NTU Singapore study shows major earthquakes can affect current sea-level projections in Southeast Asia
A weak, slowly flowing mantle layer beneath the region deforms after major tremors, causing the ground above to continue moving and sinking for decades
image:
(Right) ASE Chair Prof Emma Hill with EOS Research Fellow Grace Ng.
view moreCredit: NTU Singapore
Earth scientists from Nanyang Technological University, Singapore (NTU Singapore) have published an international study showing that major earthquakes in Southeast Asia can affect regional relative sea-level projections.
The findings show that large tremors can trigger long-term sinking of the land. If this post-earthquake ground movement is not accounted for in sea level modelling, coastal flood risks in low-lying areas could be underestimated.
The team found that a weak layer of hot rock in the upper mantle beneath the Sumatran backarc – the region behind Sumatra’s chain of volcanoes – deforms after major earthquakes.
Although this layer is solid, it can move slowly over time. This allows the ground above to keep shifting and sinking years after a major tremor.
The study, published in Communications Earth & Environment, a Nature Portfolio journal, was led by NTU’s Earth Observatory of Singapore (EOS) Research Fellow Dr Grace Ng, Asian School of the Environment (ASE) Asst Prof Lujia Feng, and Chair of ASE, Professor Emma Hill, who is also the Interim Director of EOS.
Sinking land affects sea-level estimates
While global sea-level rise is driven by climate factors like melting ice sheets and ocean warming, relative sea level is determined by how the local land moves. When the land sinks, local relative sea levels rise faster.
This study shows that major earthquakes do not just cause temporary shaking but also initiate decades-long land sinking – known as land subsidence – across Southeast Asia.
As these long-term ground movements have become better understood only in the past decade, they may not be fully included in existing sea-level estimates. This means future coastal flood risks for low-lying regions could be underestimated.
This phenomenon could also occur in other subduction zones, where one tectonic plate is forced beneath another, in other parts of the world.
Senior author of the paper, Prof Emma Hill, who is the AXA-Nanyang Professor in Earth and Environmental Science, said: "Most current sea-level projections focus primarily on climate factors like ice-sheet melting and ocean warming, but we must also look at how the Earth moves beneath our feet.
“Our new study shows that post-earthquake land sinking is an important factor in regional relative sea-level change. Incorporating these deep geological movements into our models will help us improve coastal planning for low-lying cities."
What happens beneath Sumatra after major earthquakes
The NTU-led team studied up to two decades of ground movement data from Singapore, Malaysia and Thailand to understand how the region continued to move after major earthquakes.
These included the 2004 Sumatra-Andaman earthquake and the 2012 Wharton Basin earthquakes.
The data showed that the ground continued to move even in places more than 600 kilometres from where the earthquakes occurred.
This long-distance movement suggests that a weak layer deep below the region is allowing the Earth’s surface to keep adjusting after major earthquakes.
Lead author Dr Grace Ng said, “When massive earthquakes strike, they do not just shake the ground for a few minutes. They set off a slow adjustment deep within the Earth that can continue for years.
“Our study shows that a weak layer of hot rock beneath the Sumatran backarc can slowly deform after major earthquakes. This helps explain why the land above can continue to shift and sink across areas hundreds of kilometres away from the earthquake.”
To test what was happening below the surface, the researchers used computer models of the Earth’s layers and compared them with ground movement recorded by GPS stations.
They found that the observed movement could be explained only if the upper mantle beneath the Sumatran backarc was weak enough to flow slowly over time.
This gives scientists a clearer picture of how major earthquakes can continue to affect land height long after the shaking has stopped.
Co-author Asst Prof Lujia Feng, an expert in using satellite positioning data to study the Earth’s crustal motions and natural hazards, added: “This study would not have been possible without more than a decade of continuous observations from ground-based GPS networks across the region. Such long-term geodetic records are vital for revealing how the solid Earth responds to great earthquakes, and how these processes evolve over time.”
Improving coastal planning
The study highlights why sea-level projections in Southeast Asia should account for both rising seas and moving land.
For coastal planners, relative sea level matters most. This refers to how high the sea is compared with the land at a specific location.
If land is sinking, water levels can rise faster relative to the coast, increasing flood risks for low-lying areas.
By showing how post-earthquake land movement can affect relative sea levels, the NTU study provides data to improve future coastal risk models, which can then inform and support longer-term planning for flood defences, drainage systems and coastal infrastructure in Southeast Asia.
Journal
Communications Earth & Environment
Method of Research
Data/statistical analysis
Subject of Research
Not applicable
Article Title
Weak asthenosphere of Sumatran backarc revealed by long postseismic geodetic records
Article Publication Date
6-Jul-2026
Mainshock locations shape earthquake size distributions
Japanese earthquake sequences show b-values are controlled more by place than by time
Kyoto University
image:
Spatial distribution of b-values around Japan and the locations of magnitude (M) 6 and larger earthquakes. Colors indicate the spatial distribution of b-values. The results suggest that large earthquakes tend to occur in regions with relatively low b-values. The b-value map was estimated from earthquake data after excluding the specific earthquake sequences analyzed in the study.
view moreCredit: Aron Mirwald
Kyoto, Japan -- Japan is well known for its large earthquakes, but not all regions show the same patterns of earthquake activity. One way to understand which places tend to experience large or small earthquakes is the b-value, a key statistical measure long used by researchers for understanding seismicity and assessing earthquake occurrence patterns.
The b-value describes the relative numbers of small and large earthquakes in a given area. A high b-value means that small earthquakes are comparatively more frequent, whereas a low b-value indicates a relative increase in the proportion of larger earthquakes.
Many previous studies have suggested that b-values may decrease before -- or increase after -- large earthquakes, possibly reflecting changes in stress before and after rupture. Other studies have proposed that large earthquakes tend to occur in areas where b-values are low. Until now, however, it has remained unclear whether time-dependent changes or spatial differences are more important when comparing many large earthquake sequences.
To address this question, a team of researchers from Kyoto University and ETH Zurich set out to analyze the Japanese earthquake catalog, examining how b-values vary before, after, and around large earthquakes. The study was led by Aron Mirwald, formerly a short-term visiting student at Kyoto University and a doctoral student at ETH Zurich, together with Bogdan Enescu of Kyoto University and Leila Mizrahi and Stefan Wiemer of the Swiss Seismological Service (SED) at ETH Zurich.
The research team examined large earthquake sequences in Japan using the Japan Meteorological Agency earthquake catalog from January 2000 to July 2025. They carefully restricted the data by depth, distance from the coast, and space-time windows in order to reduce catalog incompleteness and possible contamination from other large earthquakes.
Their results revealed no clear evidence for a systematic decrease in b-values before large earthquakes, or for a systematic increase afterward. Instead, the team found that large earthquakes tend to occur in regions where b-values are slightly lower than in the surrounding areas. Moreover, differences in b-values among earthquake sequences could be explained mainly by the location of the mainshock.
"Our results suggest that, at broad spatial scales, b-values are controlled more by where a large earthquake occurs than by systematic changes before or after the event," says Enescu. "This highlights the importance of local geological and stress conditions in shaping earthquake size distributions."
These findings suggest that b-values may reflect local geological and physical conditions, such as rock properties and stress state. Although this conclusion does not imply that individual earthquakes can be predicted, it provides a stronger statistical basis for interpreting b-values around large earthquakes and may contribute to future studies of long-term earthquake hazard.
###
The paper "b-values of large earthquake sequences depend on their mainshock location" appeared on 12 July 2026 in Geophysical Research Letters, with doi: 10.1029/2025GL121450
About Kyoto University
Kyoto University is one of Japan and Asia's premier research institutions, founded in 1897 and responsible for producing numerous Nobel laureates and winners of other prestigious international prizes. A broad curriculum across the arts and sciences at undergraduate and graduate levels complements several research centers, facilities, and offices around Japan and the world. For more information, please see: http://www.kyoto-u.ac.jp/en
Journal
Geophysical Research Letters
Method of Research
Data/statistical analysis
Subject of Research
Not applicable
Article Title
b-Values of Large Earthquake Sequences Depend on Their Mainshock Location
Article Publication Date
12-Jul-2026
