Landslide Risk Profile with Incomplete Evidence
A district of 180,000 people in a hilly region must finalise its landslide risk profile before the monsoon arrives in six weeks. The national susceptibility map (1:50,000) places 12 villages in the highest class. A recent NGO drone survey, commissioned after three weeks of rain totalling 410 mm against a monthly average of 280 mm, found tension cracks and seepage on 31 additional slopes. The district has budget for only eight geotechnical site investigations. Community reports mention springs that have dried up since last year, but there are no local rain gauges or soil maps. The district commissioner proposes issuing a blanket relocation order for the 12 mapped villages to demonstrate action before the election. Local leaders say the drone survey is alarmist and that no one will move without compensation. A hospital, two schools and a market lie within the 12 villages.
1. Within the eight-investigation budget, how should the district decide which slopes and villages to prioritise, and what should it do about the 31 unverified slopes in the meantime?
Read the worked answer
Treat both maps as evidence, not verdicts. Build a simple prioritisation matrix for all 43 candidate locations (12 mapped plus 31 drone-identified) scoring: population and critical facilities exposed (hospital, schools, market); physical evidence such as tension cracks, seepage, recent deforestation and slope angle; antecedent rainfall and community reports of spring changes; accessibility for investigation and evacuation; and uncertainty, especially drone-only sites with no ground check. Use the eight investigations on the highest combined scores, ensuring at least two drone-only sites near the hospital or schools and one lower-score control site to test the drone method. Do not investigate only the 12 mapped villages: the drone evidence is site-specific and post-rainfall, so it may be more current than the national map. For the 31 unverified slopes, apply interim no-regret measures: trained community members record crack widths weekly, install simple rain gauges, restrict new construction on suspicious slopes, clear drainage, and agree rainfall-triggered temporary evacuation for the highest-exposure hamlets. Document assumptions and residual uncertainty rather than presenting false precision. This is proportionate precaution under incomplete data, not blanket action.
2. Should the commissioner issue a blanket relocation order for the 12 villages? Justify your advice.
Read the worked answer
No. A blanket relocation order is a high-cost, irreversible decision that should not rest on a 1:50,000 map plus an unverified drone survey. It would also likely fail: local leaders have already rejected it, compensation is unspecified, and election timing makes it look political. But the duty of care is real, so differentiate. For the highest-risk hamlets, those with cracks, seepage, critical facilities and no safe evacuation route, prepare temporary evacuation during intense rainfall, identify host sites and explain triggers. For the other mapped villages, strengthen early warning, drainage and monitoring, and commission investigations before any permanent relocation. Any relocation must follow due process: participatory risk assessment, alternative land, livelihood restoration, compensation, and free, prior and informed consultation. Frame the decision as investigate, monitor and prepare rather than do nothing; publish the criteria and review dates. If an investigation confirms imminent collapse risk, issue a targeted, time-bound evacuation for that slope, not a district-wide order.
3. What minimum data and arrangements would you put in place to set local rainfall thresholds and trigger actions before the monsoon?
Read the worked answer
Minimum package: at least three calibrated rain gauges distributed across the elevation range, plus daily manual readings by trained community volunteers; a seven-day antecedent rainfall record, because landslides here followed 410 mm in three weeks; geological survey classification of the eight investigated slopes covering soil depth, permeability, bedding and old slip surfaces; a community event calendar recording past landslides, spring changes and the rainfall associated with them; and crack and seepage monitoring on the 31 drone sites. Use an empirical intensity-duration approach, but set conservative local triggers because data are thin: for example, alert at 100 mm in 24 hours when the previous seven days exceed 200 mm, and evacuate high-risk hamlets at 150 mm in 24 hours or any rapid crack widening. Publish the triggers, test them in a tabletop exercise, and review after each rain event. Treat triggers as decision aids, not guarantees, and pair them with visual signs and clear local authority to act.