Of Moonquakes and Marsquakes

Introduction: We have investigated the seismicity of two planetary bodies. The observed deep seismicity of the Moon is triggered by external forcing and shows a spatial distribution, which was not expected during the planning of the APOLLO missions. We show the influence of the station network on the uncertainty of the resulting quake locations. A model of the expected seismicity of Mars has been developed as a tool for the planning of upcoming seismological experiments. Based on this model, we conducted simulations of the data return that can be expected from ExoMars, depending on the choice of the landing site. Moonquakes: More than 30 years ago, after a station life time of up to 8 years (A12), the Apollo passive seismic experiment was shut down. It resulted in the detection of more than 12000 seismic events. The majority of events listed in the Lunar Seismic Event Catalog has been recognized as representing clusters of tidally triggered quakes in the deep lunar interior [1]. However, the spatial distribution of these clusters still poses some questions, the most important of which is: Why does the lunar far side appear to be almost aseismic? Even a simple analysis of network geometry, in terms of the azimuthal gap as first-order proxy to expected location accuracy, clearly shows that the Apollo network has suboptimal location capabilities on the largest part of the Moon’s surface. An assessment of the error boundaries of the locations is necessary to test if the far side aseismicity is an artifact. We have re-analyzed the arrival time data of [1] using a new, non-linear adaptive grid scheme that does not search for the best-fitting hypocenter , but maps the entire volume of possible hypocenters which reproduce the observed seismic arrival times within given uncertainties (globally assumed to be ±10s, since individual estimates are not available). The seismic velocity model of [2] is used to compute theoretical arrival times. All 121 clusters with three or more arrivals were processed, and location uncertainty volumes described. For 15 previously unlocated clusters, a location volume could be found. A total of 112 out of 121 clusters are compatible with a focal depth of 930-960km. The narrowest depth range that is compatible with all processed clusters reaches from 465km to 1305km. Only 69 clusters are definitely located on the near side. The other 52 clusters may as well be situated on the far side, and most of these 52 have more than two thirds of the location volume there. Marsquakes: The cumulative seismic moment available per year for the production of marsquakes is estimated from the thermal contraction of the seismogenic lithosphere as described by [3], but with new and competing sets of parameters, based on a recent modeling of Mars’ thermal evolution [4]. The models differ in cooling rate, seismogenic thickness, shear modulus and other parameters. We obtain cumulative moments between 3.4x1016 Nm (WEAK models) and 4.8x1018 Nm (STRONG models) per Julian year. An amount of about 20-30 quakes with a seismic moment release that exceeds 1.3x1015 Nm (Mw=4) can be expected per year. We assume that the martian seismicity occurs in the shallow regions of the crust only – pressure inside Mars is too large to allow for events in a relative depth comparable to that of the Moon. The spatial distribution of epicenters is connected to the visible traces of tectonic faults on the surface of Mars. These are thought to be weak zones of the crust, and will therefore fail first under elastic stress. Using a reasonabl size-frequency distribution, the estimated moment budget and the spatial distribution of faults, we can simulate earthquake catalogs and use them as input for a statistical evaluation of detection rates of a seismometer of given sensitivity and placed at a given landing site. In the case of installing more than one lander, the layout of the network would need consideration of detection thresholds as well as location accuracy. For a single instrument, like in the ExoMars GEP, it is more important to detect as many quakes as possible. The site should therefore be chosen such that the instrument is closest to the centers of activity and the least number of events is shadowed by the core or too far away to be detected, although some events beyond the core shadow boundary are required to determine the core structure. To achieve this, we present maps of statistical properties of the fault distribution with respect to all points of the martian surface, in order to quantify the suitability of each point of the surface as seismometer site. References: [1] Nakamura, Y. (2005) J. Geophys. Res., vol. 110, doi:10.1029/2004JE002332 [2] Nakamura, Y. (1983) J. Geophys. Res., vol. 88, No. B1, 677-686 : [3] Solomon S.C. et al. (1991), LPI Tech Rep. #91-0 [4] Schumacher S. and Breuer D. (2006), J. Geophys. Res., vol 111, E2, doi:10.1029/2005JE002429.