A Study on the Increase in Measured Methane Concentration Values During the 2024 Noto Peninsula Earthquake

Abstract

This study aims to demonstrate the presence of a pronounced coseismic increase in atmospheric methane concentrations during the 2024 Noto Peninsula Earthquake and to examine whether this increase may have originated from underground natural gas release. By analyzing hourly CH₄ data from the Ministry of the Environment’s monitoring network, this study shows that significant methane increases occurred only in areas with seismic intensity of 6– or greater, and that an exceptional anomaly—reaching 29 times the standard deviation of the past year—was recorded at the Nanao station. The validity of this anomaly was confirmed through consultation with local atmospheric officer, and high-time-resolution data (6 min values) were provided, verifying continuous instrument operation. Detailed analysis further shows that two major methane peaks occurred, each rising not immediately after the main shock but synchronously with two large aftershocks approximately 8 and 44 min later. Geological and hydrogeological information indicates the presence of water-soluble gas and unsaturated hydrocarbons beneath the Nanao region, suggesting that seismic shaking may have ruptured clay layers and released accumulated gas. Analyses of public reports and interviews with local officials show that alternative explanations—such as fire smoke, pipeline rupture, instrument malfunction, and gas-cylinder damage—were unlikely. These findings indicate that the observed methane anomaly was most likely caused by earthquake-synchronous underground gas release, suggesting that methane-release risk should be considered in post-earthquake fire-hazard assessments.

1. Introduction

1.1. Historical Background: Ignition Sources of Earthquake Fires Remain Unresolved

In the 1923 Great Kanto Earthquake, approximately 70% of fatalities resulted from fires that broke out after the shaking. In Honjo Ward, about 38,000 of the 40,000 evacuees perished in a catastrophic fire only three and a half hours after the earthquake. Scientific investigations at the time left many uncertainties regarding the ignition sources [1]. Similarly, in the 2024 Noto Peninsula Earthquake, large-scale fires occurred, yet even with a century of technological advancement, the ignition sources remain unidentified [2].

1.2. Emergence of the Methane Release Hypothesis

Recent studies have suggested that the intensification of earthquake-induced fires may be linked to methane released from the subsurface [1,3]. Enomoto et al. highlighted melted iron products and platinum found after the Great Kanto Earthquake, noting that these metals do not melt at typical wooden-house fire temperatures (~1100 °C), whereas methane’s adiabatic flame temperature (1963 °C) could melt them upon direct exposure [1]. They further proposed that methane may have contributed to the large fires following the Noto Peninsula Earthquake, presenting supporting evidence from historical surveys, geological investigations, and groundwater gas analyses [3].

1.3. Outstanding Scientific Controversies

Despite these insights, several key controversies remain:

• Can seismic shaking trigger the release of subsurface gas?
• Does methane undergo seismically induced exsolution, as confirmed for CO₂?
• Is pre-existing gas accumulation sufficient to assess hazard levels?
• Does subsurface gas significantly contribute to the intensification of earthquake fires?
• Do coseismic anomalies in atmospheric methane concentration actually occur?

1.4. Theoretical Framework: CO₂ Exsolution Mechanism by Crews et al.

Crews et al. established a theoretical and experimental framework demonstrating that seismic shaking can induce exsolution of dissolved CO₂ in groundwater [4,5,6,7]. Their laboratory experiments and field observations showed that dynamic stresses from seismic waves generate excess pore-water pressure, followed by decompression-driven bubble formation. This mechanism successfully reproduced post-seismic water-level rises observed in isolated wells.

1.5. Extension to Methane and Implications for Hazard

If CO₂ undergoes seismically induced exsolution, methane—also dissolved in groundwater—may behave similarly. Moreover, strong ground motions can generate cracks and liquefaction, allowing groundwater to intrude into newly formed voids where rapid decompression promotes gas vaporization. Thus, both pre-existing free gas and newly exsolved methane may be released simultaneously, potentially overturning the conventional assumption that regions with low pre-existing gas accumulation are safe.

1.6. Knowledge Gap: No Direct Observations in the Noto Peninsula Earthquake

Despite the above theoretical and geological considerations, no direct observations of methane release immediately after the 2024 Noto Peninsula Earthquake have been reported. Furthermore, the existence, timing, and spatial characteristics of coseismic atmospheric methane anomalies remain unverified. Building upon these theoretical considerations and unresolved controversies, the present study investigates whether coseismic atmospheric methane anomalies occurred during the 2024 Noto Peninsula Earthquake.

1.7. Purpose of This Study

The purpose of this study is to identify previously unreported coseismic anomalies in atmospheric methane concentration during the 2024 Noto Peninsula Earthquake and examine whether these anomalies may be attributable to subsurface natural gas release.

1.8. Overview of the Approach

To address this objective, we analyze:

• CH₄ concentration data from the Atmospheric Environmental Regional Observation System (AEROS, “Soramame-kun”) [8],
• high-time-resolution measurements obtained from local environmental agencies,
• seismic intensity distributions [9], and
• geological and hydrogeological information of the Nanao region [10,11,12,13,14,15,16].

These datasets are integrated to evaluate the temporal synchronization between seismic events [17] and methane anomalies, and to assess the plausibility of subsurface gas release.

1.9. Contributions of This Study

This study makes the following contributions:

• It provides the first identification of coseismic methane concentration anomalies associated with the Noto Peninsula Earthquake.
• It offers a multi-disciplinary evaluation of their potential subsurface origin.
• It introduces a new perspective for assessing post-earthquake fire hazards by considering rapid methane release triggered by seismic shaking.

2. Materials and Methods

2.1. Methane Gas Data, Seismic Intensity Distribution, and Locations of Earthquake Fires During the Noto Peninsula Earthquake

2.1.1. Methane Concentration Data (AEROS/Soramame-Kun)

To investigate CH₄ concentrations following the 2024 Noto Peninsula Earthquake, data were obtained from the Atmospheric Environmental Regional Observation System (AEROS), commonly known as Soramame-kun, operated by the Ministry of the Environment [8]. Data are provided in carbon-equivalent units (ppmC), which are used throughout this study. Only preliminary values are publicly available; finalized values require approximately two years for release. The validity of the preliminary values at the Nanao station before and after the earthquake was directly confirmed with officials of the Ishikawa Prefectural Institute for Public Health and Environmental Science.

Calibration and Quality Control of CH₄ and NMHC Measurements

Both CH₄ and NMHC concentrations in the Soramame-kun system are measured using gas chromatography with flame ionization detection (GC-FID). The instruments undergo routine automatic zero and span calibrations, and periodic maintenance is conducted by certified contractors commissioned by the Ministry of the Environment. Daily quality-control checks are also performed to identify abnormal fluctuations, instrument drift, or discontinuities in the time series. Although the publicly available data are preliminary values, these calibration and QC procedures ensure that concentration variations are reliably captured as long as the instrument is operating normally.

For the Nanao station, the continuous operation of the GC-FID instrument before and after the earthquake was independently confirmed by the Ishikawa Prefectural Institute for Public Health and Environmental Science. High-resolution (6 min) data provided by the staff further verified that no instrument interruption or malfunction occurred during the period of interest.

2.1.2. Preliminary vs. Finalized Values

The downloadable Soramame-kun data are preliminary values that may be corrected after verification. Although finalized values are recommended for research, their release requires two years; therefore, preliminary values were used in this study.

2.1.3. Methane Observation Stations

Table 1. Partial preliminary methane concentration data (1 January 2024 13:00–2 January 12:00). The data were downloaded from the Soramame-kun system [8]. Units are ppmC. Values exceeding 2.1 ppmC are highlighted in yellow, and values exceeding 2.5 ppmC are highlighted in red. The mainshock of the Noto Peninsula earthquake occurred at 16:10. At Wajima (L1), data were missing after the earthquake.

Time	L1: WAJI MA	L2: NANA O	L3: UCHI NADA	L4: MATT OU	L5: KOMA TSU	R1: MATSU HAMA	R2: OO YAMA	R3: YAMA KIDO	R4: UO TSU	R5: IWA GASE
1/1 12:00	2.01	1.99	2.05	2.01	1.99	2.00	2.01	2.02	2.05	1.99
13:00	2.00	1.99	2.06	2.00	1.99	2.01	2.01	2.02	2.05	1.98
14:00	2.01	1.99	2.06	2.01	1.99	2.01	2.02	2.03	2.05	2.00
15:00	2.01	1.99	2.07	2.01	1.99	2.02	2.01	2.02	2.05	1.98
16:00	2.01	1.99	2.06	2.01	1.99	2.02	2.01	2.02	2.05	2.00
17:00		2.54	2.27	2.01	2.02	2.03	2.05	2.44	2.05	1.98
18:00		3.54	2.36	2.02	2.00	2.03	2.03	2.05	2.05	1.99
19:00		2.85	2.27	2.05	2.01	2.03	2.18	2.12	2.05	1.99
20:00		2.61	2.21	2.06	2.02	2.33	2.19	2.26	2.05	2.00
21:00		2.40	2.17	2.04	2.01	2.21	2.10	2.10	2.04	2.01
22:00		2.38	2.25	2.05	2.01	2.11	2.09	2.10	2.04	2.06
23:00		2.31	2.28	2.05	2.04	2.12	2.06	2.07	2.04	2.04
1/2 00:00		2.37	2.21	2.01	2.03	2.07	2.08	2.08	2.04	2.05
01:00		2.40	2.22	-	2.00	2.06	2.09	2.08	2.03	2.05
02:00		2.45	2.14	-	2.02	2.08	2.10	2.07	2.05	2.02
03:00		2.58	2.15	2.00	2.02	2.07	2.05	2.05	2.05	2.02
04:00		2.48	2.19	2.00	2.00	2.06	2.07	2.09	2.05	2.06
05:00		2.43	2.17	2.00	2.00	2.03	2.08	2.11	2.05	2.04
06:00		2.35	2.17	2.00	2.02	2.05	2.10	2.13	2.04	2.05
07:00		2.30	2.14	2.00	2.02	2.05	2.11	2.12	2.03	2.06
08:00		2.38	2.13	2.00	2.01	2.04	2.10	2.14	2.03	2.08
09:00		2.28	2.20	2.00	2.00	2.06	2.08	2.11	2.04	2.08
10:00		2.19	2.15	2.00	2.00	2.06	2.09	2.07	2.04	2.04
11:00		2.09	2.13	2.00	2.00	2.05	2.08	2.06	2.05	2.04
12:00		2.03	2.06	2.00	1.99		2.06	2.05	2.04	2.01

presents a portion of the preliminary methane data. Values ≥ 2.1 ppmC are highlighted in yellow, and values ≥ 2.5 ppmC in red. Figure 1 shows the locations of methane observation stations on the Noto Peninsula. Two stations are located on the peninsula: L1 (Wajima) and L2 (Nanao). Western-side stations are labeled L1–L5 from north to south, and eastern-side stations are labeled R1–R5.

2.1.4. Earthquake-Induced Fire Locations

Red squares (F1–F4) in Figure 1 indicate the locations of earthquake-induced fires reported on 1 January 2024.

2.1.5. Seismic Intensity Distribution and Earthquake Catalog

Figure 2 shows the JMA seismic intensity distribution of the earthquake [9]. Shaking of intensity 6– or greater occurred across the Noto Peninsula and parts of Niigata. Methane time series visualized from Soramame-kun data [8] were compared with the JMA earthquake occurrence list [17] to examine temporal and spatial relationships between methane anomalies and seismic events.

2.2. Possibility That the Increase in Methane Concentration Originated from Subsurface Natural Gas

2.2.1. Historical Evidence of Natural Gas in the Nanao Area

Survey materials on natural gas in central Japan [10] report that wells in the Nanao area produced gas originating from Tertiary strata, containing unsaturated hydrocarbons and exhibiting oily odors in coastal strata. Figure 3 shows the distribution of these wells, with the Nanao Observation Station added in red.

2.2.2. Distribution of Water-Soluble Gas Fields

According to the Petroleum Mining Handbook [11], the Nanao station lies directly above water-soluble gas fields (Figure 4).

2.2.3. Surface Fissures and Liquefaction

Figure 5 shows surface fissures mapped by the Geospatial Information Authority of Japan [12]. The nearest fissure (A) lies across a river approximately 100 m west of the Nanao Observation Station; numerous fissures also occur along the coastline (~800 m). Liquefaction was confirmed between points B and C (Figure 5) [13].

2.2.4. Geological Structure of the Nanao Region

Figure 6 illustrates the stratigraphy of central Nanao City, consisting of alluvial deposits, diluvial deposits, silt layers, sand layers, and underlying bedrock [15]. These layers differ in permeability and influence groundwater flow and ground stability.

2.2.5. Borehole Stratigraphy near the Nanao Station

Figure 7 shows borehole logs arranged from west to east [16]. Subsurface structure varies markedly within several hundred meters:

• Borehole 55364757001 reaches weathered rock at shallow depth (high permeability).
• Borehole 217001010034 is overlain by thick, soft sediments (likely causing strong shaking amplification).
• Site 13-N-21 exhibits intermediate characteristics. Liquefaction along the coastline suggests that methane stored beneath clay layers may have been released.

Figure 7. Geological structures and location information of boreholes near the Nanao Observation Station. The inset also shows the positions of the Nanao station [8] and boreholes [16], together with the locations of wells used for groundwater level measurements [14] and the Japan Meteorological Agency seismic intensity observation station in Nanao City, Fuchu-honmachi [18].

Figure 7. Geological structures and location information of boreholes near the Nanao Observation Station. The inset also shows the positions of the Nanao station [8] and boreholes [16], together with the locations of wells used for groundwater level measurements [14] and the Japan Meteorological Agency seismic intensity observation station in Nanao City, Fuchu-honmachi [18].

2.2.6. Groundwater Level Variations

Figure 8 visualizes seven years of groundwater level data from wells ①–③ [14]. Wells ① and ② show seasonal variations, whereas well ③ does not. After the earthquake, a sustained rise in groundwater level was observed, suggesting possible release of water-soluble gases.

2.3. Four Possible Non-Subsurface Explanations for the Increase in Methane Concentration

2.3.1. Influence of Fire Smoke Plumes

Data sources: fire locations [19,20]; JMA wind direction and wind speed data [21].

Method: fire locations were overlaid on the station map; wind data were analyzed to assess plume influence.

2.3.2. Influence of Ruptured City Gas Pipelines

Data sources: Cabinet Office report [22]; METI Gas Safety Division; Nanao City Environmental Division.

Method: reports were reviewed and agencies were contacted to confirm whether gas leakage incidents occurred.

2.3.3. Influence of Instrument Malfunction

Data sources: manufacturer’s opinion via Ishikawa Prefectural Institute; high-resolution methane data.

Method: high-resolution data were compared with earthquake timing to evaluate possible malfunction.

2.3.4. Influence of Damaged Gas Cylinders

Data sources: same high-resolution dataset.

Method: timing of CH₄ and NMHC increases and component ratios were compared with gas cylinder compositions.

3. Results

3.1. Visualization and Temporal Analysis of Methane Concentration Data

3.1.1. Time-Series at All Observation Stations (Figure 9)

Figure 9 presents time-series graphs of CH₄ and NMHC concentrations measured at observation stations in the Noto region [8]. In the graphs, the purple line represents CH₄ and the green line represents NMHC. The period covers six days before and after the mainshock. The data shown in Table 1 corresponds to a portion of these graphs. To align with the station locations indicated in Figure 1, the graphs were arranged from north to south: stations on the western side (L1–L5 in Figure 1) are shown on the left, and those on the eastern side (R1–R5) are shown on the right. Based on the earthquake list [17], vertical light-blue lines were drawn at the times of earthquakes with magnitude ≥ 5, with annotations such as “M7.6 (J7)” above them, where M denotes magnitude and J denotes maximum seismic intensity. The mainshock was highlighted in red.

From the left side of Figure 9, it can be seen that measurements at Wajima (L1) stopped immediately after the mainshock. At the Nanao station (L2), both CH₄ and NMHC concentrations increased following the earthquake and remained elevated above normal levels until around noon the next day. Both L1 and L2 are located in areas where seismic intensity exceeded 6− (see Figure 2).

At the Uchinada station (L3), a smaller increase was observed compared with Nanao. Further south, at stations L4 and L5, no increase in CH₄ concentration was detected. Stations L3 to L5 are located in areas where seismic intensity reached 5+ (see Figure 2).

From the right side of Figure 9, it can be seen that at the northern stations Matsuhama (R1) and Oyama (R2), small increases in CH₄ concentration comparable to that at Uchinada (L3) were observed immediately after the mainshock, followed by a very pronounced increase two days later. At Yamaki-do (R3), which is located near R1 and R2, a small increase was observed immediately after the mainshock, but no pronounced increase was detected two days later. These three stations are located in areas where seismic intensity reached 5+ during the mainshock. At stations further south, no increase in CH₄ concentration was observed.

Figure 9. Changes in CH₄ (purple line) and NMHC (green line) concentrations based on hourly data before and after the Noto earthquake. The ten panels are arranged to reflect the geographic layout of the monitoring network shown in Figure 1: the left column corresponds to western stations (L1–L5) and the right column to eastern stations (R1–R5), ordered from north to south. Vertical light-blue lines indicate the occurrence times of earthquakes in the Noto Peninsula, based on the list provided by Yahoo Japan [17], with annotations showing magnitude (“M”) and maximum seismic intensity (“J”). The red line and label indicate the mainshock.

Figure 9. Changes in CH₄ (purple line) and NMHC (green line) concentrations based on hourly data before and after the Noto earthquake. The ten panels are arranged to reflect the geographic layout of the monitoring network shown in Figure 1: the left column corresponds to western stations (L1–L5) and the right column to eastern stations (R1–R5), ordered from north to south. Vertical light-blue lines indicate the occurrence times of earthquakes in the Noto Peninsula, based on the list provided by Yahoo Japan [17], with annotations showing magnitude (“M”) and maximum seismic intensity (“J”). The red line and label indicate the mainshock.

3.1.2. Long-Term Variations at the Nanao Observation Station (Figure 10a)

To examine temporal variations in greater detail, long-term and short-term graphs of the Nanao observation station are shown in Figure 10a represents a long-term record of two and a half years, Figure 10b represents a short-term record of 28 h, and Figure 10c shows the increment of CH₄ concentration relative to the stable value (1.99 ppmC) together with the ratio to NMHC concentration. Vertical lines indicate the occurrence times of earthquakes with magnitude ≥ 5. The mainshock is highlighted in red, and in the long-term graph only the mainshock is annotated.

From Figure 10a, it can be seen that the pronounced increase in CH₄ concentration at the time of the earthquake had never been recorded during the two and a half years. The average concentration during past one year was 1.99 ppmC, with a standard deviation of 0.0522. The maximum concentration of 3.54 ppmC corresponds to a deviation of 1.54 ppmC from the average of the past one year, which is equivalent to 29 times the standard deviation. NMHC also exhibited a pronounced increase after the earthquake; however, unlike CH₄, even larger peaks were observed in November 2023 and October 2024. Notably, CH₄ did not increase in conjunction with the 2023 peak.

Figure 10. CH4 and NMHC concentrations at the Nanao observation station before and after the Noto earthquake, based on hourly data: (a) long-term variations over two and a half years, (b) short-term variations over 28 h, and (c) ratio changes. The purple line represents CH4, the green line represents NMHC, and the orange line represents their ratio.

Figure 10. CH4 and NMHC concentrations at the Nanao observation station before and after the Noto earthquake, based on hourly data: (a) long-term variations over two and a half years, (b) short-term variations over 28 h, and (c) ratio changes. The purple line represents CH4, the green line represents NMHC, and the orange line represents their ratio.

3.1.3. Short-Term Variations at the Nanao Station (Figure 10b)

From Figure 10b, it can be seen that CH₄ concentration rose gradually over two hours immediately after the earthquake, and then decreased simultaneously with the end of an M5 aftershock. NMHC showed a similar trend, although the increase in CH₄ appeared to precede that of NMHC by approximately one hour.

3.1.4. CH₄/NMHC Ratio Analysis (Figure 10c)

From Figure 10c, it can be seen that by 17:00 the increment in CH₄ concentration had reached seven times that of NMHC. This suggests that CH₄ may have been released earlier than NMHC following the earthquake.

3.2. Examination of Methane Emission Mechanisms at the Nanao Observation Station

3.2.1. Existence of Subsurface Gas Reservoirs

As shown in Section 2.2, a natural gas reservoir exists beneath the Nanao Observation Station [11]. Unsaturated hydrocarbons have been identified from piles, and strata distributed along the coast exhibit an oil odor [10]. In addition, ground fissures are present around the station [12], with the nearest crack located approximately 100 m to the northwest, across the Misogi River drainage channel. Liquefaction was confirmed in an area about 500 m northeast of the station after the earthquake [13]. Furthermore, a geological structure consisting of a clay layer overlying sandstone has been reported [14,15,16], and wells in the area have shown continuous rises in water level following the earthquake [14].

These facts suggest that methane may have been capped beneath clay layers in riverbeds or coastal inlets. Strong shaking of seismic intensity 6+ could have induced liquefaction and fissure formation, thereby potentially damaging the clay cap and enabling methane release. This represents the simplest and most rational mechanism to explain CH₄ emission.

In addition to the deeper natural gas reservoir, the near-surface geological environment around the Nanao Observation Station is also highly favorable for the generation and retention of biogenic methane. The coastal and river-mouth areas are characterized by alternating layers of silt, clay, and organic-rich sediments, which create anaerobic conditions that promote microbial methanogenesis in shallow strata.

The presence of thick clay and mud layers with extremely low permeability suggests that methane produced in these shallow sediments may have been effectively capped beneath these fine-grained deposits. Such clay-capped shallow gas pockets are commonly observed in deltaic and estuarine environments, where gas accumulates beneath mud layers until the cap is mechanically disrupted.

3.2.2. Clay-Cap Rupture and Shallow Gas Exsolution

During the 2024 Noto Peninsula earthquake, strong shaking (seismic intensity 6+) likely induced liquefaction and ground fissures, which could have ruptured these shallow clay caps. The sudden failure of these low-permeability layers would allow previously trapped biogenic methane to escape rapidly to the surface. This mechanism provides an additional and independent pathway for coseismic methane release, complementing the deeper reservoir-derived gas described above.

In addition to mechanical rupture of clay caps, seismic shaking itself can promote gas exsolution in shallow sediments. Microbially generated methane often exists as dissolved gas or micro-bubbles within silt and mud layers. When subjected to low-frequency ground motion (approximately 0.1–1 Hz), these micro-bubbles can expand and coalesce, accelerating the transition from dissolved to free gas. This process is analogous to shaking an opened carbonated beverage, where agitation enhances bubble formation and drives rapid degassing.

Therefore, even before the clay cap is fully ruptured, seismic vibration can increase the internal gas pressure within shallow sediments. Once liquefaction or fissuring breaks the low-permeability clay layer, the accumulated methane can be released abruptly to the surface.

3.2.3. Excess Pore-Water Pressure Hypothesis (Crews et al.)

Another possible process of subsurface methane release is the hypothesis of excess pore-water pressure proposed by Crews et al., introduced in the Introduction [5]. They demonstrated that when strong vibrations are applied to groundwater containing CO₂, degassing phenomena cause a unilateral increase in excess pore-water pressure. This phenomenon does not occur in pure water without dissolved CO₂. According to the hypothesis, vibrations from a large earthquake can induce gas exsolution in groundwater, and the formation of new voids in the ground near the surface. If the groundwater intrudes into those voids, a sudden pressure drop happens. This leads to additional gas exsolution. If methane is included in the groundwater, methane would also be emitted.

When applied to the Nanao observation station, the following conditions can be identified:

• Condition 1: Presence of dissolved gas.

• From Figure 3 and Figure 4 and reference [10,11], it has been confirmed that water-soluble gases are dissolved in groundwater. Although direct information on saturation is not available, the unilateral rise in groundwater level after the earthquake shown in Figure 8 suggests a transition from saturation to degassing.

• Condition 2: Strong shaking and frequency components.

• The Nanao area experienced shaking of seismic intensity ≥ 6 (Figure 2). Furthermore, spectral analysis shown in Figure 11 [18] revealed prominent frequency components in the range of 0.01–0.3 Hz, satisfying the necessary condition.

• Condition 3: Formation of pathways for sudden pressure drop.

• Records of ground fissures and liquefaction (Figure 5, ref. [13]) indicate that subsurface density balance was altered, making the formation of channels for sudden pressure reduction highly plausible.

Figure 11. Vibration spectra of the Noto Peninsula earthquake. NS denotes the north–south direction, EW the east–west direction, and UD the vertical direction. The location corresponds to the JMA seismic intensity observation point situated 1.5 km southeast of the Nanao station (Fuchu-honmachi, Nanao City). The spectral diagram from site [18] was excerpted, and the frequency range of 0.01–0.3 Hz was highlighted in orange.

Figure 11. Vibration spectra of the Noto Peninsula earthquake. NS denotes the north–south direction, EW the east–west direction, and UD the vertical direction. The location corresponds to the JMA seismic intensity observation point situated 1.5 km southeast of the Nanao station (Fuchu-honmachi, Nanao City). The spectral diagram from site [18] was excerpted, and the frequency range of 0.01–0.3 Hz was highlighted in orange.

3.2.4. Summary of Conditions and Integrated Interpretation

In summary, although it remains uncertain whether dissolved gases were in a saturated state, the other conditions are satisfied. As shown in Table 1, earthquakes of M5.7, M7.6, and M6.1 occurred consecutively immediately before the rise in CH₄ concentration, which may have accelerated saturation and triggered the degassing phenomenon. Investigating the abnormal rise in groundwater level after the earthquake (Figure 8) from the perspective of gas generation provides important implications for future management of combustible gas risks.

3.3. Examination of Four Non-Subsurface Possibilities for Concentration Increase

Here, it is demonstrated that the four possible causes of methane increase discussed in Section 2.3 do not apply to the Nanao Observation Station.

3.3.1. Influence of Fire Smoke

The horizontal diffusion velocity of smoke is estimated to be approximately 1 m/s. For fire smoke to reach the Nanao Observation Station by 17:00, when the observation peak was recorded, a fire would have needed to exist within 3 km of the station during the 50 min following the mainshock (16:10). Among the fire locations shown in Figure 1 [19,20], the nearest site, Nakajima Town (F2), is 12 km away, and thus none of them satisfy the arrival condition. Although F1 and F2 were detected after 17:00, which theoretically leaves room for consideration, this does not affect the conclusion.

Consideration of wind direction and velocity leads to the same result. As shown in

Table 2. Time-series data of wind direction and wind speed at the Nanao observation station [21] and locations of earthquake-induced fires [19,20].

	F1:WAJIMA, N39 km		SHIGA, WNW24 km		F3:HIMI, S21 km		L2:Nanao, 0 km
Time	Wind Direction	Speed [m/s]	Wind Direction	Speed [m/s]	Wind Direction	Speed [m/s]	Wind Direction	Speed [m/s]
16:00	N	1.4	ENE	1.3	NNW	1.9	NNE	0.8
16:10	NNW	1.3	ENE	1.0	N	1.4	NNE	0.7
16:20	W	1.4	ENE	1.0	NNW	2.2	E	0.4
16:30	WNW	1.6	ENE	1.4	NNW	1.8	SSE	0.5
16:40	S	1.4	ENE	1.6	N	1.5	SSE	0.6
16:50	S	2.1	E	1.6	NNW	1.5	SSE	0.8
17:00	SW	2.1	ESE	1.6	WNW	1.5	S	0.8
17:10	SSW	1.4	ESE	1.6	WSW	1.7	S	0.7
17:20	SSW	2.6	ESE	1.6	WSW	1.4	S	0.6
17:30	SW	2.7	E	1.5	W	1.3	SSE	0.7
17:40	SW	2.4	E	1.3	W	1.7	S	0.8
17:50	SSW	2.2	E	1.2	W	1.4	S	0.6
18:00	S	1.3	E	1.2	SW	1.3	S	0.6

and Figure 12, easterly winds prevailed around the station between 16:10 and 17:00, causing smoke from Nakajima Town (F2) to disperse westward, away from the station. In the cases of Himi City Chuo Town (F3) and Takaoka City Shugo Town (F4), northerly winds prevented smoke from reaching the station. At Wajima City Kawai Town (F1), winds temporarily blew toward the station; however, the calculated travel distance was only 2.58 km, insufficient to reach the station located 39 km away. Subsequently, winds reversed, completely eliminating the possibility of arrival.

Therefore, it can be concluded that fire smoke generated by the earthquake did not reach the Nanao Observation Station by 17:00, and the possibility that fire smoke contributed to the observed CH₄ concentration increase can be excluded.

3.3.2. Influence of Urban Gas Pipeline Rupture

The composition of urban gas (13A) is methane 89.6%, ethane 5.62%, propane 3.43%, and butane 1.35%, with methane being the dominant component. Based on these ratios, methane corresponds to approximately nine times the amount of NMHC. In contrast, Figure 10c shows that the CH₄/NMHC ratio is about seven times and decreases over time. This has led to the hypothesis that “urban gas leakage occurred, and the lighter methane rose first, causing the ratio to decline.” This section examines that hypothesis.

Central Government Data

An investigation of 52 reports on the Noto Peninsula Earthquake published by the Cabinet Office [22] revealed no incidents of urban gas leakage near the Nanao Observation Station. The final report, dated 29 October 2024, was issued 11 months after the earthquake, making the possibility of unreported cases unlikely.

Table 3 presents excerpts from the 2 January 2024 report concerning gas-related items:

• Urban gas:

• No supply disruption in general gas pipeline operations. Although gas transmission from the Naoetsu LNG terminal was suspended for safety, backup supply ensured continuity.

• Gas retail business (simplified gas):

• One simplified gas housing complex in Ishikawa Prefecture experienced landslide damage to a main branch pipeline due to the earthquake. This system was LP gas, which does not contain methane.

• Heat supply business:

• No supply disruption and no damage reported.

• LP gas:

• No damage information at the time.

Thus, the descriptions in Table 3 also confirm that no leakage due to rupture of urban gas supply pipelines occurred around Nanao City.

Furthermore, confirmation from the Gas Safety Section, Industrial Safety and Security Group, Ministry of Economy, Trade and Industry (as of November 2024) indicated, as shown in Table 4, that “Nanao City has no urban gas supply area and no supply facilities, and therefore no accidents occurred”.

The possibility of leakage from the Naoetsu LNG terminal (located 200 km east of Nanao) was examined. If leakage had occurred immediately after the earthquake, detection at Nanao by 17:00 would have required transport at approximately 33.3 m/s. However, the wind speed at that time was about 2 m/s, with adverse wind conditions, making arrival impossible.

Table 3. Gas-related information from the Cabinet Office report [22]. Ministry of Economy, Trade and Industry data as of 2 January, 7:00.

Supervised Business	Contents
City Gas	As of 7:00 a.m. on Tuesday, 2 January, there are no supply disruptions for the general gas pipeline business. At the INPEX Naoetsu LNG Terminal, gas transmission has been suspended to ensure safety following the earthquake. After the tsunami warning is lifted, the terminal will be inspected, and gas transmission is scheduled to resume. Currently, backup gas supply is being received from Shizuoka Gas. Based on the current situation, gas supply is expected to continue until around 6:00 p.m. on Wednesday, 3 January. If the restart of gas transmission from the Naoetsu LNG Terminal is further delayed beyond that point, arrangements are being made to continue gas supply by receiving backup supply from Tokyo Gas. Once the tsunami warning is lifted, the terminal will be inspected, and gas transmission is scheduled to resume.
Gas Retail Business	At one simplified gas complex in Ishikawa Prefecture, an earthquake caused a landslide that damaged the main gas pipeline. The area is currently restricted due to fire department operations, making entry impossible. Future restoration measures are undetermined.
Heat Supply Business	There are no supply disruptions, and no damage reports have been received.
LP Gas	No damage reports at this time.

Table 3. Gas-related information from the Cabinet Office report [22]. Ministry of Economy, Trade and Industry data as of 2 January, 7:00.

Supervised Business	Contents
City Gas	As of 7:00 a.m. on Tuesday, 2 January, there are no supply disruptions for the general gas pipeline business. At the INPEX Naoetsu LNG Terminal, gas transmission has been suspended to ensure safety following the earthquake. After the tsunami warning is lifted, the terminal will be inspected, and gas transmission is scheduled to resume. Currently, backup gas supply is being received from Shizuoka Gas. Based on the current situation, gas supply is expected to continue until around 6:00 p.m. on Wednesday, 3 January. If the restart of gas transmission from the Naoetsu LNG Terminal is further delayed beyond that point, arrangements are being made to continue gas supply by receiving backup supply from Tokyo Gas. Once the tsunami warning is lifted, the terminal will be inspected, and gas transmission is scheduled to resume.
Gas Retail Business	At one simplified gas complex in Ishikawa Prefecture, an earthquake caused a landslide that damaged the main gas pipeline. The area is currently restricted due to fire department operations, making entry impossible. Future restoration measures are undetermined.
Heat Supply Business	There are no supply disruptions, and no damage reports have been received.
LP Gas	No damage reports at this time.

Table 4. Response from the Gas Safety Section, Industrial Safety and Security Group, Ministry of Economy, Trade and Industry.

Date and Time Sent	Response Content
5 November, 2024, 9:09 p.m.	There are no city gas providers in Nanao City, Ishikawa Prefecture, and no city gas supply areas exist. Consequently, no city gas supply facilities (such as city gas pipelines) are present. Therefore, no city gas accidents occur.

Table 4. Response from the Gas Safety Section, Industrial Safety and Security Group, Ministry of Economy, Trade and Industry.

Date and Time Sent	Response Content
5 November, 2024, 9:09 p.m.	There are no city gas providers in Nanao City, Ishikawa Prefecture, and no city gas supply areas exist. Consequently, no city gas supply facilities (such as city gas pipelines) are present. Therefore, no city gas accidents occur.

Local Government Inquiry

An inquiry to the Environmental Section, Department of Civil Affairs of Nanao City confirmed, as shown in

Table 5. A question from AIST and an answer from the Environment Division, Citizens’ Life Department, Nanao City.

Date and Time	Question and Answer
1 November, 2024 (Question)	Are there any record of reports received by Nanao City about a city gas leak occurring between 4:00 p.m. on 1 January and 12:00 p.m. on 2 January?
5 November, 2024 (Answer)	No such information has been reported (though several cases of leaks from LP gas cylinders have been reported). To begin with, we understand that Nanao City has virtually no city gas. Primarily, gas companies and the administration do not provide widespread city gas supply to the public. (Omitted) Methane fermentation is not conducted at the sewage treatment plant. While there is a national petroleum and gas reserve base, it stores only propane and butane.

, found that no urban gas supply facilities exist and that no leakage accidents have been reported. Several cases of leakage from LP gas cylinders were reported; however, the National Petroleum Gas Stockpiling Base stores only propane and butane, and methane is not included.

3.3.3. Reliability of Methane Measurements (Instrument Error Check)

An inquiry was made to the atmospheric officer of the Ishikawa Prefectural Institute of Public Health and Environmental Science, which oversees the Nanao Observation Station (https://www.ishikawa-taiki.com/, accessed on 22 December 2025), regarding the validity of the methane measurements. The final response was as follows: “The methane concentration measuring instrument continued to operate before and after the earthquake. According to the manufacturer, a malfunction of the instrument is unlikely. At present, the source of methane emissions remains unknown”.

Furthermore, the staff provided a graph with 6 min intervals demonstrating that the instrument was operating continuously during the earthquake.

Table 6. Concentrations of CH₄ and NMHC extracted from the 6 min interval graph provided by the Ishikawa Prefectural Institute of Public Health and Environmental Science (https://www.ishikawa-taiki.com/, accessed on 22 December 2025). The URL refers to the organization to which the local atmospheric officer belongs; the data themselves were provided directly by the officer and were not downloaded from the website. Red letters indicate the hourly averages, and the values in parentheses represent the CH₄ concentrations shown in Figure 10b. Green letters indicate the occurrence time and magnitude of the earthquakes. Other comments are shown in black.

Time	CH4	NMHC	Hourly Average CH4 (CH4 in Figure 10b) Other Comments	Time	CH4	NMHC	Hourly Average CH4 (CH4 in Figure 10b) Other Comments
15:00	1.99	0		17:00	2.92	0.51	CH4, NMHC increase
15:06	1.99	0		17:06	3.36	0.92
15:12	1.99	0		17:12	3.85	1.35
15:18	1.99	0		17:18	4.25	1.75	CH4 2nd peak
15:24	1.99	0		17:24	4.11	2.1
15:30	1.99	0.01		17:30	3.77	2.55	NMHC 1st peak
15:36	1.99	0.01		17:36	3.32	1.69
15:42	1.99	0.01		17:42	3.17	1.55
15:48	1.99	0.01		17:48	3.2	1.38
15:54	1.99	0.01	1.99 (1.99)	17:54	3.3	1.08	3.535 (3.54)
16:00	1.99	0.02		18:00	3.29	0.82	18:03 M5.3
16:06	1.99	0.02	16:06 M5.7, 16:10 M7.6	18:06	3.15	0.81	18:08 M5.6
16:12	1.99	0.02		18:12	2.99	0.96
16:18	2.00	0.02	16:18 M6.1	18:18	2.9	1.15
16:24	2.25	0.02	CH4 steep increase	18:24	2.88	1.12
16:30	2.67	0.04		18:30	2.85	0.93
16:36	3.14	0.06		18:36	2.75	0.81
16:42	3.3	0.10	CH4 1st peak	18:42	2.62	0.82
16:48	3.09	0.15		18:48	2.55	0.91
16:54	2.89	0.20	2.531 (2.54), 16:56 M5.7	18:54	2.57	0.84	2.855 (2.85)

presents values extracted by the author from this graph, which include reading errors. To compare these with the officially published hourly values in Table 1, the hourly averages derived from the extracted data are shown in red in Table 6. Values in parentheses represent the published hourly data (Figure 10b and Table 1). From Table 6, it can be seen that the maximum error in the hourly averages is 0.009 ppm.

This level of error is negligible compared to the observed variation range (>1 ppmC). Therefore, the extracted values are considered sufficiently reliable for discussing the fluctuations in methane concentration associated with the seismic events.

Because Figure 10 is based on hourly data, the two short-term CH₄ peaks are not fully resolved. To examine this behavior in more detail, higher-resolution 6 min data were analyzed.

Figure 13 plots the data from Table 6, with CH₄ shown as the purple line and NMHC as the green line. As evident from both Table 6 and Figure 13, the significant increase in CH₄ concentration began at 16:24, shortly after the M6.1 aftershock at 16:18. In contrast, NMHC began to rise at 16:56, coinciding with the M5.7 aftershock at that time. CH₄ increased approximately 32 min earlier than NMHC, peaked once, and then began to decline. During this decline, NMHC gradually started to increase. As CH₄ rose again, NMHC exhibited a sharp increase at 17:00, slightly after the 16:56 aftershock.

Figure 14 illustrates the ratio of CH₄ to NMHC concentration increases derived from Figure 13. The ratio began rising sharply around 16:24 (M6.1 aftershock), peaking at 16:42, when CH₄ levels were nearly 20 times higher than NMHC. Over the following 30 min, the ratio steadily decreased to 1, forming a minor peak. These patterns suggest that the 6 min interval data provide more detailed insight into the timing and dynamics of CH₄ release.

Both CH₄ and NMHC concentrations were measured using the same flame ionization detector (FID), which operates by combusting hydrocarbons in a hydrogen flame and detecting the resulting ionized carbon atoms as an electric current. Since this method responds proportionally to the number of carbon atoms present, any malfunction in the detector would be expected to affect both CH₄ and NMHC measurements equally.

Therefore, the observed temporal discrepancy—where CH₄ increased significantly while NMHC remained stable—cannot be attributed to instrumental bias. This divergence strongly suggests that the CH₄ increase reflects a distinct physical process, rather than a measurement artifact.

These findings collectively indicate that the observed increase in methane concentration at the Nanao Observation Station is unlikely to have resulted from instrumental malfunction. The manufacturer reported that a malfunction was unlikely, and the device continued operating without interruption before and after the earthquake.

3.3.4. Influence of Damaged Gas Cylinders

Finally, the results of the investigation into the potential impact of damaged gas cylinders are presented. An inquiry to the Environment Division of the Citizens’ Life Department, Nanao City confirmed that there had been reports of leakage from LP gas cylinders (Table 5). The main components of LP gas are propane (C₃H₈) and butane (C₄H₁₀). If LP gas cylinders had been damaged, NMHC should have been detected first. However, as shown in Figure 10b and Figure 13, CH₄ concentrations at the Nanao Observation Station increased earlier, and at 16:36 the increment of CH₄ from the average reached approximately 20 times that of NMHC. At least the initial rise in CH₄ (16:18–42 in Figure 13) cannot be attributed to damage of LP gas cylinders.

4. Discussion

From the results of the previous section, it is suggested that the increase in CH₄ concentration observed at Nanao Observation Station was most likely of subsurface and natural origin. In addition, the high time-resolution data presented in Table 6 and Figure 13 enabled a more detailed examination of the relationship between the earthquake occurrence time and the onset of methane concentration increase. In this chapter, these data are used to discuss the location and timing of methane emissions from underground sources.

4.1. Range of Methane Emission Sources Recorded at the Nanao Observation Station

Figure 15 shows a map of the area surrounding the Nanao Observation Station. To the west lies the Misogi River, and to the southeast is Saiko-ji Temple.

Table 7. Wind directions in Nanao City, 10 min travel distances, and CH₄ concentrations observed at Nanao up to the first peak in Figure 13.

Time	Wind Direction	Wind Speed [m/s]	10 min Traveling Distance [m]	CH4 Concentration [ppmC]
16:00–16:10	NNE	0.7	420	16:00 (15:57–16:03), 1.99 16:06 (16:03–16:09), 1.99
16:10–16:20	E	0.4	240	16:12 (16:09–16:15), 1.99 16:18 (16:15–16:21), 2.00
16:20–16:30	SSE	0.5	300	16:24 (16:21–16:27), 2.25 16:30 (16:27–16:33), 2.67
16:30–16:40	SSE	0.6	360	16:36 (16:33–16:39), 3.14
16:40–16:50	SSE	0.8	480	16:42 (16:39–16:45), 3.3 (peak) 16:48 (16:45–16:51), 3.09

presents wind directions between 16:24 and 16:42, during which methane concentrations were rising toward the first peak, extracted from Table 2. In Table 7, the 10 min travel distances calculated from wind speed and the corresponding methane observation values have been added. As indicated in references [23,24,25,26], observational results and theoretical calculations show that methane released at ground level does not reach the upwind direction.

In practice, methane possesses buoyancy; therefore, when wind is present and the air intake of the instrument is located 3 m above ground level, emissions released too close to the intake may pass beneath it and remain unrecorded, while emissions released too far away may pass above it and also remain unrecorded. Only emissions originating from an intermediate distance are recorded by the instrument. In this section, the buoyancy effect is first disregarded, and attention is focused on whether the distance is sufficient for methane to reach the observation station. Specifically, the possibility of methane released from the fissures shown in Figure 5 reaching the Nanao Observation Station is examined.

In the following, fissures A–F shown on the right side of Figure 5 are examined.

Fissure A: The fissure closest to the Nanao Observation Station is located at point A, over 100 m to the northwest, across the Misogi River. As shown in Table 7, westerly winds were absent throughout the observation period; therefore, even if methane had been released from the fissure to the northwest, it would not have reached the Nanao Observation Station.

Fissure B: The next closest fissure is located at point B, 280 m to the north, across the Misogi River. Since northeasterly winds were blowing from 16:10, the 10 min travel distance was 420 m > 280 m. If lateral diffusion perpendicular to the wind direction was significant, methane released from this fissure could have contributed to the increase observed between 16:10 and 16:20. After 16:20, however, the wind shifted to east and south-southeast, making further transport impossible. Thus, this fissure did not contribute to the methane increase between 16:30 and 16:40.

Fissure C: The next fissure is located at point C, 547 m to the northeast. Although northeasterly winds were blowing from 16:10, the 10 min travel distance was 420 m < 547 m, insufficient to reach the Nanao Observation Station. After 16:20, the wind direction shifted to east and south-southeast, so methane from this fissure could not have reached the station.

Fissure D: The next fissure is located at point D, 694 m to the south-southeast. At 16:10 and 16:20, the wind directions did not allow transport to the Nanao Observation Station. At 16:30, the 10 min travel distance was 300 m < 694 m, still insufficient. Adding another 10 min from 16:40 yields 660 m < 694 m, which also falls short. Therefore, this fissure did not contribute to the methane increase between 16:20 and 16:40.

Fissure E: The next fissure is located at point E, 488 m to the south-southwest. At 16:10 and 16:20, the wind directions did not allow transport to the Nanao Observation Station. At 16:30, south-southeasterly winds were blowing; if lateral diffusion perpendicular to the wind direction was significant, methane could have moved toward the station, traveling 300 m in 10 min. From 16:40, the 10 min travel distance was 660 m, making it possible for methane to reach the station between 16:40 and 16:50. Thus, this fissure did not contribute to the methane increase between 16:20 and 16:30.

Fissure F: The next fissure is located at point F, 325 m to the southwest. At 16:10 and 16:20, the wind directions did not allow transport to the Nanao Observation Station. At 16:30, south-southeasterly winds were blowing, but the direction was offset by 67.5° from the station, preventing transport. Therefore, this fissure did not contribute to the methane increase between 16:20 and 16:30.

In summary, the CH₄ increase at 16:24 may have originated from fissure B, located 280 m to the north, while the CH₄ increase at 16:42 may have originated from fissure E, located 488 m to the south-southwest. The largest increases, observed at 16:30 and 16:36, cannot be explained by emissions from these fissures. Moreover, considering the low density of methane, it is questionable whether methane could have remained at a height of 3 m until coming from points B or E.

As shown in Figure 2 and Figure 11, the vicinity of the Nanao Observation Station experienced seismic intensity 6 [18], suggesting that widespread disturbances in subsurface density conditions occurred, similar to the surface fissures illustrated in Figure 5. Furthermore, as indicated in Figure 4, methane is present underground. As described in Section 3.2, the main shock triggered the formation of gas bubbles, leading to an increase in groundwater pressure and a rise in groundwater levels in wells connected to specific aquifers, as shown in Figure 8. Subsequently, the earthquake-induced fissures allowed intrusion, causing a sudden pressure drop, which may have led to widespread methane generation underground and its seepage to the surface.

This mechanism may explain why gas was not observed immediately after the main shock but began to be detected eight minutes later. Therefore, the CH₄ peaks recorded at the Nanao Observation Station are reasonably interpreted as methane released from the ground located upwind of the station, transported by wind as it rose. The emission range was not confined to a single location but likely occurred over a broad area.

The interpretation of wind-field effects is subject to several uncertainties. The wind data used in this study represent 10 min averaged values and do not capture short-lived turbulent fluctuations that may influence gas transport. Local topographic effects near the Misogi River and surrounding buildings may also modify the actual wind direction relative to the measured values. Furthermore, the analysis assumes horizontally uniform mixing and does not incorporate three-dimensional turbulence or buoyancy-driven plume rise. These uncertainties do not alter the qualitative conclusion that methane must have originated from upwind directions, but they limit the precision with which individual fissures or emission points can be identified.

4.2. Consideration of CH₄ Emission Timing at the Nanao Observation Station

4.2.1. Estimation Method of Methane Emission Timing Using the Explicit Solution of the One-Dimensional Advection–Diffusion Equation

When CH₄ reaches subsurface fissures and then diffuses and rises through the soil, it is expected that, by the time it reaches the surface, the horizontal concentration differences will have diminished, as illustrated in Figure of reference [27]. If the methane concentration distribution above the ground surface is approximated as horizontally uniform, it can be described by the one-dimensional advection–diffusion equation in the vertical direction [28]. This approximation implies that, when methane diffuses from a single source to a distant location, a considerable release concentration would be required, and thus the predicted concentration values may be lower than the actual release concentration. In other words, the values obtained represent the lower bound of the predicted concentrations.

If the temporal variation in methane concentration at the ground surface can be determined, the causal relationship between seismic motion and gas release can be investigated by comparison with the earthquake occurrence time. In order to reproduce the observed methane concentrations at 3 m above ground shown in Figure 13, it is necessary to solve the inverse problem to obtain the temporal variation in concentration at 0 m above ground. The simplest procedure is as follows:

• STEP A: Assume the temporal variation in methane concentration at 0 m above ground.
• STEP B: Calculate the temporal variation in methane concentration at 3 m above ground.
• STEP C: Repeat Steps A → B until the early peak waveform observed in Figure 13 is reproduced.

If the waveform cannot be reproduced, the advection velocity is gradually increased, and the procedure is restarted from Step A.

When the one-dimensional advection–diffusion equation [28] is discretized using an explicit numerical scheme, the following expression is obtained in C language [29].

c[i] = c0[i]-u*dt/(2*dx)*(c0[i + 1]-c0[i − 1]) + D*dt/(dx*dx)*(c0[i − 1] − 2*c0[i] + c0[i + 1]);

(1)

Here, dt denotes the time step, dx the distance between sample points, u the advection velocity, and D the diffusion coefficient. c[i] represents the concentration at position i·dx at the current time, while c0[i] represents the concentration at position i·dx one time step dt earlier. In other words, this equation explicitly calculates the spatial distribution c[i] at the current time from the previous distribution c0[i].

First, only diffusion is considered, without upward advection. The parameters are set as follows: advection velocity u = 0 m/s, diffusion coefficient D = 0.0000196 m²/s, time step dt = 60 s, and spatial step dx = 1 m. The index i is considered from 0 to 101, and the initial concentration is set to 2 ppmC at all points. A temporal concentration waveform is applied at position i = 50, and the temporal variation in concentration is observed at position i = 53, which is 3 m above the release point. Boundary values are held equal to those of the previous time step, i.e., c [0] = c0[0] and c[100] = c0[100].

This represents the simplest one-dimensional model that neglects the presence of the ground. Since reflection effects from the ground are ignored, the concentration values are underestimated; however, this is consistent with the purpose of obtaining the lower bound of the predicted concentrations.

4.2.2. Simulation Results and Methane Release Time

Figure 16 shows the simulation results for a methane concentration of 5 ppmC released at the ground surface (x = 0 m, i = 50) at the time of the M6.1 aftershock at 16:18. In this figure, the purple circles represent the actual methane measurements at the Nanao Observation Station shown in Figure 13, the green line indicates the temporal variation in methane concentration at the release point (i = 50, x = 0 m), and the light blue line shows the concentration variation at the gas intake point (i = 53, x = 3 m), located 3 m above the release location.

As can be seen, the methane introduced at x = 0 m at 16:18 remains almost unchanged by 16:42, when the observed values (purple circles) reach their first peak. Consequently, the concentration at the 3 m point remains indistinguishable from the background level of 2 ppmC. This is because the vertical advection velocity u was set to zero, meaning that the transport was too slow for methane to reach 3 m. In the following analysis, u will be gradually increased in an attempt to reproduce the observed peak.

When the advection term u is set to a large value, divergence of the numerical solution occurs in the explicit scheme. To prevent this, the diffusion coefficient D is set to approximately 40–50% of the value of u. As a result, a diffusion coefficient larger than the actual value of 0.0000196 m²/s must be used. Although this introduces a side effect in which particles dissipate more rapidly from the location, the approximation is considered reasonable when the effect of wind velocity in advecting particles away is greater than the effect of diffusion. In situations where extremely large concentration gradients arise, this assumption breaks down, and implicit schemes would be required. In the present calculation, such situations were not considered, since the objective was to obtain the lower bound of the prediction.

By gradually increasing both the advection term u and the diffusion coefficient D, the decay portions following the first methane peak and the second methane peak can be reproduced. This indicates that the observed decrease in concentration is replicated by increasing the rate at which particles escape from the location.

Figure 17 shows the simulation results of methane concentration variation when the diffusion coefficient is set to D = 0.000796 m²/s and the advection velocity to u = 0.00196 m/s. The colors used are the same as those in Figure 16. The difference lies in the timing at which a concentration of 5 ppmC was added to the existing value at the release point: here, it was set at 16:15 (five minutes after the main shock and three minutes before the aftershock). As shown in the figure, the temporal variation in methane concentration at a position 3 m above the release point closely matches the first peak observed in the actual measurements in Figure 13.

When the advection velocity u is smaller than 0.00196 m/s, the simulated waveform produces a delayed peak compared to 16:42, and the post-peak decay becomes too gradual to reproduce the observed curve. When u is larger than this value, the decay tail becomes too steep; however, by continuously supplying methane emissions in a temporally sustained manner to compensate for the decrease, the observed curve can be reproduced. Therefore, in order to reproduce the actual measurements, a value of approximately u = 0.00196 m/s is required as the minimum necessary advection velocity.

Figure 18 shows the temporal variation in methane concentration at x = 0 m and x = 3 m under the same conditions as Figure 17, with additional methane releases at different times. At the release point (x = 0 m), concentrations of 8.4 ppmC at 16:54, 4.5 ppmC at 17:33, 2.5 ppmC at 18:05, and 1.8 ppmC at 18:36 were added. As shown in the figure, when the temporal variation in methane concentration at x = 0 m is given as indicated by the green line, the advected–diffused concentration at x = 3 m produces a waveform similar to the actual observations.

In Figure 13, the CH₄ increase appears to begin almost simultaneously with the first two aftershocks. However, when comparing the onset time of the green line at the ground surface (x = 0 m) with the aftershock times, the gas release seems to precede the aftershocks by several minutes. This result appears inconsistent with the hypothesis that the aftershocks triggered the gas release.

As the vertical advection velocity u is increased, the onset time of methane concentration at the ground surface (x = 0 m, green line) in Figure 17 approaches the release time of the aftershock (orange). When u is increased to 0.004 m/s and D to 0.002 m²/s, the timing of methane release at the ground surface occurs after the aftershock. In this case, the causal relationship becomes aftershock → surface gas release. However, this is not yet definitive, because subsurface gas diffusion precedes the release at the ground surface. In other words, when considering not only the time required for diffusion to reach 3 m above ground but also the subsurface transport time, it is suggested that diffusion may have begun prior to the occurrence of the aftershock. At present, two possibilities remain: one in which the aftershock occurs first, and another in which subsurface gas generation occurs earlier. The latter possibility has been proposed as a hypothesis in reference [5].

It should be emphasized that the above interpretation depends strongly on the assumed vertical advection velocity u. When u is relatively small, the inverse modeling can only be satisfied if methane is assumed to have been present at the ground surface several minutes before the aftershock. In contrast, when a larger value of u is assumed, the modeled time at which methane appears at the ground surface can shift to after the aftershock. However, this shift does not necessarily imply a reversal of the actual subsurface causal sequence, because the present model does not resolve subsurface transport and methane appearing at the ground surface at any time must have been generated in the subsurface beforehand. The only exception would be a scenario in which seismic shaking creates ground fissures that allow rapid venting of subsurface gas to the surface, in which case a causal sequence of aftershock-triggered methane release could be physically plausible.

At present, the depth at which methane generation initiated and the time required for the gas to migrate through the subsurface cannot be constrained. The present simulation assumes that methane release begins at the ground surface and then moves upward to the sensor height, which represents a limitation of this study. Therefore, both possibilities—aftershock-triggered gas release and earlier subsurface gas generation—remain viable within the range of plausible transport parameters.

In this context, it is noteworthy that Crews et al. [5] proposed a mechanism in which increases in excess pore-water pressure reduce the effective normal stress on a fault, thereby promoting aftershock activity. The inverse modeling presented here suggests that, for smaller values of the vertical advection velocity u, subsurface methane generation must have begun several minutes before the aftershocks, which is qualitatively consistent with the pore-pressure-driven mechanism proposed by Crews et al. Furthermore, one of the monitored wells exhibited a clear post-seismic rise in groundwater level, supporting the possibility that pore-pressure changes occurred in the subsurface.

However, because the subsurface origin and migration history of methane cannot be determined, an opposite causal sequence—aftershock-triggered methane release—also remains plausible when larger values of u are assumed. Thus, while the present observations do not allow us to determine the direction of causality, they are consistent with a scenario in which methane release and aftershock activity are both manifestations of a common subsurface process initiated by the main shock.

4.3. Possibility of Biogenic Methane and Shallow Organic Sources

In addition to thermogenic and deep-reservoir methane, the possibility of biogenic methane generated in shallow organic-rich sediments must also be considered. As described in Section 3.2.1, the coastal and river-mouth environments around the Nanao station contain alternating layers of silt, clay, and organic material that are favorable for microbial methanogenesis. Earthquake shaking could disturb these shallow sediments, enhance microbial activity, or release methane accumulated beneath low-permeability clay layers. Although the rapid and large-amplitude CH₄ anomalies observed at the Nanao Observation Station are more consistent with sudden degassing from subsurface gas pockets, shallow biogenic sources cannot be fully excluded and may have contributed to part of the observed signal.

4.4. Limitations

This study has several limitations. First, the methane data are preliminary values from the Soramame-kun system, although their reliability was independently verified through high-resolution measurements and instrument checks. Second, groundwater-level data are available only at monthly resolution, preventing detailed temporal correlation analysis with methane anomalies. Third, observations from some stations, such as Wajima, were lost due to earthquake-induced power outages, reducing spatial coverage. Fourth, the one-dimensional advection–diffusion model used here provides a lower-bound estimate of transport behavior and does not resolve subsurface migration pathways. Finally, the relative contributions of deep thermogenic gas, shallow biogenic gas, and earthquake-induced degassing cannot be fully separated with the available data. These limitations should be addressed in future studies using higher-resolution groundwater monitoring, isotopic analysis of methane, and three-dimensional atmospheric dispersion modeling.

5. Conclusions

This study has aimed to demonstrate the presence of a pronounced coseismic increase in atmospheric methane concentrations during the 2024 Noto Peninsula Earthquake and to examine whether this increase may have originated from underground natural gas release. By analyzing hourly CH₄ data from the Ministry of the Environment’s monitoring network, this study has shown that significant methane increases occurred only in areas with seismic intensity of 6− or greater, and that an exceptional anomaly—reaching 29 times the standard deviation of the past year—was recorded at the Nanao Observation Station. The validity of this anomaly has been confirmed through consultation with a local atmospheric officer, and high-time-resolution data (6 min values) verified continuous instrument operation. Detailed analysis further showed that two major methane peaks occurred, each rising not immediately after the main shock but synchronously with two large aftershocks approximately 8 and 44 min later. Geological and hydrogeological information indicated the presence of water-soluble gas and unsaturated hydrocarbons beneath the Nanao region, suggesting that seismic shaking may have ruptured clay layers and released accumulated gas. Analyses of public reports and interviews with local officials showed that alternative explanations—such as fire smoke, pipeline rupture, instrument malfunction, and gas-cylinder damage—were unlikely. These findings indicate that the observed methane anomaly was most likely caused by earthquake-synchronous underground gas release, suggesting that methane-release risk should be considered in post-earthquake fire-hazard assessments. An important implication of this study is that the observed methane anomaly was detected by an atmospheric monitoring system that was never designed for subsurface gas surveillance. The fact that such a system captured a clear coseismic signal suggests that existing environmental monitoring networks may contain untapped information relevant to subsurface processes. This highlights the value of integrating heterogeneous datasets—atmospheric observations, groundwater levels, seismic records, and wind measurements—to investigate earthquake-related gas-release phenomena. Future interdisciplinary collaboration will be essential to fully explore these possibilities. Future research should build upon the findings of this study by incorporating more specialized observational and analytical approaches. In particular, (1) continuous high-resolution groundwater monitoring, (2) isotopic and molecular analyses of methane to distinguish its origins, (3) three-dimensional atmospheric dispersion modeling, and (4) denser spatial deployment of methane sensors in the Noto region would allow a more detailed understanding of coseismic gas-release processes. Such investigations will require collaboration among geologists, geochemists, atmospheric scientists, and monitoring agencies.