Environmental and Earth Sciences

Abstract: Fossil physeteroids exhibited substantially greater diversity in body size and feeding ecology than their extant relatives, yet body-length estimates for many taxa remain dependent on incomplete skeletons and proportional scaling from a limited number of reference species. This study reassesses body-size estimates in fossil physeteroids using a cranial proportion-based framework centered on bizygomatic width (BZW) and condylobasal length (CBL). The reference body lengths of Zygophyseter varolai and Brygmophyseter shigensis were reevaluated using published vertebral measurements, preserved skeletal length, and scale-bar-based skeletal data, and were conservatively set at approximately 6.5 m and 6.0 m, respectively. These revised reference values were then used to generate updated body-length estimates for a broad sample of extinct physeteroids. Revised estimates indicate that large-bodied physeteroids were relatively uncommon. Among the non-kogiid genera examined, only seven reached or exceeded 6 m in estimated total length, whereas Livyatan and Eophyseter were the only fossil genera estimated to exceed 9.5 m. Macroraptorial physeteroids ranged from approximately 4.0–4.3 m in Acrophyseter spp. to approximately 15–16 m in the giant Livyatan melvillei. The results further suggest that large body size evolved under multiple ecological regimes. Some large-bodied physeteroids were macroraptorial predators, whereas others show adaptations consistent with suction feeding or predation on smaller prey. The temporal distribution of giant forms is therefore more consistent with repeated, independent increases in body size than with a single evolutionary trend toward gigantism. However, incomplete preservation and uncertainty in the body lengths of reference taxa remain important limitations. The framework presented here provides a more consistent basis for reassessing body-size diversity in fossil physeteroids and can be refined as more complete cranial and postcranial material becomes available.

Abstract: We present a novel simple parametric tomographic technique allowing to derive the cloud extinction coefficient fields from the measured cloud dimensions and optical thickness (COT). The currently available cloud tomographic algorithms are based on either least-square fit over multi-dimensional space of cloud spatial configurations or on a simpler technique using Radon transform. Both approaches are computatially extensive and require radiometric measurements made with high angular resolution. This raises questions about their applicability to satellite data analysis. In this study we propose a simpler and much more robust parametric alternative to these algorithms that is based on the internal structures found in LES-generated clouds. This approach is analytical, thus, computationally fast, and has a potential of application to airborne and satellite observations in real time.

Abstract: Emergency forecasts benefit from scene detail when it constrains the relevant physical process and reaches decision-makers in time. This critical review examines flood inundation, wildfire spread and marine drift through 148 sources and 16 purpose-selected study families. Within this selected corpus, four families develop dedicated 3D conversion, chiefly for hydraulic-domain or fuel-material preparation; the coding rule excludes routine terrain input and existing ocean analyses. Forecast cases mainly update states, controls or initial trajectories. Hydraulic connectivity can alter local inundation. Corrected fire fronts remain exposed to propagation error, while wider marine search regions improve coverage at increased search cost. These mechanisms explain why geometric accuracy and analysis fit alone cannot establish predictive value. We propose a framework that selects scene attributes, update targets and observation timing jointly. Matched alternatives share a future response quantity, information cutoff and resource budget; deployment-specific tolerances connect their comparison to evacuation requirements and search capacity. Physically coupled neural representations require calibrated material properties and an online delivery-time account. The proposed tests assess the forecast effect of additional scene detail, the persistence of a correction and the time left for response.

Abstract: Vegetation and snow cover over the Tibetan Plateau (TP) strongly regulate land–atmosphere energy and water exchanges and alter the thermal forcing of the TP and thereby influence regional climate. Against the background of climate change, pronounced vegetation greening has been observed and is projected to continue in the future. However, the potential impacts of increased non-growing-season vegetation, particularly exposed withered grass stems, on winter snow cover and regional climate remain poorly understood. Using numerical experiments conducted with the Community Earth System Model version 2.2 (CESM2.2), we investigate the effects of increased withered grass stems (WGS) over the TP on winter snow cover fraction, the surface energy budget, atmospheric circulation, precipitation, and near-surface air temperature. The results show that increased WGS reduces the winter-mean snow cover fraction by approximately 11.1%, with the largest decrease occurring in December (16.7%), primarily over the central and eastern TP. Greater exposure of WGS above the snow-pack lowers surface albedo and increases net shortwave radiation, thereby enhancing surface diabatic heating. The enhanced thermal forcing induces warming in the low-er-to-middle troposphere over the TP and its southern flank, but cooling to the north. Correspondingly, at 500 hPa, anticyclonic circulation anomalies develop to the south of the TP, while cyclonic anomalies occur to the north. These circulation changes further modify large-scale horizontal heat and moisture transport. Enhanced moisture convergence leads to increased winter precipitation over the Yangtze River basin, regions to its south, and the western TP, whereas enhanced moisture divergence contributes to reduced precipitation over the eastern Indochina Peninsula and adjacent regions. The cooling north of the TP is mainly associated with anomalous horizontal heat transport, whereas warming over the TP is directly related to the WGS-induced reduction in surface albedo and increase in net surface radiation. These results highlight that changes in non-growing-season vegetation can substantially modify wintertime thermal forcing over the TP through vegetation–snow–surface energy interactions and, in turn, exert considerable impacts on atmospheric circulation and regional climate over surrounding and downstream regions.

Abstract: We quantified potentially toxic elements in surface soils from Taranto (Apulia, Italy) along an industrial-to-urban gradient (0-15 km) from a large integrated steel plant. Industrial emissions have been associated with increased cancer incidence, respiratory and neurodevelopmental impacts. We performed 3,446 georeferenced in situ measurements using portable X-ray fluorescence (p-XRF) spectroscopy in five residential districts (Tamburi, Statte, Paolo VI, Taranto centre, and Talsano) at incremental distance from the point source. Concentrations of arsenic (As), copper (Cu), iron (Fe), lead (Pb), manganese (Mn), mercury (Hg), nickel (Ni) rubidium (Rb), strontium (Sr), selenium (Se) and zinc (Zn) differed significantly across districts. Fe, Hg, Mn, Ni, Pb, Se, and Zn showed a clear distance-related decrease from the plant perimeter to more distal residential locations. Exceedances of Italian residential soil limits (D.Lgs. 152/2006) were frequent for As, Cu, Ni, Pb, Se, and Zn and Hg exceeded the 1 mg/kg residential limit in all measurements. Spatial patterns were consistent with atmospheric transport and deposition of industrial particulates, with additional contributions from traffic and other local sources. These findings document widespread multi-element soil contamination in Taranto and support sustained environmental monitoring and targeted remediation, prioritizing neighborhoods closest to the industrial area and incorporating community engagement.

Abstract: In the task of building extraction from high-resolution remote sensing images, challenges remain due to large variations in building scales, confusion between buildings and complex backgrounds, and oversmoothing of boundary details. Existing deep learning methods based solely on convolutional neural networks (CNNs) or Transformers often fail to simultaneously capture local geometric details and global semantic consistency. This study proposes a dual-branch building extraction network that combines CNN-based multi-scale spatial features with Transformer-based global semantic features. An attention-guided multi-scale feature fusion module is designed to alleviate feature conflicts and to suppress complex background noise. A boundary-aware loss and a multi-scale deep supervision strategy are introduced to enforce geometric constraints on building contours and intermediate features. Experiments on the Gaofen-7 (GF-7) dataset show that the proposed method achieves an IoU of 0.8065 and an F1 score of 0.8929, outperforming the compared methods. When using only 50% of the training samples, the model still attains an IoU of 0.7701, indicating good stability and application potential under limited labeled data. The method enables fine extraction of buildings from high-resolution remote sensing images and provides technical support for geographic information applications such as land-use survey, urban planning, and disaster assessment.

Abstract: Industrial-residue valorisation through polymer-composite manufacturing can reduce reliance on virgin resources, yet the environmental safety of such materials is rarely evaluated beyond conventional physicochemical characterisation. This study provides a novel, integrated assessment of a waste-derived phenol–formaldehyde/glass-fibre composite by combining pH-dependent leaching analysis with a multispecies ecotoxicological test battery. The composite contained post-production glass fibres, paper flakes, and viscose fibres bound with phenol–formaldehyde resin. Under near-neutral extraction conditions, most analysed metals and metalloids showed limited release. In contrast, acidic conditions substantially increased the mobilisation of Ba, Cr, Zn, Cu, and Pb, while the phenol index increased approximately 28-fold, from 0.25 ± 0.06 to 6.9 ± 1.7 mg L⁻¹. Biological responses differed markedly among taxa. Hordeum vulgare and Eisenia andrei showed limited short-term effects, whereas the marine diatom Skeletonema marinoi exhibited pronounced concentration- and time-related inhibition, with no cells detected after three days of exposure to the highest extract loading. The Aliivibrio fischeri assay also revealed concentration-dependent acute toxicity, with the strongest effects observed during the initial exposure period. The pronounced algal response could not be fully explained by the targeted chemical analyses, indicating a possible contribution of unidentified leachable constituents or mixture effects. The novelty of this study lies in demonstrating that technical functionality, waste diversion, and low concentrations of selected regulated contaminants do not alone ensure the environmental safety of waste-derived polymer composites. Integrating stress-condition leaching with sensitive biological endpoints offers a practical framework for safe-by-design formulation, application selection, and end-of-life management. Further studies should address material ageing, repeated leaching, particle release, and non-target identification of bioactive constituents.

Abstract: County-level hurricane power outage risk prediction faces a mismatch in spatial support among hazard rasters, infrastructure exposure, and outage labels. Conventional county-mean aggregation can weaken the spatial association between local hazard peaks and transmission corridors. To address this issue, we propose the Raster-to-Grid Spatiotemporal GeoAI Network (R2G-STGeoNet). It first combines Sentinel-1 SAR, precipitation, wind fields, and geographic environmental information to encode multi-source hazard states. Raster-to-Grid Exposure Projection (R2GEP) then transforms continuous hazard fields into features including county exposure, along-line exposure, local peaks, and lengths of above-threshold exposure. A dual-relation spatiotemporal graph combining geographic adjacency and transmission-corridor relationships jointly predicts outage occurrence probability and outage rate. Using hurricanes as case studies, we evaluate matched-input baselines, structural ablations, cross-event transfer, and spatial holdout, accounting for missing labels and data latency. The full model achieves an AUROC of 0.913; replacing R2GEP with county means lowers AUROC to 0.867. With identical R2G inputs, the proposed model outperforms R2G-ST-GNN (AUROC = 0.895). The results indicate that spatial support transformation and infrastructure relation modeling improve county-level outage risk prediction under extreme weather.

Abstract: The cork oak (Quercus suber) is an ecologically and economically significant species, supporting forest biodiversity, carbon sequestration, and the global cork industry. Climate-driven pathogens, specifically Biscogniauxia mediterranea and Phytophthora cinnamomi, increasingly threaten cork oak forests with charcoal and bleeding cankers, respectively, leading to rapid forest decline. Early, reliable identification of these pathogens in the field is critical for effective management, but distinguishing visually similar diseases remains challenging for non-specialist personnel. This study investigates whether a small set of representative bark images can provide an interpretable, field-oriented approach to cork oak disease classification. Using a dataset of 1,198 bark images including healthy, B. mediterranea-infected, and P. cinnamomi-infected samples, we extracted features with MobileNet-v3 Small and applied K-Means clustering to identify pseudo-centroids, or real, representative images closest to each cluster’s center. These pseudo-centroids serve as interpretable reference points that field workers can visually compare with new samples, rather than relying solely on "black box" model outputs. A k-Nearest Neighbor (KNN) classifier evaluated against these pseudo-centroids achieved 74.90% accuracy across the three classes. To assess resilience to human labeling errors, a common source of error in field-collected datasets, we simulated noise by randomly swapping pseudo-centroid labels, which resulted in only a 5.44% decline in accuracy, demonstrating that the framework degrades gracefully rather than catastrophically under noisy conditions. This robustness, combined with the interpretability of the pseudo-centroid approach, makes the framework practical for forest rangers and conservation workers to prune or treat infected limbs based on direct visual similarity ranking, without requiring deep machine learning or biological expertise. Given the environmental and economic importance of cork production, this approach offers a scalable, field-ready tool for monitoring the health of cork oak forests. Future work will expand the framework to larger, more diverse datasets and additional pathogens, broadening its applicability across cork-producing regions.

Abstract: Soybean biodiesel plays a pivotal role in local bioeconomy strategies and global transport decarbonization. However, carbon footprint assessments often rely on national default values or exclude soil organic carbon (SOC) dynamics. This study provides a comprehensive cradle-to-wheel Life Cycle Assessment (LCA), unifying a six-crop-season agricultural inventory (2019/20–2024/25) in Córdoba province with primary industrial data from four local biodiesel processing plants. At the farm gate, the weighted average carbon footprint of soybean was 135.1 kg CO₂eq t⁻¹ excluding soil dynamics and 331.4 kg CO₂eq t⁻¹ (+145.3%) when incorporating Tier 2 SOC changes. For 1 MJ of soybean biodiesel burned in engines, total life cycle GHG emissions ranged from 14.36 to 21.16 g CO₂eq MJ⁻¹ for local combustion in Córdoba and from 18.02 to 24.82 g CO₂eq MJ⁻¹ when exported to Europe. These figures represent a 74% to 81% GHG emission reduction compared to the European Union fossil fuel benchmark (94 g CO₂eq MJ⁻¹). Agricultural production and transport accounted for 23%–42% of total emissions, oil extraction for 7%–11%, and biodiesel industrialization for 47%–70%. Results demonstrate that regional, small-scale biodiesel production in Argentina achieves environmental performances highly competitive with large agro-export facilities, fulfilling strict international RED II/III sustainability criteria.

Abstract: An integrated signal-processing, performance-characterization, and atmospheric-measurement framework is presented for an all-fiber coherent Doppler lidar operating near 1.5 µm for wind sensing and atmospheric boundary-layer observations. The lidar uses pulsed heterodyne detection at a pulse repetition rate of 20 kHz, with atmospheric return signals sampled at 400 MS/s and preprocessed in real time using a field-programmable gate array (FPGA). Two embedded processing architectures are consolidated: direct range-gated fast Fourier transform (FFT) processing with accumulated power spectra, and digital in-phase/quadrature demodulation followed by down-sampling and autocorrelation accumulation. The autocorrelation approach preserves flexibility for subsequent range-dependent processing. Receiver performance is characterized by normalizing measured spectra to a reference local-oscillator shot-noise spectrum to compensate for the non-flat receiver response. For 10,000-pulse accumulation, the measured normalized noise fluctuation is approximately 0.015–0.0155, compared with an idealized shot-noise value near 0.014, and a normalized detection threshold of approximately 0.024–0.025 is used. Range-dependent coherent return is interpreted through heterodyne efficiency and loss of spatial coherence caused by distributed aerosol scattering and atmospheric refractive-index fluctuations. Range-corrected coherent-lidar measurements are compared with direct-detection lidar observations, while coordinated microwave-radiometer measurements provide thermodynamic context. The combined results show that FPGA processing, statistical receiver characterization, range-dependent coherence analysis, and multi-sensor observations form a unified framework for quantitative coherent Doppler lidar measurements of urban boundary-layer wind and turbulence.

Abstract: The rapid expansion and structural complexity of applied clay science create a pressing need for process-oriented educational frameworks to replace traditional descriptive instruction. To address this challenge, this technical note introduces an illustrative, process-based approach supported by generative artificial intelligence (AI) tools to clarify clay mineral modification pathways. Utilizing AI-enhanced prompt workflows integrated with literature synthesis, we developed a systematic "toolbox" framework that visually categorizes key physical and chemical pathways, including acid/alkali etching, thermal activation, mechanochemical milling, organo-functionalisation, and nanoparticle decoration; They link directly to property outcomes and target applications. Furthermore, we map the innovation pathway from laboratory synthesis to commercial production, detailing critical high-attrition failure points along the translation process. This AI-powered visual strategy bridges fundamental clay mineralogy with real-world technological, environmental, and agricultural applications.

Abstract: It may be argued that forest monitoring based only on land-cover conversion can miss functional decline inside areas that remain spectrally forest-like. This study presents a label-free remote-sensing framework for mapping candidate hidden forest degradation from annual satellite data. The method combines physically interpretable Sentinel-2 vegetation, moisture and burn-sensitive indices with Landsat land surface temperature and year-to-year AlphaEarth embedding drift. These signals are normalised and fused into an annual hidden functional degradation score, then separated into two pathways: candidate spectral cover-loss/conversion pixels and candidate hidden degradation within pixels that still satisfy relaxed forest-like spectral conditions. Dynamic World probabilities are excluded from model construction and used only afterwards as a public land-cover transition reference, which avoids direct label leakage while retaining the limitation that both Dynamic World and part of the model input use Sentinel-2 information. It should be noted that, across six annual transitions from 2019 to 2025, AlphaEarth drift and the combined score separated Dynamic World transition pixels from stable-forest reference pixels with AUC values of 0.948–0.999 and 0.939–0.999, respectively. At the top 2% threshold for 2024–2025, F1 reached 0.937 for AlphaEarth drift and 0.916 for the combined score. These validation results apply to the full-domain score layers rather than to direct confirmation of the hidden stable-forest pathway itself. The combined score does not always outperform AlphaEarth drift on F1, but it provides a more physically interpretable candidate early-warning layer for forest-condition assessment and follow-up validation.

Abstract: This study investigates the thermal behavior of three Lufft CHM15k ceilometers at ACTRIS observational platforms and evaluates statistical associations between housekeeping temperatures and attenuated backscatter. The instruments are deployed at RADO-Galati and RADO-Bucharest (Romania) and AGORA Granada (Spain). External, internal, optical-module, and detector temperatures from 14 July 2023 to 14 July 2025 were analyzed using Pearson and sliding-window correlations, while time–range correlations assessed associations with the SNR-screened and smoothed attenuated backscatter coefficient. Long-term analysis showed strong external–internal thermal coupling, with clearer seasonal modulation at the Romanian sites than at AGORA–Granada. Five warm- and five cold-season days from 2024 were analyzed per platform. Across the three sites, 51 contiguous regions with strong correlations (|PCC| ≥ 0.8) and at least 50 connected hour–range cells were identified; 47 occurred during warm-season cases and four during cold-season cases. Significant associations were concentrated mainly at lower measurement ranges and extended farther during warm-season cases. These correlations do not demonstrate direct temperature-induced modification of atmospheric backscatter because common diurnal atmospheric variability may contribute. Systematic housekeeping-parameter monitoring can support quality control, preventive maintenance, and resource-efficient operation of long-term atmospheric remote-sensing networks, strengthening sustainable environmental monitoring and resilient research infrastructure relevant to the Sustainable Development Goals (SDGs).

Abstract: Coastal water levels during storms result from interacting tides, storm surges, waves, currents, bathymetry, and coastal geometry, producing spatially heterogeneous responses that remain difficult to characterize from point observations alone. The Surface Water and Ocean Topography (SWOT) mission provides a new opportunity to investigate this complexity through wide-swath High-Resolution (HR) observations of sea surface height (SSH). Here, SWOT HR water levels are first evaluated against five tide gauges in the English Channel and subsequently analyzed at three contrasting sites, Cherbourg, Villers-sur-Mer, and Étretat, using complementary spatial and scale-resolved approaches. SWOT HR shows strong agreement with tide gauges at the best-performing sites, with RMSE values of approximately 0.11 m and correlations exceeding 0.99, while accuracy remains site dependent. Beyond point-based validation, the two-dimensional HR fields reveal pronounced spatial heterogeneity, including localized and elongated SSH anomalies with distinct geometries and orientations. EOF decomposition shows that more than half of the variance remains beyond the first six modes at all three sites, highlighting the complexity of the HR spatial signal. These residual structures are spatially organized rather than random, while their point-by-point relationship with radar backscatter (σ⁰) remains negligible. Two-dimensional Morlet wavelet analysis identifies a recurrent energetic range of approximately 0.5–1.0 km across the three domains, showing strongly contrasting intensity, localization, geometry, and persistence across neighboring scales. Comparison with independent wave, surface-current, and bathymetric fields further shows that this variability occurs within markedly different hydrodynamic and morphological environments, consistent with the combined influence of storm forcing and local coastal configuration rather than a single controlling process. The complementarity of these analyses demonstrates the added value of SWOT HR for resolving the two-dimensional, multiscale complexity of storm-time coastal SSH, providing new spatial information for understanding coastal dynamics and for high-resolution model validation and data assimilation.

Abstract: Warm season heat may be associated with infant mortality, but temperature definitions and county social context can influence estimates. We linked county mortality, daily maximum temperatures, and American Community Survey estimates in an exploratory ecological study. Complete samples comprised 1,483 overall county years during 2019–2024, 1,721 White county years and 963 Black county years during 2017–2024. Poisson mean models included county and year effects, live birth offsets, and county clustered uncertainty. Per 10 additional May–September days at or above 85°F, simultaneous adjustment for poverty, education, and Black population share yielded rate ratios of 1.0126 (95% CI, 1.0025–1.0227) for overall mortality, 1.0152 (95% CI, 1.0038–1.0267) for White mortality, and 1.0088 (95% CI, 0.9926–1.0252) for Black mortality. The overall 80°F association also persisted (rate ratio, 1.0135; 95% CI, 1.0018–1.0254). Formal Black–White comparisons were inconclusive. More warm days were associated with higher overall and White mortality after measured contextual adjustment. Exploratory cutoff selection and annual exposure aggregation limit interpretation of a minimum harmful temperature.

Abstract: Smart agricultural technologies can improve input efficiency and environmental performance, but their benefits depend on farmers’ knowledge and use. This study assessed knowledge and use of new agricultural technologies among Greek crop farmers, perceptions of sustainable agriculture and environmental impacts, and the usefulness of agricultural education. A cross-sectional survey was conducted from September 2023 to May 2024 among 1,054 crop farmers across all 13 Greek regions using convenience sampling. Pearson’s chi-square tests and Cramer’s V examined associations between technology knowledge and age, education, and agricultural experience. Overall, 56.8% reported knowledge of new technologies, 29.7% reported technologically equipped machinery, and 17.9% reported using such technologies. Knowledge was significantly associated with age (χ²(4) = 64.41, V = 0.249), education (χ²(5) = 204.02, V = 0.444), and experience (χ²(9) = 182.51, V = 0.420; all p < 0.001). Furthermore, 70.1% reported no knowledge of sustainable agriculture, while the perceived usefulness of agricultural education was very high (mean = 4.73/5). The findings indicate a gap between reported technological knowledge and actual use and support practical training in smart machinery, digital skills, environmental protection, and agronomic practices. Given the cross-sectional design, convenience sampling, and self-reported measures, the observed associations should not be interpreted causally.

Abstract: Cultural tourism drives rural revitalization in China, yet traditional villages with comparable heritage resources differ sharply in the number of visitors they attract. Focusing on the single spatial variable of openness, this study examines nine comb-shaped Cantonese villages in the Pearl River Delta, classifies their layouts into typical-comb, curved-comb and mesh-comb types, and builds a three-tier openness evaluation system of ten AHP-weighted indicators covering external accessibility, internal connectivity and public-space configuration. External transportation accessibility dominates the weighting (0.653). Openness correlates strongly with annual tourist volume (r = 0.826, p < 0.01) and explains 63.7% of its variation (adjusted R² = 0.637), whereas layout type alone is not significant (p = 0.238). Mesh-comb villages attain the highest openness (0.683), and typical- and curved-comb villages the lowest. An interval correction coefficient [0.64, 1.285] converts point prediction into interval prediction, absorbing non-spatial influences. For Huitong Village, renovation raised openness from 0.576 to 0.675, and the observed 800,000 visitors fell within the predicted interval, a deviation of only 0.6%. These findings establish openness as a quantifiable predictor of cultural-tourism attractiveness and yield a “limited porosity” strategy for suitability evaluation, scheme comparison and management decisions.

Abstract: Burned area mapping (BAM) provides essential spatial information for fire monitoring, emissions accounting, and ecological impact assessment, but reported accuracy cannot be interpreted independently of mapping targets, reference data, validation design, and application context. We systematically reviewed 1339 remote sensing BAM studies published from 1999 to 2026 to examine technological advances and the evidence supporting validation, transferability, and downstream use. Existing studies address multiple targets, including burned extent, total burned area, fire perimeter/event, burn date, and fire progression, each of which requires distinct evaluation criteria. High-spatial-resolution data, dense time series, multi-source fusion, and advanced learning methods have expanded mapping capabilities, yet small-fire omission, mixed pixels, cloud and smoke interference, confusion with non-fire agricultural disturbances, reference-data limitations, and long-term nonstationarity remain important sources of error. Most studies emphasize local validation, whereas explicit external-transfer tests are comparatively scarce, and random pixel or patch splits do not demonstrate transfer across fire events, regions, years, or sensors. Evidence that BAM errors alter downstream outcomes is even more limited. BAM reliability should therefore be interpreted in relation to the mapping target, tested conditions, transfer evidence, and task-specific fitness for use rather than a single accuracy metric.

Abstract: How far can reservoir operation displace a river's calendar? We address the question with 92 years of monthly discharge on the Naryn, the transboundary artery of the Syr Darya basin, together with 46 years of release data and two years of daily reservoir balances. Annual flow shows no trend, yet the centre of timing of flow moved from late June to mid January, a displacement of 165 days (bootstrap 95% interval 158 to 173) that inverts the hydrograph and far exceeds the week-scale timing shifts reported for climate-driven change. A phase and concentration decomposition separates this rotation from an earlier flattening; change points near 1986, 1990 and 1994, confirmed by independent segmentation methods, coincide with the documented transition from irrigation-oriented to energy-oriented operation. Roughly 4.8 km³ per year (4.0 to 5.6) now arrives outside the growing season. The evidence identifies reservoir operation as the dominant driver of the inversion and quantifies the seasonal redistribution at the centre of a thirty year allocation dispute.