Earthquakes often affect places with steep topography such as mountains or coastlines, where the shaking can cause the steep slopes to fail as landslides. These landslides contribute substantially to the immediate damage associated with earthquakes. They can also have less well-recognized but equally pernicious effects over longer time, for example by producing loose sediment material that generates debris flows during subsequent storms, or filling river valleys downstream and increasing the chances of flooding. Because they are important in generating the sediment that gets picked up and removed by rivers, earthquake-triggered landslides also determine how mountains form and change shape over geologic time. Yet, because large earthquakes only happen infrequently, much is not known about why earthquake-triggered landslides occur where they do, and how much sediment material they produce. Important lessons have come from past work studying landslide-affected areas years after earthquakes occurred, but much of the information is erased soon after these events, for example during rain storms that start to wash away sediment. The purpose of this rapid-response grant was to take advantage of the earthquake that struck Nepal in April 2015, known as the Gorkha earthquake. By allowing us to research this event immediately after it occurred, this grant made it possible to capture the kind of information that is only recorded for a brief time in the immediate aftermath.

The first part of the research involved assembling photographs collected by satellites and painstakingly analyzing these to identify landslides. We drew lines around each landslide, taking advantage of their clear contrast compared to vegetation-covered slopes. We individually mapped over 20,000 landslides caused by the Gorkha earthquake (see companion figure). We shared this dataset with agencies working to help the Nepalese in their recovery and rebuilding from the earthquake, and the data are now publicly available through the US Geological Survey web server. The landslide map also allowed us to look at patterns of where landslides occurred, telling us what was most important in causing slopes to fail during this earthquake.

In the second part of this project, we visited the areas in Nepal affected by the Gorkha earthquake to collect data and images needed to characterize individual landslides and to measure the amount of loose sediment material that they produced. Despite the logistical challenges of visiting areas damaged by the earthquake, we were able to make observations of many of the landslides from the ground. We also collected images using unmanned aerial vehicles (UAVs, or drones) equipped with high-resolution cameras, giving us information about areas not accessible from the ground. By combining our field observations together with the map we generated from satellite images, we are able to assemble the information to build better models for predicting landslides caused by earthquakes in the future, in Nepal and elsewhere around the world.

We also used the UAV photographs to generate 3-dimensional models of landslides (see companion figure). Using these models, we can measure the amount of sediment material produced in each landslides. During our field expeditions, we measured the size of the sediment material produced by several of the landslides, by digging pits and weighing the sediment that passed through sieves of different sizes. The information about sediment amounts and the sizes of the sediment particles tells us how much material we can expect to enter rivers, clogging them and increasing risks from flooding. Finally, we also collected samples of this sediment material and returned these samples to our laboratories in the US, where we analyzed the material for chemical indicators that tell us about longer-term rates of sediment production and removal from this landscape. This additional information helps us to understand whether the observations we have made from this one event can be expected for other events, both in the past and in the future.

This project developed a new collaboration between researchers at the University of Southern California, the University of Michigan, and Tribhuvan University in Kathmandu, Nepal. This network allowed us to conduct the post-earthquake fieldwork that would not have been possible otherwise. Several students were involved in this research, including two PhD students from USC, two PhD students from Michigan, several undergraduate students from both institutions, and a master’s student from Nepal. Research outcomes have been communicated in classrooms and at scientific conferences, in the scientific literature, through public release of the landslide dataset, and through publicly released videos (for example, see: https://www.youtube.com/watch?v=_cZLpBGWOsw) .

Last Modified: 12/09/2017
Modified by: A Joshua West