Acta Geophysica最新文献

{"title":"Understanding flood susceptibility in a monsoon-influenced catchment through interpretable deep learning and hydro-climatic controls","authors":"Dillip Kumar Ghose,&nbsp;Bibhu Prasad Mishra","doi":"10.1007/s11600-026-02001-4","DOIUrl":"10.1007/s11600-026-02001-4","url":null,"abstract":"<div><p>Flooding is one of the most common and severe hydro-meteorological hazards, posing a significant risk to water resources, infrastructure, and socio-economic security, especially in monsoon-prone river catchments. Accurate identification of flood-prone areas is critical for effective water risk management and flood-prone catchment planning. This study proposes a data-driven framework for flood susceptibility mapping in the Baitarani River Basin, eastern India, by incorporating hydro-geomorphic, hydro-climatic, and hydro-soil factors, which are critical for understanding the complex flood mechanisms in the catchment. The proposed framework incorporates a novel feature selection method, the μ-Relief algorithm, which selects critical flood predictors and removes redundancy in high-dimensional flood datasets. A novel interpretable deep-learning model specifically designed for tabular data has been developed and improved by evolutionary optimisation for improved performance and reliability. The performance of the proposed framework was evaluated using multiple performance metrics, including AUROC, F1-score, Jaccard index, and Hamming loss, in multiple validation experiments to improve the reliability and robustness of the proposed framework. The proposed framework achieved promising predictive performance, with an AUROC of 0.93 and an F1-score of 0.888, outperforming the baseline models. Spatial analysis revealed that 29% of the basin is classified as moderately to highly flood-susceptible, predominantly occurring in very-low lying floodplains, river confluences, and rapidly urbanising areas. The interpretability analysis has shown that elevation and rainfall are the dominant flood predictors. In contrast, wetness, land use, and soil factors also contribute significantly to flood susceptibility, which is critical for understanding flood mechanisms in the Baitarani River Basin. The proposed framework has demonstrated its applicability in other catchments and provides a tool to enhance flood resilience in flood-prone catchments in response to changing hydro-climatic and hydro-geomorphic conditions.</p></div>","PeriodicalId":6988,"journal":{"name":"Acta Geophysica","volume":"74 5","pages":""},"PeriodicalIF":2.1,"publicationDate":"2026-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148927089","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":"Alarming increase in uncomfortable hot hours in the cities of Kerala, India","authors":"M. S. VishnuPriya,&nbsp;V. Agilan","doi":"10.1007/s11600-026-02005-0","DOIUrl":"10.1007/s11600-026-02005-0","url":null,"abstract":"<div><p>The surface temperature in India has experienced an increase as a result of climate change in recent decades. When the ambient temperature fluctuates, the human body strives to maintain equilibrium; however, if the human body fails to achieve that balance, we feel uncomfortable. Temperatures above 32 °C can be uncomfortably hot for humans in India. Thermal discomfort has adverse impacts on various sectors, such as human health, labour productivity and power. This study investigates the temporal variations in the annual, seasonal and monthly counts of the Number of Hours in which the Temperature exceeds 32 °C (NHT32) for eight Indian cities: Ahmedabad, Bangalore, Chennai, Kolkata, Kozhikode, Mumbai, New Delhi and Thiruvananthapuram. Initially, the trend in the annual, seasonal and monthly counts of NHT32 is examined using the Mann‒Kendall (MK) test. Further, the change in the Percentage of Hours in which the Temperature exceeds 32 °C (PHT32) is assessed. The results show that the NHT32 counts of Kozhikode and Thiruvananthapuram are increasing at an alarming rate among the cities considered in this study. In Kozhikode, the PHT32 increased by 7.32% in a year, from 2.69 to 10.01%. The most substantial increase in the PHT32 is observed for Kozhikode during March (from 5.65 to 22.2%).</p></div>","PeriodicalId":6988,"journal":{"name":"Acta Geophysica","volume":"74 5","pages":""},"PeriodicalIF":2.1,"publicationDate":"2026-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148927020","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":"Seismicity patterns and b-value variability in the Santorini–Amorgos system within the South Aegean volcanic arc","authors":"Gulten Polat,&nbsp;Günseli Erdem Altın","doi":"10.1007/s11600-026-02007-y","DOIUrl":"10.1007/s11600-026-02007-y","url":null,"abstract":"&lt;div&gt;&lt;p&gt;The Santorini–Amorgos system is one of the most active volcanic–tectonic settings of the South Aegean Volcanic Arc, where crustal extension, active faulting, volcanism, and recurrent seismicity interact within a complex geodynamic environment. This study investigates the spatial and temporal characteristics of seismicity using the homogeneous National Observatory of Athens–Hellenic Unified Seismological Network (NOA-HUSN) earthquake catalogue for the period from 17 March 2017 to 3 February 2026. An initial catalogue of 38,137 earthquakes was extracted within 33.5°–37.5° N and 18.8°–29.5° E. Following quality control and the application of focal-depth and magnitude criteria, 36,195 crustal earthquakes with depths of 2–40 km and local magnitudes of ML 1.0–6.3 were retained. Magnitude completeness (Mc), Gutenberg-Richter &lt;i&gt;b&lt;/i&gt;-values, and Omori–Utsu &lt;i&gt;p&lt;/i&gt;-values were evaluated using frequency–magnitude statistics, sliding-window analyses, spatial gridding, node-specific completeness estimates, and bootstrap uncertainty analysis. The maximum-curvature method yielded Mc = 2.15, and a conservative reference threshold of Mc₀ = 2.35 was adopted for the fixed-threshold analyses. Spatially smoothed &lt;i&gt;b&lt;/i&gt;-value estimates range from approximately 0.4–1.8 and reveal pronounced regional heterogeneity in earthquake magnitude–frequency characteristics. Relatively low &lt;i&gt;b&lt;/i&gt;-values occur within several earthquake-dense and structurally organized seismic corridors, whereas moderately elevated values form discontinuous patches in several offshore sectors surrounding the central tectonic corridor. No specific anomaly is assigned directly to the immediate Santorini volcanic complex or Kolumbo volcanic centre, where valid grid estimates are absent or spatially incomplete. Bootstrap-derived uncertainties and the spatial distribution of Mc indicate that the principal patterns are generally stable and unlikely to result primarily from sampling variability or magnitude truncation. Temporal analyses show moderate short-term fluctuations in Mc and &lt;i&gt;b&lt;/i&gt;-values but no persistent long-term monotonic trend. Comparisons between fixed and window-specific completeness thresholds demonstrate that time-dependent Mc estimates can amplify apparent temporal variations in &lt;i&gt;b&lt;/i&gt;-values through methodological coupling, whereas the fixed-threshold approach provides a more stable representation of long-term seismicity evolution. The estimated Omori–Utsu &lt;i&gt;p&lt;/i&gt;-values exhibit limited spatial variability, remaining close to unity (approximately 0.94–1.03). Within the adopted spatially smoothed nearest-neighbour methodology, these estimates represent regional post-mainshock temporal decay characteristics rather than parameters of individually identified aftershock sequences. Accordingly, earthquake magnitude-frequency characteristics exhibit substantially greater regional variability than the estimated temporal decay parameter. These results provide new quantitative","PeriodicalId":6988,"journal":{"name":"Acta Geophysica","volume":"74 5","pages":""},"PeriodicalIF":2.1,"publicationDate":"2026-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148926969","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":"Comprehensive geophysical and petrographic approach in seismic hazard assessment: case study of Ransiki Fault Area, South Manokwari Regency, Eastern Indonesia","authors":"Richard Lewerissa,&nbsp;Darma Santi,&nbsp;Laura A. S. Lapono,&nbsp;Restu Tandirerung,&nbsp;Jacson V. Morin,&nbsp;Herlambang Hudha","doi":"10.1007/s11600-026-01987-1","DOIUrl":"10.1007/s11600-026-01987-1","url":null,"abstract":"<div><p>The Ransiki fault zone in South Manokwari Regency, West Papua, Indonesia, poses significant seismic hazards to residents and infrastructure in the region. This study assessed these hazards using a multidisciplinary approach that integrated Probabilistic Seismic Hazard Analysis (PSHA), Horizontal-to-Vertical Spectral Ratio (HVSR) analysis from microtremor recordings, Vertical Electrical Sounding (VES), and petrographic analysis. The PSHA was conducted at 72 points (node spacing 0.1°, ~ 11 km) with 18 HVSR and VES control points using a logic tree with three ground motion prediction equations (GMPE). The results show spectral acceleration (<i>S</i>_a) of 0.049–1.185 g at <i>T</i> = 0.2 s and 0.006–0.129 g at <i>T</i> = 2.0 s for a 10% Probability of Exceedance (PoE) in 50 years (~ 475-year recurrence), with the highest hazard along the Ransiki Fault. HVSR analysis at 18 stations revealed natural frequencies (<span>(f_{0})</span>) of 0.350–9.480 Hz, amplification factors (<span>(A_{0})</span>) of 1.970–5.460, and dominant periods (<span>(T_{0})</span>) of 0.105–2.860 s. HVSR curves from Raipawi Village (<span>(f_{0})</span> = 0.35 Hz; <span>(A_{0})</span> = 5.46) and Wariab Village (<span>(f_{0})</span> = 1.5 Hz; <span>(A_{0})</span> = 4.70) show site response contrasts that reflect subsurface geological heterogeneity. VES surveys at four earthquake-affected locations identified water-saturated sediment layers and potential aquifer zones at depths of 5–60 m, with resistivity values of 2.62–2554.73 Ohm m, indicating high liquefaction risk in the southern coastal plains. Petrographic analysis identified andesite and sublitharenite as the dominant lithologies, each with different seismic loading susceptibility. The integrated dataset produced a detailed seismic hazard zoning map that established a vulnerability gradient from north to south. This study is the first geophysical and petrographic seismic hazard assessment in the Ransiki District, providing a scientific foundation for earthquake-resistant infrastructure design and spatial planning in high-risk Eastern Indonesia.</p></div>","PeriodicalId":6988,"journal":{"name":"Acta Geophysica","volume":"74 5","pages":""},"PeriodicalIF":2.1,"publicationDate":"2026-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148926985","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":"Spatio-temporal evaluation of gridded precipitation products over Balochistan province in Pakistan","authors":"Aman Jan,&nbsp;Kamal Ahmed","doi":"10.1007/s11600-026-02000-5","DOIUrl":"10.1007/s11600-026-02000-5","url":null,"abstract":"<div><p>Gridded precipitation (<i>P</i>) products provide a promising substitute for ground-based measurements in data-scarce regions such as Balochistan province in Pakistan. This study evaluated the performance of five state-of-the-art gridded <i>P</i> products (CHIRPSv2, MSWEPv2.2, PERSIANN-CDR, ERA5, and 5-KM) in the study area. The <i>P</i> products were evaluated through a point-to-pixel comparison at the monthly, annual, and seasonal temporal scales using data from 13 gauges. The modified Kling-Gupta efficiency (KGE’), along with its individual components, the linear correlation (<i>r</i>), bias ratio (<i>β</i>), and variability ratio (<i>γ</i>), and the root mean square error (RMSE), were used as statistical indices to evaluate the performance of the <i>P</i> products. The results of the study revealed that generally, MSWEPv2.2 (daily gauge-corrected) presented the best spatio-temporal KGE’ performance over most of the stations and across the province. The reanalysis dataset of ERA5 closely followed MSWEPv2.2 and was the best-performing product at the annual temporal scale. CHIRPSv2 (monthly gauge-corrected), PERSIANN-CDR (relies on secondary sources of gauge data), and 5-KM (daily gauge-based gridded product) showed relatively poor KGE’ performance. However, 5-KM presented the best performance for the <i>β</i> component of the KGE’, overall. RMSE results were different than those of the KGE’. Generally, CHIRPSv2 and PERSIANN-CDR outperformed MSWEPv2.2 and ERA5 at the temporal scales. Similar results were obtained for the spatial scale at the wetter stations that are located in eastern Balochistan. However, MSWEPv2.2 and ERA5 presented better spatial RMSE performance in the western drier stations. As far as 5-KM is concerned, it generally showed the weakest RMSE performance at all the spatio-temporal scales across the province. The results of this study confirm that it is difficult to point out a single best-performing <i>P</i> product that performs the best across regions with different climates and topographical gradients and at different temporal scales. However, overall, MSWEPv2.2 and ERA5 are found to be most acceptable in the study area.</p></div>","PeriodicalId":6988,"journal":{"name":"Acta Geophysica","volume":"74 5","pages":""},"PeriodicalIF":2.1,"publicationDate":"2026-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148926910","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":"Present-day kinematics and strain rate field along North Africa from seismic moment tensor summation","authors":"Abdelouahab Issaadi,&nbsp;Abdelhakim Ayadi,&nbsp;Farida Ousadou","doi":"10.1007/s11600-026-01992-4","DOIUrl":"10.1007/s11600-026-01992-4","url":null,"abstract":"<div><p>The geodynamics of northern Africa was examined by several studies following various approaches. All these studies concluded to a NW–SE convergence of the African plate towards the Eurasian plate with different local kinematic pattern. Seismic moment tensor summation method was adopted to study the seismic deformation and present-day kinematic along Northern Africa. In the study area, we have an interaction between the Eurasian, the Arabian and the African plates, exhibiting: (1) a collision movement in the Maghreb, Libya and Egypt with a shortening in the NW–SE direction, (2) a rifting along the Red Sea and (3) a transform fault zone in the Gulf of Aqaba. The seismic activity in Northern Africa is characterized by moderate seismic activity with strong earthquakes occurring episodically, most of the strong events are located in the Maghreb area, especially in Algeria and Morocco, the others are in the Gulf of Aqaba along the Dead Sea Fault (DSF). The seismic moment tensor summation method we adopted, leads to calculate the strain rates in several zones using the available seismic catalog of source parameters of earthquakes for Morocco, Algeria, Tunisia, Libya, and Egypt. The catalog was compiled from several local and global databases such as Harvard CMT, ETHZ, IGN, and RCMT and others from published papers, covering the period from 1954 to 2022. The results show a counterclockwise rotation of Africa with respect to Eurasia. The general direction of movements varies from northward (N10°) in the Gulf of Aqaba to westward (N285°) in the Moroccan Rif. The resulting velocity vectors are compared to the results given by other studies using other approaches.</p></div>","PeriodicalId":6988,"journal":{"name":"Acta Geophysica","volume":"74 5","pages":""},"PeriodicalIF":2.1,"publicationDate":"2026-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148927151","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":"Understanding flood susceptibility in a monsoon-influenced catchment through interpretable deep learning and hydro-climatic controls","authors":"Dillip Kumar Ghose,&nbsp;Bibhu Prasad Mishra","doi":"10.1007/s11600-026-02001-4","DOIUrl":"10.1007/s11600-026-02001-4","url":null,"abstract":"<div><p>Flooding is one of the most common and severe hydro-meteorological hazards, posing a significant risk to water resources, infrastructure, and socio-economic security, especially in monsoon-prone river catchments. Accurate identification of flood-prone areas is critical for effective water risk management and flood-prone catchment planning. This study proposes a data-driven framework for flood susceptibility mapping in the Baitarani River Basin, eastern India, by incorporating hydro-geomorphic, hydro-climatic, and hydro-soil factors, which are critical for understanding the complex flood mechanisms in the catchment. The proposed framework incorporates a novel feature selection method, the μ-Relief algorithm, which selects critical flood predictors and removes redundancy in high-dimensional flood datasets. A novel interpretable deep-learning model specifically designed for tabular data has been developed and improved by evolutionary optimisation for improved performance and reliability. The performance of the proposed framework was evaluated using multiple performance metrics, including AUROC, F1-score, Jaccard index, and Hamming loss, in multiple validation experiments to improve the reliability and robustness of the proposed framework. The proposed framework achieved promising predictive performance, with an AUROC of 0.93 and an F1-score of 0.888, outperforming the baseline models. Spatial analysis revealed that 29% of the basin is classified as moderately to highly flood-susceptible, predominantly occurring in very-low lying floodplains, river confluences, and rapidly urbanising areas. The interpretability analysis has shown that elevation and rainfall are the dominant flood predictors. In contrast, wetness, land use, and soil factors also contribute significantly to flood susceptibility, which is critical for understanding flood mechanisms in the Baitarani River Basin. The proposed framework has demonstrated its applicability in other catchments and provides a tool to enhance flood resilience in flood-prone catchments in response to changing hydro-climatic and hydro-geomorphic conditions.</p></div>","PeriodicalId":6988,"journal":{"name":"Acta Geophysica","volume":"74 5","pages":""},"PeriodicalIF":2.1,"publicationDate":"2026-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148926965","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":"Spatio-temporal evaluation of gridded precipitation products over Balochistan province in Pakistan","authors":"Aman Jan,&nbsp;Kamal Ahmed","doi":"10.1007/s11600-026-02000-5","DOIUrl":"10.1007/s11600-026-02000-5","url":null,"abstract":"<div><p>Gridded precipitation (<i>P</i>) products provide a promising substitute for ground-based measurements in data-scarce regions such as Balochistan province in Pakistan. This study evaluated the performance of five state-of-the-art gridded <i>P</i> products (CHIRPSv2, MSWEPv2.2, PERSIANN-CDR, ERA5, and 5-KM) in the study area. The <i>P</i> products were evaluated through a point-to-pixel comparison at the monthly, annual, and seasonal temporal scales using data from 13 gauges. The modified Kling-Gupta efficiency (KGE’), along with its individual components, the linear correlation (<i>r</i>), bias ratio (<i>β</i>), and variability ratio (<i>γ</i>), and the root mean square error (RMSE), were used as statistical indices to evaluate the performance of the <i>P</i> products. The results of the study revealed that generally, MSWEPv2.2 (daily gauge-corrected) presented the best spatio-temporal KGE’ performance over most of the stations and across the province. The reanalysis dataset of ERA5 closely followed MSWEPv2.2 and was the best-performing product at the annual temporal scale. CHIRPSv2 (monthly gauge-corrected), PERSIANN-CDR (relies on secondary sources of gauge data), and 5-KM (daily gauge-based gridded product) showed relatively poor KGE’ performance. However, 5-KM presented the best performance for the <i>β</i> component of the KGE’, overall. RMSE results were different than those of the KGE’. Generally, CHIRPSv2 and PERSIANN-CDR outperformed MSWEPv2.2 and ERA5 at the temporal scales. Similar results were obtained for the spatial scale at the wetter stations that are located in eastern Balochistan. However, MSWEPv2.2 and ERA5 presented better spatial RMSE performance in the western drier stations. As far as 5-KM is concerned, it generally showed the weakest RMSE performance at all the spatio-temporal scales across the province. The results of this study confirm that it is difficult to point out a single best-performing <i>P</i> product that performs the best across regions with different climates and topographical gradients and at different temporal scales. However, overall, MSWEPv2.2 and ERA5 are found to be most acceptable in the study area.</p></div>","PeriodicalId":6988,"journal":{"name":"Acta Geophysica","volume":"74 5","pages":""},"PeriodicalIF":2.1,"publicationDate":"2026-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148927088","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":"Integrating machine learning-estimated Swir for robust rock typing and 3D modeling of heterogeneous carbonate reservoirs","authors":"Iman Zahmatkesh,&nbsp;Samaneh Barbari,&nbsp;Akbar Heidari,&nbsp;Khaled Marufi,&nbsp;Maedeh Eslami","doi":"10.1007/s11600-026-01966-6","DOIUrl":"10.1007/s11600-026-01966-6","url":null,"abstract":"<div><p>Conventional log-based rock typing often fails to capture the complex pore systems of heterogeneous carbonate reservoirs, particularly in transition and water-bearing zones where fluid saturation masks true rock quality. To resolve this uncertainty, this study implements a physics-guided machine learning workflow in the Bangestan reservoir of the Ab-Teymur field, SW Iran. First, irreducible water saturation (Swir) was selected as a direct proxy for pore geometry and estimated across 60 wells using a <i>K</i>-nearest neighbor (KNN) algorithm. This model, trained on physically filtered data, achieved high predictive accuracy (<i>R</i>2 = 0.93) and was rigorously validated against core-derived saturation measurements. The estimated Swir log was then integrated into a self-organizing map (SOM) to guide the clustering process. The resulting model identified five distinct electrofacies that correspond to specific hydraulic units, ranging from high-quality dissolution-enhanced packstones (mean core-derived Swir ~ 17%) to tight mudstones (mean core-derived Swir ~ 62%). Finally, 3D sequential indicator simulation (SIS) was employed to model the spatial distribution of these facies and explicitly quantify uncertainty via probability maps. By anchoring machine learning predictions to physical rock measurements, this approach effectively links static well data to dynamic reservoir properties, providing a robust framework for reducing geological uncertainty.</p></div>","PeriodicalId":6988,"journal":{"name":"Acta Geophysica","volume":"74 5","pages":""},"PeriodicalIF":2.1,"publicationDate":"2026-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148927019","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":"Integrating machine learning-estimated Swir for robust rock typing and 3D modeling of heterogeneous carbonate reservoirs","authors":"Iman Zahmatkesh,&nbsp;Samaneh Barbari,&nbsp;Akbar Heidari,&nbsp;Khaled Marufi,&nbsp;Maedeh Eslami","doi":"10.1007/s11600-026-01966-6","DOIUrl":"10.1007/s11600-026-01966-6","url":null,"abstract":"<div><p>Conventional log-based rock typing often fails to capture the complex pore systems of heterogeneous carbonate reservoirs, particularly in transition and water-bearing zones where fluid saturation masks true rock quality. To resolve this uncertainty, this study implements a physics-guided machine learning workflow in the Bangestan reservoir of the Ab-Teymur field, SW Iran. First, irreducible water saturation (Swir) was selected as a direct proxy for pore geometry and estimated across 60 wells using a <i>K</i>-nearest neighbor (KNN) algorithm. This model, trained on physically filtered data, achieved high predictive accuracy (<i>R</i>2 = 0.93) and was rigorously validated against core-derived saturation measurements. The estimated Swir log was then integrated into a self-organizing map (SOM) to guide the clustering process. The resulting model identified five distinct electrofacies that correspond to specific hydraulic units, ranging from high-quality dissolution-enhanced packstones (mean core-derived Swir ~ 17%) to tight mudstones (mean core-derived Swir ~ 62%). Finally, 3D sequential indicator simulation (SIS) was employed to model the spatial distribution of these facies and explicitly quantify uncertainty via probability maps. By anchoring machine learning predictions to physical rock measurements, this approach effectively links static well data to dynamic reservoir properties, providing a robust framework for reducing geological uncertainty.</p></div>","PeriodicalId":6988,"journal":{"name":"Acta Geophysica","volume":"74 5","pages":""},"PeriodicalIF":2.1,"publicationDate":"2026-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148926909","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}