人工神经网络在地震中期预报中的应用

将ＢＰ神经网络用于地震中期预报。使用一些常用的地震学指标作为神经网络的输入,而将ＢＰ神经网络的输出作为表征地震活动增强的特征参数Ｗ１,并将其用于华北地区进行空间扫描。结果表明,中强地震前１～３年未来震中周围通常出现明显的Ｗ１值中期异常区,该方法具有较好的中期预报效果。     

人工神经网络在地震中短期预报中的应用

本文将ＢＰ神经网络用于地震中短期预报。作者把一些常用的地震学指标作为神经网络的输入,而将ＢＰ神经网络的输出作为表征地震活动平静的特征参数Ｗｑ,井将其用于华北地区进行空间扫描,结果表明中强地震前１年左右或稍长时间,未来震中周围一般都开始出现Ｗｑ值的中短期异常区,证明本方法具有限好的中短期预报效果。     

小震活动增强图像与1996年新疆两次强震的中期预报

阐述了小震活动增强图像的物理基础,资料分析处理方法和地震活动水平等级划分标准、预报判据及回顾性检验结果。还介绍了１９９６年新疆阿图什６．７级和喀喇昆仑７．１级２次强震的中期预报过程。震例分析表明;多数目标地震发生前１－２年内出现区域小震活动增强异常图像,若取异常结束后１２个月作为预报时段,则异常对频率为０．３５,有震报准率为０．７０,通过Ｒ值检验,中期预报效果较好。     

新疆地震年度趋势预报效能的统计评价

介绍了统计检验的基本原理和方法及地震预报效能的统计评分方法，并据此对新疆维吾尔自治区地震局年度趋势会商预报的效能进行了统计评价。结果表明，从总体上说地震中期(年度)预报能通过检验，预报评分值约为0．25-0．3，显示出新疆地震局年度会商会的预测结果对地震中期预报有一定能力。由于各年的评分涨落很大，不宜依据一、二年的评分值来说明预报能力或水平的提高或下降。     

人工神经网络在地震综合预报中的应用

多级判别、聚类分析等统计模式识别方法在地震综合预报中的应用已有多篇文献报道(石绍先、沈斌,1986;1990;王学仁,1982).多年预报实践表明,模式识别方法对于地震预报是有效的.随着国际上神经网络研究的兴起,该方法也被用于模式识别领域,但尚未见其用于地震综合预报方面.本研究运用“反向传播”神经网络,对滇西南地震区17个震例进行了计算机分析,建立了地震综合预报的专家系统,从而为进一步研究决定地震强度的物理因素,准确地预报将来的未知地震提供参考信息.     

地震拟合优度在青海中强地震预报中的应用

运用地震拟合优度的计算公式,对青海省中东部的6个ME≥6．0地震从震前数年开始进行地震拟合优度计算,发现这6个地震震前一段时间地震拟合优度都有一定程度的下降。再对共和7．1级地震震中附近一定区域内的地震进行拟合优度计算,进而提取震中周边300kmME≥5．8以上地震的中期预报指标。并对这一指标进行预报水平评分。结果表明,这一方法对共和地震震中胤边300km ML≥5．8以上地震的中期预报有一定的应用价值。另外,对拟合优度计算参数的选取进行分析,确定了最优的计算参数。     

1996年新疆两次强烈地震

１９９６年新疆两次强烈地震戈澍谟（新疆维吾尔自治区地震局，乌鲁木齐８３００１１）主题词：新疆强震中期预报长期预报１９９６年是新疆地区地震强烈活动的一年。在新疆境内发生了两次强烈地震，它们是３月１９日阿图什６．７级地震和１１月１９日和田西南的７．１级地...     

遗传BP网络在地震和爆破识别中的应用

将遗传算法(GA)和反向传播算法(BP算法)相结合成为GA-BP算法，以此建立了遗传BP神经网络．并将以BP算法为基础的BP神经网络及以GA-BP算法为基础的遗传BP神经网络用于对地震和爆破的识别中．得到的结果表明:遗传BP网络比BP网络对事件的识别能力略好些．      

地震学资料的综合性中期预报

论文简介了用于中期（1～2年）地震预报的一些地震学方法。其中有重复率曲线斜率γ值（译者注：即b值）、弱震活跃—平静法（RTE）、纵横波波速比（VP／VS）方法、中等地震（M≥3．0活跃—平静法及据此得到的前兆显示时间与震级之间的经验关系式、地震活跃迁移法及据此得到的震中距（R）及前兆持续时间（△T）与震级的经验关系式。文章最后给出了综合适用上述方法编制的最近1～2年内哈萨克斯坦及邻区（吉尔吉斯斯坦等）可能发生5＜M＜6．5级地震的中期预报危险区图。     

对地震危险区的探索——以南黄海5．3级地震等预报为例

1991年11月5日在江苏射阳发生了4．7级地震，1992年1月23日在南黄海北部发生了5．3级地震，10月22日射阳东海域发生了4．8级地震，11月4日南黄海北部发生了4．6级地震，对上述4次地震笔者等曾作出了一定程度的年度中期预报。本文对该地震震前的预报概况，预报基本思路和主要科学依据作了概述，并讨论了给予的启示。认为在一定条件下对某些地震作出某种程度的科学预报还是可能的。     

地震预报之谜

2010年,普林斯顿大学出版社（Princeton University Press）出版了Susan Hough撰写的Predicting the Unpredictable——The Tumultuous Science of Earthquake Prediction一书,现将Christopher H Scholz对该书的评介翻译介绍给读者。     

The application of neural networks to comprehensive prediction by seismology prediction method

IntroductionThe theory of artificial neural netWorks has been used in some fields for recent years such asearthquake damage prediction (Shi, Liu, 1991), earthquake intensity (Wang, 1993), earthquakecomprehensive prediction (Wang, Dai, 1997), and so on. The initial Studies indicate that someresults are prevail over classical statistical pattern recognition and fuZZy recognition methods.Neural network system is a high adaptive nonlinear dynamical system. It can extract causalitythrough a ple…     

On numerical earthquake prediction

Can earthquakes be predicted? How should people overcome the difficulties encountered in the study of earthquake prediction? This issue can take inspiration from the experiences of weather forecast. Although weather forecasting took a period of about half a century to advance from empirical to numerical forecast, it has achieved significant success. A consensus has been reached among the Chinese seismological community that earthquake prediction must also develop from empirical forecasting to physical prediction. However, it is seldom mentioned that physical prediction is characterized by quantitatively numerical predictions based on physical laws. This article discusses five key components for numerical earthquake prediction and their current status. We conclude that numerical earthquake prediction should now be put on the planning agenda and its roadmap designed, seismic stations should be deployed and observations made according to the needs of numerical prediction, and theoretical research should be carried out.     

Regional daily baseflow prediction

The purposes of this study are to identify the bias of applying the analysis of a log–log plot of baseflow and to derive an equation to describe successive regional mean baseflow. The function ?dQ/dt = a Q^b has been used to describe baseflow in many studies that obtain the values of a and b from the log–log plot. According to analysis in this study, the value of 1 can be assigned to b in two boundary conditions, but the parameter a is proved to be related to the depth of water table and starting time of recession and thus different values of a may be found for different recession events. This paper points out that no single regression line can be obtained by plotting all baseflow data on a log–log diagram. Instead, there should be parallel lines, and each for a recession event. It implies that no single set of parameters a and b can be applied to predict baseflow. Thus, a new equation describing the relationship between three successive mean baseflows was derived in this study. The bias in the analysis of the log–log plot and the ability of the derived equation to predict baseflow were verified for five watersheds in Taiwan. Results indicate that the formula of mean baseflow prediction can provide reasonable estimates of flows with a leading time of 6 days. Furthermore, stream flows of the Tonkawa creek watershed in USA were used to verify that using average flows can result in better predictions than using instantaneous flows. Copyright © 2004 John Wiley & Sons, Ltd.     

同位素地球化学在地震研究方面的作用

主要论述流体同位素地球化学的基本原理及其在地震活动过程中的变化特征、异常判识的同位素地球化学方法及其在地震预测方面的应用。同位素地球化学方法在地震监测预报方面有重要的作用，可以用来判定流体的物质来源和循环过程、判断流体异常的原因、建立古地震和历史地震活动的时间序列、推断地震活动规律、预测未来可能发生的地震。研究认为，目前在地震监测研究方面除^222Rn和^14C应用普遍之外，其他同位素地球化学方法应用较少，从而使得许多异常难以判定，许多古地震得不到较准确的年龄。     

寂静的前震与地震预测

寂静地震的研究具有重要的理论与实际意义。统计表明，与寂静地震有关的前驱波比前震有更大的普遍性，出现的时间主要集中在主震前7天之内，从一部分中等地震到特大地震之前都已观测到前驱波，这就说明大的脆性破裂之前，较小的缓慢破裂在自然界是存在的而且比较普遍。根据1969年渤海地震和1976年唐山地震的前驱波记录进行定位试验，证明在观测条件具备时，概略的定位是可行的，而且表明前驱波是来源于大震震源附近的信号。可为地震预报特别是短临预报提供重要的前兆信息。     

地震预报与地震预警

在回顾了地震预报历史沿革的基础上,认为虽然地震预报的难度比较大,但成功地预报地震不是不可能的.地震预警系统已被实践证明是一套有效的避震措施,将在未来的防灾减灾工作中发挥巨大作用,但地震预警并不能代替地震预报.     

Testing an earthquake prediction algorithm

A test to evaluate earthquake prediction algorithms is being applied to a Russian algorithm known asM8 TheM8 algorithm makes intermediate term predictions for earthquakes to occur in a large circle, based on integral counts of transient seismicity in the circle. In a retroactive prediction for the period January 1, 1985 to July 1, 1991 the algorithm as configured for the forward test would have predicted eight of ten strong earthquakes in the test area. A null hypothesis, based on random assignment of predictions, predicts eight earthquakes in 2.87% of the trials. The forward test began July 1, 1991 and will run through December 31, 1997. As of July 1, 1995, the algorithm had forward predicted five out of nine earthquakes in the test area, which success ratio would have been achieved in 53% of random trials with the null hypothesis.     

An empirical earthquake prediction model

We have monitored seismic activity induced by impoundment of Lake Jocassee in northwest South Carolina for about two years. Low-level shallow activity was recorded. The larger felt events (2.0 ? M_L ? 2.6) were found to be associated with precursory changes in one or more of the following; number of events, t_S/t_p ratio values and radon concentrations in groundwater.The microearthquakes in the precursory period were accurately located in time and space, and their location pattern was used to develop an empirical earthquake prediction model.The precursory period consists of two phases; α-phase or a period of slow (or no) increase in seismicity, and β-phase, a period when the activity increase is more rapid. The main shock was found to be located within a cluster, a “target” area defined by the location of events in the β-phase. There is a general absence of seismic activity in the “target” area in the α-phase. The main shock occurred soon after a period of quiescence in the seismic activity in the β-phase. The magnitude of the shock, M_L is given by: M_L = 2 log D ? 0.07, where D is the duration of the precursory period in days.The model was successfully tested with data for a magnitude 2.3 event on February 23, 1977 which was also accompanied by radon and t_s/t_p anomalies.