Previously, inadequate earthquake catalogs and the lack of method made it challenging, if not impossible, to determine the dominant period of earthquake occurrence. With the advent of densely distributed seismic stations in Japan since 2002, 410 642 M1.0+ earthquakes have been cataloged under the mainland within 30-km depth, which provide a superb dataset to identify the periodicity of earthquakes. We processed this catalog using time-frequency analyses and daily stacking, which are powerful in extracting weak periodical signal from strong random noises. We concentrated on the time sector from 0:00 to 6:00 (i.e. from midnight to daybreak), which is a reliable time window for much higher detectability of weak earthquakes, since it has the lowest affects from cultural noises. We successfully observed two apparent periodicities of 12- and 24-hour, which are much smaller than the fortnightly periodicity presented previously in the literature. Synthetic earth tides, after intentionally ignoring the contribution from the Moon, present similar dominant periods as the earthquakes. This may indicate that the dominant period of earthquakes is statistically associated with the Sun rather than the Moon. The daily stacking number of earthquakes using a 15-minute or 1-hour interval shows a peak around 1:30, rather than usually expected 3:00 to 4:00. In addition, bigger earthquakes show more evident variations in the stacking results, and the trend is very consistent for various lower limits of earthquake magnitude from M1.0 to M4.0. These discoveries settled the disputes on the existence of the periodicity of earthquakes since 1886 and may open a window to unravel the mystery of earthquakes.

中文翻译：

---

地震从午夜到黎明的昼夜周期性的证据

以前，由于地震目录不完整和方法的缺乏，确定地震发生的主要时期具有挑战性，甚至是不可能的。随着日本自 2002 年以来密集分布的地震台站的出现，在大陆下 30 公里深度内已编录了 410 642 次 M1.0+地震，为识别地震的周期性提供了极好的数据集。我们使用时频分析和每日叠加来处理这个目录，这在从强随机噪声中提取弱周期信号方面非常强大。我们专注于从 0:00 到 6:00（即从午夜到黎明）的时间段，这是一个可靠的时间窗口，可以提高弱地震的可探测性，因为它受文化噪音的影响最小。我们成功地观察到了两个 12 小时和 24 小时的明显周期性，这比以前文献中提出的每两周一次的周期小得多。人造地球潮汐在有意忽略月球的贡献后，呈现出与地震相似的主导时期。这可能表明地震的主要时期在统计上与太阳有关，而不是与月亮有关。使用 15 分钟或 1 小时间隔的每日地震叠加次数在 1:30 左右出现峰值，而不是通常预期的 3:00 至 4:00。此外，地震越大，叠加结果的变化就越明显，从M1.0到M4.0的各个震级下限变化趋势非常一致。这些发现解决了1886年以来关于地震是否存在周期性的争论，可能为揭开地震之谜打开一扇窗。