Mass movement and erosion hazard patterns by multivariate analysis of landscape integrated data: the Upper Orcia River Valley (Siena, Italy) case

Abstract

A landscape integrated survey was carried out on the badlands area in the Upper Orcia River Valley (Siena, Italy). A methodology has been investigated to explain the soil response to dynamic and anthropic factors from an erosion hazard point of view. The Upper Orcia River Valley is one of the many neogenic basins linked to the Pliocene sea ingression, in Tuscany and all over Italy. Sediments are mainly fine silty-clay and the region is characterized by an estensive net of joints, faults and fractures due to the neotectonics in the area. The area is undergoing a fast land degradation characterized by “biancane” (domes), “calanchi” (very steep and deep gullies with a typical dendritic drainage net), gullies, several surficial erosional forms and mass movement phenomena (creeping, landslides, solifluctions and anthropic levelling). Seventeen Land Units and 46 Types were recognized according to the relationships between Elementary Landform and slope (EL), Hydrographic Net Patterns (HNP), River Bed (RB), Geomorphic Dynamics (GD), Erosion Forms (EF), Mass Movements (MM), vegetation and land use (VLU), Cultural Management systems (CM) and Morphological Sequences (MS). Lithology and soils were not taken into account as the former is to be considered almost homogeneous, while the latter are of the Entisols and Inceptisols orders ( Soil Survey Staff, 1975) in the Typic Xerorthent and Xerochrept Subgroups. The soil pattern is mainly linked to geomorphic dynamics and appears extremely leopardized, every landform being characterized by various toposequences. PCA and clustering based on Euclidean distance (similarity measure) were utilized. A normalization of the variables to the standard deviation of the population was used. The previous PCA and Varimax rotation reduced the number of the variables from the original 9 to the 5 more weighing ones (1-2-5-6-7). Clusters were created to verify the Land Types groups as well as the possibility of unifying Land Units with the higher similarity level. The cluster analysis resulted in 15 groups that can well explain different land degradation patterns and which, in some cases, confirmed the Land System Analysis. The Authors came to the conclusion that multivariate analysis of landscape characteristics can only partially substitute the landscape integrated survey but can be considered a helpful means to analyse the distribution in the area of mass movements and erosion phenomena.