The research team, led by Yanglong Chen and Chaojun Ouyang of the Chinese Academy of Sciences, compiled a database of 32,714 landslide records drawn from public inventories around the world, including thousands of landslides triggered by the 2008 Wenchuan earthquake in China and the 2015 Gorkha earthquake in Nepal. The collection spans rainfall-triggered rock slopes in Hong Kong, hurricane-induced soil slides in Puerto Rico, coseismic failures in Greece and the Nepal Himalaya, and many other settings. Each record contains geometric descriptors such as landslide volume, source area, vertical drop height, slope angle, and various width measurements, along with the observed runout distance. The sheer scale of the database is itself a milestone, but its heterogeneity is what makes it scientifically interesting and computationally challenging.

That heterogeneity takes several forms. Different data sources cover different regions, report different subsets of variables, and describe vastly different numbers of events. Rainfall-triggered landslides account for 17,637 cases in the compiled dataset, earthquake-triggered events for 14,391, while landslides triggered by combined factors number only 67. Rock landslides dominate with 27,824 samples, compared with 4,533 soil landslides and just 285 mixed-material events. Because some inventories report volume but not slope angle, and others record source width but not total area, every possible combination of input factors corresponds to a different subset of usable samples. A model trained on one configuration may therefore be evaluated on data that differ systematically from the data behind another configuration, confounding any comparison of which factors truly matter.

The researchers argue that this is precisely where most previous machine learning studies of landslide runout have fallen short. Work has typically emphasized predictive accuracy for a single landslide type or region, treating model selection as a secondary concern. But different algorithms trained on the same data can latch onto different relationships, and when similarly performing models disagree about which factors drive runout, the popular SHAP interpretation technique, which attributes predictions to individual input variables, will produce different answers depending on which model happens to be chosen. In other words, factor importance inferred from a single model may reflect the quirks of that algorithm rather than any stable property of the underlying physics.

To break this dependence, the team built a selective ensemble framework. Six base learners representing distinct modeling paradigms were trained under identical preprocessing, factor configurations, and data splits: a multilayer perceptron, Random Forest, Support Vector Regression, XGBoost, and two attention-based deep learning architectures for tabular data, TabNet and FT-Transformer. Performance was assessed with five-fold cross-validation, and only models whose relative performance fell within five percent of the best learner, or three percent for very small sample groups, were admitted to the ensemble. The selected models were then combined using non-negative least squares, a constrained optimization that assigns each model a non-negative weight, preventing predictions from canceling one another and keeping every contributor’s influence directly interpretable. Crucially, the ensemble is retained only if its cross-validation performance exceeds that of the best individual model; otherwise the framework falls back to the single best learner. This combination of screening, constrained aggregation, and a validation-based fallback rule is what allows the method to reduce model-selection variability without blindly averaging everything together.

The results are striking. When the vertical drop height H was included among the predictors, the ensemble achieved coefficients of determination between roughly 0.65 and 0.98 across the seven landslide types analyzed, with earthquake-triggered rock landslides reaching an R-squared of 0.96533 on independent test sets and soil landslides 0.95341. For four of the seven types, optimal performance was achieved with only two input variables, height and volume, echoing a familiar lesson from empirical runout formulas: a small number of well-chosen geometric factors often outperforms a crowded feature list, especially when adding variables shrinks the usable sample size. The ensemble matched or exceeded the best individual model in most cases, and its cross-validation performance was never lower, confirming improved stability rather than merely marginal accuracy gains.

The framework was also used to quantify what happens when a key factor becomes unavailable. Vertical height is often difficult to estimate accurately before a landslide occurs, so the team compared models with and without it. The consequences varied dramatically by landslide type. Rainfall-triggered landslides suffered the largest performance collapse, with R-squared falling from about 0.944 to roughly 0.377 when height was removed and remaining factors re-screened. Earthquake-triggered landslides were far more resilient, dropping only from 0.965 to 0.914, because substitute variables such as source height could partially compensate. This sensitivity analysis turns an abstract question about factor importance into a practical guideline: for some landslide types, investing in better pre-event height estimates pays enormous dividends, while for others, existing topographic proxies are nearly sufficient.

The ensemble-level SHAP analysis then revealed how the same physical factors behave differently depending on the triggering mechanism and material. Vertical height emerged as the dominant or near-dominant factor in nearly every category, showing a consistent nonlinear signature: its contribution is negative at low values, rises rapidly as height increases, and then plateaus, indicating diminishing marginal returns on elevation difference. Volume, by contrast, behaved in a trigger-dependent way. For rainfall-triggered rock landslides its contribution generally decreased with increasing volume, while for earthquake-triggered rock landslides it increased, a contrast the authors interpret cautiously given the strong correlations among volume, source area, and source width in the rainfall subset. Slope angle showed the starkest divergence: its contribution declined with steeper angles for rainfall-triggered and soil landslides, consistent with the idea that steep rain-fed slopes store and infiltrate less water, but it turned sharply positive above roughly 50 degrees for earthquake-triggered rock slopes, where seismic shaking can destabilize near-vertical faces directly.

The authors are careful to frame these patterns as statistical associations learned by the models rather than proof of physical causation, noting that correlated predictors share attribution and that data-processing differences among source inventories introduce uncertainties that cannot be fully eliminated. They also acknowledge gaps: multiple factor-triggered landslides remain severely underrepresented, and type-specific variables such as rainfall intensity are inconsistently documented across databases. Even so, the study delivers a template that other hazard domains could adopt. By screening diverse learners under uniform conditions, combining only the competitive ones with transparent weights, and interpreting factors at the ensemble level, the framework converts an unruly, heterogeneous compilation of 32,714 landslides into a stable basis for deciding which variables matter, for which landslide types, and at what cost in data availability. For regional-scale risk assessment, where thousands of potential failures must be screened quickly and cheaply, that combination of robustness, interpretability, and modest data requirements may prove as consequential as any single accuracy record.

Subject of Research: Machine learning prediction of landslide runout distance using a multi-source heterogeneous dataset and a selective ensemble framework with SHAP-based factor evaluation

Article Title: A selective ensemble framework for reliable factor evaluation and landslide runout prediction under data heterogeneity

Article References: Chen, Y., Ouyang, C., Zhao, B., Yang, W., & Wang, F. (2026). A selective ensemble framework for reliable factor evaluation and landslide runout prediction under data heterogeneity. Results in Engineering, 32, Article 113218. https://doi.org/10.1016/j.rineng.2026.113218

Image Credits: AI Generated

DOI: 10.1016/j.rineng.2026.113218

Keywords: landslide runout, machine learning, selective ensemble, SHAP, data heterogeneity, hazard assessment, Wenchuan earthquake, Gorkha earthquake, XGBoost, Random Forest, vertical drop height, regional risk mapping

The research, led by Anjela A. Ilagan of the Philippine Atmospheric, Geophysical and Astronomical Services Administration (DOST-PAGASA) and the University of the Philippines Diliman, together with Olivia C. Cabrera and Marcelino Q. Villafuerte II, tackles a persistent blind spot in Philippine hazard science. Most previous investigations of the country’s storm surges have concentrated either on individual catastrophic events or on numerical simulations of specific tropical cyclones. What was missing was a systematic, observation-based reconstruction spanning the whole archipelago and many decades. Because the Philippines is made up of more than seven thousand islands with an extraordinarily long and complex coastline, local bathymetry, shelf width, and coastal orientation can dramatically alter how high the sea rises when a typhoon makes landfall. A national picture assembled from actual measurements, rather than models alone, is therefore an essential foundation for hazard assessment.

The team’s approach hinged on a deceptively simple but technically demanding manipulation of tide gauge data. Tide gauges record the total water level at the coast, which is a mixture of the predictable astronomical tide, longer-term sea level variations, and the meteorological contribution from winds and pressure changes associated with storms. To isolate the storm signal, the researchers first removed the tidal component using harmonic analysis with the UTide software package, which fits the known tidal constituents to the observed record and allows the tide to be subtracted out. What remains is the residual sea level, the non-tidal part of the water level that carries the fingerprint of storm-driven water pile-up. The raw residuals were then smoothed with a low-pass filter to suppress high-frequency noise, ensuring that short-lived fluctuations unrelated to genuine surge events would not contaminate the detection.

Identifying storm surge events from residuals required a detection threshold, and here the researchers made a methodologically important choice. Rather than applying a single fixed cutoff across the entire country and record, they used an annually varying threshold, which accounts for the fact that baseline sea levels and seasonal conditions differ from year to year and from station to station. This approach, consistent with skew surge methods used by established sea level monitoring facilities, allows the detection algorithm to remain sensitive to genuine surge events even in years or locations where the background water level is naturally higher. Applying this threshold to the de-tided records across multiple coastal stations, the team identified 133 distinct storm surge events over the 77-year study period, the most comprehensive instrument-based storm surge inventory ever assembled for the Philippines.

The statistics that emerged from this inventory carry a message with direct implications for coastal safety. The majority of detected storm surges, measured as pure surge residuals, reached heights of only 20 to 50 centimeters, and only about 5 percent of events exceeded one meter. On its face, that might suggest that surges in the Philippines are generally modest. But the picture changes sharply when the researchers examined storm tide, the combined water level produced when the storm surge coincides with the astronomical tide. In approximately 39 percent of events, the storm tide exceeded one meter, nearly eight times the proportion seen for surge alone. The lesson is unambiguous: the timing of a typhoon’s arrival relative to the tidal cycle can transform a manageable rise in water into a genuinely dangerous coastal flood. A 40-centimeter surge arriving at high tide can inundate areas that the same surge at low tide would leave untouched.

The single largest surge in the record was measured at Pasacao, in the province of Camarines Sur, during Typhoon Molave in 2020, when the water level anomaly reached 211 centimeters, more than two meters above the expected tide. Molave, known locally as Quinta, swept across the Bicol region in late October 2020, and the Pasacao observation stands as a stark reminder of what even a storm that is not among the most intense on record can deliver when its winds align favorably with coastal geography. The study’s case analyses of Molave and Super Typhoon Goni, which struck the same general region just days later in early November 2020, highlighted how the orientation of the coastline relative to the storm’s track and the phase of the tide jointly determine the severity of coastal impacts. Two storms hitting within a week of each other, with different tracks and tidal timings, produced markedly different surge outcomes.

Mapping the geographic distribution of the 133 events revealed a pronounced regional pattern that reflects the underlying physics of surge generation. High storm surges clustered along the coasts of Catanduanes, Camarines Sur, Eastern Samar, Cagayan, Batanes, and Palawan. What these disparate locations share is a shallow and wide continental shelf, a bathymetric configuration that favors surge amplification. When storm winds push ocean water across a broad, shallow shelf, friction with the seabed constrains the water’s vertical escape and forces it to pile up at the coast, a mechanism that deep water adjacent to steep coasts does not permit. Eastern Samar’s presence in this list will resonate with anyone familiar with Haiyan, whose catastrophic surge in Leyte Gulf was amplified by exactly this kind of shelf-driven shoaling. Equally telling is where surges were largely absent: Mindanao, the large southern island, registered few significant events, a pattern consistent with its exposure to a less typhoon-frequented portion of the western North Pacific basin.

Beyond its scientific findings, the database itself may prove to be the study’s most consequential product. A validated, observation-based catalogue of historical surge events spanning nearly eight decades provides exactly the kind of ground truth that storm surge forecasters, hazard mappers, and model developers need. Numerical surge models, from the high-resolution unstructured-mesh systems used in operational hurricane forecasting to global reanalyses of extreme sea levels, depend on historical observations for calibration and validation. Extreme value statistics, the branch of statistics used to estimate the probability of rare events such as hundred-year surges, likewise require long, homogeneous observational records to produce trustworthy return-period estimates. By anchoring such analyses in real tide gauge measurements from Philippine waters, the new inventory reduces the reliance on extrapolations from other coastlines and improves the credibility of national hazard assessments, evacuation planning, and coastal infrastructure design.

The study also arrives at a moment when the stakes of surge forecasting in the Philippines are rising. Mean sea level is climbing globally, and even a modest long-term rise raises the baseline upon which every future storm surge builds, effectively converting yesterday’s moderate surge into tomorrow’s damaging flood. The researchers’ use of an annually varying threshold partly reflects this shifting baseline, and their 1947-to-2024 record offers a valuable window into how surge occurrences have unfolded across a period of both changing climate and changing observation technology. The authors acknowledge the National Mapping and Resource Information Authority’s Hydrography Branch for providing the tide gauge data, and the work was carried out without dedicated external funding, a testament to the value of careful analysis of existing national observational assets.

For a country that sits squarely in the most active tropical cyclone basin on Earth, with roughly twenty storms entering its area of responsibility in a typical year, the ability to look back through 77 years of measured sea level data is more than an academic exercise. It is a way of letting the ocean itself testify about which coasts are most exposed, how badly the tide can conspire with the wind, and what the worst-case scenarios look like when the next Molave or Goni forms over the warm waters of the Pacific. The 133 events now catalogued in this inventory are, in effect, 133 rehearsals for future disasters, and the researchers hope that forecasters, planners, and coastal communities across the archipelago will use them to be better prepared when the next rehearsal becomes the real thing.

Subject of Research: Historical storm surge events in the Philippines reconstructed from tide gauge observations

Article Title: Investigating historical storm surge occurrences in the Philippines from tide gauge observations

Article References: Ilagan, A. A., Cabrera, O. C., & Villafuerte, M. Q., II (2026). Investigating historical storm surge occurrences in the Philippines from tide gauge observations. Natural Hazards, 122(18), Article 625. https://doi.org/10.1007/s11069-026-08360-x

Image Credits: AI Generated

DOI: 10.1007/s11069-026-08360-x

Keywords: storm surge, storm tide, Philippines, tide gauge, tropical cyclone, Typhoon Molave, Super Typhoon Goni, sea level, coastal flooding, hazard assessment, UTide, Natural Hazards

A new study published in Ocean Dynamics by Feidong Zheng of PowerChina Kunming Engineering Corporation Limited and colleagues, including researchers at Nanjing Hydraulic Research Institute and Nanjing University of Information Science and Technology, now offers a predictive framework that spans the full spectrum of wave energy. The work extends the team’s earlier high-energy investigations with a fresh set of controlled laboratory experiments focused specifically on low-energy responses, allowing the researchers to stitch the two regimes into one coherent model. The result is a suite of equations that predicts not only how tall the waves will be, but how they attenuate with distance, how their periods and wavelengths evolve through dispersion, and how the energy of the collapsing granular mass is transferred into the water.

The physics of the problem is deceptively intricate. A subaerial granular landslide is not a rigid block; it is a porous, deformable collection of grains that disintegrates as it slides, impacts the free surface, and continues to move underwater. The wave that emerges depends on the slide’s volume, velocity, thickness, and impact angle, as well as the water depth and the geometry of the reservoir. Classical approaches, such as the widely used model of Heller and Hager developed at ETH Zurich, condense these variables into an impulse product parameter that captures the strength of the landslide impact. That framework has proven robust for the largest, most energetic waves, but its performance for gentler, Stokes-like waves, in which the free surface oscillates in a nonlinear but still wave-like fashion, had not been systematically tested.

Zheng and his co-authors began by doing exactly that: validating the Heller and Hager model against their new low-energy data across the entire energy spectrum. The model held up remarkably well. For maximum wave height, the mean deviation between prediction and measurement was about 15 percent, a figure the authors describe as confirming the reliability of the classical formulation from the weakest oscillatory responses up to the most violent surges. In practical terms, this means engineers can continue to use the established height model with confidence, knowing it does not silently fail when a landslide is smaller or slower than the laboratory benchmarks on which it was originally calibrated.

But wave height alone does not tell the whole hazard story. The maximum amplitude, the single highest excursion of the water surface above still-water level, is often the quantity that determines whether a wave overtops a dam or floods a shoreline bench. Here the team went beyond validation and derived a novel maximum amplitude model of their own, and the improvement is striking: prediction errors fell to 7.80 percent, roughly half the deviation of the height model. The tighter fit is not merely a statistical achievement. The new equation is consistent with a specific physical interpretation of how Stokes-like waves are born, namely that the vertical, kinematic displacement of the free surface, driven by the slide pushing water upward and outward at the impact zone, dictates amplitude generation, rather than the deep transfer of mass momentum through the water column.

That distinction matters because it changes how one thinks about the hazard. If amplitude were controlled by bulk momentum injected deep into the water, then the mass and submerged motion of the landslide would dominate, and mitigation strategies would focus on the underwater runout. If instead the kinematic free-surface displacement is the governing mechanism, as the new model suggests for Stokes-like regimes, then the geometry and speed of the slide at the moment it crosses the shoreline become the critical controls. The authors’ laboratory observations support the latter picture, giving hazard assessors a clearer target for the parameters that most influence the worst-case water-surface excursion near the impact site.

The framework does not stop at the wave crest. Zheng and colleagues developed type-specific governing equations for wave attenuation, describing how the impulse waves lose height as they propagate away from the impact zone, and for the dispersive evolution of wave period and wavelength, capturing how an initially impulsive disturbance stretches into a longer, more regular wave train as frequency components separate with distance. These propagation characteristics are essential for translating a near-field measurement or prediction into an estimate of what a wave will look like when it reaches a dam face or a lakeside community kilometers downstream. A wave that is modest at the source may still carry dangerous energy if its period lengthens in a way that resonates with the geometry of a narrow reservoir arm.

Perhaps the most conceptually interesting component is the team’s energy model, which captures what the authors call mass-dependent momentum coupling. In their experiments, the porous granular mass does not behave as an impermeable piston. Instead, water can percolate into and through the disintegrating slide, and the granular skeleton buffers the interaction between the solid and fluid phases. The model demonstrates physically how this buffering regulates the efficiency of energy transfer from the landslide to the water: a highly porous, rapidly disintegrating mass dissipates more of its kinetic energy internally and in grain-fluid friction, delivering less to the wave, while a denser, more coherent mass couples its momentum more effectively into the free surface. This provides a mechanistic explanation for why granular landslides of equal volume and speed can produce markedly different waves depending on their internal structure and grain-size distribution.

The experimental campaign behind these results sits within a broader body of work by the same group, which has previously examined impulse waves across a broad spectrum of grain diameters, waves generated by subaerial cylinders, and the influence of rigid vegetation on wave characterization, as well as numerical simulations using coupled smoothed particle hydrodynamics and discrete element methods. By anchoring the new low-energy experiments to that earlier high-energy dataset, the authors avoid the common pitfall of building a model that fits one regime beautifully and another not at all. The unified framework is explicitly empirically derived, meaning its coefficients come from measurement rather than assumption, yet each equation carries an explicit physical interpretation, a combination that should make it easier for practitioners to judge when the model can be extrapolated and when it cannot.

For reservoir operators and geohazard agencies, the practical payoff is a set of tools that can be applied across the full range of plausible landslide scenarios, from small rockfalls that generate gentle oscillations to large collapses that launch near-field surges. The research was supported by Yunnan Fundamental Research Projects, a POWERCHINA Science and Technology Project, and Yunnan’s Technology Innovation Center for Digital Water Engineering, reflecting the acute relevance of landslide-generated waves to the hydropower reservoirs of southwestern China, where steep terrain, seismicity, and monsoonal loading keep slopes in a state of perpetual tension with the water below. With validated height predictions, a high-accuracy amplitude model, propagation and dispersion equations, and an energy-transfer framework grounded in the physics of granular porosity, the study moves the field closer to the long-sought goal of a single, trustworthy forecast: given a landslide on the slope, how big will the wave be, how fast will it decay, and how much of the slide’s energy will the reservoir ultimately absorb.

Subject of Research: Predictive modeling of impulse waves generated by subaerial granular landslides in reservoirs

Article Title: Unified predictive framework for subaerial granular landslide-generated stokes-like waves across energy regimes

Article References: Zheng, F., Liu, Q., Xue, Z., Li, C., Yang, Y., Yao, C., Huang, T., Liu, G., Xu, J., & Lin, X. (2026). Unified predictive framework for subaerial granular landslide-generated stokes-like waves across energy regimes. Ocean Dynamics, 76(10), Article 104. https://doi.org/10.1007/s10236-026-01862-z

Image Credits: AI Generated

DOI: 10.1007/s10236-026-01862-z

Keywords: landslide-generated impulse waves, Stokes-like waves, subaerial landslides, reservoir geohazards, wave amplitude prediction, wave attenuation, energy conversion, granular flows, laboratory experiments, Heller and Hager model, Ocean Dynamics, hazard assessment

Zhejiang presents an ideal and demanding test case. This coastal province in southeastern China is home to roughly 66.7 million people and an economy exceeding 9 trillion CNY in 2024, yet nearly 75 percent of its land is covered by hills and mountains, with only about 20 percent plains and a sliver of rivers and lakes. Rapid urbanization compounds the challenge, since weather stations are sparse, unevenly distributed, and subject to relocations, instrumentation changes, and gaps in the historical record. Ground observations offer accuracy but limited spatial coverage; satellite land surface temperature products offer detail but generally begin only in the early 2000s and are vulnerable to cloud cover, terrain shading, and atmospheric interference; reanalysis products such as ERA5 provide continuity but at coarse spatial resolutions, typically around 0.25 degrees or coarser, far too blunt to resolve the fine thermal texture of mountainous landscapes.

The team’s solution is an elegant two-part construction they call a spatial-background residual framework. First, they built a high-resolution climatological baseline from an existing hourly, 1-kilometer gridded temperature dataset covering 2008 to 2018, which had itself been developed using the INCA data-fusion framework with reanalysis fields and dense automatic weather station observations. This baseline serves purely as a spatial background, encoding how temperature varies across the terrain on each calendar day of the year. Second, daily temperature residuals—the departures of each station observation from that climatological expectation—were calculated for every meteorological station across the full 1961 to 2020 period. Because these residuals are computed directly from observed temperatures, they inherently preserve the long-term warming trend and interannual variability, while the baseline contributes the terrain-driven spatial detail. Summing the interpolated residual field with the baseline yields the finished reconstruction.

A critical methodological question was how best to interpolate those daily residuals across space. The researchers evaluated three schemes representing different levels of topographic correction and complexity: plain inverse distance weighting, or IDW, which relies only on spatial proximity; a lapse-rate-adjusted version of IDW, in which station temperatures are first corrected to grid-cell elevation using a fixed adiabatic lapse rate of 6.0 degrees Celsius per kilometer; and a multiple linear regression incorporating longitude, latitude, and elevation as predictors. Using leave-one-out cross-validation, in which each station is successively withheld and predicted from the others, the team assessed performance with mean absolute error, root-mean-square error, and the coefficient of determination across decades, seasons, and elevation zones.

The verdict was striking: the simplest method won. IDW without any lapse-rate correction consistently delivered the lowest errors and highest skill, achieving the best performance at roughly 68 percent of stations for daily minimum temperature, 70.7 percent for mean temperature, and 74.7 percent for maximum temperature. In a representative example from 1971 to 1980, IDW reconstructed minimum temperatures with a mean absolute error of just 0.73 degrees Celsius and an R-squared of 0.987, comfortably beating both rivals. The reason lies in the behavior of the lapse rate itself. Analysis of 60 years of observations revealed that near-surface temperature lapse rates in Zhejiang are strongly non-stationary: they peak in summer, with minimum-temperature lapse rates exceeding 7.0 degrees Celsius per kilometer in mountainous areas during July and August, yet collapse toward zero or even turn negative in winter lowlands, where temperature inversions prevail. Applying a fixed correction therefore risks systematic, elevation-related biases—a caution with implications well beyond Zhejiang.

The errors that do remain follow clear and intelligible patterns. Reconstruction accuracy improved steadily from the 1960s onward as station density grew, and summer months outperformed winter months because spatial temperature gradients are weaker in warm weather. Low-elevation areas below 400 meters consistently yielded smaller errors than high-elevation zones, where complex terrain and sparse instrumentation conspire against interpolation. Spatially, larger uncertainties cluster in the mountainous southwest, including parts of Lishui and western Wenzhou, while the plains around Hangzhou, Shaoxing, and Jinhua show excellent agreement, with most stations achieving R-squared values above 0.90 and many above 0.95. Across the entire 60-year span, the annual mean error for all three temperature variables stayed within plus or minus 0.1 degrees Celsius, with no systematic drift across decades—a testament to the temporal stability of the method.

Perhaps the most compelling validation came from real disasters. The team tested ZJ-DAT against two extreme events from 2007, using more than a thousand automatic weather stations as independent ground truth while deliberately excluding any stations that had contributed to the reconstruction. During the cold wave of 4 to 9 March 2007, ZJ-DAT tracked the south-to-north advance of the cold air, accurately reproducing the observed cold centers around Lishui, with R-squared values of 0.65 to 0.79 and root-mean-square errors of 1.01 to 1.67 degrees Celsius. By comparison, the CDAT national dataset managed only moderate agreement, while ERA5 performed poorly, with near-zero or negative correlations and errors approaching 3 degrees Celsius. The heat wave of 30 June to 10 July 2007 told the same story: ZJ-DAT best resolved the core hot zones above 37 degrees Celsius over Jinhua, Shaoxing, and Ningbo and the inland-coastal thermal contrast, while CDAT smoothed away local extremes and ERA5 drifted with warm biases and excessive homogenization. Case studies of cold and heat events in January and July 2020 at four environmentally distinct stations—an island, a mountain site, and two plain stations—further confirmed the reconstruction’s fidelity, with discrepancies generally under 2 degrees Celsius.

Beyond validation, the dataset enabled a first-of-its-kind hazard assessment for the province. Using Gumbel distribution analysis of return periods, the researchers mapped the intensity of extreme cold and heat expected at 5-, 20-, and 50-year recurrence intervals. The results expose stark geographic contrasts in climate risk. Extreme low-temperature hazards concentrate in the northwestern inland regions, where 50-year minimum temperatures plunge below minus 15 degrees Celsius, while the southeastern coast stays comparatively mild. Extreme heat hazards show the opposite pattern, dominated by low-altitude basins in central and northern Zhejiang, where 50-year maximum temperatures climb above 43 to 44 degrees Celsius—figures that carry sobering weight given projections of accelerating heatwave duration under global warming. These maps, grounded in kilometer-scale temperature data rather than coarse reanalysis, offer planners a far sharper picture of where adaptation investments are most needed.

The authors are candid about limitations. Anchoring the reconstruction to a climatology drawn from 2008 to 2018 means the reference field does not reflect the climate state of earlier decades, though because it functions only as a spatial scaffold while temporal signals come from station residuals, warming trends and variability remain intact. The team suggests that future refinements could employ temporally adaptive reference fields. The broader significance, however, is clear: ZJ-DAT demonstrates that a simple, computationally efficient interpolation of station anomalies, layered onto a modern high-resolution climatology, can outperform more elaborate schemes in complex terrain—provided the scheme is chosen with local lapse-rate physics in mind. The framework, and the publicly available dataset released through Zenodo, is designed to be transferable to other topographically complex, observation-limited regions, offering a practical foundation for climate monitoring, extreme-event risk assessment, and adaptation planning as the planet continues to warm.

Subject of Research: High-resolution daily near-surface air temperature reconstruction for Zhejiang Province, China, from 1961 to 2020 using statistical residual interpolation

Article Title: A high-resolution daily temperature reconstruction for Zhejiang Province during 1961–2020 using statistical residual interpolation

Article References: Li, Y., Guo, H., Dong, M., Wu, J., Deng, F., Chen, Y., & Chen, F. (2026). A high-resolution daily temperature reconstruction for Zhejiang Province during 1961–2020 using statistical residual interpolation. Theoretical and Applied Climatology, 157(10), Article 632. https://doi.org/10.1007/s00704-026-06521-3

Image Credits: AI Generated

DOI: 10.1007/s00704-026-06521-3

Keywords: Zhejiang Province, air temperature reconstruction, inverse distance weighting, spatial interpolation, temperature lapse rate, extreme temperature events, heat wave, cold wave, climate monitoring, hazard assessment, ERA5, high-resolution gridded dataset