Multidimensional Landslide Vulnerability and Risk Assessment Using Principal Component Analysis in Lumbini Province, Nepal

In the rugged landscapes of the Himalayas, the ground is never truly still. The mountains are young, the rock is often fragile, and the rains can be violent, creating a perfect storm for landslides. For communities living on these slopes, a slide is not just a geological event; it is a threat to their homes, their farms, and their lives. Scientists have long known how to map where the ground is likely to give way, using tools to measure steepness, rock type, and rainfall. But knowing where a landslide might happen is only half the story. The real danger, or risk, depends on what is actually there when the ground moves. A steep, unstable hillside with no people on it poses little threat, while a gentle slope crowded with houses and roads can be devastating. To truly understand safety, experts must look at three things together: the physical likelihood of a slide, the assets exposed to it, and the vulnerability of the people living there—how poor they are, how well their houses are built, and how far they are from help.

A team of researchers from Tribhuvan University and local consulting firms set out to apply this complete picture to Lumbini Province in southwestern Nepal. This region is a microcosm of the country's diversity, stretching from the flat, fertile plains in the south to the high, snow-capped peaks in the north, passing through the fragile, eroding hills in between. The researchers wanted to move beyond simple maps of unstable ground and create a detailed risk assessment for every small administrative unit in the province, known as a ward. By combining data on the terrain with information about buildings, roads, and the socio-economic conditions of the people, they aimed to show exactly where the danger is highest and why. Their work reveals that the places most likely to slide are not always the places most at risk, and that the true threat is a complex mix of geology and human circumstance.

To build their assessment, the team first created a highly accurate map of landslide susceptibility. They analyzed thousands of past landslides, verified by field visits, and compared them against the landscape features that caused them, such as slope angle, soil type, and distance to roads. This allowed them to predict where future slides were most likely to occur. They found that more than 19.6 percent of the province's land area is highly or very highly susceptible to landslides. These dangerous zones are concentrated in the high mountain districts of the north, particularly in Rukum East, as well as in the mid-hill districts of Gulmi, Rolpa, and Pyuthan, and the fragile hills of Palpa and Arghakhanchi. However, the researchers knew that a map of potential slides was not enough. They needed to know what was sitting on those slopes.

The team then looked at exposure, identifying the physical assets that would be hit if a slide occurred. They counted the houses and measured the roads that fell within the high-risk zones. In the mountainous districts, settlements and infrastructure are often built directly on or near unstable slopes, meaning a large number of people and assets are exposed. In contrast, the flat plains in the south, while densely populated, sit on stable ground and face almost no landslide threat. But even with high exposure, the danger is not uniform. The researchers realized that two villages with the same number of houses on a risky slope could face very different fates depending on the people living there.

This led to the most innovative part of their study: measuring vulnerability. Instead of guessing which factors mattered most, the researchers used a statistical method to let the data speak for itself. They gathered information on sixteen different aspects of life in each ward, including literacy rates, poverty levels, the quality of housing, access to markets, and dependence on farming or forests. The analysis sorted these factors into four main groups that define how vulnerable a community is. The first group is socio-economic disadvantage, covering issues like low education and poverty. The second is environmental and livelihood exposure, which measures how much a community relies on the land for food and income. The third is institutional and infrastructure access, looking at how far people are from hospitals, government offices, and main roads. The fourth is structural and economic stability, which considers the strength of their homes and their ability to earn a living.

The results of this analysis were revealing. The researchers found that the biggest driver of vulnerability in Lumbini Province is not poverty alone, but the dependence on the land itself. Nearly 45 percent of the variation in vulnerability came from how much communities rely on agriculture, forests, and water resources. When a landslide destroys a farm or blocks a stream, it doesn't just damage a building; it destroys the means of survival for the entire household. Access to services and the stability of housing and income also played major roles, while pure socio-economic factors, though important, contributed less to the overall variation in this specific dataset. This means that a community might be poor but resilient if they have strong social networks and easy access to markets, while a slightly wealthier community could be extremely vulnerable if they are isolated and entirely dependent on a single, fragile crop.

Finally, the team combined these three elements—the chance of a slide, the assets exposed, and the vulnerability of the people—to create a single risk score for every ward. The picture that emerged was striking. While the physical danger of sliding is highest in the mountains, the actual risk is a product of that danger meeting human reality. The study identified that 16.2 percent of all wards in the province face high or very high risk. These areas are concentrated in the districts of Rolpa, Rukum East, Pyuthan, Gulmi, and Palpa. In these places, the steep, unstable terrain coincides with settlements that are exposed, and the people living there often lack the resources or access to services needed to recover quickly.

The researchers used statistical tests to prove that looking at just one factor is misleading. They showed that while the physical landscape explains most of where landslides happen, it explains very little about the final risk. The location of a district and its specific social and economic conditions are far more important in determining the final danger level. A ward with a moderately risky slope but poor housing and no road access can be far more dangerous than a ward with a steep slope but strong buildings and good connectivity. This finding challenges the old way of thinking that simply mapping unstable ground is enough for safety planning.

The study concludes that effective safety strategies must be tailored to the specific mix of hazards and human conditions in each ward. In the high-risk districts, protecting the land through better farming practices and stabilizing slopes is critical, but it must be paired with efforts to improve housing, build better roads, and ensure that remote communities have access to emergency services. By understanding that risk is a combination of the earth's instability and human fragility, local leaders in Lumbini Province can now prioritize their resources more effectively, saving lives and livelihoods in the places that need it most.