← Latest papers
🔬 physics

A Referenced Empirical Framework for Regional adjustment of Ground-Motion Models in the Indian Himalayas

This study proposes and validates a reference-based empirical framework for developing regionally adjusted Ground Motion Models in the Indian Himalayas using a dataset of 724 recordings, demonstrating that constrained and simplified adjustment models significantly improve hazard estimates over global models in this data-limited region.

Original authors: Shikha Sharma, . Omkar, Utsav Mannu, Sanjay Singh Bora

Published 2026-11-19
📖 4 min read☕ Coffee break read

Original authors: Shikha Sharma, . Omkar, Utsav Mannu, Sanjay Singh Bora

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Earthquakes are not uniform events; the shaking they produce depends heavily on where the rupture happens, how the energy travels through the earth, and the specific geology of the ground beneath a building. To keep people safe, engineers rely on mathematical maps called ground-motion models. These tools predict how strongly the ground will shake at a specific location for a given earthquake size, allowing them to design structures that can withstand the force. In well-studied regions like California or Japan, scientists have built these maps using thousands of recordings from actual earthquakes, creating highly reliable guides. However, in other parts of the world, such as the Indian Himalayas, the data is sparse. The region is tectonically complex, formed by the massive collision of two continental plates, and lacks the dense network of sensors found in the West. When engineers in India try to use models built for California, the predictions often miss the mark, either overestimating or underestimating the danger, because the way seismic waves travel through the Himalayan crust is different.

A team of researchers from the Indian Institute of Technology Gandhinagar and Earth Sciences New Zealand has developed a new way to fix this mismatch without needing a massive new dataset. Instead of trying to build a completely new model from scratch, which would require more earthquake recordings than currently exist, they used a "referenced empirical" approach. Imagine taking a detailed, reliable map of a familiar city and using a few local landmarks to redraw the streets of a neighboring town that looks similar but has different traffic patterns. The researchers started with five of the most trusted global models used worldwide. They then compared the predictions from these global models against a specific collection of 724 actual earthquake recordings from the Himalayas, gathered between 1986 and 2018. These recordings came from 189 different earthquakes and 197 stations, covering a wide range of sizes and distances, though the data was still thin for very large, nearby quakes.

By analyzing the difference between what the global models predicted and what the sensors actually recorded, the team identified a consistent pattern of error. The global models were not capturing the unique way the Himalayan ground absorbs and transmits energy. To correct this, the researchers derived specific adjustment factors. They tested three different ways to apply these corrections. The first was a complex model that tried to account for every variable. The second was a "constrained" model that forced the predictions to match the global models very closely for distances within 10 kilometers, a zone where data is scarce and predictions are most uncertain, while allowing adjustments for farther distances. The third was a simplified model that relied almost entirely on a linear correction based on distance.

The results showed that the complex model was not necessarily better. In fact, the constrained model and the simplified distance-only model performed just as well, if not better, in matching the real-world observations. The constrained approach proved particularly effective because it maintained stability in the near-source zone, where the ground motion is most violent and difficult to predict, while still correcting the errors that appeared at intermediate and far distances. The simplified model demonstrated that a straightforward adjustment based on distance was sufficient to capture the dominant regional effects, avoiding the risk of overcomplicating the math with terms that the limited data could not support.

This work provides a practical, data-driven framework for updating seismic hazard assessments in the Indian Himalayas. By anchoring global models to local reality, the researchers have created a set of tools that engineers can use to design safer buildings without waiting for decades of new data to accumulate. The study confirms that while the Himalayas have unique characteristics that global models miss, these differences can be systematically corrected. The findings suggest that for this region, a simpler, distance-focused adjustment is often more robust than a complex one, offering a reliable path forward for reducing earthquake risk in one of the world's most seismically active and populated areas.

Drowning in papers in your field?

Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.

Try Digest →