Sinkhole Risk and Route Choice on the Otavi Dolomite
A consulting team must advise the Roads Authority on the alignment of a 48 km gravel-road upgrade between two towns in the Otavi Valley. Alignment A follows the valley floor for its whole length, across dolomite and limestone. Three sinkholes have formed within 500 m of the road in 20 years, the largest 8 m across and 12 m deep, and the water table lies 35–45 m down and swings about 6 m between wet and dry seasons. Alignment A would cost N$310 million. Alignment B swings 6 km east onto schist and quartzite, crossing a 400 m band of shale; it needs two cuttings up to 14 m deep and would cost N$390 million. Traffic is 400 vehicles a day, a quarter of them heavy trucks hauling lime and cattle feed. Rainfall is 560 mm, almost all between November and March. A clinic 300 m from Alignment A has developed new cracks in its walls since the last heavy season. The Authority wants a decision within three months, and its budget speech has already promised that no route will cost more than N$350 million.
1. Explain, using rock type and process, why sinkholes form along Alignment A and why the shale band on Alignment B is also a hazard. Then recommend an alignment and justify your choice.
Read the worked answer
Alignment A (dolomite/limestone): these are carbonate rocks, slightly soluble in rainwater made acidic by dissolved carbon dioxide. Water moving down joints widens them into cavities; soil is washed into the cavities (piping), leaving a thin roof. Two things make this route dangerous here. First, the 6 m seasonal water-table swing: when the table falls, cavities lose buoyant support and the road plus 100 heavy vehicles a day must be carried by a thinner, weaker roof. Second, construction dewatering would lower the table further and trigger collapse. The clinic cracks 300 m away prove the process is active off the road line, not just beneath it - collapse cannot be predicted from a map.
Alignment B (shale): shale contains swelling clay minerals and has bedding planes. A 14 m cutting removes the toe that was holding the slope, and rain soaking in raises pore water pressure along the bedding planes, reducing shear strength so the slope slips.
Recommendation: Alignment B, but only with engineered cuttings (benching, cut-off drains, toe retaining structures, rockfall netting) and a slope-drainage design. Its hazard is predictable and designable; karst collapse location is not. The N$80 million difference (26%) must be weighed against whole-life costs on A: grouting, geogrid rafts, repeated reinstatement, and closure of the only lime-haulage route if a truck drops into a void. If the N$350 million ceiling is fixed, I would not simply take A - I would value-engineer B (realign one cutting) before accepting a karst route.
2. The Authority says it cannot afford Alignment B. Set out the investigations and conditions you would require before approving Alignment A, and the monitoring regime you would attach.
Read the worked answer
Approval should be conditional, not a yes/no. Before design: a geophysical survey along the whole corridor (microgravity or 2D electrical resistivity imaging) to locate cavities, followed by boreholes at no more than 50 m spacing to 60 m depth to confirm the rockhead and cavity pattern; a repeat survey after the first wet season, because a single dry-season picture will miss water-filled voids. Design conditions: no permanent dewatering - the water table is the roof's support, so pumping during construction must be banned or time-limited with a monitored trigger level; ponded runoff must be led away in lined drains so it does not recharge and dissolve rock under the road; high-risk sections to be spanned by a reinforced geogrid raft rather than simply filled; a grouting programme priced into the contract. Monitoring: precise levelling or crack gauges on the clinic and on any structure within 500 m, read quarterly and after every rain event over 50 mm; repeat resistivity every two years; a traffic limit and axle-load control; a written closure protocol with the hauler so a suspected void stops traffic immediately. Contractual: a retention or bond to cover probable future collapse repairs. Only if the geophysics comes back with no cavities within 30 m of the line would I approve A on technical grounds; otherwise the cost comparison is false economy.
3. The Authority argues the geology map is enough to decide. Explain what seasonal and operational evidence could still change your recommendation, and why.
Read the worked answer
A geology map shows where soluble rock occurs, not where a void is now. Three kinds of evidence could overturn the choice.
Seasonal evidence: 560 mm falling in roughly five months means a sharp recharge pulse. That pulse both accelerates dissolution (fresh acidic water reaching the rock) and raises the water table, while the late dry season reverses it. If monitoring showed the sinkholes cluster in the two months after the rains, the karst hazard is a recurring annual event, not a one-off, and A's whole-life cost rises sharply. Equally, if the shale band stays saturated through the wet season, pore pressures on B could stay high for months and require heavier drainage than designed.
Operational evidence: actual axle loads and truck numbers. If lime haulage grows beyond 400 vehicles a day, the loading on any thin karst roof increases and the case for B strengthens; if traffic is falling, the cost difference dominates.
Monitoring evidence: an InSAR or levelling result showing subsidence over a 200 m stretch would make A indefensible regardless of cost. Conversely, geophysics showing competent, cavity-free rock beneath A would let the cheaper route stand with grouting only at the clinic side. The point is that the decision should be tied to a trigger - a subsidence rate or a cavity count - not to the map alone.