In many areas, storm surges caused by tropical or extratropical cyclones are the main contributors to critical extreme sea level events. Storm surges can be simulated using numerical models that are based on the underlying physical processes, or by using data-driven models that quantify the relationship between the predictand (storm surge) and relevant predictors (wind speed, mean sea-level pressure, etc.). This study explores the potential of data-driven models to simulate storm surges globally. A multitude of predictors (obtained from remote sensing and climate reanalysis) along with predictands (from tide gage observations and storm surge reanalysis) are utilized to train and validate data-driven models to simulate daily maximum surge for the global coastline. Datadriven models simulate daily maximum surge better in extratropical and subtropical regions [average correlation and root-mean-square error (RMSE) of 0.79 and 7.5 cm, respectively], than in the tropics (average correlation and RMSE of 0.45 and 5.3 cm, respectively). For extreme events, the average correlation decreases to 0.54 (0.33) and RMSE increases to 14.5 (13.1) cm for extratropical (tropical) regions. Models forced with remotely sensed predictors showed a slightly better performance (average correlation of 0.69) than models forced with predictors obtained from reanalysis products (average correlation of 0.68). Results also highlight a significant improvement (i.e., average correlation increases from 0.54 to 0.68; RMSE reduces from 11 to 7 cm) over the Global Tide and Surge Reanalysis (GTSR), derived from the only global hydrodynamic model. For approximately 70% of tide gages, mean sea-level pressure is the most important predictor to model daily maximum surge. Our results highlight the added value of datadriven models in the context of simulating storm surges at the global scale, in addition to existing hydrodynamic numerical models.

Coastal communities across the world are already feeling the disastrous impacts of climate change through changes in extreme sea levels 1 . These changes reflect the combined effect of sea-level rise and changes in storm surge activity. Understanding the relative importance of these two factors in altering the likelihood of extreme events is crucial to the success of coastal adaptation measures. Existing analyses of tide gauge records 2-10 agree that sea-level rise has been a major driver of trends in sea-level extremes since at least 1960. However, the contribution from changes in storminess remains unclear, owing to the difficulty of inferring this contribution from sparse data and the consequent inconclusive results that have accumulated in the literature 11,12 . Here, we analyse tide gauge observations using spatial Bayesian methods 13 to show that, contrary to current thought, trends in surge extremes in Europe since 1960 were comparable to the rate of sea-level rise. We determine that the trend pattern of surge extremes reflects the contributions from a dominant north-south dipole associated with internal climate variability and a single-sign positive pattern related to anthropogenic forcing. Our results demonstrate that both external and internal influences can considerably affect the likelihood of surge extremes over periods as long as 60 years, suggesting that the current coastal planning practice of assuming stationary surge extremes 1,14 might be inadequate.Floods resulting from extreme sea levels are among the costliest natural hazards, causing tens of billions of dollars in economic losses globally each year 1 . Without adaption, such losses are certain to worsen in the decades ahead as sea level rises 15,16 . Cost-effective adaptation plans are key to reducing this vulnerability while also avoiding costly overprotection measures 17 . However, their success relies on robust understanding of how changes in mean climate affect the likelihood of extreme sea-level events. This effect can occur primarily through (omitting tides and waves) changes in storminess affecting the occurrences of storm surges, and changes in mean sea level (MSL) raising or lowering the baseline level for storm surges. While most studies agree that sea-level rise 18,19 has made extreme sea-level events more likely across the world since at least the mid-20 th century 2-10 , the contribution from changes in surge extremes is subject to debate 11,12 .Clarifying this debate is a priority because, in most countries, current practice for assessing future coastal flood risk assumes that the probability of surge extremes is the same now as in the future 1,14 (i.e., only sea-level rise is considered). Should this assumption turn out to be invalid, this could cause adaptation plans to be ineffective.Numerical models generally predict that storm surge activity will change this century in many places as the climate warms [20][21][22][23][24][25][26] , albeit with varying intensity depending on the study. Yet observational evidence for such c...

Storm surges are among the deadliest coastal hazards and understanding how they have been affected by climate change and variability in the past is crucial to prepare for the future. However, tide gauge records are often too short to assess trends and perform robust statistical analyses. Here we use a data-driven modeling framework to simulate daily maximum surge values at 882 tide gauge locations across the globe. We use five different atmospheric reanalysis products for the storm surge reconstruction, the longest one going as far back as 1836. The data that we generate can be used, for example, for long-term trend analyses of the storm surge climate and identification of regions where changes in the intensity and/or frequency of storms surges have occurred in the past. It also provides a better basis for robust extreme value analysis, especially for tide gauges where observational records are short. The data are made available for public use through an interactive web-map as well as a public data repository.

We address the challenge, due to sparse observational records, of investigating long-term changes in the storm surge climate globally. We use two centennial and three satellite-era daily storm surge time series from the Global Storm Surge Reconstructions (GSSR) database and assess trends in the magnitude and frequency of extreme storm surge events at 320 tide gauges across the globe from 1930, 1950, and 1980 to present. Before calculating trends, we perform change point analysis to identify and remove data where inhomogeneities in atmospheric reanalysis products could lead to spurious trends in the storm surge data. Even after removing unreliable data, the database still extends existing storm surge records by several decades for most of the tide gauges. Storm surges derived from the centennial 20CR and ERA-20C atmospheric reanalyses show consistently significant positive trends along the southern North Sea and the Kattegat Bay regions during the periods from 1930 and 1950 onwards and negative trends since 1980 period. When comparing all five storm surge reconstructions and observations for the overlapping 1980–2010 period we find overall good agreement, but distinct differences along some coastlines, such as the Bay of Biscay and Australia. We also assess changes in the frequency of extreme surges and find that the number of annual exceedances above the 95th percentile has increased since 1930 and 1950 in several regions such as Western Europe, Kattegat Bay, and the US East Coast.

No abstract