地震はより頻繁に起きているのか?

The Myth: Earthquakes Are Becoming More Frequent

After any active period of seismicity — a run of large earthquakes around the Pacific 環太平洋火山帯世界の地震のおよそ90%が発生する、太平洋を取り囲む馬蹄形の地帯。総延長は40,000kmに及び、452の火山を含む。, a cluster of damaging events in weeks or months — media coverage inevitably raises the question: are earthquakes becoming more frequent? Social media amplifies this perception further, with crowdsourced reports and rapid news cycles making every significant earthquake globally visible within minutes of occurrence. The feeling that "there seem to be a lot of earthquakes lately" is real and understandable. But the scientific record tells a different story.

What the Earthquake Record Actually Shows

The USGS maintains comprehensive statistics on global earthquake frequency. For large earthquakes — magnitude 7.0 and above — the global average has been remarkably consistent at approximately 15-20 events per year over the period of modern seismic monitoring. For magnitude 8.0 and above, the average is about one per year. Year-to-year variation is substantial due to the random nature of earthquake occurrence, but there is no statistically significant long-term trend toward increasing frequency when examined over the full instrumental record.

The 世界地震観測網（GSN）世界の地震活動を包括的に監視する、150以上の広帯域地震観測点からなる世界規模のネットワーク。USGS・NSF・IRISが共同で運用している。, which became truly global and standardized in the 1960s, provides the most reliable comparison baseline. Analyzing this record using the グーテンベルク・リヒター則地震の頻度とマグニチュードの関係を示す統計法則で、マグニチュードが1単位増えるごとに、地震の発生頻度はおよそ10分の1になる。 relationship, seismologists find that the frequency-magnitude statistics have remained stable. There are active years and quiet years, but no systematic upward trend for large earthquakes.

The Detection Bias Problem

For smaller earthquakes, the picture is genuinely different — but not because they are more common. The number of recorded small earthquakes has increased enormously over the past several decades, but this reflects dramatic improvements in 地震観測網地震活動を継続的に監視する、連携した地震観測点の集合。世界地震観測網（GSN）は150以上の観測点から構成され、世界規模の観測範囲を提供する。 coverage and sensitivity rather than actual increases in earthquake occurrence. In 1960, the global seismic network consisted of a few hundred stations, many of them old instruments with limited sensitivity. Today, thousands of high-quality broadband stations operate worldwide, supplemented by dense regional networks and even smartphone-based crowdsourced sensing.

More stations with better sensitivity detect smaller and more distant events. A magnitude 2.0 earthquake in a previously unmonitored region that would have been invisible in 1970 is now routinely recorded and catalogued. This creates an apparent increase in earthquake frequency that is entirely an artifact of improved monitoring. When analysts apply magnitude completeness thresholds — focusing only on events large enough to have been reliably detected throughout the historical record — the frequency trends flatten out.

Normal Statistical Clustering

Even when earthquake frequency is steady on average, the actual occurrence of earthquakes is not evenly distributed in time. Earthquakes cluster. A major earthquake increases stress on surrounding fault systems, triggering 余震本震と同じ断層領域で本震の後に続く、より規模の小さい地震。余震活動は数週間から数年続くことがあり、最大の余震は通常本震よりマグニチュードが1.0〜1.2小さい。 sequences that can last years and occasionally produce M6+ events. When a major subduction zone ruptures, it may set off a sequence of large events on adjacent fault segments over a period of years to decades. The 2004 Indian Ocean earthquake was followed by the 2005 Nias earthquake, the 2007 Bengkulu earthquakes, and other large events on the Sunda subduction zone. This is seismically normal behavior, not an escalating trend.

Conversely, periods of relative global seismic quiet also occur and are equally unremarkable. The randomness inherent in earthquake processes means that clusters and gaps are expected features of any earthquake catalog, not signals of fundamental change.

Why It Feels Like More

Several cognitive and social factors make it seem like earthquakes are becoming more common even when they are not. Global news coverage of earthquakes has increased enormously since the advent of the internet and 24-hour news cycles. An earthquake that would have received a two-paragraph wire service notice in 1985 now generates hours of live television coverage, thousands of social media posts, and real-time data visualizations. The 「揺れを感じましたか？」（DYFI）地震後に一般市民から震度に関する報告を収集し、市民参加型の震度マップを作成するUSGSのプログラム。地震を感じた誰もが報告を投稿できる。 system alone has sensitized millions of people to earthquakes they would previously have ignored.

Population growth in earthquake-prone regions also means more people are affected by earthquakes of a given マグニチュード地震が放出した総エネルギー量を表す単一の数値。整数値が1増えるごとに、放出エネルギーはおよそ31.6倍になる。, generating more news coverage and more personal accounts. The 1906 San Francisco earthquake struck a city of 400,000; the same rupture today would affect a metropolitan population of 4.7 million. The earthquake itself would be identical in physical terms, but its human visibility would be vastly greater.

The Exception: Induced Seismicity

There is one genuine exception to the "no trend" finding: 誘発地震活動水圧破砕（フラッキング）、排水注入、採掘、貯水池の湛水など、人間活動によって引き起こされる地震。ほとんどは小規模（M4未満）だが、M5.5を超えたケースもある。 — earthquakes caused by human activities such as wastewater injection from oil and gas operations. In Oklahoma, for example, earthquake rates jumped dramatically between 2009 and 2015, correlating closely with the expansion of wastewater disposal wells from hydraulic fracturing operations. This was a real increase in earthquake frequency in a specific region, caused by a specific human activity, and was documented by detailed analysis of seismic catalogs and injection well records. Oklahoma went from about 1-2 M3+ earthquakes per year historically to over 900 in 2015. Regulatory interventions reducing injection volumes subsequently decreased earthquake rates.

This is localized industrial seismicity, not evidence of a global trend, but it represents a genuine documented increase in earthquake frequency that serves as a reminder that the "no trend" conclusion applies to tectonic earthquakes, not all seismicity.

What Seismologists Actually Monitor

Rather than tracking frequency trends, seismologists focus on understanding fault systems well enough to estimate the probability of future large events on specific faults. The 確率論的地震ハザード解析（PSHA）起こりうるすべての地震の発生源・マグニチュード・地震動レベルを考慮し、特定の揺れの水準を超える確率として結果を表現する、地震ハザードの定量評価手法。 (probabilistic seismic hazard analysis) framework integrates fault geometry, すべり速度断層に沿った変位の平均速度で、通常は年間ミリメートル単位で測定される。すべり速度が高いほど、一般に地震発生頻度とリスクが高いことを示す。 estimates, recurrence interval data from 古地震学断層トレンチ、隆起した段丘、津波堆積物といった地質学的証拠を通じて、先史時代の地震を研究する学問。地震の記録を数千年前まで遡らせる。, and グーテンベルク・リヒター則地震の頻度とマグニチュードの関係を示す統計法則で、マグニチュードが1単位増えるごとに、地震の発生頻度はおよそ10分の1になる。 statistics to produce probabilistic forecasts. These forecasts are the scientific foundation for 耐震基準建物の最低限の耐震安全性を確保するための、設計・建設に関する法的要件の体系。大地震で新たな脆弱性が明らかになるたびに更新される。 requirements, 地震ハザードマップ特定の期間内に地震の揺れが指定の水準を超える確率を示した地図。技術者・都市計画者・保険会社が地震リスクを評価するために用いる。 products, and long-term risk planning — and they rely on the stationarity of seismicity rates, not on detecting trends.

The Bottom Line

Earthquakes are not becoming more frequent in any meaningful scientific sense for the earthquakes that matter most — large magnitude events driven by tectonic forces. The perception of increasing frequency is a product of better detection, greater media coverage, and growing population in hazard zones. Understanding this should provide neither false reassurance (the hazard remains real and serious) nor unnecessary alarm. The scientifically appropriate response is continued investment in monitoring infrastructure, seismic hazard research, and 耐震設計地震力に耐えられるよう構造物を設計する手法。現代の耐震設計は、大地震における構造被害はある程度許容しつつ、倒壊の防止と人命の保護を目指す。 standards rather than anxiety about an apparent trend that does not exist.