AI-enhanced framework can track coastal resilience against climate change

Coastal zones, the researchers note, are among the most dynamic and densely populated regions on Earth, yet they remain highly exposed to sea-level rise, storm surges, and erosion. These forces are altering coastlines faster than ever before, eroding not just land but also livelihoods, infrastructure, and biodiversity.

A multinational team of scientists has developed a novel method to evaluate how coastlines respond to climate change, leveraging satellite data, field observations, and artificial intelligence to better predict and manage risks in vulnerable coastal zones. The study, published in the Journal of Marine Science and Engineering, introduces a new global standard for tracking the physical and ecological resilience of coasts over time.

Titled “Monitoring Resilience in Coastal Systems: A Comprehensive Assessment,” the study provides the first truly global synthesis of resilience assessment frameworks under the CRESTE (Coastal Resilience Using Satellites) network. It examines coasts in France, the Netherlands, Mexico, Australia, and China, establishing a harmonized approach for comparing shoreline dynamics, sediment budgets, and ecosystem stability using remote sensing and AI-based data analytics.

Climate change and the urgency of measuring coastal resilience

Coastal zones, the researchers note, are among the most dynamic and densely populated regions on Earth, yet they remain highly exposed to sea-level rise, storm surges, and erosion. These forces are altering coastlines faster than ever before, eroding not just land but also livelihoods, infrastructure, and biodiversity.

The study argues that resilience, the ability of a coastal system to resist, absorb, and recover from environmental disturbances, can no longer be understood through isolated indicators. Instead, it requires a multi-scale, data-driven framework that integrates physical, ecological, and socio-economic factors into a unified model.

The CRESTE framework proposed in this study focuses on three central questions:

1. How can resilience be objectively measured and compared across different coastlines?
2. What physical and environmental indicators best capture coastal vulnerability and recovery?
3. How can satellite data and AI analytics be integrated with field observations to support long-term resilience management?

By answering these, the authors seek to move global resilience research from theoretical modeling toward operational monitoring, enabling consistent assessment across diverse coastal systems.

While resilience has become a central concept in climate policy, its quantitative evaluation remains inconsistent, the study says. Different regions employ varied metrics and timescales, making international comparison difficult. The proposed framework, therefore, aims to harmonize the measurement process, allowing for direct benchmarking of coastal conditions between nations and ecosystems.

A multi-scale, AI-enhanced framework for monitoring coastal change

The study is based on a hybrid analytical approach that integrates satellite-based remote sensing, dynamic numerical modeling, and machine learning algorithms to deliver high-resolution assessments of coastal resilience. The CRESTE network leverages open-access satellite datasets from programs like Sentinel-2, Landsat, and MODIS, merging them with on-site measurements and modeling tools to analyze shoreline evolution, sediment transport, and morphological shifts over time.

The approach is tested across five contrasting study sites:

• The Bay of Biscay (France) – representing high-energy Atlantic coastlines dominated by wave-driven erosion.
• The Netherlands – a low-lying engineered landscape shaped by centuries of coastal defense and land reclamation.
• Batemans Bay (Australia) – a mixed natural-urban coastal system under pressure from estuarine sedimentation and human modification.
• Yucatán Peninsula (Mexico) – an ecologically rich but climate-vulnerable region balancing tourism development and conservation.
• Hangzhou Bay (China) – an industrialized delta facing compounded risks from rapid urbanization and subsidence.

Across these locations, the study measures resilience through two core dimensions: resistance (the capacity to withstand disturbances like storms and sea-level rise) and recovery (the ability to return to equilibrium afterward). By quantifying these parameters through satellite time series and local datasets, the model allows scientists to classify coastlines into categories of high, moderate, or low resilience.

Artificial intelligence plays a central role in refining this classification. Machine learning algorithms trained on historical coastal imagery can automatically detect shoreline changes, sandbar movements, and vegetation shifts, providing consistent results even in data-scarce environments.

The researchers stress that the use of AI and cloud-based geospatial platforms like Google Earth Engine enables scalable, replicable assessments across countries. This integration of technology reduces data processing time and ensures continuous updates, a major step toward real-time resilience monitoring.

Furthermore, the study underscores that the spatial and temporal resolution of satellite imagery is critical. By combining high-resolution optical data with radar and LiDAR sources, CRESTE can analyze not just surface morphology but also subsurface sediment dynamics and hydrodynamic influences, key to understanding the causes of erosion and accretion cycles.

From local measurements to global policy: Building a shared standard for coastal resilience

The authors argue that the lack of standardized resilience indicators hampers effective coastal governance, as local studies often rely on short-term or site-specific datasets. The CRESTE framework provides a scalable structure for comparing resilience across administrative and ecological boundaries, thus informing national adaptation strategies and international climate agreements.

In the case of Batemans Bay and Hangzhou Bay, the research highlights how human interventions, such as dredging, seawalls, and land reclamation, have altered sediment transport pathways, reducing natural recovery rates after storms. The study recommends that resilience management should prioritize restoration-oriented interventions over purely protective infrastructure, emphasizing the importance of maintaining natural buffers like wetlands, dunes, and mangroves.

The integration of social and economic data is another key advancement. The framework recognizes that coastal resilience is not purely physical, it also depends on governance capacity, community preparedness, and equitable resource distribution. By merging environmental indicators with socio-economic metrics, policymakers can identify areas where environmental vulnerability overlaps with social fragility.

The study also calls for enhanced data sharing and collaboration between governments, research institutions, and international organizations. Open data access, standardized metadata, and interoperable platforms would allow coastal monitoring systems to evolve from fragmented projects into coordinated global observatories.

Importantly, the authors highlight the potential of satellite constellations to sustain long-term observation even in politically or logistically inaccessible regions. When paired with artificial intelligence, these satellites can track coastal morphology at unprecedented precision, allowing near real-time identification of erosion hotspots and recovery patterns after extreme weather events.

Another emerging area addressed in the paper is the role of nature-based solutions in restoring resilience. By quantifying the ecological contribution of reefs, marshes, and seagrasses to wave attenuation and sediment retention, the framework can help justify investments in ecosystem restoration as a form of natural infrastructure. This evidence base is crucial for directing funding toward sustainable adaptation projects.