Magnitude distribution complexity and variation at The Geysers geothermal field
Magnitude distribution complexity and variation at The Geysers geothermal field
复制标题
间歇泉地热田强度分布的复杂性和变化
DOI：
10.1093/gji/ggaa208
复制
发表时间：
2020
影响因子：
2.8
通讯作者：
K. Leptokaropoulos
中科院分区：
文献类型：
--
作者：
K. Leptokaropoulos
地震震级（大小）分布是地震危险性评估的重要组成部分，因此，准确确定其函数形状和变化是一项至关重要的任务。虽然经常被认为是固定的，但在特定地点的震级分布可能会随着时间和空间的变化而显着变化。在这项研究中，著名的古腾堡-里希特（GR）的法律，这是被广泛认为是描述地震震级分布，测试的情况下，在间歇泉（CA，美国）地热田的流体注入引起的地震活动研究。统计测试的开发和应用，以表征一个高品质的目录，包括地震活动直接与两个注入威尔斯，在西北部的间歇泉的震级分布。事件的大小分布变化的空间，时间，流体注入和震级截止标准进行了研究。一个彻底的时空分析进行定义地震活动集群展示特征的震级分布显着不同于附近的集群。整个地震人口的震级分布，以及个别集群的复杂性进行了测试，在指数，多模态和多凸结构。然后，进一步处理所识别的聚类，并结合喷射速率波动来确定它们的特性。分析的结果清楚地表明，整个震级分布肯定是复杂的和非指数的，而随后的时期表现出显着不同的震级分布被确定。区域地震活动人口被分为三个主要的家庭，其中之一的震级分布的指数明显被拒绝，而对于其他两个GR法b值是成正比的流体注入。此外，这些家庭的b值似乎是显着的震级依赖，这是一个事实，是非常重要的地震危险性评估的实施。总之，强烈建议，震级指数必须进行测试之前进行任何b值计算，特别是在人为地震活动的情况下，复杂的和时间可变的过程发生。
Earthquake magnitude (size) distribution is a major component required for seismic hazard assessment and therefore, the accurate determination of its functional shape and variation is a task of utmost importance. Although often considered as stationary, the magnitude distribution at particular sites may significantly vary over time and space. In this study, the well-known Gutenberg–Richter (GR) law, which is widely assumed to describe earthquake magnitude distribution, is tested for a case study of seismicity induced by fluid injection at The Geysers (CA, USA) geothermal field. Statistical tests are developed and applied in order to characterize the magnitude distribution of a high quality catalogue comprising seismicity directly associated with two injection wells, at the north western part of The Geysers. The events size distribution variation is investigated with respect to spatial, temporal, fluid injection and magnitude cut-off criteria. A thorough spatio-temporal analysis is performed for defining seismicity Clusters demonstrating characteristic magnitude distributions which significantly differ from the ones of the nearby Clusters. The magnitude distributions of the entire seismic population as well as of the individual Clusters are tested for their complexity in terms of exponentiality, multimodal and multibump structure. Then, the Clusters identified are further processed and their characteristics are determined in connection to injection rate fluctuations. The results of the analysis clearly indicate that the entire magnitude distribution is definitely complex and non-exponential, whereas subsequent periods demonstrating significantly diverse magnitude distributions are identified. The regional seismicity population is divided into three major families, for one of which exponentiality of magnitude distribution is clearly rejected, whereas for the other two the GR law b-value is directly proportional to fluid injection. In addition, the b-values of these Families seem to be significantly magnitude dependent, a fact that is of major importance for seismic hazard assessment implementations. To conclude, it is strongly suggested that magnitude exponentiality must be tested before proceeding to any b-value calculations, particularly in anthropogenic seismicity cases where complex and time changeable processes take place.