Dataset of signals acquired from MEMS accelerometers during seismic events in Central Italy

Dataset of acceleration signals acquired from a low-cost Wireless Sensor Network (WSN) during seismic events occurred in Central Italy. The WSN consists of 5 low-cost sensor nodes, each embedding an ADXL355 tri-axial MEMS accelerometer with a fixed sampling frequency of 250 Hz. Continuous data was acquired from February 2023 to the end of June 2023.
The continuous data was then trimmed around the origin time of seismic events that occurred near the installation site, close to the city of Pollenza (MC), Italy, during the acquisition period. A total of 67 events were selected from the Italian Istituto Nazionale di Geofisica e Vulcanologia (INGV) Seismology data center. The waveform data was then further analyzed and annotated by analysts from INGV. Annotations include a pick time for the S and P wave, and an uncertainty level for the annotations.

The data consists of two datasets, one containing earthquake traces, the other containing noise-only traces. There are two folders: the dataset_earthquakes folder contains seismic traces; the dataset_noise folder contains noise-only traces.

The earthquake dataset consists of 328 3x25001 arrays, each related to a seismic event and with its own metadata. The dataset follows the Seisbench format, in which each trace follows the convention 'bucket0$trace_number;:n_dimensions;:n_samples', where 'bucket0' indicates the block to which the trace belongs; 'trace_number' indicates the trace' index within the block; 'n_dimensions' denotes the number of measurement axes; and 'n_samples' represents the number of samples in the trace. The waveforms are included in the the waveforms.hdf5 file of the earthquake_dataset folder, while the metadata is in the metadata.csv file in the folder. For each trace in the waveforms.hdf5 file there is an associated row in the metadata.csv file at the same index (indicated by 'trace_number' in the trace name).

The original miniSEED files that were analyzed by the INGV analysts are made available. They are contained in the miniseed_files folder. Each file name follows the format '_eventID_originTime_WS.POZA.Sx.DNy.MSEED' where eventID is the ID of the event that is recorded in the trace, originTime is the origin of the event in UTC time (expressed with the YYYY-MM-DDThh:mm:ss.ssssss format), x is a number that is used to identify the sensor that recorded the trace, y indicates the measurement direction of that trace, named '1', '2', 'Z'. For each trace in the waveforms.hdf5 file, the name of the miniSEED files that comprise the trace are in the metadata row for that trace, under the ‘trace_name_original_1’, ‘trace_name_original_2’, and ‘trace_name_original_Z’ fields in the metadata.csv file.

The dataset_noise folder follows the same convention. It contains a waveforms.hdf5 file with waveforms without seismic activity. The metadata_csv file has the metadata associated to each noise trace. The miniSEED_files_noise folder contains the original miniSEED files of the noise traces.

本数据集收录了在意大利中部地震事件期间，由低成本无线传感器网络（WSN）采集的加速度信号。该WSN由5个低成本传感器节点组成，每个节点内嵌ADXL355三轴MEMS加速度计，采样频率固定为250 Hz。数据采集自2023年2月至6月底。连续数据在采集期间对位于意大利波伦察市（MC）附近安装地点的地震事件起源时间进行了裁剪。从意大利国家地球物理与火山学研究所（INGV）地震数据中心共选取了67次事件。随后，波形数据由INGV的分析师进行了进一步分析和标注。标注内容包括S波和P波的拾取时间以及标注的不确定性等级。

数据集包含两个数据集，一个包含地震波形数据，另一个包含仅包含噪声的波形数据。存在两个文件夹：dataset_earthquakes文件夹包含地震波形数据；dataset_noise文件夹包含仅包含噪声的波形数据。

地震数据集由328个3x25001维数组组成，每个数组与一次地震事件相关，并具有相应的元数据。数据集遵循Seisbench格式，其中每个轨迹遵循以下约定：‘bucket0$trace_number;:n_dimensions;:n_samples’，其中‘bucket0’表示轨迹所属的块；‘trace_number’表示轨迹在该块中的索引；‘n_dimensions’表示测量轴的数量；‘n_samples’表示轨迹中的样本数。波形数据包含在earthquake_dataset文件夹中的waveforms.hdf5文件中，而元数据则包含在文件夹中的metadata.csv文件中。对于waveforms.hdf5文件中的每个轨迹，都存在一个与之关联的行在metadata.csv文件中，其索引与轨迹名称中的‘trace_number’相对应。

由INGV分析师分析的原miniSEED文件也已提供。它们包含在miniseed_files文件夹中。每个文件名遵循以下格式：‘_eventID_originTime_WS.POZA.Sx.DNy.MSEED’，其中eventID是记录在轨迹中的事件ID，originTime是事件起源的UTC时间（以YYYY-MM-DDThh:mm:ss.ssssss格式表示），x是一个用于识别记录轨迹的传感器的数字，y表示该轨迹的测量方向，命名为‘1’、‘2’、‘Z’。对于waveforms.hdf5文件中的每个轨迹，构成该轨迹的miniSEED文件的名称位于metadata.csv文件中该轨迹的元数据行下，在‘trace_name_original_1’、‘trace_name_original_2’和‘trace_name_original_Z’字段中。

dataset_noise文件夹遵循相同的约定。它包含一个包含无地震活动波形的waveforms.hdf5文件，metadata_csv文件包含与每个噪声轨迹相关的元数据。miniSEED_files_noise文件夹包含噪声轨迹的原miniSEED文件。