Nigeria Gas Flare Detection & Emissions Analysis Using VIIRS Satellite Data (2012–2024)

Abstract

Gas flaring remains one of the most visible and persistent environmental failures in Nigeria’s oil and gas sector. Nigeria is responsible for approximately 10% of global gas flaring, concentrated in the Niger Delta — a region spanning Delta, Rivers, Bayelsa, Imo, and Akwa Ibom States. Between 2012 and 2024, Nigerian operators flared 96.46 billion cubic metres (BCM) of gas, releasing an estimated 270.1 million tonnes of CO&sub2; equivalent into the atmosphere — a profound environmental and economic loss. This study draws on field-level annual data from the World Bank Global Gas Flaring Reduction Partnership (GGFR), derived from VIIRS (Visible Infrared Imaging Radiometer Suite) satellite night-fire observations, covering 57 operators across 13 years.

This project applies exploratory data analysis, geospatial mapping, operator benchmarking, and CO&sub2; estimation to characterise the structure and trajectory of Nigeria’s flaring problem. Particular analytical focus is given to Seplat Energy’s Flares Out programme — an explicit commitment to eliminate routine gas flaring across its operations in OMLs 4, 38, 40, and 41 — and to the strategic implications of the 2024 acquisition of MPNU (now SEPNU), formerly ExxonMobil’s Nigerian shallow-water subsidiary. A Random Forest regression model is deployed to identify the structural drivers of flaring intensity, providing a machine learning demonstration directly relevant to the analytical and technical competencies required by the 2026 Seplat Energy Technical Graduate Trainee Programme.

Dataset Variables

1. National Flaring Trends

Nigeria flared 9.62 BCM in 2012, declining 45% to 5.32 BCM by 2022 — driven by government regulation under the Associated Gas Framework Agreement, the NNPC/Seplat joint venture’s gas commercialisation investments, and increasing domestic gas utilisation for power generation. The 2017 spike reflects well-documented Niger Delta pipeline disruptions that curtailed gas evacuation infrastructure, forcing increased on-site flaring. However, 2023 and 2024 show an uptick to 6.48 BCM, signalling that structural barriers to full gas utilisation remain — particularly as the newly acquired SEPNU offshore assets bring additional flaring volumes under Nigerian operator management. These data provide an independent, satellite-derived verification of sector performance that is not reliant on self-reported operator disclosures.

Figure 1. Nigeria total annual gas flaring volume (BCM, left axis, red bars) and CO&sub2; equivalent emissions (million tonnes, right axis, blue line) from 2012 to 2024. The dashed vertical line marks 2022, the historical low point. The 2023–2024 uptick is annotated. Data source: World Bank GGFR / VIIRS satellite observations.

Operator-level analysis reveals a highly concentrated market structure. ExxonMobil (16.53 BCM cumulative) was by far the largest single flaring entity in Nigeria over the 2012–2024 period, followed by Eni (12.94 BCM), Chevron (10.08 BCM), and Shell (8.91 BCM). These four international oil companies (IOCs) account for 50.1% of total flaring over the period. The December 2024 completion of Seplat Energy’s acquisition of MPNU (renamed SEPNU — Seplat Energy Production Nigeria Unlimited) fundamentally changes this picture: Seplat now assumes responsibility for the assets previously operated by the single largest flaring entity in the dataset. Historically, Seplat itself ranks 8th with a cumulative 3.48 BCM (for the SEPLAT Petroleum Development entity), underscoring both the scale of the integration challenge and the opportunity for the Flares Out programme to drive a step-change reduction.

Figure 2. Top 10 Nigerian gas flaring operators by cumulative BCM, 2012–2024. Seplat Energy is highlighted in green. The annotation marks the strategic significance of the December 2024 MPNU acquisition (now SEPNU). ExxonMobil’s 16.53 BCM cumulative figure represents the flaring liability now integrated into Seplat’s portfolio.

The distribution of flaring by intensity level and operational environment reveals important structural features. Medium-classification flares dominate by volume, accounting for the majority of Nigeria’s flared gas — a pattern that reflects the prevalence of chronic, ongoing associated gas flaring at producing fields, rather than a small number of large discrete events. This has direct implications for abatement strategy: addressing medium flares through gas utilisation infrastructure (pipelines, compression, processing) yields greater cumulative reductions than eliminating episodic large flares alone. Onshore operations account for the majority of flaring by volume, though offshore flaring is proportionally significant given the smaller number of offshore fields.

Figure 3. Left: distribution of total flaring volume (BCM) by VIIRS intensity classification (Small / Medium / Large) across 2012–2024. Right: onshore versus offshore total flaring volume with percentage share. Medium-intensity flares represent the dominant category by volume, underscoring the importance of chronic flare abatement over episodic interventions.

2. Seplat Energy — Flares Out Programme

Seplat Energy Plc, Nigeria’s leading indigenous energy company listed on the Nigerian Exchange (NGX: SEPLAT) and London Stock Exchange (LSE: SEPL), has committed to eliminating routine gas flaring across its onshore operations through the Flares Out programme. The satellite data provides independent, third-party verification of this commitment, disaggregated at the individual field level — a degree of transparency that self-reported corporate disclosures alone cannot offer.

Seplat’s peak flaring year was 2013, at 0.498 BCM. By 2024, this had declined to 0.252 BCM — a 49% absolute reduction. This decline outpaces Nigeria’s national average (45% over the same period), reflecting targeted investments including the Sapele Integrated Gas Plant (SIGP), commissioned in 2024, which diverts previously flared gas to domestic power generation. The ANOH Gas Plant (targeting 300 MMscfd, having achieved first gas in 2025 under the Assa North-Ohaji South JV between Seplat and NNPC’s subsidiary) will further accelerate flaring reduction by commercialising associated gas from OML 53. A notable rebound is visible between 2018–2020, which corresponds to production ramp-up at new fields prior to full gas infrastructure commissioning — a transient increase that is consistent with the expansion phase of the Flares Out programme timeline.

Figure 4. Seplat Energy annual flaring (BCM, right axis, green) versus Nigeria national total (BCM, left axis, red), 2012–2024. The shaded green area under the Seplat line emphasises the volume trajectory. The 2013 peak reference line and the 49% reduction annotation provide context for the Flares Out programme’s progress. The light-green background shading marks the 2022–2024 Flares Out acceleration period.

2.1 Field-Level Analysis

Sapele (37.3%, 1.733 BCM) and Oben (35.1%, 1.630 BCM) fields together account for 72.4% of Seplat’s historical flaring, concentrated in OML 38 and OML 4 respectively. Robertkiri (OML 40) contributes a further 14.4% (0.667 BCM). These three fields are the primary targets for Seplat’s ongoing flaring reduction investments and collectively account for 86.8% of cumulative flaring. The remaining fields — Idama, Amukpe, Inda, and Ubima — contribute minimally and are likely already approaching near-zero-routine-flaring status. The stacked area chart for the top three fields illustrates how Sapele and Oben have historically dominated the portfolio, with both showing measurable downward trends in the 2022–2024 period as the SIGP and gas evacuation infrastructure investments take effect.

Figure 5. Left: Seplat Energy fields ranked by cumulative flaring (BCM), 2012–2024, with a green gradient from highest to lowest. Right: stacked area chart of the top three Seplat fields (Sapele, Oben, Robertkiri) showing annual BCM contributions, illustrating the dominant role of Sapele and Oben within the Flares Out programme target portfolio.

3. CO&sub2; Emissions & Climate Impact

Gas flaring in Nigeria released an estimated 270.1 million tonnes of CO&sub2; equivalent between 2012 and 2024 — a figure equivalent to the annual greenhouse gas emissions of approximately 58 million passenger vehicles. This represents a substantial and quantifiable contribution to atmospheric forcing from a single country’s oil sector. Beyond the climate cost, this volume represents enormous untapped commercial value: at prevailing LNG spot prices during the period, the 96.46 BCM flared over 13 years could conservatively have generated approximately $9–12 billion USD in revenue for Nigerian gas commercialisation channels, including domestic gas supply for power generation, LPG production, and export-grade LNG. Seplat’s Flares Out programme is therefore simultaneously an environmental commitment and a commercial strategy — converting waste gas into revenue and contributing to Nigeria’s Decade of Gas initiative.

Seplat’s CO&sub2; trajectory closely mirrors its flaring volume trend, peaking in 2013 at approximately 1,395 thousand tonnes and declining to approximately 707 thousand tonnes by 2024. The downward trend line fitted to Seplat’s annual CO&sub2; data confirms a statistically consistent reduction trajectory — an important signal for ESG investors on the London Stock Exchange and for the company’s alignment with net-zero transition pathways.

Figure 6. Left: Nigeria’s annual CO&sub2;-equivalent emissions from gas flaring (million tonnes), 2012–2024, with a cumulative total annotation. Right: Seplat Energy’s annual CO&sub2; trajectory (thousand tonnes) with a linear trend line, demonstrating a consistent downward trajectory consistent with the Flares Out programme objectives.

4. Geospatial Distribution

The Niger Delta accounts for virtually all of Nigeria’s gas flaring, concentrated in a roughly 400 km coastal zone spanning Delta, Rivers, Bayelsa, Imo, and Akwa Ibom States. The 2024 satellite map reveals that Seplat’s operational fields (Sapele, Oben, Robertkiri — within OMLs 4, 38, 40, and 41) cluster in the central-northern Niger Delta onshore zone, while major offshore flaring from ExxonMobil (now SEPNU), Chevron, and Shell occurs in the coastal and shallow-water zones. The spatial concentration of flare sites in specific geographic clusters reflects the structural relationship between oil field locations, pipeline infrastructure, and gas utilisation capacity — a relationship that the machine learning analysis in Section 5 quantifies with precision. With the SEPNU acquisition, Seplat now carries flaring responsibilities across both onshore and offshore environments for the first time, requiring an extension of the Flares Out methodology to shallow-water production operations.

Figure 7. Geospatial distribution of all active gas flare sites in Nigeria’s Niger Delta for 2024, as detected by VIIRS satellite observations. Seplat Energy sites are marked with green stars; large, medium, and small flares by other operators are shown as red, orange, and yellow circles scaled by intensity. Seplat field labels are annotated directly on the map. The dashed rectangle outlines the core Niger Delta flaring zone. Coordinates in decimal degrees (WGS-84).

5. Machine Learning — Predictive Modelling

A Random Forest Regression model was trained on the full dataset (n = 2,264 observations) to predict individual flare site annual volume (BCM) from geographic and operational features. The model provides quantitative insight into the structural drivers of flaring intensity and demonstrates the potential for supervised machine learning to support operational decision-making, regulatory monitoring, and investment prioritisation in the oil and gas sector.

The most significant finding from the feature importance analysis is that geographic coordinates (latitude and longitude) together account for 86.3% of predictive importance — latitude contributing 43.15% and longitude contributing 43.19%. This confirms that gas flaring in Nigeria is not randomly distributed but is structurally concentrated in specific spatial clusters within the Niger Delta, reflecting the underlying geology, infrastructure, and operational history of the region. Year contributes 12.3% of importance, capturing the declining trend in flaring over the 2012–2024 period. Field type and location type (onshore/offshore) contribute less than 1.5% combined, suggesting that which company operates a field, or whether it is on- or offshore, matters far less to predicted flare volume than where that field is located geographically.

Figure 8. Left: Random Forest feature importance scores (%), confirming that geographic coordinates (latitude + longitude) dominate at 86.3%, with year contributing 12.3%. Right: scatter plot of actual versus predicted BCM values on the 20% held-out test set (n = 453), with the perfect-prediction diagonal shown in red. The model achieves R² = 0.7675 and MAE = 0.01548 BCM.

6. Key Findings & Industry Implications

7. Methodology

• Data Source World Bank Global Gas Flaring Reduction Partnership (GGFR), VIIRS satellite-derived field-level estimates, 2012–2024. Covers all identified flare sites in Nigeria with annual BCM volumes, operator attribution, and geographic coordinates. Resolution: individual flare sites; temporal: annual aggregates.
• CO&sub2; Estimation 1 BCM of flared gas ≈ 2.8 million tonnes CO&sub2; equivalent, applying the IPCC Sixth Assessment Report (AR6) conversion factor for associated gas combustion (assuming predominantly methane with characteristic NGL composition). This factor accounts for incomplete combustion and is consistent with World Bank GGFR methodology.
• Machine Learning scikit-learn RandomForestRegressor (200 estimators, max_features='sqrt', random_state=42). Categorical features (Location, Field Type) label-encoded prior to training. Evaluation metrics: coefficient of determination (R²) and mean absolute error (MAE) computed on a held-out 20% test set (random split, random_state=42).
• Visualisation All figures generated with matplotlib (version 3.x) at 150 DPI for high-resolution web display. Geospatial scatter plot uses decimal-degree coordinates with manually defined Niger Delta extent bounds (lon: 4.0–8.8, lat: 3.2–6.8). No third-party chart libraries are used in the web page itself.
• Limitations VIIRS cloud cover during the Niger Delta wet season (May–October) may reduce satellite detection sensitivity, potentially underestimating flaring during peak production months. World Bank GGFR volume estimates carry an inherent uncertainty of approximately ±9.5% (Elvidge et al., 2016). Company attribution is based on field operator records and may not fully reflect complex joint venture arrangements or operatorship changes during the period.

Citation

@misc{alagbe2026nigeriaflare,
  author       = {Alagbe, Olalekan Joshua},
  title        = {Nigeria Gas Flare Detection and Emissions Analysis
                  Using VIIRS Satellite Data (2012--2024)},
  year         = {2026},
  howpublished = {GitHub Pages,
                  \url{https://OlalekanAlagbe.github.io/nigeria-flare-analysis/}},
  note         = {Data Science Portfolio Project; aligned with
                  Seplat Energy 2026 Technical Graduate Trainee Programme.
                  Data: World Bank GGFR / VIIRS satellite observations.}
}