Search

The instrument package SEIS (Seismic Experiment for Internal Structure) with the two co-located seismometers VBB and SP is installed on the surface of Mars as part of NASA's InSight mission. When compared to terrestrial installations, SEIS is deployed in a very harsh wind and temperature environment that leads to inevitable degradation of the quality of the recorded data. The daily atmospheric temperature variations of approx. 80K are attenuated by different insulation layers to approx. 15K peak-to-peak at the sensor level. Typical wind speeds vary between 0 and 5 m/s leading to a diurnal variation in the broad-band rms noise level by two orders of magnitude. One ubiquitous artifact in the raw broad-band data is an abundance of one-sided, transient pulses often accompanied by high-frequency spikes. We show that these pulses, which we term "glitches", can be modeled as the response of the instrument to a step in acceleration, while the spikes can be modeled as the response to a simultaneous step in displacement. We attribute the glitches primarily to intermittent stress relaxation events internal to SEIS caused by the large diurnal temperature variations to which the instrument is exposed during a Martian sol. Only a small fraction of glitches correspond to a motion of the SEIS package as a whole caused by minuscule tilts of the instrument. Whilst such kind of data disturbances are typically discarded when occurring in terrestrial data, this is no option for the data returned from the Red Planet. We therefore do not only demonstrate their effects on the seismic data and analyze their origins, but also propose algorithms that are able to detect and remove many of these (mostly) non-seismic signals. We further published our codes (both Python and MATLAB) so that interested researchers can make their own choices on how to treat the data and to which extent.

AGU Fall Meeting in San Francisco, 9-13 December 2019, The SEIS seismometer package was successfully deployed on Mars by February 2019. Routine operations are split into two services: the Mars Structure Service (MSS) and the Marsquake Service (MQS), which are responsible for defining interior structure models and seismicity catalogs, respectively. Initial model delivery from MSS is based on a limited dataset of two Marsquakes with a clear P and S arrivals. Different inversion algorithms were developed by the MSS team in order to retrieve the first 1D averaged model of Mars. Two complementary approaches are considered. One set of models (called M1) is parameterized in seismic velocity and density as a function of depth. A second set of models (called M2) is obtained by parameterizing with geodynamical constraints like temperature and composition. We use Bayesian inversion techniques to obtain robust probability density functions of seismic velocity profiles. Different types of data are considered for these inversions including body waves, surface waves and receiver functions. To characterize what we could learn about Mars¿ interior structure with only one station and with the first seismic event, we performed inversions of synthetic data following a blind test process, where the interior model and the Marsquake parameters (location, depth, origin time, and moment tensor) were unknown to all team members carrying out data analysis and inversion. In this presentation we will discuss the results of this blind test in terms of structure and compare different methods developed by the MSS. We will then show results from investigations of the first, real seismic data due to quakes on Mars recorded by SEIS in terms of the structure and quake locations. We will especially focus our investigation on joint inversions made not only with the arrival time, but also with secondary seismic data extracted from the detected events, including apparent attenuation rate and with receiver functions. Of course, much more detailed analysis will be made if Mars seismicity provide us in the near future larger quakes with body wave phases and first orbit surface wave dispersion, and/or one event large enough to record multiple orbit surface waves, and will augment future interiors models of Mars.

AGU Fall Meeting 2019 in San Francisco, 9-13 December 2019, Exploring the internal structure and the dynamics of our solar system is mandatory to understand the behavior of our universe and its origin. One of the tools chosen by NASA is seismology particularly in order to constrain the parameters of the deep interior structure of the red planet via the Insight (Interior Exploration using Seismic Investigations, Geodesy, and Heat Transport) mission. InSight was successfully landed on November 26th, 2018 in Elysium Planitia with geophysical instruments a short-period seismometer and a broadband seismometer (SEIS, Seismic Experiment for Interior Structure). Both seismometers are now installed directly on Mars surface and enable to analyze the continuous seismic signal.But, before making the structure inversion, we need to extract the features from SEIS data. However, those features may nevertheless be hidden into noise, or may escape from analysis due to the limitations imposed by the current methodologies.Therefore, the aim of this study is to overcome this problem by well extracting, recognizing and classifying the instrument signals using Machine Learning and Deep Learning new strategies inspired from the Deep scattering network.This is very promising for the SIES data as, we¿re going to be able not only to detect the familiar signals, but the exciting part is the unseen or the unknown ones. This technique is used to clean the data from the glitches. In fact, this tool has recently proved to be powerful in signal processing, data automatic feature extraction and may even be helpful to detect new types of signals. Those new signals can reveal unknown processes and lead to new discoveries about Mars physical processes.The method used in this study is divided into three fundamental steps. The first one, to make an automatic feature extraction using the Deep scattering transform which is a convolution neural network that computes a cascade of wavelets calculations and filtering operations to get a stable waveform representation stable to local deformations and overlapping at multiple times and frequencies.. The second step is to use those features for signal classification using Machine Learning classifier Gaussian Mixture Network. Finally, we update the wavelet mother bank depending on the results of the classification error minimization using Adam stochastic gradient descent.

AGU Fall Meeting 2019 in San Francisco , 9-13 December 2019, EIS/InSIght team, InSight is the first planetary mission with a seismometer package, SEIS, since the Apollo Lunar Surface Experiments Package. SEIS is complimented by APSS, which has as a goal to document the atmospheric source of seismic noise and signals. Since June 2019, SEIS has been delivering 6 axis 20 sps continuous seismic data, a rate one order of magnitude larger originally planned. More than 50 events have been detected by the end of July 2019 but only three have amplitudes significantly above the SEIS instrument requirement. Two have clear and coherent arrivals of P and S waves, enabling location, diffusion/attenuation characterization and receiver function analysis. The event¿s magnitudes are likely ¿ 3 and no clear surface waves nor deep interior phases have been identified. This suggests deep events with scattering along their final propagation paths and with large propagation differences as compared to Earth and Moon quakes. Most of the event¿s detections are made possible due to the very low noise achieved by the instrument installation strategy and the very low VBB self-noise. Most of the SEIS signals have amplitudes of spectral densities in the 0.03-5Hz frequency bandwidth ranging from 10-10 m/s2/Hz1/2 to 5 10-9 m/s2/Hz1/2. The smallest noise levels occurs during the early night, with angstrom displacements or nano-radian tilts. This monitors the elastic and seismic interaction of a planetary surface with its atmosphere, illustrated not only by a wide range of SEIS signals correlated with pressure vortexes, dust devils or wind activity but also by modulation of resonances above 1 Hz, amplified by ultra-low velocity surface layers. After about one half of a Martian year, clear seasonal changes appear also in the noise, which will be discussed. One year after landing, the seismic noise is therefore better and better understood, and noise correction techniques begun to be implemented, either thanks to the APSS wind and pressure sensors, or by SEIS only data processing techniques. These data processing techniques open not only the possibility of better signal to noise ratio of the events, but are also used for various noise auto-correlation techniques as well as searches of long period signals. Noise and seismic signals on Mars are therefore completely different from what seismology encountered previously on Earth and Moon.

The instrument package SEIS (Seismic Experiment for Internal Structure) with the three very broadband and three short‐period seismic sensors is installed on the surface on Mars as part of NASA's InSight Discovery mission. When compared to terrestrial installations, SEIS is deployed in a very harsh wind and temperature environment that leads to inevitable degradation of the quality of the recorded data. One ubiquitous artifact in the raw data is an abundance of transient one‐sided pulses often accompanied by high‐frequency spikes. These pulses, which we term “glitches”, can be modeled as the response of the instrument to a step in acceleration, while the spikes can be modeled as the response to a simultaneous step in displacement. We attribute the glitches primarily to SEIS‐internal stress relaxations caused by the large temperature variations to which the instrument is exposed during a Martian day. Only a small fraction of glitches correspond to a motion of the SEIS package as a whole caused by minuscule tilts of either the instrument or the ground. In this study, we focus on the analysis of the glitch+spike phenomenon and present how these signals can be automatically detected and removed from SEIS's raw data. As glitches affect many standard seismological analysis methods such as receiver functions, spectral decomposition and source inversions, we anticipate that studies of the Martian seismicity as well as studies of Mars' internal structure should benefit from deglitched seismic data., Plain Language Summary: The instrument package SEIS (Seismic Experiment for Internal Structure) with two fully equipped seismometers is installed on the surface of Mars as part of NASA's InSight Discovery mission. When compared to terrestrial installations, SEIS is more exposed to wind and daily temperature changes that leads to inevitable degradation of the quality of the recorded data. One consequence is the occurrence of a specific type of transient noise that we term “glitch”. Glitches show up in the recorded data as one‐sided pulses and have strong implications for the typical seismic data analysis. Glitches can be understood as step‐like changes in the acceleration sensed by the seismometers. We attribute them primarily to SEIS‐internal stress relaxations caused by the large temperature variations to which the instrument is exposed during a Martian day. Only a small fraction of glitches correspond to a motion of the whole SEIS instrument. In this study, we focus on the detection and removal of glitches and anticipate that studies of the Martian seismicity as well as studies of Mars's internal structure should benefit from deglitched seismic data., Key Points: Glitches due to steps in acceleration significantly complicate seismic records on Mars. Glitches are mostly due to relaxations of thermal stresses and instrument tilt. We provide a toolbox to automatically detect and remove glitches., Centre National d'Etudes Spatiales (CNES), InSight PSP Program, Agence Nationale de la Recherche http://dx.doi.org/10.13039/501100001665, ANR‐19‐CE31‐0008‐08