Rising Seas Threaten Thousands of Heritage Sites in China’s Greater Bay Area

More than 3,500 historic temples, gardens, ancestral halls, and ancient structures in one of China’s most economically dynamic regions could be flooded by a single extreme storm surge, according to a new study that has, for the first time, mapped the climate threat facing every registered immovable cultural heritage site in the Guangdong–Hong Kong–Macao Greater Bay Area. The research, published in the International Journal of Disaster Risk Science, evaluated 11,220 heritage sites across the eleven cities of the Pearl River Delta under two storm surge scenarios and found that nearly a third of them lie within the footprint of a once-in-a-century flood event. As typhoons intensify with warming seas, the findings expose a troubling gap between the region’s celebrated cultural wealth and its preparedness for the rising water.

The Greater Bay Area, which encompasses Guangzhou, Shenzhen, Zhuhai, Foshan, Huizhou, Dongguan, Zhongshan, Jiangmen, and Zhaoqing along with the Hong Kong and Macao Special Administrative Regions, is both a powerhouse of global commerce and a repository of centuries of southern Chinese heritage. Its inventory includes two UNESCO World Heritage properties, 105 national-level key cultural relics protection units, 324 provincial-level units, more than 2,000 municipal and county-level sites, and nearly 8,700 unclassified heritage structures. Yet the same low-lying, estuary-fringed geography that made the delta a hub of trade also makes it a trap for storm water. Between 2013 and 2022, fifty storm surges struck the region, triggering twenty-six recorded disasters and economic losses exceeding 25.8 billion yuan. The region’s heritage has already felt the consequences: Typhoon Mujigae in 2015 caused structural failures at Yuyin Shanfang, a renowned classical garden in Guangzhou, and in 2016 successive super typhoons damaged the World Heritage site of Gulangyu in neighboring Fujian along with dozens of other protected sites.

To quantify the danger, the research team, led by Juanyu Wu of South China University of Technology, coupled the Intergovernmental Panel on Climate Change’s hazard–exposure–vulnerability framework with high-resolution hydrodynamic modeling using the Advanced CIRCulation model, or ADCIRC. The model’s unstructured triangular mesh, generated with OceanMesh2D, spans a domain stretching from 104°E to 124°E and 14°N to 26°N, with grid resolution tightening from 30 kilometers at the open ocean boundary to just 200 to 500 meters along the Greater Bay Area’s intricate coastline—a total of 164,909 grid cells and 102,795 grid points. This refinement matters because surge behavior in narrow, branching estuaries like the Pearl River differs dramatically from open-coast dynamics, and coarser models such as SLOSH sacrifice precisely the local detail that heritage risk assessment demands.

The team constructed two synthetic worst-case scenarios by combining the track of Typhoon Mangkhut, the devastating 2018 storm that remains the benchmark for regional surge modeling, with the intensity parameters of Typhoon Rammasun for the 100-year return period and Typhoon Usagi for the 20-year event. Rammasun made landfall in 2014 with a minimum central pressure of 910 hectopascals, while Usagi struck Shanwei in 2013 at 935 hectopascals. Maximum wind speeds and wind circle radii were derived from empirical relationships linking central pressure to wind structure, and the coupled Holland–ADCIRC wind-and-water model was validated against tide gauge observations from Typhoon Kalmaegi, with simulated errors contained within roughly 24 centimeters. The team also confirmed through copula-based joint probability analysis that the track-intensity combination exceeds a 100-year recurrence interval, ensuring the scenario is physically plausible rather than an artificial exaggeration.

The simulated results are stark. Under the 100-year scenario, surge heights reach six meters at the Guangzhou and Huangpu tide stations and around 5.5 meters at Haiqin, with the peak concentrated in a window of less than twelve hours. Nearly 19 percent of the region’s land area is inundated, with maximum flood depths reaching 59.4 meters in terrain pockets. Zhongshan, Foshan, and Zhuhai fare worst in proportional terms, with 71.04, 68.16, and 55.50 percent of their land areas respectively submerged. Intersecting the flood footprint with the heritage database identified 3,587 affected sites—31.97 percent of the total—of which nearly a quarter experienced the most severe hazard level. Even the more frequent 20-year scenario floods 2,413 sites, or 21.51 percent, demonstrating that heritage damage is not a tail risk but a recurring threat.

The exposure analysis reveals that danger is not distributed evenly across the landscape. Exposure, measured through land use type and the spatial density of heritage assets, is driven primarily by urbanization: sites embedded in dense residential and commercial districts face greater surrounding infrastructure and population pressure. Guangzhou emerges as the most exposed city in both scenarios, with more than half of its affected sites in the highest exposure class, followed closely by Foshan, where a large heritage stock overlaps heavily with inundation-prone ground. The researchers applied a value-equivalence conversion method to make sites of different protection grades comparable, weighting a national-level unit as equivalent to eighty unclassified sites, which allowed the density calculations to reflect cultural significance rather than raw counts.

Vulnerability, by contrast, proved moderate across most of the region. The team scored each site according to construction material, age, and the economic capacity of the surrounding city, weighting material sensitivity highest at 0.5. Ancient sites scored the highest material vulnerability at 0.8150, followed by ancient buildings and tombs, while rock carvings and stone inscriptions scored lowest. Because most affected structures date from the Ming-Qing or modern periods, their structural resilience tempers the overall vulnerability picture, though clusters of highly vulnerable sites persist in Zhuhai, Huizhou, Zhaoqing, and Zhongshan. The integrated risk calculation—hazard multiplied by exposure multiplied by vulnerability—then translated these dimensions into four risk tiers for each scenario, classified using the natural breaks method to capture the distinct statistical distributions of the two flood intensities.

Perhaps the most consequential finding concerns the mismatch between official protection status and actual risk. Under the 100-year scenario, national-level heritage units mostly fall into the lowest risk tier, benefiting from clear protection boundaries and stronger management, while provincial, municipal, county-level, and unclassified sites cluster in the highest tiers. More than 70 percent of high-risk relics are unclassified units—structures that lack formal designation and, presumably, dedicated conservation resources. This regulatory blind spot means the very sites most likely to be damaged by intensifying storm surges are those with the weakest institutional safeguards, a pattern the authors argue demands a risk-informed reallocation of conservation funding before irreversible losses occur.

The spatial pattern of risk also shifts between scenarios, complicating protection planning. Under the extreme 100-year event, high-risk clusters concentrate in Guangzhou and Foshan, where dense heritage and severe inundation converge, making these cities the clear priority for emergency-oriented defenses. Under the more frequent 20-year scenario, the risk hotspots migrate toward Shenzhen and Dongguan, where more than half of the higher-risk sites are located, while Hong Kong, Huizhou, and Zhaoqing emerge as low-frequency but high-impact vulnerable zones. The authors recommend a zoned, prioritized management strategy: reinforced disaster prevention in the core delta cities for extreme events, enhanced routine monitoring in Shenzhen and Dongguan, and resilience-building in the peripheral cities. They also acknowledge limitations, including the absence of vertical dike data in the simulation, the coarse 30-meter resolution of the elevation model relative to centimeter-level surge calculations, and the exclusion of urban pluvial flooding, which would compound coastal inundation. Even so, the study delivers a sobering message: as the IPCC projects more frequent and intense extreme sea-level events, the Greater Bay Area’s cultural legacy now stands, quite literally, in the water’s path.

Subject of Research: Storm surge risk to immovable cultural heritage in the Guangdong–Hong Kong–Macao Greater Bay Area

Article Title: Storm Surge Risk Assessment of Immovable Cultural Heritage in the Guangdong–Hong Kong–Macao Greater Bay Area

Article References: Storm Surge Risk Assessment of Immovable Cultural Heritage in the Guangdong–Hong Kong–Macao Greater Bay Area. (n.d.). https://doi.org/10.1007/s13753-026-00772-4

Image Credits: AI Generated

DOI: 10.1007/s13753-026-00772-4

Keywords: storm surge, cultural heritage, Greater Bay Area, ADCIRC model, climate change, flood risk, heritage conservation, typhoon, risk assessment, Pearl River Delta, IPCC framework, coastal inundation