Geologic and geomorphic factors of the landslides triggered in the Cardoso T. basin (Tuscany, Italy) by the 19th June, 1996 intense rainstorm.

On June 19, 1996 exceptionally intense rainfall fell in the southern Apuan Alps (Tuscany, Italy): ca. 474 mm rainfall was recorded within 12 hours (21% of the annual rainfall) and the maximum rainfall intensity recorded was 158 mm/hour. The rainstorm hit a territory ca. 150 km2 wide in the area of the main watershed between the Versilia and the Serchio River basin. It caused hundreds of landslides, flooding of wide areas, 14 fatalities and very heavy structural damage. Here, the results of the studies on the landslides occurred in the Cardoso Torrent basin are summarized. This basin, 13 km2 wide, corresponds to the most severely damaged area, where almost 400 landslides were triggered (ca. 30 landslides/km2). The morphology is typically mountainous, with deeply cut valleys and very steep, woody mantled slopes, underlain by mainly metamorphic, impervious rocks. The most significant parameters of the landslides were analysed, in order to identify the factors which most influenced their activation. Moreover, the total amount of mobilized material was estimated. The most common type of landslide movement (more than 80%) was complex, very or extremely rapid, debris slide-debris flow, with a high length to breadth ratio (3<L/B<20). Because of their fluidity, the sliding and flowing masses often poured into the riverbeds, mixed with the runoff and overloaded streams with sediments and wood. Most of the landslides were probably first time landslides; ca. 95% of them involved the colluvium cover of slopes (0.5-2.5 m thick), while the bedrock was rarely involved. The studies in the landslide sites also highlighted many geomorphic and geological recurrent factors, summarized as follows. 90% of landslides occurred on rather steep slopes (30É-45É), in first-order basins and hollows. In these situations the concave geometry of the colluvium/bedrock interface favoured the convergence of groundwater flow and the build-up of pore pressure, leading to failure. In landslide sites, a frustum-conic concave shape of the surface and a rectilinear profile of the slope were frequent (respectively ca. 56% and 76% of landslides). The bedrock was generally made up of impervious or scarcely pervious rocks; among them, the Pseudomacigno Formation (metamorphic sandstone with interbedded phillyte) was the most common one (8.4% of landslide density). The presence of a main discontinuity in the bedrock (bedding or schistosity) dipping downslope was significant in many landslide sites: it was ascertained in more than 50% of the landslides. According to some witnesses, many landslides were activated in a short time, when ca. 250 mm rainfall within 8 hours had been recorded (ca. 31 mm/hour). This rather high value may represent the landslide-triggering rainfall threshold in the study area. The total surface involved in landslides on June 19, 1996 was estimated at ca. 0.55 Km2, 4.2% of the total basin surface. More than 81% of this surface was covered by chestnut trees: thus, ca. 4,400 chestnut trees were uprooted by the landslides and fell into the riverbeds. This significantly contributed to the extent destruction and blockage of bridge spans (with flooding). The total volume of mobilized material was estimated at ca. 850,000 m3: most of this volume (ca. 755,000 m3) poured into the riverbeds, while the rest remained on the slopes involved. Ca. 640,000 m3 were mobilized by stream erosion in the riverbeds. Therefore, ca. 1,500,000 m3 was the total volume of the material mobilized by both slope and stream processes on June 19, 1996 in the Cardoso T. basin. Research is still in progress although many rainfall, hydrogeological and geotechnical data are still lacking. These data would help us to evaluate better slope stability conditions and the critical landslide-triggering rainfall thresholds. Nevertheless, the present results (through the knowledge of the recurrent factors in landslide activation) enable us to perform risk scenarios and also plan works of prevention, also in neighbouring areas, where risk conditions are similar.