The Lavina Di Roncovetro landslide (RE, Italy)

dott. geol. Giovanni Bertolini
Rrgione Emilia-Romagna                       
Servizio Tecnico dei Bacini Affluenti del Po

Download Technical Report

1) History
Almagià, an important precursor of landslide researchers, in publishing the first Italian inventory of landslides, coined the term of “Lavina di Roncovetro”, where Lavina meant (and means also today) Landslide in local idiom. It was the year 1907 and the landslide was already well known. For what we know, the morphological features of this landslide were not so evident in more ancient times, since the oldest existing cartographies, dating in 1821, 1828 and 1858, did not represent it (see figures).
The rapid evolution of this landslide during the last century is also demonstrated by recent documents and images, as shown in the figures. In particular, the rapid retrogression of the crown, now reaching the mountaintop, is evident.

Contrary to what is observed in the majority of large earth flows in Northern Apennine, the many core borings here performed did not find ancient woods or organic matter buried inside its body. The maximum observed “conventional age” by 14C method was 119 years BP, which means “modern” in calibrated years BP.
Making an assumption, we can guess that this earth flow is quite young, having begun its formation during the period between the middle and the end of the XIX century and not in more ancient times (e.g. in the middle of Holocene), as is usual for similar landslides in the Northern Apennine.
The rapid retrogression of the crown also results from the comparison of aerial photography in the last 20 years (see figures). In that period of time, in fact, we may estimate that about 100 k cubic metres of bedrock descended from the main scarp into the landslide body.

2) Features
The 2,5 km long Roncovetro / Vedriano landslide carves the southern slope Monte Staffola from its top to the Tassobbio River where it causes the formation of a small lake. The slope is formed by sub-ligurian clayey calcareous-arenaceous flysch. The clay fraction is dominant from the mechanical point of view, causing the landslide to behave in its upper portion as a very active fluid-viscous mudflow, with maximum velocities up to 10 m/day.
The total volume of the landslide is about 3 M of cubic metres. The landslide reactivated completely in autumn 1993, it has since slowed during the summer months but has never completely stopped. The average speed of the upper lithosome L1 is on the order of hectometres per year, while the lower (L2) moves a few decametres per year. A unique feature is the long and narrow channel, 30 to 40 m wide and deeply carved into the flysch bedrock, which links the depletion zone to the accumulation zone.
In the middle-lower part of the slope, the upper portion (L1 in the figures) is superimposed upon the thick landslide body called here L2, built by several more ancient mudflows.

3) Upper part L1
The upper landslide body (L1) is 1,5 km long and is nourished by surficial mudflows (about 5-10 meters thick) coming from the main scarp that forms the “apparent” crown area (660-670 m asl). These mudflows are continually replenished by mud and highly mineral water (sulphate) coming out from a long series of springs located over the altitude of 660 m a.s.l. in the main scarp. From here to the top of the mountain (680 m asl) the bedrock is moving valleyward through rotational surfaces of rupture emerging even on mountaintop and even on the northern slope beyond the summit itself. The carpet of surficial mudflows hides the tip line of these en-masse sliding blocks, which may be coming out near the base of apparent main scarp (660 m). The rate of movements is here in the order of magnitude of centimetres/decimetres per year. The depth of these surfaces was measured in over 20 meters by a series of inclinometers, showing clear single-shear surface of displacement (“mature” stage). These instruments went out-of-service in the nineties. This mechanism of movement locally alternates with “deep creep” deformations, well represented by few inclinometer showing displacements gradually decreasing with depth and going to zero in about 20 meters (“incipient” stage).

4) The middle channel
At the altitude of about 550 metres a.s.l. the landslide material enters into the narrow channel -about 1 km long, 30 m wide and 20 m deep- carved inside the flysch bedrock. This channel has no other function than transport the mud from the depletion to the deposition area, maintaining in confined conditions the fluidity of the mud, well demonstrated by the flat surface represented in the figure. The state of fluidity here varies in time and space, causing varying flow velocities inside the moving mass. Because o this, the narrow median channel is often empty and when this occurs the source and accumulation areas appear separated.

5) Lower part L2
More downward, at about 415 m a.s.l., at the exit of the natural channel, the muddy flow brakes, losing water and spreading on the lower and older lithosome (L2), circa 700 m long, where forms a series of superimposed thin layers, circa 1 meter thick.
From there, the material evolves in more plastic physical behaviour, well represented by the convex shape of the lithesome L2. From here to the valley bottom, sliding is the main mechanism of movement, as demonstrated by inclinometer I-12. The thickness of this portion of the landslide is about 15 metres.  Reaching the Tassobbio River, the landslide material forms a dam and a seasonal small lake. The damming of the river occurs by bulging and rising of the thalveg, due to the subterranean collision of the tip of the Roncovetro landslide with the tip of another landslide moving in opposite direction on the other side of the valley.

6) Triggers
The evident fluidity that characterizes the upper lithesome (L1) is remarkable: the permanent fluid state of consistency is caused by high porewater pressures maintained by percolating highly mineralised groundwater mixed with methane. The presence of methane is demonstrated by explosions occurred during the drilling operations, reported burning of water wells in proximity and by evidence of bubbling through groundwater, even if several sampling attempts failed. The important role of waters coming from the subsoil is evident in the graph of figure XXX, which represents a daily record of the water table depth with respect to the ground level. The open-pipe piezometer was situated on the crest of the mountain and the water table is found at the unexpected shallow depth of only 4.2 – 1.5 m. The graph shows sudden ground water table - level (GWLT) rises, of several metres, which cannot be directly related to the amount of precipitation (Bertolini & Gorgoni, 2001). As a hypothesis, the methane lowers the groundwater’s density and facilitates its upward percolation.

7) WiGim: 2014 survey and studies
The upper part of the Roncovetro Landslide (“Lavina di Roncovetro” Auctt) appear almost ideal to the implementation of the experimental monitoring network WiGim. The most important factors are:
1)    the landslide shows almost continuous movements;
2)    as a consequence of the existing different mechanisms of rupture (deep creep, sliding, flowing), the upper area of the landslide shows a very large range of displacements;
3)    the concave shape of the landslide allows the inter-visibility of the monitoring network nodes;
4)    there is the possibility to implement a parallel system of monitoring (automated geodesy) by positioning the total station (planned in the project) on the concrete building of the aqueduct reservoir rising on the very top of Monte Staffola (680 m a.s.l.) where the site stability is guaranteed and cross-checkable. From there, the whole monitoring area is visible.
In consequence of that, as planned, during the early months of year 2014 a series of surveys have been performed in order to characterise this slope, aiming to implement the experimental monitoring:
1)    a new terrestrial survey was performed using direct GIS mapping techniques and reporting the survey data on Google Earth, which was chosen because the more recent available topographic maps go back to the seventies and were too obsolete to be used as a cartographic basis. They are visible in some figure here.
2)    two complete new airborne photo-geological surveys were carried out by using a Light Sport Aircraft. Observations were made inflight and about 600 images were taken, both vertical (zenital) and prospective. Based on these images, a provisional digital elevation model was edited (see below). These images have been compared with available older images that were taken in a similar way in previous years, so allowing a detailed analysis of the recent evolution of the landslide.


ALMAGIA’ R. 1907: Studi geografici sulle frane in Italia. Mem. Soc. Geogr. It, 13(1), Roma.

BERTOLINI G. (con un contributo di GORGONI C.) 2001: La Lavina di Roncovetro (Vedriano, Comune di Canossa, Provincia di Reggio Emilia). Quaderni di Geologia Applicata, 8-2, Pitagora ed.