--- title: "Tibet Earthquake Hit Homes That Open Map Data Doesn’t Show" url: "https://wherobots.com/todays-news/tibet-earthquake-exposure/" description: "Who lay in the Tibet earthquake's strong shaking? Open map data shows no building within 1 km of 40 of the 42 settlements there. Homes may stand unmapped." published: "2026-10-23" modified: "2026-10-24" --- # Tibet Earthquake Hit Homes That Open Map Data Doesn’t Show At 22:33 on October 22, 2026, a magnitude 7.0 earthquake struck Baqên County in Nagqu, northern Tibet. At least 16 people died, China’s state news agency Xinhua reported. [CNN](https://www.cnn.com/2026/10/22/asia/earthquake-chinas-tibet-latam-intl), [CBS News](https://www.cbsnews.com/news/powerful-earthquake-tibet-china/) and [Al Jazeera](https://www.aljazeera.com/news/2026/10/22/magnitude-7-earthquake-kills-16-people-in-chinas-tibet-xinhua-reports) carried the report. Nagqu’s media office said at least 339 homes and 20 livestock sheds collapsed, [Outlook India](https://www.outlookindia.com/international/tibet-earthquake-atleast-16-killed-54-injured-as-339-homes-collapse-in-nagqu) reported. Xinhua photos carried by the [Associated Press](https://www.timesunion.com/news/world/article/a-magnitude-7-earthquake-shakes-china-s-tibet-22489279.php) show families sleeping in tents because their homes were unsafe. Some of them slept at a primary school. News reports quickly gave the quake’s magnitude, its epicenter and how hard the ground shook. But they leave open a more practical question. Who and what lay in the strong shaking, and do open building maps show them? With Wherobots, we counted the mapped buildings and settlements inside the strong shaking. We also read US Geological Survey (USGS) estimates of the people who lived there. **What we found:** the October 22 quake hit homes that open map data does not show. Open map data (Overture Maps) shows no building within 1 km of 40 of the 42 settlements in the quake’s strong shaking (intensity VII or more, as modeled by USGS). That counts mapped buildings, and homes may stand there unmapped. No coverage we found has reported this gap. Intensity measures how strongly the ground shook at a place, on the Modified Mercalli scale. USGS says most homes in the area are adobe or unreinforced brick laid in mud, which collapse easily in strong shaking. ![Map of the October 22, 2026 Tibet earthquake's strong shaking around the epicenter in Baqên County. 40 of the 42 settlements inside it have no building mapped within 1 km in open maps.](https://wherobots.com/wp-content/uploads/2026/10/tibet-earthquake-exposure-oct22-v3-25554df2ce462a20.webp?ver=1791575225) *Settlements (villages, township and county seats) in the October 22 quake’s strong shaking, and whether open maps show any building within 1 km. Shaking: US Geological Survey (USGS) ShakeMap. Settlements and buildings: Overture Maps. Analysis: Wherobots.* Key takeaways - At least 16 people died, by state media reports, and the toll may rise. At least 339 homes collapsed across Nagqu, and families slept in tents. - Open map data shows no building within 1 km of 40 of the 42 settlements in the strong shaking, as modeled by USGS. This counts mapped buildings, not buildings on the ground. - USGS estimates that 6,348 people lived in that shaking, about one sixth as many as in the strong shaking at Tingri in January 2025. These are model estimates, and they miss herders who move with their animals. - There are two possible reasons fewer died than the 126 at Tingri (Chinese state media). Far fewer people were in the strong shaking, and violent shaking, intensity VIII or more, covered 48% as much land. Both come from USGS models, and the October 22 map may change. Our comparison cannot say how much each reason counts. - How we did it: one Wherobots job matched Overture buildings, places and settlements to the USGS shaking maps of both quakes. The maps hold 338,922 cells, each about 1 km across, and the job took about 3 minutes. ## How Wherobots makes this possible We needed five layers that are usually stored apart: modeled shaking, population estimates, building outlines, mapped places and settlement locations. In Wherobots, we matched each building, place and settlement to its spot on the shaking map and gave it that spot’s shaking level. We then counted what lay in the strong shaking and added up the area at each shaking level. ## What we found about the Tibet earthquake ### 1. Do open maps show the buildings in the strong shaking? Strong shaking reached 42 settlements: villages and township and county seats. We checked each one for buildings in Overture Maps, an open map that draws on OpenStreetMap and other open sources. - **40 of the 42 settlements** have no building mapped within 1 km. - **70 buildings** in all are mapped in the strong shaking. - **75 of 142 settlements** had no building mapped within 1 km in the strong shaking of the 2025 Tingri quake. So the gap is not new to this quake. A damage estimate built from open maps would count no building within 1 km of almost every settlement the strong shaking reached. Buildings may stand there unmapped. ### 2. How many people lived in the strong shaking? USGS estimates that: - **6,348 people** lived where the shaking reached intensity VII or more. - **2,652 of them** lived where it reached intensity VIII or more. - **USGS’s headline figure, about 110,000 people,** also counts lighter shaking, from intensity V up. We compare the people in strong shaking, intensity VII or more, with Tingri. The headline 110,000 also counts lighter shaking. These figures estimate how many people lived at each shaking level, not how many were hurt. They come from a global population model, which does not count herders who move with their animals. ### 3. Why did fewer people die than in the 2025 Tingri earthquake? In January 2025, a quake of similar size at Tingri, also in Tibet, killed 126 people, [PBS NewsHour](https://www.pbs.org/newshour/world/tibet-earthquake-kills-at-least-126-people-damages-hundreds-of-houses) and [Al Jazeera](https://aljazeera.com/news/2025/1/7/earthquake-hits-tibets-shigatse-what-we-know-so-far) reported. Coverage of the October 22 quake calls the area remote, as in [AccuWeather’s headline](https://www.accuweather.com/en/weather-news/magnitude-7-0-earthquake-strikes-remote-region-of-tibet-china/1949713). So we compared what lay in the strong shaking of the two quakes. ![Two maps at the same scale. The October 22, 2026 Tibet earthquake put an estimated 6,348 people (USGS model) in 4,441 km² of strong shaking. The January 2025 Tingri earthquake put an estimated 37,641 in 4,862 km².](https://wherobots.com/wp-content/uploads/2026/10/tibet-earthquake-exposure-compare-v3-ae49f135f301a10c.webp?ver=1791575225) *Strong shaking (intensity VII or more) and the settlements inside it, October 22, 2026 (left) and January 7, 2025 (right), at the same scale. Shaking: USGS ShakeMap. People: USGS Prompt Assessment of Global Earthquakes for Response (PAGER), its estimate of the people exposed. Settlements and buildings: Overture Maps. Analysis: Wherobots.* - **People:** USGS estimates 6,348 people in the strong shaking on October 22 and 37,641 at Tingri, about one sixth as many. - **Area:** the strong shaking on October 22 covered 91% as much land as at Tingri. - **Violent shaking:** intensity VIII or more covered 48% as much land as at Tingri. Only Tingri reached the next level, intensity IX. So the strong shaking covered about as much land but held far fewer people. That fits the coverage calling the area remote. But violent shaking covered about half as much land this time, so it may explain part of the lower toll too. These comparisons come from USGS models and cannot say how much each reason counts. Counts of collapsed homes by village would show which reason counts more. ## From shaking maps to real-world facts: the technology behind the analysis We brought together several types of information in Wherobots: - **Shaking data:** the USGS ShakeMap grid of modeled shaking for each quake. - **Population data:** the USGS Prompt Assessment of Global Earthquakes for Response (PAGER) estimates of the people at each shaking level. - **Map data:** buildings, places and settlements from Overture Maps. In one Wherobots job, we loaded each ShakeMap grid as a table. The grids for both quakes hold **338,922 cells**, each about 1 km across. Matching every building and settlement to them took **about 3 minutes**. This way of matching places on a map is called a [spatial join](https://wherobots.com/discover/spatial-join/). The same job can run on another quake’s ShakeMap. ## Why this matters beyond the Tibet earthquake Emergency management agencies and aid groups can see which settlements in strong shaking have no mapped buildings. This shows where an estimate built from open maps may count too little. For insurers and reinsurers, the same check shows where open building maps may have gaps. For utilities and infrastructure operators, the same match of shaking to assets could show what may need inspection. The same kind of matching works for storms. For [Super Typhoon Champi](https://wherobots.com/todays-news/typhoon-champi-path/), we matched mapped buildings to the storm’s forecast wind areas. This work moves from knowing **where the ground shook** to seeing **what open maps and USGS estimates show inside the strong shaking**. So far only a Nagqu-wide total of 339 collapsed homes is reported. For the families in Baqên County’s villages, the toll may still rise as rescuers reach them. ## Data sources and methodology We used USGS ShakeMap shaking models, USGS PAGER population estimates and Overture Maps buildings, places and settlements. Settlements are Overture localities. Deaths are as reported by Chinese state media, and the toll may rise. Reported injuries rose from 32 ([NBC News](https://www.nbcnews.com/world/china/70-earthquake-hits-chinas-tibet-killing-16-injuring-32-rcna605100)) to 76 ([Press Trust of India](https://theprint.in/world/16-killed-76-injured-as-magnitude-7-quake-hits-tibet-over-12000-evacuated/3079476/)). We followed the USGS PAGER method (Jaiswal and Wald, 2010), which counts the people at each shaking level. Like PAGER, we round the modeled intensity to the nearest whole level. Any modeled value of 6.5 or more therefore counts as intensity VII. USGS updates ShakeMaps as data arrives, and we used version 3 for October 22. **How we checked this.** An earlier run counted only modeled values of 7.0 or more as intensity VII. Its comparisons between the two quakes came out the same way. We switched to PAGER’s rounding rule before writing. We also compared the 70 mapped buildings in the strong shaking with 3,782 at Tingri. That comparison cannot be relied on, because the two areas are not mapped alike, as the settlement counts in question 1 show. Mapped places such as schools, clinics and shops numbered 2 on October 22 and 13 at Tingri. Both counts are too small to rely on. We had not planned to compare the areas of violent shaking. We report it because it suggests fewer people may not explain the lower toll alone. ## Frequently asked questions ### What happened in the October 22 Tibet earthquake? A magnitude 7.0 earthquake struck Baqên County in Nagqu, northern Tibet, at 22:33 local time on October 22, 2026. At least 16 people died, China’s state news agency Xinhua reported. Injury counts rose from 32 to 76 as reports came in. ### Do open maps show the buildings in the earthquake zone? Almost none near the settlements. Strong shaking, intensity VII or more as modeled by the US Geological Survey (USGS), reached 42 settlements: villages and township and county seats. Overture Maps draws on OpenStreetMap and other open sources. It shows no building within 1 km of 40 of those settlements. Buildings may stand there unmapped. In the strong shaking of the 2025 Tingri quake, it shows no building within 1 km of 75 of 142 settlements. ### What does intensity VII mean? Intensity measures how strongly the ground shook at a place, on the Modified Mercalli scale from I to XII. USGS calls intensity VII “very strong” shaking. Damage is slight in well-built buildings but considerable in poorly built ones. USGS says most homes in this area are adobe or unreinforced brick laid in mud. In this article, strong shaking means intensity VII or more. ### Why did fewer people die than in the 2025 Tingri earthquake? There are two possible reasons, and both come from USGS models. First, far fewer people lived in the strong shaking: 6,348 against 37,641, USGS estimates, though it covered nearly as much land as at Tingri (91%). Second, intensity VIII or more covered 940 km², about half of the 1,946 km² it covered at Tingri. These comparisons cannot say how much each reason counts. ### How did we measure this? In one Wherobots job, we loaded the USGS shaking maps for both earthquakes. We matched each Overture building, mapped place and settlement to its shaking level. Mapped places include schools, clinics and shops. We then read the USGS population estimates and compared the two quakes.