When people think about future coastal flooding, they usually picture one process: the sea rising. And that is certainly part of the story. Global sea level is increasing, and coastal cities around the world are already feeling the consequences through higher tides, storm surges, and more frequent extreme water levels.
But in many coastal regions, the story does not stop at the shoreline.
Sometimes, the land itself is sinking. And when the sea rises while the ground sinks, flood risk can grow much faster than ocean change alone would suggest. That idea became the starting point of our recent study on the Guangdong-Hong Kong-Macao Greater Bay Area (GBA), one of the world’s most densely populated and economically dynamic coastal regions.
The hidden half of relative sea-level rise
The GBA is home to more than 86 million people and includes major cities such as Guangzhou, Shenzhen, Zhuhai, Hong Kong, and Macao. It is also a region shaped by decades of urban expansion, infrastructure development, and coastal land-use change. Much of its coastline lies on low-lying deltaic land made of soft sediments — a setting where subsidence can become an important factor in coastal flood risk.
Using Sentinel-1 satellite radar observations, we mapped vertical land motion across the GBA from 2017 to 2024. What we found was highly uneven. Some inland areas are relatively stable, while several coastal and peri-urban zones show higher subsidence rates, with localized rates exceeding 40 mm per year. These are not just abstract geophysical signals. In our analysis, higher subsidence rates are found in some reclaimed areas, agricultural delta plains, and coastal districts undergoing rapid land-use change.
That matters because coastal flood risk is driven not only by how much the sea rises, but also by where the land is losing elevation.
Figure 1 | Spatial pattern of vertical land motion across the Guangdong-Hong Kong-Macao Greater Bay Area, together with tide-gauge-based sea-level records and future sea-level projections. The map highlights spatially heterogeneous vertical land motion, with higher subsidence rates observed in some coastal and deltaic zones.
Building a better picture of coastal flooding
A major motivation for this work was that many existing flood assessments have limited ability to represent this part of the story. Some focus mainly on climate-driven sea-level rise, while others rely on simplified representations of coastal water levels or flood propagation.
We wanted to improve that.
Instead of relying on a single uniform coastal water level, we used the high-resolution DTU25 mean sea surface product, which benefits from recent advances in satellite altimetry, including observations from the Surface Water and Ocean Topography (SWOT) mission. This helped us represent near-shore sea-level variability more realistically — an important step in a region with a highly complex coastline and shallow estuarine waters.
We then combined this improved coastal water-level information with satellite-derived land subsidence, sea-level rise projections from the IPCC, estimates of extreme water levels from tides and storm surges, coastal elevation data, and a two-dimensional hydrodynamic flood model. In other words, rather than treating flooding as water simply “filling a bathtub”, we simulated how floodwater can actually move through channels, reclaimed coastal land, and low-lying urban areas.
Bringing all of these datasets together was one of the biggest challenges of the study. They come from different sources, resolutions, time periods, and vertical reference systems. Much of the effort went into making them consistent enough to work together in a single flood-risk framework.
What changes by 2050?
We compared present-day flood exposure in 2023 with a near-term future scenario for 2050, because this is the timescale most relevant to decisions being made now about infrastructure, urban planning, and coastal adaptation.
Our results show that by 2050, the combined effects of land subsidence and sea-level rise could expose an additional ~222 km2 of land in the GBA to coastal flooding. This corresponds to about 228,900 additional people and 27.1 billion CNY (US $3.9 billion) in additional exposed GDP.
But the key message is not just the magnitude of the increase. It is what drives it.
By separating the effects of sinking land and rising seas, we found that land subsidence accounts for about 51% of the increase in exposed population, while sea-level rise accounts for about 43%. The remaining share comes from the interaction between the two. In other words, for the near-term future of the GBA, human-induced subsidence should not be treated simply as a small correction to sea-level rise. In our scenario analysis, it contributes slightly more than sea-level rise to the projected increase in exposed population.
Figure 2 | Simulated coastal flood inundation in the Greater Bay Area under the baseline year 2023 and projected scenario 2050. The simulations indicate that projected flood exposure mainly occurs in low-lying coastal zones and areas connected to the Pearl River Delta channel network, including some reclaimed coastal areas.
Why this matters for adaptation
This distinction matters because it changes how we think about solutions.
Sea-level rise is a global process, and reducing its long-term magnitude depends on climate mitigation worldwide. Subsidence, however, often has strong local causes. In some places, it can be slowed through better groundwater management or more cautious land-use planning. In others — especially reclaimed coastal areas or places affected by long-term sediment compaction — it may be harder to reverse and must be directly accounted for in adaptation planning.
We also estimated how much additional coastal land elevation would be needed to keep future flood exposure at roughly today’s level. The average value across the GBA is about 0.45 m by 2050, and about two-thirds of that requirement is attributable to subsidence rather than sea-level rise.
That does not mean every coastline should simply be raised by the same amount. Instead, it offers a practical way to identify where adaptation pressure will be greatest and where flood defence enhancement, improved subsidence monitoring and management, or targeted planning may be most urgently needed.
A lesson beyond the Greater Bay Area
Although our study focuses on southern China, the broader message extends far beyond the GBA. Many coastal megacities and deltas around the world are built on soft sediments, shaped by land reclamation, affected by groundwater use, and exposed to rising seas. In these settings, future flood risk is not only a sea-level story — it is also a land-subsidence story.
To understand and manage that risk, we need to look both outward to the ocean and downward to the ground beneath our feet.