Morphology, structure and kinematics of a rainfall controlled slow-moving Andean landslide, Peru Swann Zerathe, 1,2 * Pascal Lacroix, 1 Denis Jongmans, 1 Jersy Marino, 3 Edu Taipe, 3 Marc Wathelet, 1 Walter Pari, 2 Lionel Fidel Smoll, 2 Edmundo Norabuena, 4 Bertrand Guillier 1 and Lucile Tatard 1 1 ISTerre – IRD/CNRS, Université de Grenoble, Saint Martin d’Hères, France 2 Instituto Geológico, Minero y Metalúrgico INGEMMET, Lima, Peru 3 Instituto Geológico, Minero y Metalúrgico INGEMMET, Arequipa, Peru 4 Instituto Geofísico del Peru IGP, Lima, Peru Received 21 June 2015; Revised 26 January 2016; Accepted 26 January 2016 *Correspondence to: Swann Zerathe, ISTerre – IRD/CNRS, Université de Grenoble, 1381 rue de la piscine, 38400 Saint Martin d’Hères, France. E-mail: swann.zerathe@ird.fr ABSTRACT: The large slow-moving landslide of Maca is located in the upper Colca valley (southern Peru), a region characterized by a well pronounced rainy period, and intense and recurrent sustained seismicity. The landslide, developed in deep lacustrine de- posits, has recently accelerated, threatening the Maca village. This work aims at understanding the rupture mechanism and the causes of the recent landslide reactivation/acceleration. We present a multidisciplinary characterization of the Maca landslide that includes: (i) geological and morphological mapping in the field; (ii) remote sensing analysis using an historical aerial photograph of 1955 and the Pléiades satellite images (2013); (iii) global positioning system (GPS) including time-series of surveys over 13 years, and continuous measurements over 14 months; (iv) a geophysical campaign with deep electrical resistivity tomography profiles ac- quired across the landslide mass. Our study shows that this 60 Mm 3 landslide, which can be classified as a clay/silt compound land- slide, moved by 15 m between 2001 and 2014 with a large inter-annual velocity variation (up to a factor of 500) depending on the rainfall intensity. We suggest that these dramatic changes in velocity are the result of the combination of a threshold mechanism and the short intense rainy season in Peru. This study reveals three main driving factors acting at different timescales: (i) over several de- cades, the river course has significantly changed, causing the Maca landslide reactivation in the 1980s due to the erosion of its toe; (ii) at the year scale, a minimum amount of rainfall is required to trigger the motion and this amount controls the landslide velocity; (iii) transient changes in slide velocity may occur anytime due to earthquakes. This study particularly highlights the non-linear behav- iour of the motion with rainfall. Copyright © 2016 John Wiley & Sons, Ltd. KEYWORDS: slow moving landslide; lacustrine deposits; forcing factors; monitoring; Peru Introduction Slow moving landslides are common in mountainous areas and constitute a major geological hazard, damaging facilities such as buildings, roads, railways, bridges, dams (Guzzetti et al., 2004; Mansour et al., 2011; Handwerger et al., 2013). Although their velocity usually ranges between a few centimetres per year (cm/yr) to a few metres per year (m/yr), they may exhibit sudden acceleration phases and flows that are generally difficult to predict, and can result in loss of life (Iverson et al., 1997; Petley et al., 2002; Jongmans et al., 2009; Van Asch and Malet, 2009). The kinematics of these landslides is found to be primarily driven by precipitation and water infiltration, resulting in increased pore water pres- sure (Iverson and Major, 1987; Iverson, 2000; Hilley et al., 2004; Van Asch et al., 2007; Schulz et al., 2009; Strozzi et al., 2010; Handwerger et al., 2013). Another effect of rain- fall on landsliding is to increase the erosional efficacy of flu- vial processes, undercutting the slope toe (Huggel et al., 2012), as lateral river erosion can be a key driver of landslide activity on hillslopes (Larsen and Montgomery, 2012). Several studies of clay slides affecting riverbanks have pointed out the importance of river erosion in promoting slope failures (Eilertsen et al., 2008; Kohv et al., 2009). Other sources of ex- ternal forcing include earthquakes (Keefer, 2002; Scheingross et al., 2013; Lacroix et al., 2014), glacier retreats (Strozzi et al., 2010) or human activity (Mansour et al., 2011). All these factors act on different timescales varying from seconds (earthquakes) to several hundreds or thousands years (glacier retreat). In high seismicity regions with rainy seasons, both earthquake and climatic triggering mechanisms are relevant, with complex interactions possible, in which one mechanism could prepare the conditions for landslide triggering by alter- natively fracturing/weakening the medium or increasing the pore pressure (Lin et al., 2008; Tatard et al., 2010). Time delay is regularly observed between the forcing event and the initiation (or reactivation) of landslide movement for clay slopes (Handwerger et al., 2013) or earthflows (Iverson and Major, 1987), leading to a lag of weeks to months. For pre- cipitation episodes, the generation of cracks and fissures in EARTH SURFACE PROCESSES AND LANDFORMS Earth Surf. Process. Landforms (2016) Copyright © 2016 John Wiley & Sons, Ltd. Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/esp.3913