pure and applied geophysics最新文献

{"title":"Seismic Random Noise Suppression with Wavelet-Inspired Invertible Network","authors":"Yan Zhou,&nbsp;Ying Rao,&nbsp;Zhencong Zhao,&nbsp;Yongfeng Zou","doi":"10.1007/s00024-026-04056-2","DOIUrl":"10.1007/s00024-026-04056-2","url":null,"abstract":"<div><p>Random noise attenuation is an essential process in seismic data processing, with  a major impact on subsequent workflows such as seismic interpretation. Because of the variety of frequency bands and apparent velocities, suppressing seismic random noise is challenging. Conventional techniques for suppressing seismic random noise include vector-median filtering, low-rank approximation and sparsity-domain based techniques. Degradation of effective seismic signals is inevitable, especially in areas with extremely complex structures. Machine learning is widely used in seismic data processing, offering high efficiency and automation. However, the black-box nature of machine learning algorithms limits their application in seismic data processing. To address this problem, we propose applying the multi-scale Wavelet-inspired Invertible Network (WIN) to seismic random noise suppression. The network has the following advantages: (1) It combines multiple lifting invertible neural networks, which ensures perfect reconstruction of the original signal; (2) The multiple lifting invertible neural networks act as multiple scales of the original signal and resemble denoising in the wavelet domain in a more efficient way; (3) The denoising performance of the proposed network can be controlled by choosing the noise level. A numerical study on field seismic data shows the advantage of the neural network over conventional denoising algorithms.</p></div>","PeriodicalId":21078,"journal":{"name":"pure and applied geophysics","volume":"183 8","pages":"3907 - 3920"},"PeriodicalIF":2.1,"publicationDate":"2026-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148751626","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Indice-Based Characterization of Precipitation Variability Using Multiple Trend Analysis Methods in the High-Rainfall Eastern Black Sea Basin, Türkiye","authors":"Cihangir Koycegiz,&nbsp;Kemal Saplıoğlu,&nbsp;Meral Buyukyildiz","doi":"10.1007/s00024-026-04035-7","DOIUrl":"10.1007/s00024-026-04035-7","url":null,"abstract":"<div><p>Climate change-driven shifts in precipitation regimes have intensified the frequency of meteorological disasters, directly impacting water resources and agricultural sustainability. This study investigated the spatial and temporal variability of precipitation in the Eastern Black Sea Basin (EBSB), the most humid basin in Türkiye, using data from nine meteorological stations for the 1965–2021 period. Five key indices were evaluated: the Precipitation Concentration Index (PCI), Modified Fournier Index (MFI), Rainfall Anomaly Index (RAI), Simple Daily Intensity Index (SDII), and the annual count of days with daily precipitation ≥ 20 mm (R20). Temporal trends were analyzed using the classic Mann–Kendall (MK) test alongside two innovative approaches: the Combination of Wilcoxon Test and Scatter Diagram (CWTSD) and the Onyutha Trend Test (OTT). The findings demonstrate that precipitation patterns align closely with index variability. Eastern stations recorded the highest MFI, SDII, and R20 values, reflecting higher rainfall totals, whereas the lowest values were observed in the central part of the basin. Western stations exhibited intermediate levels, surpassing the central region but remaining below eastern thresholds. PCI exhibited an opposite pattern, with maximum values in the central of the basin and lower values in the east and west. RAI lacked a distinct spatial distribution. Trend analysis revealed a consistent downward trend at the Akcaabat station for both precipitation and indices, while most other stations exhibited increasing trends. The results from both innovative and classical methods were largely consistent, enhancing the reliability of the findings. Significant upward trends (at the 5% level) were detected in MFI, RAI, and annual precipitation, especially across in eastern and western stations. These outcomes emphasize the pronounced regional variability in precipitation regimes and provide critical insights for mitigating hydrological risks under a changing climate.</p></div>","PeriodicalId":21078,"journal":{"name":"pure and applied geophysics","volume":"183 8","pages":"4337 - 4359"},"PeriodicalIF":2.1,"publicationDate":"2026-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148752231","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Modal Filtering of Depth-Dependent Stress Drop in a Finite Seismogenic Layer: An Analytical Source Model for Effective Rupture Thickness","authors":"Antonino D’Alessandro","doi":"10.1007/s00024-026-04052-6","DOIUrl":"10.1007/s00024-026-04052-6","url":null,"abstract":"<div><p>The finite thickness of the seismogenic layer is a fundamental constraint on earthquake rupture and seismic moment release. In many source models, however, this constraint is incorporated in a largely geometric manner, implicitly assuming a direct mapping between depth-dependent stress drop and fault slip. Here we develop an analytical earthquake source model that explicitly accounts for the elastic boundary conditions governing deformation within a finite-thickness seismogenic layer. The problem is formulated as a two-dimensional antiplane elastic boundary-value problem with a traction-free surface and an elastically coupled basal boundary. Depth-dependent stress-drop distributions are prescribed along the fault plane, while the resulting slip field is obtained through modal decomposition of the vertical elastic problem. This formulation reveals that imposed stress-drop profiles are mechanically filtered by a discrete set of vertical eigenmodes whose structure is controlled by the boundary conditions. We show that the fundamental vertical mode dominates depth-integrated quantities such as slip and seismic moment, whereas higher-order modes contribute oscillatory components that are progressively suppressed by elastic stiffness. Consequently, different depth-dependent stress-drop tapers, despite sharing the same geometric thickness and stress-drop scale, produce systematically different slip distributions and seismic moments. To quantify this effect, we introduce the concept of an effective rupture thickness, which represents the fraction of the seismogenic layer that efficiently contributes to coseismic deformation. The resulting framework generalizes classical seismic moment scaling relations by demonstrating that the thickness term entering the scaling law is an emergent mechanical property rather than a purely geometric parameter.</p></div>","PeriodicalId":21078,"journal":{"name":"pure and applied geophysics","volume":"183 8","pages":"3939 - 3955"},"PeriodicalIF":2.1,"publicationDate":"2026-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s00024-026-04052-6.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148752228","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Estimation of the Maximum Earthquake Magnitude, mmax—Choosing the Correct Time Interval for the Maximum Observed Magnitude","authors":"Andrzej Kijko,&nbsp;Petrus Johannes Vermeulen","doi":"10.1007/s00024-026-04057-1","DOIUrl":"10.1007/s00024-026-04057-1","url":null,"abstract":"<div><p>This study complements a prior study by Kijko (2004) that provides a statistical estimator of the area-characteristic maximum earthquake magnitude, <span>({m}_{max})</span>. This estimator is highly regarded and widely used. However, as this study explains, such estimators are often misapplied owing to the misinterpretation of the time period of the occurrence of <span>({m}_{max}^{obs})</span>, which is the largest <i>observed</i> earthquake. Accordingly, here, we examine the sources of such misunderstandings, which can lead to misuse of the procedure and, therefore, significantly affect the estimates of <span>({widehat{m}}_{text{m}text{a}text{x}})</span> and the overall seismic hazard indicators. As we show in an illustrative example, an assessment of <span>({m}_{max})</span> for the area surrounding Cape Town in South Africa, there is a substantial difference between correct and incorrect estimates of the maximum possible earthquake magnitude resulting from the use of the incorrect time interval. Furthermore, we discuss some insights about the variability of the estimated <span>({m}_{max})</span>, noting that the misapplication considered here can lead to overestimation, whereas a lack of data can lead to underestimation. In addition, we provide recommendations for discerning when over- or underestimation could occur. We conclude that the best practice is to incorporate prior information and apply a Bayesian formalism when such information is available.</p></div>","PeriodicalId":21078,"journal":{"name":"pure and applied geophysics","volume":"183 8","pages":"3893 - 3906"},"PeriodicalIF":2.1,"publicationDate":"2026-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s00024-026-04057-1.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148752230","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Full-Waveform Inversion Cycle-Skipping Mitigation with Dynamic Time Warping, Part 1: A Method for a Proximity Analysis Between the Current and the Observed Data","authors":"Claus Eikmeier,&nbsp;Jaime Souza,&nbsp;Carlos A. M. Chaves,&nbsp;Ernani Volpe","doi":"10.1007/s00024-026-04058-0","DOIUrl":"10.1007/s00024-026-04058-0","url":null,"abstract":"<div><p>Full-Waveform Inversion (FWI) is a seismic imaging technique employed to achieve detailed subsurface representations by leveraging seismic data. The method solves an optimisation problem to minimise discrepancies between modelled and observed data. However, FWI poses inherent challenges due to the complexity of its non-linear optimisation process and the non-convex nature of the objective function. Typically, gradient-based optimisation algorithms, such as L-BFGS-B, are used to search for the global minimum. Nevertheless, there exists no guarantee of convergence to the global minimum, potentially yielding inaccurate results. To circumvent the risk of converging to local minima, either an initial model closely resembling the true solution must be provided, or strategies to prevent cycle skipping should be employed. This paper is Part 1 of a study that explores the potential of using the Dynamic Time Warping (DTW) methodology to address cycle-skipping with a focus on underinvestigated approaches. We introduce the Dynamic Time Warping Proximity Analysis (DTW-PA), a procedure designed to provide a qualitative estimation of the current model’s suitability for FWI, while in Part 2, we present a method for mitigating cycle-skipping itself. Here, a time-domain multiscale strategy was employed to tackle cycle-skipping challenges. Our results indicate that DTW-PA offers low qualitative insights into the adequacy of the initial model for running FWI without cycle-skipping. However, it facilitates informed decision-making regarding the application of frequency filters for multiscale step building, reducing the number of steps needed and minimising time and computational costs.</p></div>","PeriodicalId":21078,"journal":{"name":"pure and applied geophysics","volume":"183 8","pages":"3957 - 3976"},"PeriodicalIF":2.1,"publicationDate":"2026-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s00024-026-04058-0.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148752229","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"On the Possibility of Ionospheric Memory Over the Low-Latitude Mexican Region","authors":"M. A. Sergeeva,&nbsp;I. D. Orrala-Legorreta,&nbsp;E. Rodriguez-Orozco,&nbsp;J. H. Castro-Chacon,&nbsp;O. A. Maltseva,&nbsp;J. A. Gonzalez-Esparza,&nbsp;P. Corona-Romero,&nbsp;V. J. Gatica-Acevedo,&nbsp;M. Pazos,&nbsp;M. E. Gonzalez-Alvarez,&nbsp;R. Bonifaz,&nbsp;A. Melgarejo-Morales,&nbsp;E. Cabral-Cano,&nbsp;F. Amezcua,&nbsp;N. A. Barboza-Tello,&nbsp;C. I. Castellanos-Velazco,&nbsp;E. Andrade-Mascote,&nbsp;P. Villanueva","doi":"10.1007/s00024-026-04061-5","DOIUrl":"10.1007/s00024-026-04061-5","url":null,"abstract":"<div><p>Features of possible regional ionospheric memory manifested themselves at low-latitudes under solar eclipse conditions.</p></div>","PeriodicalId":21078,"journal":{"name":"pure and applied geophysics","volume":"183 8","pages":"4179 - 4185"},"PeriodicalIF":2.1,"publicationDate":"2026-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s00024-026-04061-5.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148751627","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Full Waveform Inversion of DAS Crosswell Data to Monitor HT-ATES: A Synthetic Feasibility Study","authors":"Clara E. Fraile,&nbsp;Emmanuel Gaucher,&nbsp;Christophe Barnes","doi":"10.1007/s00024-026-04051-7","DOIUrl":"10.1007/s00024-026-04051-7","url":null,"abstract":"<div><p>Monitoring underground reservoirs is essential to ensure their efficient and sustainable operation. This study investigates the detectability and imaging, using seismic methods, of modifications in elastic properties caused by cyclic injection and production of geothermal fluid in high-temperature aquifer thermal energy storage (HT-ATES) systems. In such systems, the induced variations in elastic properties are expected to be very small, requiring high-sensitivity methods for reliable detection. The work is conceived as a feasibility study in a setting where field data are not yet available. Synthetic modelling is used to evaluate the effectiveness and sensitivity of Full Waveform Inversion (FWI) to reservoir variations. FWI is selected for its ability to exploit the full recorded wavefield, which is critical for resolving subtle elastic changes in thin, low-contrast reservoirs. The DeepStor HT-ATES research project in Karlsruhe, Germany, where layers thinner than 10 m at 1.3 km depth are targeted, serves as a case study. Two elastic models were developed: one unperturbed and one incorporating anomalies expected after five years of cyclic operations. Due to the thickness and depth of the targets, surface seismic methods were excluded in favour of cross-well configurations using Distributed Acoustic Sensing (DAS). Active sources are adopted to ensure controlled and repeatable illumination required for time-lapse monitoring. Results indicate that geometries with receivers intersecting the reservoirs significantly improve imaging and parameter estimation. This study evaluates the applicability of established FWI approaches in thin, stratified geothermal systems and highlights the importance of survey design for successful HT-ATES monitoring.</p></div>","PeriodicalId":21078,"journal":{"name":"pure and applied geophysics","volume":"183 8","pages":"4085 - 4111"},"PeriodicalIF":2.1,"publicationDate":"2026-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s00024-026-04051-7.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148752271","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Spectral Characterization of Coastal Tsunami Response in Suruga Bay, Japan, Using Stochastic Simulation Ensembles for Nankai Trough Earthquakes","authors":"Takuya Miyashita,&nbsp;Tung-Cheng Ho,&nbsp;Nobuhito Mori,&nbsp;Tomoya Shimura","doi":"10.1007/s00024-026-04060-6","DOIUrl":"10.1007/s00024-026-04060-6","url":null,"abstract":"<div><p>Coastal tsunami impacts are strongly influenced by local bathymetric features, which can induce resonance and prolong wave activity after the initial arrival. While previous studies have characterized resonant behavior using observations or deterministic simulations, the extent to which such frequency-dependent responses can be systematically related to offshore tsunami characteristics remains unclear. In this study, we investigate coastal tsunami response in Suruga Bay, Japan, using a large ensemble of numerical simulations based on stochastic earthquake source models for the Nankai Trough. For each scenario, we analyze the spectral characteristics of simulated wavefields at offshore and coastal locations and estimate frequency-dependent response functions associated with bathymetric features. These analyses reveal consistent amplification patterns at specific period ranges that are associated with local topography. Building on this, we examine statistical relationships between offshore spectral characteristics and the amplitude of later-arriving coastal waves. The results indicate that certain frequency bands show moderate but consistent correlations with the time-averaged amplitude of post-arrival oscillations, suggesting that offshore spectral information may provide useful indicators of prolonged coastal response under similar source conditions. The objective of this study is not to fully resolve the physical coherence of tsunami sources or nonlinear resonance processes, but rather to provide a statistical framework for linking offshore wave characteristics with site-specific coastal responses. The proposed approach offers a complementary perspective for interpreting tsunami-induced oscillations and may contribute to improving the assessment of prolonged tsunami effects in coastal regions with complex bathymetry.</p></div>","PeriodicalId":21078,"journal":{"name":"pure and applied geophysics","volume":"183 8","pages":"3869 - 3891"},"PeriodicalIF":2.1,"publicationDate":"2026-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s00024-026-04060-6.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148752264","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Real-Time Storm Surge Forecast and Its Validation for Severe Cyclonic Storm ‘Dana’","authors":"Siva Srinivas Kolukula,&nbsp;P. L. N. Murty,&nbsp;J. Padmanabham,&nbsp;V. Chandrasekhar,&nbsp;Monica Sharma,&nbsp;M. Mohapatra,&nbsp;T. M. Balakrishnan Nair","doi":"10.1007/s00024-026-04064-2","DOIUrl":"10.1007/s00024-026-04064-2","url":null,"abstract":"<div><p>The coastal regions of India are highly vulnerable to storm surge flooding during tropical cyclones. On 22 October 2024, at 05:30 IST, a severe cyclonic storm named Dana formed in the Bay of Bengal. It made landfall between 01:30 and 02:30 IST on 25 October 2024, with wind gusts reaching 120 km/h at Bhitarkanika and Dhamra ports in Odisha. Cyclone Dana caused heavy rainfall, strong winds, and inundation in low-lying areas of Kendrapara, Bhadrak, and Balasore districts, resulting in significant damage and disruption across Odisha and West Bengal. Approximately 500,000 people were evacuated from coastal Odisha by the National Disaster Response Force. The current study evaluates the performance of the Storm Surge Early Warning Centre at the Indian National Centre for Ocean Information Services during Cyclone Dana. The ADCIRC hydrodynamic model was forced with Holland parametric wind fields derived from 24-h and 48-h India Meteorological Department forecast tracks and compared against the best track. Atmospheric forcing was validated using Sentinel-1 Synthetic Aperture Radar winds and in-situ observations from the Ocean Moored Buoy Network for the Northern Indian Ocean. Simulated surge heights were validated against tide gauge records from Dhamra and Paradip, while inundation extents were compared with Sentinel-1 flood maps. Results indicate that the 24-h forecast accurately reproduced wind fields, surge heights, and inundation extents, whereas the 48-h forecast underestimated the intensity and spatial coverage. Although best-track simulations provide the most accurate results, real-time 24–48-h forecasts remain valuable for early warning and risk reduction. Future improvements will focus on incorporating river discharge, rainfall inputs, and land-use information to enhance forecast reliability.</p></div>","PeriodicalId":21078,"journal":{"name":"pure and applied geophysics","volume":"183 8","pages":"4411 - 4430"},"PeriodicalIF":2.1,"publicationDate":"2026-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148752265","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"SPEI-Based Drought Regionalization and Cluster-Specific Machine Learning Forecasting in the U.S. Black Belt","authors":"Negin Nahidi,&nbsp;Wesley C. Zech,&nbsp;Mohammad Ismaeil Kamali,&nbsp;Rouzbeh Nazari","doi":"10.1007/s00024-026-04045-5","DOIUrl":"10.1007/s00024-026-04045-5","url":null,"abstract":"<div><p>This study evaluates the historical spatial variability of drought in the Black Belt region of the southeastern U.S. by classifying the study area into drought-variation clusters and assesses the performance of three machine learning models for short-term drought forecasting. The Black Belt is a distinct agricultural subregion of the southeastern U.S., characterized by fertile Blackland soils, strong dependence on precipitation-based farming, limited irrigation infrastructure, uneven climatic monitoring coverage, socioeconomic challenges, and a history of repeated severe drought events that have caused substantial agricultural and water-supply impacts…Monthly gridded Standardized Precipitation Evapotranspiration Index (SPEI) [~ 5 km (3.1 miles) spatial resolution] was computed from precipitation and temperature data from the National Oceanic and Atmospheric Administration (NOAA) National Centers for Environmental Information (NCEI) nClimGrid-Daily dataset, based on Global Historical Climatology Network-Daily (GHCN-Daily) observations, and used as the primary drought indicator for the 1990–2025 study period. SPEI includes precipitation and temperature-driven evapotranspiration, which strongly influences drought onset in warm subtropical regions such as the Black Belt. The 6-month accumulation period (<span>({text{S}text{P}text{E}text{I}}_{6})</span>) is suitable for capturing agricultural drought by representing soil moisture conditions related to crop stress. Since the Black Belt has sparse and unevenly distributed weather stations, high-resolution gridded NOAA data provides spatially continuous, quality-controlled coverage (NOAA NCEI, n.d.). The Silhouette method was used to select influential <span>({text{S}text{P}text{E}text{I}}_{6})</span> statistical parameters and validate three clustering algorithms (i.e., K-means, Hierarchical, and Expectation–Maximization) to create drought clusters from <span>({text{S}text{P}text{E}text{I}}_{6})</span> data. K-means performed best, identifying three distinct drought-characteristic clusters. Three machine-learning models, Support Vector Regression (SVR), the M5P model tree (a decision-tree algorithm that fits linear regressions at its terminal nodes), and a Hybrid ensemble model combining SVR and M5P outputs, were trained for each cluster for 1-, 2-, and 3-month forecasting horizons using lagged <span>({text{S}text{P}text{E}text{I}}_{6})</span> time series. Model performance declined for longer horizons and for the extreme-like drought characteristics, while the Hybrid model achieved the best results. These findings highlight the effectiveness of combining drought clustering with region-specific machine learning models for improved short-term drought forecasting in heterogeneous environments such as the U.S. Black Belt.</p></div>","PeriodicalId":21078,"journal":{"name":"pure and applied geophysics","volume":"183 8","pages":"4361 - 4396"},"PeriodicalIF":2.1,"publicationDate":"2026-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s00024-026-04045-5.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148752267","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}