The stakes could hardly be higher. More than 40 percent of the global population lives in coastal regions, and sea-level rise alone is projected to erase up to half of the world’s sandy coastlines by the end of the century. The hazards now converging on those shores are not isolated events but compound extremes, in which multiple drivers strike simultaneously and amplify one another. Hurricane Harvey offered a devastating demonstration in 2017, when record-breaking rainfall, river discharge, and surface runoff combined with a prolonged storm surge to produce catastrophic compound flooding in Houston, causing approximately 152.5 billion US dollars in damage and ranking as the second-costliest natural disaster in United States history. Research cited in the review shows that precipitation is the dominant contributor to changes in compound flooding frequency across 83 percent of global coastlines.

The physical backdrop to this challenge is shifting in ways that conventional models struggle to capture. Assessments from the Intergovernmental Panel on Climate Change indicate with high confidence that both the frequency and intensity of heavy precipitation have increased across most land regions since the mid-twentieth century, and CMIP6 climate model projections point to substantial further increases in extreme precipitation under medium- to high-emission scenarios, particularly across northern Europe and adjacent seas. Storm surges and extreme sea levels, meanwhile, exhibit pronounced spatial heterogeneity: surge magnitudes are projected to increase in regions such as the Baltic Sea in response to shifting wind patterns, while other basins, including parts of the Mediterranean and Atlantic, remain uncertain or strongly modulated by large-scale climate oscillations such as the North Atlantic Oscillation. These nonlinear, non-stationary dynamics are precisely the kind of multivariate complexity that machine learning was designed to untangle.

To understand how the field has responded, researchers led by Kamran Tanwari of the University of Szczecin in Poland conducted a PRISMA-compliant systematic search of the Scopus database, screening 666 initial records down to 73 studies that applied artificial intelligence or machine learning to coastal climatic extremes. Each study was assessed across seven methodological domains adapted from Cochrane risk-of-bias guidance and the PROBAST+AI framework for prediction models. The results reveal a field in rapid but uneven growth: most studies appeared between 2021 and 2025, and the overwhelming majority originated from just two countries, the United States with 24 studies and China with 18, leaving the most vulnerable regions of the Global South strikingly underrepresented. Storm surges and extreme sea levels dominated the hazard landscape, examined in 44 studies, while compound flooding and heavy precipitation each appeared in only 11.

Deep learning was the clear methodological favorite, used in 57 of the 73 studies, with artificial neural networks, convolutional neural networks, and long short-term memory networks the most common architectures. The appeal is fundamentally computational. Process-based coastal models such as ADCIRC and MIKE 21 solve depth-integrated shallow-water equations, while spectral wave models like SWAN track wave propagation, refraction, shoaling, and dissipation, all of which demand enormous computing resources. AI surrogates, by contrast, can generate forecasts in a fraction of a second, making them highly attractive for early warning systems. Reported storm surge lead times typically range from 1 to 24 hours, but the review found a nearly universal pattern: model skill degrades significantly, sometimes dropping to random chance, once the prediction window extends beyond 48 to 72 hours, because data-driven models excel at interpolating within their training data but frequently fail to extrapolate to unprecedented extremes.

The review also uncovered a troubling pattern of contradictory findings that suggests the field lacks standardized benchmarking. In shoreline detection, one study found that a traditional object-based image analysis using random forest outperformed a convolutional neural network approach, directly contradicting earlier results. In surge prediction, some authors argue that standard LSTM networks are superior for handling temporal dependencies in water level sequences, while others found that hybrid ConvLSTM architectures, which combine spatial feature extraction with temporal regression, provide a more robust balance, and still others demonstrated that simple neural networks or even linear regression can perform comparably well at specific tide gauges with short or patchy records. This may indicate that the rush toward ever-deeper, more complex hierarchical models sometimes outpaces actual operational necessity at individual sites.

Perhaps the most consequential finding concerns what the models do not understand. When explainable AI techniques were applied to high-performing architectures, including Vision Transformers, researchers discovered a startling contradiction: despite achieving high statistical accuracy, some models primarily learned the autocorrelation of historical water levels rather than the actual physical mechanisms of wind-surge coupling. For mitigation design, where AI is increasingly used to optimize dynamic seawalls, revetments, and drainage infrastructure through reinforcement learning and Bayesian networks, this mechanistic shortcoming is a serious liability, because models that do not grasp physical drivers cannot reliably extrapolate the stress loads of unprecedented extreme events. The review also cautions that highly optimized hard defenses can promote a false sense of security, potentially leading to catastrophic disasters when climate extremes inevitably exceed design thresholds.

The gap between prediction and practice extends across the entire application spectrum. The review classified AI applications into six progressive domains: monitoring, prediction, mitigation, adaptation, resilience, and uncertainty quantification. Monitoring and prediction dominate the literature, while resilience emerged as the most critically underrepresented domain, requiring the integration of land use, elevation, soil, and post-event reconnaissance data through sophisticated architectures such as causal spatio-temporal graph neural networks. Only 11 studies explicitly addressed multi-decadal climate trajectories, and there is strong consensus that advanced adaptation models, which use reinforcement learning to navigate deep uncertainty and identify flexible policy pathways, remain largely confined to academic journals with very little uptake by actual coastal policymakers. The algorithms frequently optimize policies in a theoretical vacuum, overlooking conflicting political interests, municipal funding obstacles, and the time lags required to pass coastal zoning laws.

Uncertainty quantification, the domain the authors identify as the bridge to operational trustworthiness, is practiced in only a quarter of the reviewed studies. The most sophisticated approaches now distinguish aleatoric uncertainty, the inherent noise in meteorological inputs, from epistemic uncertainty, the internal ignorance of the neural network itself, using Bayesian hierarchical models, Gaussian process regression, and Monte Carlo simulations. But these frameworks carry a cruel irony: generating the massive ensembles of synthetic data required for robust uncertainty propagation effectively neutralizes the primary advantage of AI, its computational speed. Moreover, explainable AI tools such as attention mechanisms and permutation feature importance currently validate only statistical correlation, falling short of proving true physical consistency or causality.

The review’s authors close with four priorities for advancing coastal AI: developing multivariate frameworks capable of representing compound and cascading extremes; integrating physics-informed neural networks and hybrid hydrodynamic-AI architectures so that models obey fundamental physical laws when extrapolating beyond their training data; incorporating socioeconomic scenarios and Shared Socioeconomic Pathways to shift the focus from short-term prediction toward long-term resilience; and applying transfer learning to adapt models developed in data-rich settings to the highly vulnerable, data-scarce regions of the Global South. They also point to the rapid rise of machine-learning weather emulators such as GraphCast, Pangu-Weather, FourCastNet, and GenCast, which could supply atmospheric forcing for coastal models but require rigorous assessment of local accuracy and performance during rare extremes. For now, the verdict is clear: purely data-driven models must be considered complementary to, not a substitute for, physics-based hydrodynamic models, especially when navigating the deep uncertainty of a changing climate.

Subject of Research: Application of artificial intelligence and machine learning to modelling coastal climate extremes such as storm surges, extreme sea levels, and compound flooding

Article Title: Artificial intelligence for modelling coastal climate extremes: a global systematic review

Article References: Tanwari, K., Shi, X., Śledziowski, J., Giza, A., & Terefenko, P. (2026). Artificial intelligence for modelling coastal climate extremes: a global systematic review. Regional Environmental Change, 26(4), Article 207. https://doi.org/10.1007/s10113-026-02689-6

Image Credits: AI Generated

DOI: 10.1007/s10113-026-02689-6

Keywords: artificial intelligence, coastal hazards, storm surge, compound flooding, extreme sea levels, machine learning, deep learning, climate change, sea-level rise, uncertainty quantification, explainable AI, coastal adaptation

The core insight of the study is deceptively simple. Tides are not random. They are governed by the gravitational interplay between the Earth, the Moon and the Sun, producing regular cycles of rising and falling water that can be calculated years in advance. In many coastal locations, the tide determines whether a given storm surge is sufficient to push water levels above established flood thresholds. When a moderate surge coincides with a high tide, flooding follows; when the same surge arrives at low tide, water levels may remain safely below the threshold. As Haigh explains, even storm-driven flooding often inherits the timing of the tides, because the tide acts as the gatekeeper that decides whether the surge crosses the critical line.

What makes the new work distinctive is its focus on the hour of the day rather than the season or the year. Previous research on flooding has concentrated heavily on seasonal patterns, examining, for example, why river floods cluster in certain months or why coastal floods are more frequent during particular times of the year. Haigh notes that surprisingly little attention has been paid to what hour of the day flooding actually occurs. Part of the reason, he argues, lies in how flooding is measured. Flood impacts are commonly summarized as the number of flood days per year, and once an event is compressed into a single day, the information about whether the flooding happened at 3 a.m., 8 a.m. or 3 p.m. simply disappears. By returning to the underlying records and preserving the temporal detail, the research team was able to recover patterns that conventional flood statistics had erased.

The regional results are striking in their specificity. In Boston, floods occur predominantly around noon and midnight, mirroring the timing of that harbor’s high tides. In Southern California, flooding is more common in the morning, again tracking the local tidal cycle. In Florida, the picture is more varied, because tidal behavior differs substantially from one stretch of coastline to another. Along the Atlantic Coast and down toward the Florida Keys, coastal flooding is moderately predictable in its timing, while along the Gulf Coast, where tidal ranges are smaller and more complex, flooding is less predictable. These differences matter, because they show that the daily fingerprint of flooding is not a universal pattern but a local one, shaped by the specific tidal regime of each coastline.

The study also establishes that the phenomenon is not confined to places with semidiurnal tides, the familiar pattern of two high tides and two low tides each day that characterizes much of the U.S. East Coast and the U.K. The researchers found that predictable daily timing also emerges in locations with diurnal tidal systems, which experience only one high tide and one low tide per day. The reason lies in the way different tidal signals interact. Tides are generated by multiple astronomical constituents, each with its own period, and their superposition produces the observed water-level cycle. Whether a coastline experiences two highs a day or one, the resulting pattern remains regular enough that flood events, which require water levels to cross a threshold, cluster around the recurring peaks of that cycle.

The technical implication is that flood risk is not evenly distributed across the hours of a day, even when the weather is. A storm system that lingers offshore for twelve hours does not pose a uniform threat throughout its passage. Its most dangerous moments are those that coincide with high tide, and in many locations those moments can be identified in advance with the same precision used to publish tide tables. This reframing has consequences for how flood forecasting is constructed. Traditional forecasts emphasize the magnitude of expected water levels, but the timing of when those levels are most likely to be exceeded is equally important for operational decisions, from closing roads to repositioning emergency assets.

Haigh argues that this matters because timing influences impacts. The same flood can have very different consequences depending on whether it occurs during the morning commute, during business hours or in the middle of the night. A flood that peaks at rush hour can strand commuters, snarl traffic and endanger drivers who attempt to cross inundated roadways. A flood of identical height at 3 a.m. may cause comparable physical damage but far fewer casualties and disruptions, simply because fewer people are exposed. Understanding these temporal patterns, the study suggests, could improve flood forecasting, emergency planning, infrastructure operations and long-term coastal adaptation. Utilities, hospitals, schools and transit agencies could all incorporate the known tidal windows into their contingency planning, aligning preparedness measures with the hours when flooding is statistically most likely.

The research arrives at a moment when the baseline for coastal flooding is shifting. As sea levels continue to rise, coastal flooding is increasingly occurring without storms to spur it on, a phenomenon often described as nuisance or sunny-day flooding, in which exceptionally high tides alone are enough to inundate low-lying streets. Haigh points out that higher mean sea levels mean it takes only a relatively small weather contribution, or sometimes simply a very high astronomical tide, to cross a flood threshold. In such conditions, the timing of the tides becomes even more decisive, and understanding exactly when those critical tidal windows occur could become an increasingly important component of coastal flood forecasting and preparedness.

The practical payoff of the work could be substantial. Haigh says the findings could make existing flood forecasts more predictable, providing more warning before disasters occur. That added lead time can help residents, businesses and emergency managers anticipate disruption and plan accordingly, whether that means scheduling deliveries around high-tide windows, warning drivers before morning inundations or pre-positioning pumps and barriers. Because the tidal component of flood timing is deterministic, it can be computed far in advance, offering a layer of foresight that weather-driven forecasting alone cannot match. The uncertainty in a flood forecast then narrows to the meteorological contribution, while the tidal contribution is already known.

The study, titled Coastal flooding at predictable hours and published in Nature Communications on 28 July 2026, represents a collaboration among international researchers and reflects Haigh’s long career in sea level and coastal impacts. A professor who joined the University of Central Florida in 2025 as director of the Center for Integrated Coastal Research, Haigh previously held the chair in sea level and coastal impacts at the University of Southampton and directed the Center for Doctoral Training for Resilient Flood Futures. His team’s work suggests a quiet but consequential shift in coastal science: the hazard that communities have treated as random now reveals a schedule, and learning to read that schedule may prove one of the most practical tools available as rising seas make the ocean’s high-water moments an ever more frequent visitor to the world’s coastlines.

Subject of Research: Predictable daily timing of coastal flooding driven by tidal cycles

Article Title: Study shows coastal flooding occurs at predictable times

Article References: Study shows coastal flooding occurs at predictable times. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: coastal flooding, tides, sea level rise, flood forecasting, storm surge, tidal cycles, Nature Communications, University of Central Florida, emergency planning, nuisance flooding, tide gauges, coastal adaptation

The modelling rests on a future scenario developed by the Intergovernmental Panel on Climate Change known as SSP2-4.5. This intermediate scenario, combining moderate greenhouse gas emissions with a partial global policy response, points to roughly 2.7 degrees Celsius of average global warming above pre-industrial levels by 2100. That figure is broadly in line with where existing national pledges and policies are steering the planet, but it sits well above the 1.5 degree threshold that world leaders enshrined in the Paris Agreement as the aspirational safe limit for climate change. In other words, the study’s headline losses are not a worst-case projection; they are close to business as usual.

To translate rising water into rising bills, the research team followed a deliberately systematic approach. They first mapped the areas likely to be inundated by combining sea-level projections with estimates of storm surge and high-resolution elevation data across the Australian continent. Within those flood-exposed zones, they then identified every affected property and parcel of land, drawing on a wide range of data sources and empirical measures to estimate the value of homes, farms, roads, utilities, ecosystems and agricultural acreage. The result was a granular national picture of exposure that no previous Australian study had assembled at this scale.

The second stage of the analysis converted exposure into expected harm. Using statistical modelling, the researchers estimated how severely the assets inside the mapped flood zones could be damaged over time as sea levels climb and storm surges push further inland. The physical damages projected by the model were then translated into economic losses, yielding a nationwide damage function that captures the compounding interaction between higher seas, extreme coastal water levels and the value of what lies in the water’s path. The method is a computational simulation grounded in observed relationships, designed to give policymakers a defensible estimate of future costs rather than an abstract warning.

The geographic distribution of the projected losses is striking. Every state and the Northern Territory will bear significant economic costs, the study finds, ranging from 230.5 billion dollars in Western Australia down to 7.9 billion dollars in Tasmania. Queensland follows close behind Western Australia with projected losses of 214.5 billion dollars, then Victoria at 167 billion dollars, New South Wales at 151 billion dollars, South Australia at 49.2 billion dollars and the Northern Territory at 35.2 billion dollars. The pattern reflects the combination of long, low-lying coastlines, dense coastal development and the sheer value of the property and infrastructure concentrated along Australia’s populated edges.

Property counts tell a parallel story of exposure. Queensland has the greatest number of properties at risk, with 93,157 identified in the flood-exposed zones, followed by New South Wales with 71,210 and Western Australia with 51,366. Yet within these state-wide totals, a single urban hotspot stands out. The Gold Coast alone faces 84.4 billion dollars in projected economic losses, making it the most exposed urban area in the entire country. The finding will resonate with residents who watched Tropical Cyclone Alfred tear away stretches of beach in a single night, an event that cost the City of Gold Coast 35 million dollars in emergency beach repairs and left communities along the east coast still recovering.

Despite the enormous sums, the study’s authors stress that their estimate is conservative in several important respects. The modelling uses a middle-of-the-road emissions scenario, holds storm intensity constant rather than allowing it to intensify with warming, and excludes losses from coastal erosion altogether. Perhaps most significantly, the figures do not reflect the possibility of rapid collapse of the Greenland and West Antarctic ice sheets, which could add up to two metres of global sea level rise by the end of the century. If any of those factors worsen, the true bill for coastal Australia could climb far beyond the 855 billion dollar floor that the study establishes.

The human dimension behind the numbers is considerable. The majority of Australians live within 50 kilometres of the coastline, meaning sea level rise may physically affect a large share of the national population. Millions of hectares of land are at risk by 2100 on current projections, and the damages extend well beyond housing to essential infrastructure, ecosystem services and valuable agricultural land. Professor Tom Kompas of the University of Melbourne, who led the study and co-authored the Climate Council’s accompanying Rising Seas Rising Bills report, notes that because carbon emissions have remained too high, a certain amount of sea level rise is already locked into the system. Adaptation, he argues, is no longer optional; the least costly option available is to avoid building new development in the places where risks are known to be highest, while restoring coastlines and reforming construction practices to reduce future exposure. These are hard decisions, he acknowledges, but they are ones communities are already confronting.

Climate scientists involved with the accompanying report emphasise that the mechanics of the threat are deceptively simple. Storm surges do not need to become more frequent or more violent to cause escalating damage; they simply ride on higher baseline seas, pushing floodwaters deeper into neighbourhoods that once stayed dry. Adjunct Professor Andrew Watkins, a Climate Councillor and co-author of the report, warns that powerful storm surges are already putting coastal homes and infrastructure at greater risk, and that economic losses can skyrocket into hundreds of billions of dollars if action is delayed. He describes sea level rise as a slow-onset disaster, one that demands faster cuts to greenhouse gas pollution alongside serious preparation of vulnerable communities, and cautions that the rapid loss of ice sheets and glaciers could push oceans even higher than the study’s already sobering projections.

What makes the research a landmark for Australia is its completeness. Previous assessments have tended to examine individual regions, cities or asset classes, leaving policymakers without a coherent national figure to weigh against adaptation investments. By modelling every state and territory under a single consistent methodology and anchoring the results to an internationally recognised emissions scenario, the study provides a common baseline for planning debates about coastal development, insurance, infrastructure design and managed retreat. The 855 billion dollar figure is, in effect, a price tag on the coast as it exists today, and a measure of how much cheaper the future could be if emissions fall faster and the most dangerous development choices are avoided. As the century advances and the seas continue their measured climb, the study suggests that the costliest decisions will not be about whether to protect Australia’s coastline, but how quickly the nation accepts what the water is already telling it.

Subject of Research: Nationwide economic impacts of sea level rise and storm surge from global warming in Australia

Article Title: Rising sea levels could cost Australians at least $855 billion by 2100

Article References: Rising sea levels could cost Australians at least $855 billion by 2100. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: sea level rise, storm surge, Australia, coastal flooding, economic losses, climate change, IPCC scenarios, SSP2-4.5, Gold Coast, Scientific Reports, coastal adaptation, property risk