Collaborative Research: GeoGONAF: Analysis of active deformation and strain transfer along the Izmit Bay-Marmara Sea segment of the North Anatolian Fault

世界上很少有地方比土耳其的马尔马拉海地区更接近加州旧金山湾区的地震危险潜力。在这两个地方，里氏7级的地震都可能撼动一个海上大都市。伊斯坦布尔目前的人口接近1500万，每年的旅游人数为300万，未来20年发生7.1级地震的概率估计为35-70%。早在1766年，一场类似于预期的地震就摧毁了人口不足50万的伊斯坦布尔。有些人认为，这次地震即将重演。GeoGONAF是一个国际合作项目的缩写，用于监测地震前地下和海底应变的变化。它的科学团队并不期望预测未来地震的精确时间，但他们乐观地认为，密集的监测仪器阵列正在密切关注即将到来的破裂，将捕捉到丰富的信号，这些信号将表明北安那托利亚断层的几个分支中哪一个最有可能破裂。在城市周围和近海岛屿上安装这些仪器花了近十年的时间，但现在，一股数据流流入土耳其、德国和美国的实验室，原则上使这些合作国家的科学团队能够识别出即将破裂的信号特征。土耳其的经验教训将直接应用于旧金山湾海沃德断层下发生的类似前兆过程的解释。在GeoGONAF（地球物理观测站研究伊斯坦布尔附近的北安那托利亚断层的大地测量部分）的最后阶段，在钻孔中安装了六台应变仪。应变计可以检测到10e-12（1皮应变）的应变变化，相当于地球直径的变化，相当于人类头发直径的五分之一。在地下30米至100米的深度插入这些超敏感仪器后，稳定阶段几乎超过（1-3年），并且以100 Hz采样率（截至2016年6月）的6个应变数据流现在补充了由GeoGONAF的德国，美国和土耳其合作者安装的钻孔地震仪和GPS仪器的数据。目前的项目计划首先利用潮汐和传入地震波校准应变仪，然后监测马尔马拉地区出现的偶发和连续信号。具体来说，计划在伊斯坦布尔南部和东南部的海底断层上搜索慢滑事件。这些缓慢的蠕变事件对地震仪来说是不可见的，但在GPS数据上可以看到增量，在应变仪上可以看到持续时间为几分钟到几天的独特的随时间变化的信号。2016年3月18日，在即将投入使用的6个应变仪中的第一个上，一个大型慢滑事件被非常详细地捕捉到，其重要性在于，这种事件逐渐将已经准备好即将破裂的岩石推向未来的地震。尽管每一次增量都使地震更接近发生，但要知道伊斯坦布尔以南三条断层中的哪一条会发生地震还为时过早。拟议的项目将每天检查传入的应变和GPS数据，以搜索传感器阵列上同步发生的相干信号，以识别断层的滑动块及其滑动的面积和程度。初步的解释表明，大型蠕变事件每年在断层的任何一块上不会发生超过一次，但在马尔马拉海下方和附近的整个相互连接的断层网络中，蠕变事件可能相当频繁。作为对断层系统更大规模研究的一部分，将安装额外的地表蠕变仪，专门用于确定沿马尔马拉以东的北安那托利亚断层2-16毫米振幅蠕变事件的传播特征。

There are few places in the world that more closely match the seismic hazard potential of the San Francisco Bay Area of California than the Marmara Sea area of Turkey. In both places a magnitude 7 earthquake threatens to shake a maritime megacity. The current population of Istanbul is approaching 15 million, with an annual tourist population of 3 million, and the probability of an M7.1 earthquake in the next 20 years estimated as 35-70%. In 1766 a previous earthquake, similar to the one anticipated, destroyed Istanbul when its population was less than 0.5M. Some believe that a repeat of this earthquake is imminent. GeoGONAF is an acronym for an international collaborative project to monitor the subterranean and submarine strain changes that will precede the earthquake. Its scientific team does not expect to predict the precise time of this future earthquake, but they are optimistic that the dense array of monitoring instruments now listening closely to the approaching rupture will capture a rich harvest of signals that will indicate which of the several branches of the North Anatolian fault is most likely to rupture. It has taken the better part of decade to install these instruments around the city and on offshore islands, but now a stream of data flows into laboratories in Turkey, Germany and the US, that in principle permit the scientific teams of these collaborating nations to identify signals characteristic of approaching rupture. Lessons learned in Turkey will have direct application in the interpretation of similar precursory processes occurring beneath the Hayward fault located in the San Francisco Bay.Six strain-meters were installed in boreholes during the final phase of GeoGONAF (the Geodetic component of the Geophysical Observatory to study the North Anatolian Fault near Istanbul). The strain-meters can detect strain changes of 10e-12 (1 picostrain), equivalent to a change in the diameter of the Earth of one fifth of the diameter of a human hair. The stabilization phase following inserting these ultra-sensitive instruments at depths of 30 m to 100 m underground is almost over (1-3 years) and six streams of strain data at 100 Hz sampling rate (as of June 2016) now complement data from the borehole seismometers and GPS instruments installed by the German, US and Turkish collaborators of GeoGONAF. The current project plans first to calibrate the strain-meters using tides and inbound seismic waves, and then to monitor episodic and continuous signals occurring in the Marmara region. Specifically a search for slow slip events on the submarine faults south and southeast of Istanbul is planned. These slow creep events are invisible to seismometers but are seen as increments on GPS data and as distinctive time-evolving signals on strain-meters with durations of minutes to days. On 18 March 2016 a large slow slip events on the first of the six strain-meters to be commissioned was captured in remarkable detail, its importance being that such events incrementally advance rocks already poised for imminent rupture towards a future earthquake. Although each increment brings an earthquake closer, it is too early to know on which of three faults south of Istanbul this will occur. The proposed project will examine incoming strain and GPS data on a daily basis to search for coherent signals occurring synchronously on the array of sensors, in order to identify slipping patches of faults and the area and extent of their slip. Preliminary interpretations indicate that large creep events cannot occur more than once per year on any given patch of a fault but that creep events may be quite frequent throughout the interconnected network of faults beneath and near the Sea of Marmara. As part of a larger study of the fault system additional surface creep-meters will be installed specifically to identify the propagation characteristics of 2-16 mm amplitude creep-events along the North Anatolian fault to the east of Marmara.