Investigations into the rail crash in northern Italy last week will focus on the origin of water that triggered the landslide which swept the train off its tracks.
Nine people died and 30 were injured when a mudslide set off by a sudden water flow slammed into a commuter train near the town of Merano in the Bolzano province in the foothills of the Dolomites (News last week).
Investigators will examine whether the mudslide was intiated by the failure of an irrigation pipe in an orchard upslope of the railway line. They will also consider whether exceptional rainfall leading up to the event was a factor, or even if uncharted aquifers played a part.
Local press reports say the orchard owner has admitted that water had leaked from its irrigation system, but that natural aquifers in the subsoil could have caused long term instability in the ground.
Bolzano prosecutor Guido Rispoli said the investigation, which started this week, would focus on establishing the origin of water that saturated the upper part of the slope by creating a reconstruction of the debris flow.
“The landslide was no more than 300m³ to 400m³ in size but had a huge effect. Preliminary investigations will last 15-20 days”
Rinaldo Genevois, University of Padua
He has appointed University of Padua geoscientist professor Rinaldo Genevois, Vincenzo Bixio, also from the University of Padua, and professor Alberto Mazzucato from Venice’s University Iuav as technical advisors. “We are at the very beginning of our investigations,” said Genevois.
“The landslide was no more than 300m³ to 400m³ in size but had a huge effect. I think that preliminary investigations will last about 15 to 20 days; on the basis of those results other investigations could be planned.”
Genevois specialises in debris flow and landslide modelling. Bixio specialises in land irrigation, land reclamation, irrigation and hydraulic engineering and Mazzucato is a geotechnical engineer who has carried out research into the mechanical behaviour of granular soils and water seepage.
The terrain in the area is a glacial moraine or till, which is rocky debris accumulated and deposited by a glaicer. The soil is made up of rocks and gravel, is not consistant or easy to move. This makes it prone to debris slides, according to Paolo Cortini, a consultant geologist to the Bolzano province.
Cortini could not comment on the geology of the specific area of the disaster but Italian Ordnance Survey mapping shows the area to comprise an allivual cone. This is sharply inclined alluvial deposits formed where a stream emerges onto a lowland after its descent from steep upland area.
Alluvial cones are steeper than the more common alluvial fans and show a greater average particle size. Fans are more likely to generate mudslides similar to that observed in last week’s crash.
Imperial College London professor of geotechincal engineering Richard Chandler, who has investigated a similar landslide in the area (see box below), predicted that investigators would find the cause to be a reactivation of historic ground movement, a wet winter, a burst water pipe or a combination of all three.
“Unless there was heavy rainfall at, or just before the event, it is likely that the burst irrigation pipe caused the debris slide.” Rainfall was not recorded on the date of the accident.
The Stava Valley disaster
The glacial moraine or till found across the mountainous Bolzano province makes it highly vulnerable to landslides.
Tragedy previously struck the region on 19 July 1985 when 268 people died after a tailings dam constructed for a fluorine mine collapsed in the Stava valley.
The Stava valley is just 8km north of the train crash site.
Disaster struck when the upper basin of the dam gave way and collapsed into the lower basin. This produced a muddy mass of approximately 180,000m³ of sand, slime and water.
The landslide reached a speed of 90km/h completely destroying three hotels, 54 homes, six industrial buildings and condemning eight bridges.
The collapse deposited a thick layer of mud, 200mm to 400mm deep over a 4.2km area.
The cause of the collapse was found to be the chronic instability of the dam and the low factors of safety used in its design.
The dam was assessed in 1974, and although a number of vital checks were missed, the assessment concluded the dam had been “taken to its limit”.
Four main structural reasons were cited in the collapse:
- The deposited slime from the fluorine mine had not settled because of the marshy nature of the soil on which the dams were built.
- Excessive height and inclination of the dams
- The decision to enlarge the bank according to the “upstream” method, which was the quickest and most inexpensive, but also the most dangerous
- Drainage pipes were installed incorrectly
Catch fence cleared
Initial speculation that the catch fence installed to retain rock fall was failed to do its job may be unfounded, engineers said this week.
Imperial College London professor of geotechnical engineering Richard Chandler said it was unlikely that the catch fence would have been designed to prevent such an accident.
“The catch fence would have only been intended to catch the odd boulder and would not have been expected or possible to stop the debris slide,” he said. The catch fence was constructed in the 1970s by the Italian state railway company.
Railway infrastructure operator Infrastructture Ferroviarie Alto Adige (SBA) said that neither the retaining wall next to the railway line nor the catch fence were altered during the 2005 reconstruction of the line.
“The retaining wall was constructed when the railway was originally built 100 years ago,” said a spokesman. “The fencing was constructed by the Italian state railways in the 1970s. SBA did not alter the wall when reconstructing the line” BAM Ritchies business manager David Gibson, who has recently installed landslide flow barrier on the A83 in Scotland, said loadings differ significantly depending on the type of event.#
“With a rock fall you have a limited mass travelling at high velocity, while a mudslide usually has substantially more material travelling at a slower velocity. The loading of the barrier will differ between events, which is why specialised barriers have been developed recently, he said.
Corinna Wendler, civil engineer at barrier specialist Geobrugg, agreed. “On a rockfall event there’s only a single impact, whereas with mudslides the loads are more spread out.”
Chandler added that the wall would not be to blame. “Cutting at the toe of a slope is very dangerous,” he said. “However debris slides are typically unexpected failures that occur on a natural slope and cutting at the toe generally has nothing to do with it.”
By Declan Lynch and Gemma Goldfingle