‘Fire count gives a skewed picture of stubble burning’

Context

• Govt claims: 92% reduction in fire counts in Punjab and 90% reduction in Haryana (since 2021).

• New study by iForest questions this, arguing fire counts undercount actual stubble burning.

• Alternative metric — Burnt Area — shows only ~30% reduction.

• Highlights risk of policy misinterpretation due to dependence on single-source satellite data.

Key Issue

Fire Counts ≠ Actual Stubble Burning

• Fire counts = number of active fires detected by satellites.

• Burnt area = actual land surface that shows evidence of burning.

iForest Findings

• Burnt area in Punjab + Haryana:

  • 31,500 sq km (2022) → 19,700 sq km (2025)

  • ≈30% decline, not 90%+.

This mismatch suggests structural under-detection and flawed monitoring.

Why Fire Counts Are Misleading

a) Limited detection by Polar-Orbiting Satellites

India uses data from:

• MODIS (NASA’s Terra & Aqua)

• VIIRS (Suomi-NPP)

Limitations

• Observe India only during fixed time window: 10:30 a.m. – 1:30 p.m.

• Can detect only active flames at that moment → not 24-hour coverage.

• Miss small fires and short-duration fires.

• Seasonal haze or clouds further reduce detection

b) Time-of-Day Bias

• iForest analysis (using Meteosat SEVIRI):

  • Most stubble fires since 2022 occur in the evenings.

  • This is outside MODIS/VIIRS overpass times.

• Result → systematic undercounting of fires.

c) Resolution Issues

• MODIS: 1 km resolution → misses small field fires.

• VIIRS: 375 m resolution → better but still inadequate for scattered small fires.

What is ‘Burnt Area’? Why is it Better?

Burnt Area = Land surface that shows spectral signature of past burning

—not just active flames.

iForest used:

• Sentinel-2 (MSI instrument)

  • 100m × 100m resolution

  • Detects burnt residue, char, ash

  • Provides a more comprehensive picture

  • Passes over every 5 days; data available with 8–15 day lag

Despite limitations, it captures:

• Total area burned

• Fires missed by active detection sensors

Hence a more robust metric, especially for policy evaluation.

Role of Geostationary Satellites (Meteosat 8 & 9)

Advantages:

• Continuous monitoring (every 15 minutes)

• Detects fire timing patterns (e.g., evening peaks)

Limitations:

• Not suitable for precise fire count

• Lower spatial resolution → not ideal for small fires

Key Insight from SEVIRI:

• Majority of post-2022 fires occur after sunset
  → not captured by MODIS/VIIRS.

Implications for Air Quality & Policy

a) Undercount = Underestimated Emissions

Missed fires:

• Lead to underreported PM2.5, NOx, SOx emissions

• Distort air pollution models for Delhi NCR

• Affect emergency response & long-term mitigation planning

b) Policy Success May Be Overstated

• Govt measures (e.g., DSR promotion, ex-situ residue management, Happy Seeder) may appear more successful than they are.

c) Budget & incentive misallocation

Wrong data →
Wrong hotspots →
Ineffective targeting of:

• Machinery subsidies

• Enforcement drives

• Crop residue management funds

Broader Governance Problem

Over-reliance on a single metric

Using fire counts alone → Unidimensional indicator → Risk of data-driven misgovernance

Need for integrated monitoring

Combine:

• Fire counts (MODIS/VIIRS)

• Burnt area (Sentinel-2)

• Temporal monitoring (SEVIRI)

• Ground surveys

A multi-sensor approach ensures reliability.

Conclusion

• Fire counts provide a partial and skewed picture.

• Burnt area estimation, despite lags, gives a truer reflection of ground realities.

• For effective policy and accurate pollution assessment, India needs:

  • Multi-sensor satellite integration

  • Transparent data sharing

  • Ground-level validation

  • Policy recalibration based on actual burnt area

Related Posts

Rediscovering Assam’s Unsung Hero on Stage

Posted on September 2, 2023

Erdogan’s Neo-Ottoman Foreign Policy

Posted on June 11, 2025

Stalemate Over Sharing Of Krishna Water

Posted on May 23, 2023