Nepal's Ongoing Seismic Risk: Kathmandu Valley

At 11:56 AM on April 25, 2015, a Saturday when many people were outdoors in the midday sun, a M7.8 earthquake ruptured the Main Himalayan Thrust approximately 80 kilometers northwest of Kathmandu. The rupture propagated eastward along the fault for roughly 150 kilometers, reaching closest approach to the capital within the first minute of shaking. When it was over, 8,964 people were dead. Another 21,952 were injured. Approximately 604,000 houses were destroyed and another 288,000 damaged — a total of nearly 900,000 residential structures affected across 14 districts of Nepal. The historic Dharahara tower, a 19th-century landmark in central Kathmandu that had survived two centuries of smaller earthquakes, collapsed. Temples across the Kathmandu Valley that had stood for 500 years were reduced to rubble. In mountain villages north of Kathmandu, entire settlements vanished under avalanches triggered by the shaking — Langtang village was buried completely by a seismic avalanche that killed 250 people in seconds.

The 2015 Gorkha earthquake was one of the deadliest natural disasters of the decade and the worst earthquake to strike Nepal in 81 years. It shocked the world, generated an enormous international relief response, and prompted a substantial rethinking of earthquake preparedness and building standards in one of the world's most seismically exposed countries. And yet — and this is the fact that most post-disaster accounts failed to communicate clearly — seismologists studying Nepal's earthquake history described the 2015 earthquake not as the catastrophe they had feared but as a partial and in some ways fortunate release of energy on a fault system capable of something considerably more destructive. The Kathmandu segment of the Main Himalayan Thrust — the section of the fault that runs directly beneath the valley floor where 3 million people live — did not rupture to the surface in 2015. The paleoseismic record suggests it last did so in 1255 CE and possibly in a sequence that also involved the great 1934 M8.0 Bihar-Nepal earthquake. The elastic strain accumulated since those events — the stored seismic energy representing 500 to 770 years of India-Eurasia convergence — remains locked beneath the valley, waiting.

This post covers both the earthquake Nepal has already experienced — the 2015 Gorkha event, its science, its damage, and its aftermath — and the earthquake Nepal's seismologists have been forecasting for decades: a full rupture of the Kathmandu segment of the MHT that would be, by virtually every scenario projection, the deadliest earthquake to strike any major city in living memory.

The Himalayan Collision: Earth's Greatest Continental Convergence

The Himalayan mountain range — the highest on Earth, with 14 peaks exceeding 8,000 meters — is the product of the ongoing collision between the Indian subcontinent and the Eurasian plate, a collision that began approximately 50 million years ago when the Tethys Ocean that had separated them closed completely and the two continental masses began to collide. India is moving northward into Asia at approximately 40–50 mm per year — one of the fastest convergence rates between any two major tectonic plates. About half of this convergence is accommodated by the thickening of the Tibetan Plateau (the Himalayan-Tibetan orogenic system is roughly twice the normal crustal thickness), and the other half is accommodated seismically — by repeated great earthquakes on the Main Himalayan Thrust and its subsidiary fault systems.

The Main Himalayan Thrust (MHT) is the primary fault interface between the Indian plate descending northward beneath Tibet and the overlying Himalayan wedge. It dips gently northward at approximately 5–15° — a remarkably shallow angle for such a geologically important structure — and extends from the surface at the Himalayan front (the Main Frontal Thrust where the mountains meet the Indo-Gangetic plain) northward beneath the Himalayan ranges to depths of 20–30 km. The locked portion of the MHT — the seismogenic zone where elastic strain accumulates between great earthquakes — lies primarily at depths of 10–20 km, directly beneath the urban corridor of Nepal including Kathmandu, Pokhara, and the Terai lowlands.

The Paleoseismic Record: A History Written in Trenches

The geological record of great Himalayan earthquakes extends far beyond the two centuries of European historical documentation of Nepal. Paleoseismic trenching across the Main Frontal Thrust — where the MHT reaches the surface at the base of the Siwalik Hills — has documented a series of major rupture events extending back over the past millennium, providing a picture of recurrence that is both sobering and scientifically important for understanding the current hazard state.

The work of Roger Bilham, Vimal Gaur, Peter Molnar, and subsequent Nepali and international collaborators has established a framework for understanding the Himalayan earthquake cycle. The paleoseismic record documents at least four major earthquake events in the past 1,000 years that ruptured the surface trace of the MHT in Nepal: a circa 1100 CE event, the 1255 CE earthquake (which contemporary Nepali chronicles record as killing one-third of Kathmandu's population including the reigning king Abhaya Malla), the 1505 CE western Nepal earthquake (estimated M8.5+, one of the largest Himalayan earthquakes in the historical record, though its rupture extent east of the Gorkha area is debated), and the 1934 M8.0 Bihar-Nepal earthquake (which killed approximately 10,600 people in Nepal alone and destroyed much of Kathmandu's historic building stock for the first time in the modern era).

The 1934 Bihar-Nepal Earthquake: The Modern Calibration Event

The January 15, 1934 Bihar-Nepal earthquake — M8.0, epicentered in the Bihar plains of India near the Nepal border — is the closest historical analog to the anticipated "Kathmandu scenario" earthquake and represents the primary calibration event for scenario development. The 1934 earthquake ruptured the MHT in the eastern Nepal segment east of Kathmandu, generating surface rupture along the Main Frontal Thrust and producing catastrophic damage in both Nepal and Bihar. In Kathmandu Valley, the earthquake destroyed 20% of all buildings and killed approximately 4,300 people in the valley alone — in a city with a population then of perhaps 150,000, a fraction of today's 3 million.

The proportional scaling from the 1934 event to today's population and building stock — even before accounting for the accumulated strain since 1934 and the possibility of a larger-magnitude rupture — produces casualty estimates in the tens of thousands for the valley alone. And the 1934 earthquake, like the 2015 Gorkha event, is thought to have ruptured primarily east of the Kathmandu segment — meaning the central Kathmandu zone of the MHT may not have experienced a full surface-rupturing event since the 1255 CE earthquake nearly 770 years ago.

The 2015 Gorkha Earthquake: What Happened and Why It Wasn't the Worst Case

The April 25, 2015 Gorkha earthquake ruptured a 150-km segment of the MHT from a hypocenter northwest of Kathmandu eastward toward the valley but stopping short of rupturing the central Kathmandu segment to the surface. The rupture was predominantly a deep slip event — the maximum slip occurred at depths of 10–15 km, well below the surface — with relatively limited shallow rupture that would have produced the most intense near-field ground motions. This geometry had two important consequences for the damage distribution.

First, the Kathmandu Valley received ground motions that were substantial but below the maximum possible for the MHT scenario — peak ground accelerations in the valley ranged from approximately 0.15–0.20g, well below the 0.4–0.8g that a full surface-rupturing MHT event directly beneath the valley would generate. This relatively moderate valley-floor shaking is one reason that the toll in Kathmandu itself — while devastating in absolute terms — was lower than the pre-earthquake scenario projections for a direct valley rupture. Second, the rupture propagated eastward during its approximately 50-second duration, which directed much of the radiated energy away from the mountain villages to the northwest and toward the densely populated valley and plains to the east — concentrating damage in Kathmandu, Bhaktapur, and Lalitpur rather than distributing it more evenly across the mountain regions.

Why Mountain Villages Were Destroyed When Cities Survived

The damage pattern of the 2015 earthquake revealed a geographic paradox that surprised some observers: while Kathmandu suffered catastrophic damage to historic structures, many of the modern concrete buildings in the valley survived — yet mountain villages like Barpak (the epicentral village), Gorkha, and communities in Rasuwa and Sindhupalchok districts were obliterated. Two factors explain this pattern. First, the mountain villages were in the near-field of the rupture — within 20–40 km of the maximum slip zone — where ground motions were dramatically higher than in the valley 80 km to the east. Second, mountain village construction — predominantly dry-stone masonry, earthen walls, and heavy slate roofs — is among the most seismically vulnerable building technology in existence. The heavy roofs of stone and slate that protect against cold and snow in the mountain environment become lethal in an earthquake: the walls fail, and the roof mass collapses onto the building's occupants.

The Kathmandu Valley: A City on Ancient Lake Sediments

The Kathmandu Valley occupies a former lake basin — the ancient Paleo-Kathmandu Lake, which filled the intermontane valley to depths of hundreds of meters before the lake drained when rivers cut through the surrounding hills, approximately 10,000–30,000 years ago. The lake sediments — fine lacustrine clays, silts, and sands that filled the basin during the lake's existence — remain as the primary geological substrate beneath the valley floor, and they constitute one of the most seismically amplifying geological settings in Asia.

The Lacustrine Sediment Amplification

The Kathmandu Valley's lake sediments (locally called the Kalimati and Lukundol formations) are characterized by very low shear wave velocities — 100–200 m/s in the shallowest layers — compared to the bedrock velocities of 1,500–2,500 m/s in the surrounding hills. This velocity contrast produces impedance ratios that drive amplification factors of 3–8 at the periods most damaging to Kathmandu's predominant 2–4 story construction (periods of 0.3–1.5 seconds). The spatial variation of amplification across the valley is significant: the thickest, softest sediments in the valley center produce the highest amplification; areas at the valley margins, where bedrock is shallower, experience less amplification but still dramatically more than the surrounding rock outcrops.

The 2015 earthquake provided the first comprehensive set of strong motion recordings in the Kathmandu Valley from a major MHT event — data that dramatically refined the understanding of valley amplification. Analysis of the strong motion records confirmed that the valley floor experienced ground motions 2–4 times higher than the surrounding bedrock sites for the same earthquake source, and that the amplification was most pronounced at periods of 0.5–1.0 seconds — precisely the period range that dominates the response of the 2–4 story reinforced concrete frame buildings that constitute the majority of post-1990 Kathmandu construction. This finding has direct implications for the scenario projections: a full MHT rupture directly beneath the valley, generating bedrock ground motions 3–4 times larger than 2015, would produce valley-floor shaking that exceeds the design capacity of virtually every building type in Kathmandu.

The Building Stock: A City Built Without Seismic Design

Kathmandu's building stock is one of the most seismically vulnerable concentrations of urban construction in the world — a product of rapid, largely uncontrolled urbanization over the past four decades in a country with limited regulatory capacity, endemic corruption in building permit and inspection processes, and a tradition of informal construction that predates any seismic building code.

Unreinforced Masonry: The Primary Killer

The historic building stock of the Kathmandu Valley — the brick masonry temples, palaces, and residential buildings that make Kathmandu one of UNESCO's most important cultural heritage sites — is predominantly unreinforced fired-brick masonry (locally called "Newari architecture" for the Newar people who developed it over centuries). This construction type, while structurally sophisticated in its traditional proportions and earthquake-resistant in some ways compared to simpler masonry, lacks the ductility and tensile strength to resist the lateral forces of a major earthquake without extensive cracking and collapse. The 2015 earthquake destroyed or severely damaged a large fraction of the valley's historic brick masonry buildings — including most of the UNESCO World Heritage monument zones in the Kathmandu, Patan, and Bhaktapur Durbar Squares — and demonstrated unambiguously that the traditional building stock cannot survive the design-level earthquake.

Non-Engineered Reinforced Concrete: The Modern Vulnerability

The rapid urban growth of the past three decades has filled the Kathmandu Valley with a new building type that presents a different but equally serious seismic vulnerability: non-engineered or poorly-engineered reinforced concrete frame construction. These buildings — typically 3–5 story concrete frame structures with brick masonry infill walls — were built by local contractors without structural engineering input, using whatever reinforcing steel, cement mix, and construction practice was available locally. The result is a building stock that appears modern but lacks the ductile detailing — the hook-ended stirrups, the continuous column ties, the adequate beam-column connection reinforcement — that allows reinforced concrete to deform without fracturing under seismic loading.

The 2015 earthquake's performance of these non-engineered RC frames was deeply concerning to post-earthquake reconnaissance teams. Many buildings experienced partial column failures, soft-story collapse mechanisms, and pancake collapses — exactly the failure modes that proper seismic detailing is designed to prevent. The widespread presence of this building type throughout the valley means that a larger, more proximal earthquake would produce collapse rates among the modern building stock that would dwarf those of 2015.

The Scenario: What a Full Kathmandu Segment Rupture Would Do

The earthquake scenario that Nepali seismologists and international researchers have been developing for decades — a full rupture of the Kathmandu segment of the MHT, comparable in magnitude to the 1934 Bihar-Nepal earthquake but with the rupture zone displaced westward to run directly beneath the valley — is the defining worst-case planning scenario for Nepal and one of the most consequential single-event earthquake scenarios on Earth.

NSET (the National Society for Earthquake Technology — Nepal), the primary Nepali earthquake preparedness organization, and international collaborators including GeoHazards International and the USGS have developed detailed Hazus-based loss estimates for this scenario. The headline projections — recognizing the enormous uncertainty in these numbers — range from approximately 20,000 to 100,000 deaths in the Kathmandu Valley alone, depending on time of day and assumptions about building performance. The range is wide because the dominant uncertainty is not in the earthquake source parameters but in the fraction of the building stock that would collapse — and that fraction depends critically on whether the high seismic detailing requirements of Nepal's National Building Code (NBC 105), adopted in 1994, have been applied to buildings that were nominally built under it.

The post-2015 engineering assessment of building performance provided a sobering answer to that question: even buildings notionally designed to NBC 105 provisions showed systematic deficiencies in seismic detailing, suggesting that the gap between code requirements and actual practice is large. If that gap means that a significant fraction of "modern" buildings perform like non-engineered construction, the upper end of the casualty range becomes more likely. Conversely, if the post-2015 building improvements — including the government's declared requirement for engineering certification for all new construction — are effectively implemented, the lower end may be achievable.

NSET and Nepal's Preparedness Efforts

Nepal's primary earthquake preparedness institution — the National Society for Earthquake Technology (NSET) — has been working on the Kathmandu scenario for more than 25 years, since before the 2015 earthquake brought global attention to Nepal's seismic risk. NSET's work spans the full preparedness spectrum: community-level earthquake preparedness training in Kathmandu Valley neighborhoods, school building retrofitting programs, hospital seismic assessment and improvement, building code development and training for engineers and masons, and the development of emergency response frameworks for post-earthquake search and rescue in the dense urban fabric of the valley.

The School Retrofit Program

NSET's most visible and replicable achievement has been the progressive retrofit of school buildings in the Kathmandu Valley — a program that began in the late 1990s and has since been replicated across Nepal and throughout South Asia. The program addresses the highest-priority life-safety concentration in the pre-code building stock: the government school buildings, most of which are unreinforced brick masonry construction from the 1950s through 1980s, filled with students and teachers during the 90% of working hours when an earthquake is statistically most likely to occur. The retrofit technique — center-core reinforced concrete columns tied to the existing masonry walls with horizontal stainless steel anchors — is low-cost, constructable by local artisans with limited training, and demonstrably effective at preventing collapse while allowing significant cracking and damage.

In the 2015 earthquake, every school building in the Kathmandu Valley that had been retrofitted through NSET's program survived without collapse — even in the most heavily shaken areas. The unretrofitted school buildings showed the full range of collapse behavior expected for unreinforced masonry. This before-and-after performance comparison — directly observable in the 2015 damage survey data — is one of the clearest empirical demonstrations anywhere in the world that low-cost masonry retrofit is effective at the scale of a major earthquake.

The Secondary Hazards: Landslides, Aftershocks, and Infrastructure Failure

In the mountain terrain of Nepal, the primary earthquake hazard cannot be separated from the secondary hazards that a major event would trigger simultaneously. The 2015 earthquake triggered more than 10,000 individual landslides in the mountain districts north of Kathmandu — including the catastrophic Langtang avalanche — blocking roads, damming rivers, and isolating hundreds of communities from emergency access for days to weeks. A full Kathmandu segment rupture, with ground motions substantially higher than 2015 in the mountain flanks of the valley, would trigger a landslide inventory that the most advanced slope stability models project at multiple times the 2015 total — potentially blocking every road access route into and out of the Kathmandu Valley simultaneously.

The valley's road network has limited redundancy: the primary routes connecting Kathmandu to the Terai lowlands and to the Indian border pass through narrow gorges cut by the Bagmati and Trishuli rivers through the valley rim hills. These gorges are prime sites for seismically triggered landslides and slope failures, and their blockage after a major earthquake would effectively seal the valley — isolating 3 million people from outside resupply and from the heavy equipment needed for large-scale search and rescue. Nepal's one international airport — Tribhuvan International — sits on the valley floor on lacustrine sediments, adjacent to active fault traces, and would need immediate post-earthquake structural assessment before it could safely operate the heavy relief aircraft that would be the primary external resupply mechanism.

The Convergence of Population Growth and Seismic Hazard

One dimension of Nepal's seismic risk that is rarely discussed in earthquake science literature but is critical to understanding the future trajectory of the hazard is the ongoing and rapid urbanization of the Kathmandu Valley. The valley's population has grown from approximately 150,000 in 1934 to 300,000 in 1970 to 1 million in 2000 to more than 3 million today — a 20-fold increase in the 90 years since the last M8.0 earthquake. This population growth has been concentrated precisely in the areas of highest seismic amplification: the soft-sediment valley floor that was largely undeveloped agricultural land in 1934 is now a dense urban fabric of non-engineered reinforced concrete frames and unreinforced masonry buildings housing the majority of the valley's population.

The physics of the seismic hazard has not changed since 1934. The MHT is accumulating strain at the same rate it always has. What has changed — dramatically and in a direction that only increases the consequence of the next major rupture — is the number of people and the number of buildings sitting above it. Every decade that passes without the Kathmandu segment rupturing is a decade of additional population growth, additional building construction, and additional seismic strain accumulation. The risk is not static. It is growing.

Conclusion

Nepal's seismic situation is, in some ways, the starkest example in this entire series of the gap between scientific knowledge and human consequence. The earthquake science is exceptionally clear: the Kathmandu segment of the Main Himalayan Thrust is locked, is accumulating strain at a measurable and well-constrained rate, has not ruptured in full since at least 1255 CE by the paleoseismic record, and is capable of an M8.0–8.5 earthquake that would generate ground motions in the Kathmandu Valley dramatically exceeding those of 2015. Three million people live in that valley on ancient lake sediments that amplify every incoming wave. A large fraction of their buildings were built without seismic design. The access routes into the valley will likely be blocked within minutes of the mainshock. The scenario casualty range begins at 20,000.

Against this picture stands a genuine preparedness effort — NSET's 25-year program of school retrofits, community training, and code development; the post-2015 code reforms; the growing cadre of Nepali engineers trained in seismic design; and an international scientific community that has made Nepal one of the most-studied earthquake hazard zones in the world. The question is not whether these efforts are real and valuable — they are. The question is whether their pace, given the rate of urbanization and building construction on the valley floor, is sufficient to reduce the vulnerability of the existing building stock faster than the population is adding new vulnerable structures above the locked fault. That race between preparedness and accumulating risk is the defining tension of Nepal's earthquake future — and the next great Himalayan earthquake will determine which side is winning.