Debris-Flow Risk on a Greater Caucasus Road Alignment
A 14 km road reconstruction is planned in a Greater Caucasus valley in Georgia between 1,450 m and 2,100 m. The slopes above the alignment are 35-45 degrees and are underlain by fractured Jurassic shale and limestone. Mean annual precipitation is about 1,150 mm, and May-June snowmelt normally raises stream discharge. In June, 42 mm of rain fell in nine hours. A debris flow buried 200 m of the existing road, blocked a 3 m culvert, and deposited boulders up to 1.8 m across. The municipality must choose between Route A, which follows the valley floor and needs six bridges, and Route B, which uses a mid-slope bench and requires eleven cut slopes. The decision is needed before next spring's thaw. Residents report that small slides have become more frequent since a 2019 forest harvest on the upper slope. A weather station exists 9 km away, but there is no local rain gauge or ground-motion sensor.
1. Explain why the June event was a debris flow rather than a simple river flood, and identify the factors that made the slope vulnerable.
Read the worked answer
The rainfall total was not extreme: 42 mm in nine hours is about 4.7 mm/h. A river flood alone would mainly raise turbid water, not transport boulders of 1.8 m. The debris flow requires a slope failure that saturated and mobilised colluvium. Three factors combine. First, May-June snowmelt had already raised soil moisture, so the June rain fell on wet ground. Second, fractured shale produces fine matrix material, while limestone contributes blocks; on 35-45 degree slopes this mixed material can fail as a viscous flow. Third, the 2019 forest harvest reduced interception and root cohesion on the upper slope, increasing water delivery and reducing slope strength. The blocked 3 m culvert is a symptom, not the main cause: the flow exceeded the channel capacity and then deposited on the road. The event should therefore be interpreted as a rainfall-triggered slope-channel cascade, not as an unusually large flood.
2. Compare Route A and Route B. Which would you recommend, and what design changes or additional data could reverse your choice?
Read the worked answer
Route A is exposed to debris-flow runout and burial because it sits on the valley floor, and its six bridges can act as choke points. Route B avoids the main runout zone but places eleven cuts into fractured shale, which can create new slides and intercept groundwater. My recommendation is Route B only if it is redesigned as a low-cut alignment with viaducts or clear-span bridges at debris fans, deep drainage above cut faces, and no side-casting of spoil. If that is unaffordable, Route A is preferable only with raised embankments, debris-deflection walls, oversized culverts, and a maintenance budget for clearing. Additional data that could reverse the choice are LiDAR bare-earth imagery, engineering-geology boreholes, and a local rain gauge with antecedent-moisture records. If LiDAR shows that Route B crosses several deep-seated landslide bodies, Route A with robust protection may become the lower-risk option.
3. Design a monitoring and early-warning plan for the chosen route, including thresholds and institutional responsibilities.
Read the worked answer
The plan should combine rainfall, hydrological and ground-motion monitoring. Install at least two rain gauges in the valley and one pressure transducer in the main torrent; add extensometers or tiltmeters on the upper slope and an acoustic flow sensor at the culvert. Because the nearest station is 9 km away, local gauges are essential. Use intensity-duration thresholds rather than a single rainfall total: for example, alert at 25 mm in 12 hours when antecedent rainfall in the previous five days exceeds 40 mm, and evacuation at 40 mm in 12 hours with ongoing snowmelt. Community observers should report new cracks, muddy springs and unusual rumble. The municipality should own the alert protocol, the road agency should close the route, and the geology service should inspect after each event. A 2019 forest-harvest area should be a priority inspection zone. Data should be reviewed annually and thresholds revised after any event.