前震:能预测主震吗?

What Are Foreshocks and How Are They Identified?

A 前震在同一地区先于主震发生的地震。前震只能在事后被识别——目前尚无可靠方法能事先将其与普通地震区分开来。 is a smaller earthquake that occurs before a larger event on the same 断层（地质学）岩石沿其发生位移的破裂面。断层长度从数毫米到数千公里不等。会引发地震的主要断层称为活动断层。 or fault system, in the same geographic area. The word "foreshock" is inherently retrospective — an earthquake can only be identified as a foreshock after the larger 主震一次地震序列中震级最大的地震，决定了该事件的整体震级。此前有时会出现前震，之后必然伴随余震。 has occurred. In real time, any moderate earthquake is simply an earthquake; whether it will be followed by something larger is unknown.

This retrospective definition creates a fundamental challenge for earthquake 地震预报与地震预测的区别地震预报是指明确指出未来地震的确切时间、地点和震级——目前尚无法实现。地震预测则是对某一时间段内地震发生可能性的概率性估计。. Every earthquake is a potential foreshock. The vast majority are not — they are simply isolated events or the mainshocks of their own sequences. Only by waiting to see whether a larger event follows can scientists classify an earthquake as a foreshock. This ambiguity is not a deficiency of current science; it reflects a genuine physical uncertainty about whether any given fault rupture will trigger further failure on adjacent sections of the same fault.

The Statistical Reality: Only 5-10% of Earthquakes Have Foreshocks

Statistical studies of large earthquake catalogs consistently find that only about 5 to 10 percent of significant earthquakes are preceded by recognisable foreshocks in the days immediately preceding them. This means that for the remaining 90–95 percent, no elevated seismic activity precedes the mainshock that would have allowed warning. Even in sequences with foreshocks, the foreshocks often occur only hours before the mainshock — too brief a time for meaningful evacuation of large populations.

The b-value analysis of 地震丛集现象地震倾向于以丛集形式（主震—余震序列或地震群）而非在时间上随机发生的特性，与地震独立随机发生的常见假设相悖。 in the months and years before major events sometimes reveals subtle statistical signals, but translating these into actionable warnings has proven extremely difficult. The 地震观测网由若干地震台站协同组成、持续监测地震活动的系统。全球地震台网（GSN）拥有150多个台站，提供全球范围的观测覆盖。 data that would reveal such patterns in real time requires dense station coverage, sophisticated algorithms, and — most critically — a reliable way to distinguish "this is a foreshock sequence" from "this is just normal background seismicity." That reliable distinction does not currently exist.

Famous Foreshock Sequences: 2011 Tohoku and 1975 Haicheng

The 2011 Tohoku earthquake was preceded by a magnitude 7.3 earthquake two days earlier, on March 9. At the time, this event was treated as a significant but self-contained earthquake — not as a warning of the Mw 9.0 catastrophe to follow. In retrospect, it was a foreshock, but nothing about its character unambiguously marked it as one. Seismologists later identified a subtle increase in small earthquakes in the rupture zone in the weeks before March 11, but this increase was not dramatic enough to trigger unusual concern in real time.

The 1975 Haicheng earthquake in China is the most famous case where 前震在同一地区先于主震发生的地震。前震只能在事后被识别——目前尚无可靠方法能事先将其与普通地震区分开来。 activity actually led to an evacuation. In the weeks and days before the February 4, 1975 magnitude 7.0 earthquake, unusual swarms of small 地震群在数天至数月内发生于局部区域、且无明显主导主震的一系列地震，常与火山活动或流体注入相关。s occurred in the region. Local officials, acting on these observations along with ground deformation and animal behaviour reports, ordered an evacuation of the city hours before the mainshock struck. The evacuation is estimated to have saved tens of thousands of lives. However, this success was partly serendipitous — the foreshock sequence was unusually clear, and the following year, the magnitude 7.8 Tangshan earthquake struck with no warning and killed approximately 250,000 people.

Why Prediction vs Forecasting Remains Unsolved

The distinction between earthquake 地震预报与地震预测的区别地震预报是指明确指出未来地震的确切时间、地点和震级——目前尚无法实现。地震预测则是对某一时间段内地震发生可能性的概率性估计。 is critical. Prediction implies a specific statement: "An earthquake of at least magnitude X will occur within Y kilometres of location Z within T days." Forecasting is probabilistic: "The probability of an earthquake of magnitude 6+ in this region during the next month is 3 percent, elevated from a background of 0.5 percent." Current science can forecast; it cannot predict.

The inability to predict 主震一次地震序列中震级最大的地震，决定了该事件的整体震级。此前有时会出现前震，之后必然伴随余震。s from 前震在同一地区先于主震发生的地震。前震只能在事后被识别——目前尚无可靠方法能事先将其与普通地震区分开来。 activity stems from a fundamental physical problem. Fault systems are governed by highly nonlinear dynamics — small differences in initial conditions (stress distribution, fluid pressure, fault roughness) can lead to dramatically different outcomes. A small stress perturbation that produces a foreshock and triggers a great earthquake in one case might produce only an isolated small event in another. The fault system does not "know in advance" that it will produce a great earthquake; the outcome depends on minute details of the stress field that cannot be measured at adequate resolution.

The Haicheng Success Story: Lucky or Skillful?

The 1975 Haicheng evacuation is often cited as proof that earthquake prediction is possible. A closer examination suggests the story is more complicated. The key foreshock activity in Haicheng was unusually prominent — a swarm of hundreds of small to moderate events in the days before the mainshock. Many scientists consider this level of precursory activity atypically clear. Additionally, the evacuation decision involved multiple factors beyond seismology: ground deformation measurements, water level changes in wells, and — controversially — reports of unusual animal behaviour.

Most importantly, the Haicheng success has not been reproducible. The 1976 Tangshan earthquake, which killed perhaps ten times as many people as Haicheng would have without evacuation, produced no recognisable foreshock activity. The 1994 Northridge earthquake and the 1995 Kobe earthquake, both devastating urban events, had no significant foreshocks. Globally, large earthquakes with clear foreshock sequences are the exception; earthquakes with no precursory seismicity are the rule. The scientific consensus remains that short-term deterministic earthquake prediction — predicting the time, location, and magnitude of a specific future mainshock — is not currently achievable.