Borehole Seismogram Calibration and Deconvolution for Ground Motion and Earthquake Source Studies: Applications to the ICDP-GONAF project in NW Turkey

在项目最初的1.5年期间，我们重点分析了在ICDP-GONAF项目框架内实施的第一个垂直地震仪阵列的地震记录，该阵列位于伊斯坦布尔东南部的图兹拉半岛。由于低噪声条件和由此产生的井下台站与地面台站相比的高信噪比，有可能检测到地面阵列未检测到的许多微震（M<1）事件。在此处提议的扩展框架内，我们打算进一步扩大马尔马拉东海的事件数据库，方法是整合最近在阿尔穆特卢半岛完成的GONAF钻孔阵列（2014年夏季完成）以及目前正在钻探的最多四个阵列，这些阵列将于2015年春季投入使用（例如在王子群岛）。为了调查图兹拉站点的近地表特性，例如速度结构和衰减，我们采用了几种不同的技术。同样的方法也将应用于新的国家地质观测站，以评估我们在图兹拉观测站的观测结果是否具有特定的现场性，或是否也可转移到其他地质构造。不同场地的比较分析对于制定一个数据分析方案具有重要价值，该方案还可以校正场地效应，这是一个改进例如地震源参数估计所需的程序。此外，我们计划应用垂直地震剖面的处理技术，对马尔马拉东海以下的NAFZ断层分支进行断层带成像。为此，我们将选择距GONAF台站20公里震中距离（深度范围5- 20公里）内的当地地震活动，并使用不同深度的钻孔传感器。将研究首次部署在300 m深度的3分量1Hz MARK地震仪的波形记录，并有望为研究区域的剪切波特征提供新的见解。

During the initial 1.5-year period of the project we focused on analyzing seismic recordings from the first vertical seismometer array implemented in the frame of the ICDP-GONAF project, located on the Tuzla peninsula in SE Istanbul. Due to the low-noise conditions and resulting high signal-to-noise ratios at the downhole stations compared to surface stations it was possible to detect numerous microseismic (M<1) events which went undetected by surface arrays. In the frame of the here proposed extension we intend to further enlarge the event data base for the eastern Sea of Marmara by integrating additional recently completed GONAF borehole arrays on the Armutlu peninsula (completed in summer 2014) as well as up to four more arrays which are currently drilled and that shall be in operation by spring 2015 (e.g. on the Princes Islands). For the investigation of the near-surface properties at the Tuzla site such as the velocity structure and attenuation we applied several different techniques. The same methods will be applied to the new GONAF stations to evaluate whether our observations at the Tuzla site are site-specific or transferable also to other geological formations. The comparative analysis of different sites will be of great value for developing a data analysis scheme that also corrects for site effects, a procedure needed to improve for example earthquake source parameter estimations. Furthermore we plan to apply processing techniques from vertical seismic profiling for fault-zone imaging of the NAFZ fault branches below the eastern Sea of Marmara. To do so we will select local seismicity within 20 km epicentral distance (depth range 5-20km) to the GONAF stations and use the different depth levels of the borehole sensors. The waveform recordings of 3-component 1Hz MARK seismometers deployed at 300 m depth for the first time will be studied and are expected to provide new insights in shear-wave characteristics of the study area.