论文部分内容阅读
对某些主震事件来说,地震预测也许是可能的。Varnes(1989),Bufe和Varnes(1990)提出的时间—破坏法利用前兆事件(前震)来确定能量加速释放曲线。通过公式与资料的拟合,可计算出预测破坏时间和震级。到目前为止,该方法已应用于构造活动区的仅有几项研究中,而且是针对中强主震事件。新马德里地震带(NMSZ)的微震台网资料对于研究区来说,其震级≥1.5的地震资料是相当完整的,使用这种资料,用该方法对过去小到m_b=3.5的事件进行了预测。本次估算所用的新马德里地震带台网资料覆盖的时间段为1974年6月29日至1995年7月20日。在此21年间(这一时期台网一直在运行),共发生震级≥3.5的地震36个。应用时间—破坏方法需要前兆事件,而1980年前发生的主震没有足够的资料来充分地确定能量加速释放曲线。因此,我们利用1980年以后发生的震级≥3.5的26个地震,并对其中的16次主震事件进行了模拟。在大多数情况下,由前兆序列得到的预测破坏时间和震级与实际主震的值相当接近。余下的主震,包括1980年以前发生的那些事件,由于(1)没有足够的事件来充分确定前兆序列或(2)存在干扰能量加速释放曲线的干扰事件,而不能进行模拟。此外,使用Nuttli目录(Nuttli,1979)模拟了两次事件,还将两个目录合并模拟另一个事件。用时间—破坏法对19个震级≥3.5的地震进行了演算。用时间—破坏法最初计算出的预测结果具有较大的误差范围,所预测的震级基本无上限。参数之间的一个经验关系帮助限制了所预测震级的范围,同时,在较小的程度上限制了估算的破坏时间。该经验公式可改进时间—破坏方程,并获得具有上限的震级预测值。另一个经验关系说明主震矩的对数随前兆事件搜索直径的对数而线性地增大。该区域的相对地震活动性同样会影响用于寻找前兆事件的最佳搜索直径。除了估算与主震有关的加速序列外,还要对无震加速序列(不会发生主震的加速序列)的产生进行评估。通过随机选择潜在主震位置,进行了无震加速序列的初步分析。结果获得无震加速序列的发生率为2%。将经验关系与时间—破坏法相结合,新马德里地震带的未来中期地震预测的前景是鼓舞人心的。
Earthquake prediction may be possible for some mainshock events. The time-damage method proposed by Varnes (1989), Bufe and Varnes (1990) uses the precursor events (pre-shock) to determine the energy-acceleration release profile. By fitting the formulas and data, the predicted damage time and magnitude can be calculated. So far, this method has been applied to only a few studies on the construction of activity zones, but also to the mid-strong mainshock events. For the study area, seismic data with a magnitude of ≥1.5 are quite complete. Using this data, we predicted the past events as small as m_b = 3.5 using this data . The estimated period covered by the network data of the New Madrid Seismic Belt covered by this survey is from June 29, 1974 to July 20, 1995. During this 21 years (during this period the Taiwan network has been running), a total of 36 earthquakes ≥ 3.5 earthquake occurred. The application of the time-destruction method requires precursory events, whereas the mainshock that occurred before 1980 does not have sufficient information to adequately determine the energy-acceleration release curve. Therefore, we used 26 earthquakes ≥3.5 after 1980 and simulated 16 of them. In most cases, the predicted damage time and magnitude from the precursor sequence are fairly close to the actual mainshock value. The remaining mainshocks, including those that occurred prior to 1980, can not be simulated because (1) there are not enough events to adequately determine the precursor sequence, or (2) there is an interference event that accelerates the release of the disturbing energy curve. In addition, two events were simulated using the Nuttli directory (Nuttli, 1979) and the two directories were merged to simulate another event. 19 earthquakes ≥3.5 were calculated using the time-damage method. The prediction result originally calculated by the time-destruction method has a large error range, and the predicted magnitude is basically no upper limit. An empirical relationship between the parameters helps to limit the extent of the magnitude predicted and, to a lesser extent, the estimated time of destruction. This empirical formula improves the time-damage equation and obtains a magnitude forecast with an upper bound. Another empirical relationship shows that the logarithm of the principal moments increases linearly with the logarithm of the diameter of the precursory event search. The relative seismicity of the area also affects the optimal search diameter for finding precursor events. In addition to estimating the acceleration sequence associated with the mainshock, the generation of the quake-free acceleration sequence (accelerating sequence where the mainshock will not occur) is also evaluated. Through the random selection of the potential mainshock location, a preliminary analysis of the earthquake-free acceleration sequence was carried out. Results The rate of no-shock acceleration was 2%. The combination of empirical relationships with the time-destruction approach gives the prospect of a prospective mid-term earthquake prediction in the New Madrid seismic zone encouraging.