Health and environmental consequences of crop residue burning correlated with increasing crop yields midst India’s Green Revolution

The study investigates how India’s Green Revolution (GR)—which markedly increased production of staple crops such as wheat and rice—has influenced environmental pollution and human health via crop residue burning. While the GR alleviated food deficits and supported population growth, it also intensified practices like open-field burning of post-harvest residues and expanded the use of agrochemicals, contributing to environmental degradation. The research question is whether and to what extent GR-driven increases in crop yields led to higher emissions of benzo[a]pyrene (BaP), a carcinogenic polycyclic aromatic hydrocarbon, and elevated incremental lifetime cancer risk (ILCR) from inhalation across India, particularly in high-yield regions such as the Indo-Gangetic Plain (IGP). The purpose is to quantify historical BaP emissions and concentrations attributable to agricultural activities since the mid-1960s, evaluate associated lung cancer risks, and identify effective mitigation strategies. This is important because biomass burning is a dominant PAH source in India and may significantly impact public health in densely populated agricultural regions.

Prior work identifies biomass burning as a major source of PAHs in India, with the country ranked second globally for PAH emissions after China. Studies indicate that biogenic and biomass fuel use contributes the large majority of PAH emissions, with crop residue burning (both indoor and open-field) accounting for roughly 20–25% of total PAH emissions—exceeding contributions from industry and transportation. The IGP, particularly northwestern states (Punjab, Haryana), has been extensively studied for PAH contamination and source–receptor relationships, with residue burning during rice–wheat rotations implicated in poor air quality. Household biomass burning remains prevalent, especially in rural areas, and contributes substantially to indoor and ambient PAH burdens. Despite numerous studies on GR’s health implications and on PAH pollution episodes, gaps remain linking long-term, GR-induced agricultural changes to environmental BaP trends and health risks nationally over multiple decades.

Modeling: The study employs a modified regional version of the CanMETOP (Canadian Model for Environmental Transport of Organophosphorus Pesticides) framework adapted for persistent organic pollutants, configured over India and surrounding regions at 1/4° × 1/4° resolution. Simulations span 1960–2014 with 1960–1964 as spin-up; analyses focus on 1965–2014. Emissions and activity data: Agricultural BaP emissions (indoor and field crop residue burning, plus minor gas/diesel combustion from agricultural activities) are from the PKU-PAH inventory (1960–2014). Crop yields and residue estimates derive from FAO data. Meteorological inputs (wind, pressure, precipitation, terrain) are from NCEP reanalysis. Gridded yield profiles are from GRADS (1965–1980) and GDHY (1981–2014). MODIS MCD14A2 (8-day fire mask) and MCD12Q1 (land cover) products are used to characterize open fires on croplands and their seasonality. Scenarios: Two scenarios isolate GR effects: (1) NO-GR: fixed agricultural BaP emissions at 1960 levels applied throughout 1965–2014; (2) GR: time-varying monthly agricultural BaP emissions (1960–2014). Differences in modeled BaP concentrations and ILCR between GR and NO-GR quantify GR-attributable impacts. Health risk assessment: Incremental lifetime cancer risk (ILCR) from inhalation of BaP is computed per 1/4° grid using EPA risk assessment methodology. Parameter distributions include age- and sex-specific body weight, inhalation rates (derived from metabolic relationships), life expectancy, and BaP cancer slope factor (CSF=26.6 kg·day−1, log-SD=0.38). Monte Carlo simulations (n=10,000) yield probabilistic ILCR estimates. Population exposure fractions above thresholds (e.g., 10^-6, 10^-5 ILCR; 1 ng m−3 concentration standard) are computed. Population data and distributions follow national census assumptions (male:female = 1:1, 2011 census). Evaluation and uncertainty: Model performance is evaluated against observed BaP air concentrations (statistical error analyses in supplementary materials). Uncertainty sources include emission inventories (variation factors ~1.5–1.7 over 1965–2014), physicochemical parameters for BaP, and ILCR inputs (body weight, inhalation rates, CSF). A first-order error propagation approach is used due to computational constraints. Emission–risk regression: An empirical regression model log(ILCR) = a·log(E) + b links annual mean BaP ILCR to emissions by sector for the IGP (1965–2013) and is validated against 2014, showing high agreement (r=0.98, p<0.001).

- Crop production and residues: Total crop yield increased from 217.0 Mt (1965) to 283.1 Mt (2014). Crop residues burned increased from 29.7 Mt to 47.5 Mt over the same period. - Emissions: Agricultural BaP emissions rose from 106.5 t (95% CI: 67.0–169.0) in 1965 to 229.5 t (95% CI: 142.1–371.0) in 2014. BaP agricultural emissions correlate strongly with total crop yields (r=0.9). - Spatial patterns: High agricultural BaP emissions and trends align with high-yield states across the IGP (Punjab, Uttar Pradesh, Bihar). Maximum modeled BaP concentration from agricultural sources reached ~55 ng m−3, accounting for ~40% of mean BaP from all sources (~140 ng m−3). - Concentrations over time: Mean BaP from crop residue burning increased from 0.29 ng m−3 (95% CI: 0.13–0.65) in 1965 to 0.66 ng m−3 (95% CI: 0.29–1.52) by 1997, remained relatively flat until ~2008, then increased to 0.74 ng m−3 (95% CI: 0.35–1.55) by 2014. - Exceedances: From agricultural emissions alone, an estimated 64.99% (95% CI: 3.28–29.14%) of land area and 6.42% (95% CI: 2.33–44.57%) of population experienced BaP above the 1 ng m−3 national standard. The abstract notes 57% of India’s population exceeded 1 ng m−3 considering overall exposure. - ILCR levels and trends: Highest ILCRs from agricultural emissions are in Punjab and Meghalaya (~5–6 × 10^−5). National mean ILCR from agricultural BaP increased from 3.8 × 10^−7 (95% CI: 1.06 × 10^−7–1.37 × 10^−6) in 1965 to 9.66 × 10^−6 (95% CI: 2.63 × 10^−7–3.54 × 10^−6) in 2014, an average annual increase of ~2.59%. - Vulnerable population: Fraction with ILCR >10^−6 from agricultural sources rose from 2.56% (95% CI: 0.47–48.94%) in 1965 to 25.75% (95% CI: 23.68–31.75%) in 2014; for ILCR >10^−5 from 0.08% (95% CI: 0.01–0.11%) to 0.10% (95% CI: 0.03–0.58%). Agricultural sources contribute ~40–60% of total ILCR in NE and NW India and ~30% nationally. - GR attribution: ILCR differences between GR and NO-GR scenarios show positive ΔILCR concentrated in major crop-producing regions, indicating GR-driven enhancements in BaP risk. - Health linkage: State-mean ILCRs correlate with non-smoker lung cancer incidence rates (r^2=0.167; p<0.001). Over the IGP, annual ILCR agrees strongly with death rates from trachea, bronchus, and lung cancer (r=0.96, p<0.001) from 1990–2014. - Seasonality and drivers: MODIS fire masks reveal biannual burning peaks (May; Oct–Nov) coincident with wheat and rice residue burning; monsoon precipitation likely reduces outdoor emissions impacts in Jun–Sep.

The findings demonstrate that GR-driven increases in crop yields have been accompanied by rising agricultural BaP emissions, higher ambient BaP concentrations, and elevated inhalation cancer risks, especially across the IGP and parts of northern and northeastern India. The strong correlation between crop yield trends, agricultural BaP emissions, and ILCR patterns supports the hypothesis that intensified agricultural practices and residue burning under the GR materially contributed to BaP burdens and health risks. The spatial and temporal agreement between modeled ILCR and observed lung cancer metrics further implicates BaP as a non-trivial contributor to lung cancer risk among non-smokers in affected regions. Policy-relevant insights include the disproportionate impact of short, intense residue-burning seasons (post-harvest rice and wheat) and the potential for targeted interventions. Measures such as curbing open-field burning during peak seasons, improving residue management (e.g., mulching, soil incorporation), adopting technologies like Happy Seeder, and accelerating transitions from indoor solid fuels to cleaner energy (e.g., biogas, advanced biomass stoves) can substantially reduce BaP exposure. While broader socioeconomic changes (e.g., increased vehicle ownership) also affect PAH emissions, inventories indicate agricultural biomass burning remains a dominant BaP source, making agricultural interventions particularly impactful.

This study quantifies, for the first time over multiple decades, the linkage between India’s Green Revolution, agricultural residue burning, BaP emissions, ambient concentrations, and inhalation cancer risk. Agricultural BaP emissions and associated ILCR rose substantially from the mid-1960s to 2014, with the largest impacts in high-yield regions such as the IGP. Modeled ILCR aligns with lung cancer incidence and mortality patterns, underscoring public health implications. Targeted reductions in open-field residue burning during rice–wheat harvest seasons and improvements in indoor fuel use and residue management would effectively lower BaP contamination and risk. Future research should (i) refine emission inventories and sectoral attributions, (ii) integrate additional carcinogenic PAHs and co-pollutants, (iii) incorporate higher-resolution exposure and demographic data, (iv) evaluate the effectiveness and adoption barriers of mitigation technologies and policies, and (v) strengthen epidemiological linkages with long-term health outcomes.

- Emission inventory uncertainties (PKU-PAH variability ~1.5–1.7 across decades) and potential sectoral misattribution. - Physicochemical parameter uncertainties for BaP and model structural limitations affecting concentration predictions. - ILCR input assumptions (body weight, inhalation rates, CSF) and demographic simplifications (e.g., sex ratio, life expectancy) introduce uncertainty despite Monte Carlo analysis. - Use of first-order error propagation rather than full sensitivity/uncertainty analysis due to computational constraints. - Limited availability and spatial coverage of long-term observed BaP concentrations and health data, especially in the IGP, constrains validation and causal inference. - Scenario simplifications (e.g., fixed 1960 emissions for NO-GR) may not capture all non-GR temporal drivers (policy changes, technology adoption, socioeconomic trends).