44th Lunar and Planetary Science Conference (2013) 1898.pdf

STATISTICS OF MASS MOVEMENTS IN VALLES MARINERIS, MARS. M. T. Brunetti1,2, M. Cardinali1, F. Fiorucci1, M. Santangelo1, F. Guzzetti1 and G. Komatsu3, 1Research Institute for Geo-Hydrological Protection – Italian National Research Council (Via della Madonna Alta 126, 06128 Perugia-Italy, Ma- [email protected]), 2University of Perugia (Piazza Università 1, 06100 Perugia-Italy), 3University of Pescara (Viale Pindaro 42, 65127 Pescara-Italy).

Introduction: On Earth, landslides contribute to Chasmata, Valles Marineris, Mars, where landslides shaping the landscapes in various geological, tectonics are abundant. Using visual interpretation of medium to and climatic regions. Besides the Earth, Mars is the high resolution optical images, we mapped and charac- planet for which more information exists on landslides, terized 219 geomorphological features, including rock and other mass wasting processes [1-3]. Rainfall, slides, complex/compound failures, rock avalanches, snowmelt, earthquakes and volcanic activity are the debris flows, and rock glacier-like features. primary natural triggers of terrestrial landslides, but the On Earth, the non-cumulative probability density of triggers of Martian landslides remain largely undeter- landslide area, p(AL), exhibits a typical shape, with the mined. We compiled a geomorphological inventory of number of landslides increasing with their size up to a mass movements in the Valles Marineris (VM), adopt- maximum value, after which the abundance decreases ing the same visual interpretation criteria used by geo- rapidly following a power law distribution with nega- morphologists to detect, map and classify terrestrial tive scaling exponent α [11-13]. The non-cumulative landslides [4,5]. The inventory covers two E-W- probability density p(AL) of 82 landslides (rock slides striking chasmata in the western sector of the VM, the and complex/compound failures) mapped in VM was Tithonium Chasma and the Ius Chasma, where slope determined through a kernel density estimation. The failures were previously recognized and are known to non-cumulative probability density p(AL) of the VM be abundant [2-3]. The study area covers 105 km2 with landslides follows a power law for most of the range of 5 2 10 2 elevation in the range from -4500 to 6200 m, and local the mapped landslides (5×10 m < AL < 10 m ), with relief exceeding 0.8 km/km. We exploited the new ge- a scaling exponent of the distribution α = -1.01 ± 0.01, omorphological inventory of landslides in VM to de- considerably smaller (in absolute value) than the scal- termine the statistics of landslide area AL and volume ing found for terrestrial slides, α ~ -2.4 ± 0.5 [14]. This VL. is the evidence of the fact that the proportion of very 7 2 Methodology: When single or multiple landslides large landslides (i.e., AL > 10 m ) is significantly high- occur in an area, they change the local topography and er on Mars than on Earth. The distinct difference ob- the radiometric properties of the land surface, leaving served in the scaling of the power law of terrestrial and discernible morphometric [6] and radiometric [5] sig- Martian landslides points to differences in the mechan- natures that can be detected and interpreted. ical properties of the materials involved by the slope To detect and map landslides and other geomorpho- failures. logical features in the Ius Chasma and the Tithonium The probability density of landslide volumes p(VL) Chasma, we analyzed visually the photographic charac- for VM slides also obeys a negative power law trend, teristics of color, tone, pattern, and texture of images with negative scaling exponent β = -1.03 ± 0.01, in the 9 3 12 3 captured by the HRSC [7] and by the CTX [8]. Both range 10 m < VL < 10 m . This value is typical of HRSC and CTX images have a ground resolution ade- terrestrial rock falls and rock slides [15]. quate to detect and map landslides on Mars with un- Joint analysis of the measurements of AL and VL for precedented detail and geographical accuracy. In addi- 49 deep-seated slides allowed determining an empirical tion, we analyzed the local topography at 463 m/pixel relationship to link the area to the volume of the VM resolution captured by the MOLA [9]. In particular, the landslides. A power law equation of the form γ MOLA digital elevation model (DEM) was used to VL = η × AL reconstruct 3D views of the landslide areas, to facilitate where η e and γ are a coefficient and the scaling mapping of the mass movements. exponent, respectively, was fitted to the empirical data, Slope failures were classified according to [10] using robust linear fitting to minimize the effects of the based on morphological similarity with terrestrial land- outliers [16]. The obtained empirical relationship is (1.17±0.02) slides. No inference was made on the geomorphic pro- VL = (4.1 ± 1.6)× AL cesses that have triggered, or modified, the slope fail- The scaling exponent is out of the range of expo- ures. nents found for terrestrial landslides, which are be- Results: We completed a systematic mapping of tween 1.332 and 1.450 [17]. This was expected, since landslides in a 105 km2 area in the Tithonium and Ius the distribution of landslide area on Mars largely dif- 44th Lunar and Planetary Science Conference (2013) 1898.pdf

fers from that on the Earth, whereas the landslide vol- [10] Cruden D. M. and Varnes D. J. (1996) In ume distribution is very similar on the two planets. Transp. Res. Board, Spec. Rep. 247 (Turner A. K., References: [1] Lucchitta B.K. (1979) JGR, 84, Shuster R. L., eds), pp 36–75. [11] Stark C. P. and 8097–8113. [2] McEwen A.S. (1989) Geology, 17, Hovius H. (2001) Geophys. Res. Lett., 28, 6 1091– 1111–1114. [3] Quantin C. et al. (2004) Planet. Space 1094. [12] Guzzetti F. et al. (2002) Earth Planet. Sc. Sci., 52 (11), 1011–1022. [4] van Zuidam R. A. (1985) Lett., 195, 169–183. [13] Malamud B. D. et al. (2004) In Aerial photo-interpretation in terrain analysis and Earth Surf. Proc. Land., 29(6), 687–711. [14] Van Den geomorphologic mapping, 442 pp. Smits Publishers, Eeckhaut M. et al. (2007) Earth Planet. Sc. Lett., 256, The Hague. [5] Guzzetti F. et al. (2012) Earth-Sci. 588–603. [15] Brunetti M. T. et al. (2009) Nonlinear Rev., 112 (1-2), 42–66. [6] Pike R. J. (1988) Geology, Proc. Geoph., 16, 179–188. [16] Venables W.N. and 20(5), 491–511. [7] Jaumann R. et al. (2007) Planet. Ripley B.D. (2002). Modern Applied Statistics with S, Space Sci., 55, 928–952. [8] Malin M. C. et al. (2007) Springer, New York. [17] Larsen I. J. et al. (2010) Nat. JGR-Planet, 112, E05S04. [9] Zuber M. T. et al. Geosci., 3, 247–251. (1992) JGR-Planet, 97, 7781–7797.