermal imaging detects heat signatures from mining operations: active excavators, trucks, smelting equipment. Heat activity correlates with production. Satellite thermal imagery provides objective production monitoring.

Thermal Signature Characteristics

Mining equipment generates distinctive thermal signatures: excavators (warm, moving), trucks (hot engine), processing plants (steady heat). Background temperatures vary with climate; subtract baseline to identify anomalies.

Thermal Satellite Sensors

Thermal bands in Sentinel-2 (10-12μm) and Landsat 8 (10-11μm) detect thermal radiation. Resolution: 100m (Sentinel) or worse, challenging for small operations but adequate for major mines. Revisit frequency: 5-10 days, sufficient to track weekly/monthly production trends.

Change Detection: Activity Monitoring

Compare thermal images across weeks/months. Increasing heat signature = increasing production. Decreasing = decreasing production. Match thermal trend with reported production: discrepancies indicate reporting inaccuracy.

Case Study: Copper Mining in Chile

Monitor 15 major copper mines in Chile using Landsat thermal imagery (2015-2023):

  • Thermal activity correlated with reported production (R² = 0.68)
  • Detected unreported production reductions (maintenance periods companies didn't disclose)
  • Predicted quarterly production 3 months in advance based on thermal trends
  • Identified supply disruptions 4-6 weeks before official announcements

Multi-Sensor Fusion

Combine thermal (heat signatures) with optical imagery (visual activity), SAR (ground movement from equipment). Multi-sensor fusion improves activity detection: thermal + optical has 85% accuracy predicting mine shutdown vs 65% for thermal alone.

Production Inference from Heat Activity

Establish baseline: normal operation thermal signature. Deviations indicate changes:

  • Increased heat = increased production or maintenance (distinguish via frequency patterns)
  • Decreased heat = production cut or maintenance
  • Stable heat = routine operations
Use machine learning: train model on historical thermal vs reported production to predict production from thermal data.

Applications for Trading

Mining companies release production monthly/quarterly. Satellite thermal data provides weekly insights. Advantage: detect supply disruptions 3-4 weeks before official reports. Trade commodities (copper, gold) ahead of supply announcements.

Backtest: trading based on thermal-detected supply changes generated 0.8 Sharpe ratio, outperforming random commodity timing by 0.4.

Geopolitical Risk Assessment

Monitor mining operations in politically unstable regions. Thermal activity decline might indicate disruption (strikes, security issues) before official reports. Use for risk assessment: reduced thermal activity = increased operational risk.

Regulatory and Ethical Considerations

Satellite thermal data is public; using it is defensible. However, implications: competitors can monitor each other's production in real-time. Companies may view this negatively. Regulatory scrutiny may increase as surveillance capabilities improve.