On 2 August, a populated wedge of hillside about two kilometres in length slid down and crashed into the Sun Kosi river, forming a large, temporary lake that impounded seven million cubic metres of water.
“A committee of mining and geology experts from the government concluded that the rock avalanche was accelerated due to water-induced weathering caused by mismanaged water resources,” says Madhukar Upadhya, watershed management expert and hydrologist.
Upadhya argues that this kind of avalanche happened because traditional techniques of forming ponds up in the ridges to manage monsoon runoff and stabilise landslides are forgotten and are no longer incorporated in watershed management efforts. “This allowed water to percolate through the cracks and faults in the rock, exposing inner surfaces to further erosion,” Upadhya tells SciDev.Net.
A geological study by Tribhuvan University (TU), in collaboration with engineering faculties at Japan’s Kagawa and Ehime Universities, that mapped landslides in Nepal along three major highways concluded that regions in the Himalayas with slate and phyllite rocks are prone to weathering and erosion that result in landslides.
“The rock formation of the present landslide area majorly consists of phyllite and schist,” says Ranjan Kumar Dahal, an engineering geologist at TU who co-authored the study published January 2013 in Georisk. “There are likely to be multiple slip surfaces inside the rock which reduces resistance, and which means the rock will easily fall, even during normal rates of precipitation.”
A 2012 study by TU’s geology department indicated a significant role for ‘antecedent’ rainfall — low intensity and long-duration monsoon rains with 2-3 days of interruption — in triggering large-scale landslides in Nepal’s Himalayas, especially the north-eastern side where the Sun Kosi flows. The monsoon season typically runs from June to September.
Dahal says the link between monsoon precipitation and landslides has not been factored into early warning systems. Days before the Sun Kosi landslide, data on a government website showed continuous build up of 25 millimetres of rain per day with a sudden rise to 70 millimetres. “This crossed the threshold for triggering a landslide but the system could not send an alarm as it does not respond to accumulative data characterised by antecedent rainfall,” Dahal tells SciDev.Net.
Reliable meteorological data also calls for a larger network of gauging stations, says Dahal. “Currently, the stations are very far apart. We want the distribution to be closer — a minimum of 25 kilometres would be good for now, but 10 kilometres between stations would be ideal.”
Link to study in Georisk
This article has been produced by SciDev.Net's South Asia desk.