## Data Processing for InSight SEIS VBB continous data

Step1. Remove tick noise (Compaire et al., 2021)
Step2. Detect and remove glitches (Xu et al., 2022)
Step3. Cut and align traces
Step4. Remove instrument response
Step5. Rorate from UVW to NEZ

References:
[1]. Compaire, N.; Margerin, L.; Garcia, R.F.; et al. 2021. Autocorrelation of the Ground Vibrations Recorded by the SEIS©\InSight Seismometer on Mars. Journal of Geophysical Research: Planets, 126.
[2]. Xu, W., Zhu, Q., & Zhao, L. (2022). GlitchNet: A Glitch Detection and Removal System for SEIS Records Based on Deep Learning. Seismological Research Letters, 93(5), 2804¨C2817.
[3]. Beyreuther, M., Barsch, R., Krischer, L., Megies, T., Behr, Y., & Wassermann, J. (2010). ObsPy: A Python toolbox for seismology. Seismological Research Letters, 81(3), 530-533.


Data Structures:

./RawDownload_LMST/
  ./RawDownload_LMST/LMST_0001.mseed
  ...
  ./RawDownload_LMST/LMST_1440.mseed
./Data_Processed/
./Utilities/
  ./Utilities/GlitchNet
  ./Utilities/tk_noise
  ./Utilities/ELYSE.dataless


#0. We first download waveform data for each sol from 'ws.ipgp.fr' or 'service.iris.edu'.
    URL = http://ws.ipgp.fr/fdsnws/dataselect/1/querynet=XB&sta=ELYSE&loc=02&cha=BH?
    For data pertaining to each sol (e.g., sol 235), the start time is one hour before the sol (i.e., sol 234, 23:00 LMST), and the end time is one hour after the sol (i.e., sol 236, 01:00 LMST).

#1. Tick noise is an electrical disturbance (cross-talk) from the acquisition of the temperature inside SEIS.
    We follow Compaire et al. (2021) to firstly obtain the tick noises on the U, V, W components respectively by stacking continuous records.
    Since the pattern and the amplitude of the tick noise remains the same during the entire mission, we remove the stacked tick noise from the marsquake records by subtraction. 

#2. We further detect and remove the glitches using GlitchNet (Xu et al., 2022). 

#3. Due to gaps in the U, V, and W components, we cut and align these traces to ensure that the start and end times are the same across the U, V, and W components for each trace.

#4. We remove the instrument response (with a pre-filtering from 0.01 to 9.0 Hz) to obtain ground velocity records.

#5. We rotate the coordinate system from the original UVW to NEZ (North-East-Z) orientation using ObsPy (Beyreuther et al., 2010). 



Jiaqi Li
lijiaqi315@pku.edu.cn (lijiaqi315@gmail.com)