Landslide volumes and landslide mobilization rates in Umbria, central Italy

Abstract

A catalogue of 677 landslides of the slide type was selected from a global database of geometrical measurements of individual landslides, including landslide area (A_L) and volume (V_L). The measurements were used to establish an empirical relationship to link A_L (in m²) to V_L (in m³). The relationship takes the form of a power law with a scaling exponent α = 1.450, covers eight orders of magnitude of A_L and twelve orders of magnitude of V_L, and is in general agreement with existing relationships published in the literature. The reduced scatter of the experiential data around the dependency line, and the fact that the considered landslides occurred in multiple physiographic and climatic environments and were caused by different triggers, indicate that the relationship between V_L and A_L is largely independent of the physiographical setting. The new relationship was used to determine the volume of individual landslides of the slide type in the Collazzone area, central Italy, a 78.9 km² area for which a multi-temporal landslide inventory covering the 69-year period from 1937 to 2005 is available. In the observation period, the total volume of landslide material was V_LT = 4.78 × 10⁷ m³, corresponding to an average rate of landslide mobilization φ_L = 8.8 mm yr^− 1. Exploiting the temporal information in the landslide inventory, the volume of material produced during different periods by new and reactivated landslides was singled out. The wet period from 1937 to 1941 was recognized as an episode of accelerated landslide production. During this 5-year period, approximately 45% of the total landslide material inventoried in the Collazzone area was produced, corresponding to an average rate of landslide mobilization φ_L = 54 mm yr^− 1, six times higher than the long term rate. The volume of landslide material in an event or period was used as a proxy for the magnitude of the event or period, defined as the logarithm (base 10) of the total landslide volume produced during the event, or period. With this respect, the new relationship to link A_L and V_L is a starting point for the adoption of a quantitative, process based landslide magnitude scale for landslide events.

Introduction

Landslides are caused by various triggers, including earthquakes, rainfall and rapid snowmelt, and are influenced by multiple factors, such as topography, soil and rock types, fractures and bedding planes, and moisture content (Crozier, 1986, Turner and Schuster, 1996). Knowing the number, area, and volume of landslides is important to determine landslide susceptibility and hazard (Soeters and van Westen, 1996, Guzzetti et al., 1999, Malamud et al., 2004a), to ascertain landslide risk (Cardinali et al., 2002, Reichenbach et al., 2005), for forestry, wildlife and ecological studies (Montgomery et al., 2000, Miller and Burnett, 2007), and to evaluate the long-term evolution of landscapes dominated by mass-wasting processes (Hovius et al., 1997, Harmon and Doe, 2001, Lavé and Burbank, 2004, Malamud et al., 2004b, Korup, 2005a, Korup, 2005b, Imaizumi and Sidle, 2007, Guzzetti et al., 2008).

The number of landslides in an area is information easily obtained where accurate and reasonably complete landslide inventory maps are available (Guzzetti et al., 2002, Malamud et al., 2004a, Galli et al., 2008). Where landslide maps are available in digital form, the number of landslide per unit area (i.e., landslide density), the area of individual landslides, and the total landslide area can be calculated. Where multi-temporal landslide inventory maps are available in digital form (Guzzetti et al., 2005, Guzzetti et al., 2006, Imaizumi and Sidle, 2007, Galli et al., 2008), statistics of the number, density and area of landslides can be calculated for different periods.

Determining the volume of a landslide is a more difficult task that requires information on the surface and sub-surface geometry of the slope failure. This information is difficult and expensive to collect. Estimating the volume of slope failures for a large population of landslides (hundreds to several thousand failures) in an area is an even more challenging task (Malamud et al., 2004a) that, at present, can be achieved only by adopting empirical relationships to link the volume of individual landslides to geometrical measurements of the failures, chiefly landslide area (Simonett, 1967, Rice et al., 1969, Innes, 1983, Hovius et al., 1997, Guthrie and Evans, 2004a, Korup, 2005b, ten Brink et al., 2006, Imaizumi and Sidle, 2007, Guzzetti et al., 2008, Imaizumi et al., 2008).

In this paper, we first describe a catalogue of 5814 landslides for which measures of the area, A_L, and volume, V_L, are available. Next, we use a subset of this catalogue listing 677 mass movements of the slide type (Cruden and Varnes, 1996) – or predominantly of the slide type – to determine an empirical relationship linking A_L and V_L, and we compare the new relationship to similar relationships in the literature (Simonett, 1967, Rice et al., 1969, Innes, 1983, Guthrie and Evans, 2004a, Korup, 2005b, ten Brink et al., 2006, Imaizumi and Sidle, 2007, Guzzetti et al., 2008, Imaizumi et al., 2008). We exploit the obtained dependency to determine the total volume of landslide material and the rate at which landslides are mobilized in an area of Umbria, central Italy, for which a detailed multi-temporal landslide inventory map is available (Guzzetti et al., 2006, Galli et al., 2008). Lastly, we propose a landslide event magnitude scale based on the total landslide volume produced during an event, or period.