Strain localization and partitioning play a critical role in the formation of shear zones and broader tectonic processes. However, our understanding of strain localization is still restricted. Various factors contribute to how strain is distributed in a shear zone, such as differences in lithology or mineralogy, grain size reduction, the presence of weak minerals, and the formation of crystallographic preferred orientations. In this study, we examine how mineral composition and grain size reduction influence strain localization within the Santa Rosa Mylonite Zone. The Santa Rosa and San Jacinto Mountains are the youngest and highest peaks within the Peninsular Ranges, which stretch southward into Baja California. Within this region lies the Santa Rosa Mylonite Zone—the focus of our study due to its intricate geologic history and its mechanical link with the San Andreas Fault system. The zone features a range of rock types, including protoliths such as granodiorite and tonalite, as well as various deformed rocks like mylonites, ultramylonites, marbles, limestones, and other tectonites. The close spatial association of these diverse rocks offers a valuable opportunity to explore how strain is partitioned and localized, and how mineral fabrics evolved from protoliths to their deformed counterparts. This study utilizes a combination of optical microscopy, scanning electron microscopy, electron backscatter diffraction, and X-ray diffraction to investigate microstructural and compositional changes in the rocks, with a particular emphasis on the deformation of plagioclase, biotite, amphibole, and quartz. The results will provide a deeper understanding of how strain is distributed within well-developed shear zones and offer broader insights into the processes driving lithospheric deformation.

Foundation models pretrained on Earth observation archives offer broad utility across geoscience tasks. Building on our DINO-MAE model, named GAIA, trained on globally harmonized longwave infrared (IR, ~11 μm) imagery from geostationary satellites (2000–present), we evaluate its transferability to key downstream applications: gap filling, precipitation estimation, and segmentation of tropical cyclones (TCs) and atmospheric rivers (ARs).

Hydrologic forecasting in large, transboundary basins such as the Laurentian Great Lakes remains a significant challenge due to complex hydrological processes, extensive surface water coverage, and widespread ungauged regions. These factors limit the predictive skill of existing models and constrain effective operational water resource management. This study evaluates and compares two modeling approaches for seasonal runoff prediction: a data-driven Long Short-Term Memory (LSTM) neural network model and the process-based Large Basin Runoff Model (LBRM). Using the Climate Forecast System Reanalysis (CFSR) as climate forcing from 1979 to 2023, both LSTM and LBRM models were developed and tested to simulate runoff across the entire Great Lakes basin. Their seasonal forecasting performance was then assessed using archived 9-month forecasts from the Climate Forecast System (CFS) as input. Additionally, total regional runoff was evaluated from a water balance perspective to assess model effectiveness for seasonal water supply forecasting. This research aims to refine hydrologic modeling frameworks using the Great Lakes as a testbed for seasonal-to-annual water supply forecasting. The results offer insights into the comparative strengths of data-driven and process-based models and highlight their potential integration for improving forecast accuracy. Ultimately, this work supports the development of more reliable forecasting tools for operational water management in the Great Lakes and other large transboundary basins worldwide.

Observations of pyroclast and hydroclast production during subaqueous volcanism at depths >1 km have become common in the literature, despite early predictions that such activity would be pressure-limited. Direct near-vent observations of active submarine eruptions (first at NW Rota-1 and West Mata in the 2000s, e.g., Murch et al., EPSL 2022) revealed multiple volcaniclastic generation modes: in-conduit magmatic fragmentation, external water-induced fragmentation at the vent, and cooling-contraction fragmentation of lava injected molten into the water column. Follow-up studies of stratigraphy and edifice morphology at these and other sites indicate that such processes are widespread and contribute significantly to volcanic structure, hydrology, and ecology. We report here on several substantial primary and secondary pyroclastic deposits from recent eruptions in Tonga’s NE Lau Basin (mainly at West Mata and Tafu volcanoes), as well as undated but likely late Holocene deposits in nearby areas, which display volumetrically significant volcaniclastic output during mafic submarine eruptions as thick vent-proximal and thinner, broad distal deposits with diverse particle morphologies. Explosive fragmentation under these conditions can also produce varied landforms and promote slope instability. Observations from 7 NSF-, NOAA-, and SOI-sponsored cruises (2009–2018), using ROVs and AUVs, reveal both new and older tephra blankets, periodic downslope transport, and distal ripple-structured deposits. Decreasing grain size and thickness generally with distance from the vent location, imply primary pyroclast deposition as well as secondary reworking; in other cases (especially at deep sites), sand-sized and smaller grained-deposits show evidence for transport bedforms (ripples) and post-depositional reworking (sinkholes). At West Mata (2016–17 eruptions; Chadwick et al. 2019) and post-2011 Tafu (Rubin et al., AGU 2020) events, pyroclastic layers up to 1–2 m thick buried lava flows, indicating explosive final phases. Sampling these deposits for grain size/shape characterization is challenging due to coarse fragments near vents and fine particles distally, but can be done using push cores with catchers and scoop bags for basic grain size and shape analysis

Starting in November 2023, the Houthi militia occupying northern Yemen has attacked ships passing through the Bab al-Mandab Strait, a chokepoint on the route between Europe and Asia via the Suez Canal. As a result, cargo ship traffic through the Red Sea has plummeted, with ships instead taking the longer route around the Cape of Good Hope. The increase in traffic in the southeast Atlantic Ocean and decrease in the Red Sea is readily apparent in satellite retrievals of column nitrogen dioxide. Within the stratocumulus deck covering much of the southeast Atlantic, a previously detectible cloud microphysical perturbation due to ship pollution had largely disappeared following the International Maritime Organization’s sulfur-limiting regulations in 2020 but returns during 2024 due to the increase in ship traffic despite the lower cloud brightening efficacy per ship. Because nitrogen dioxide pollution per unit of fuel oil burned is not affected by switching to low-sulfur fuel or installing sulfur scrubbers, quantifying the ratio of shipping-enhanced cloud droplet number and nitrogen dioxide concentrations before and after the fuel sulfur limits went into effect provides a strong constraint on the cloud changes from the regulations. We find that the ~80% reduction in sulfur emissions leads to a 66% (95% confidence: 25-88%) reduction in the increase in cloud droplet number concentration per unit marine fuel oil burned.

Changing magnetic field configurations experienced by Titan as it orbits Saturn can induce currents in the conductive ionosphere of Titan. These induced currents in turn generate ionospheric magnetic fields that deflect the incident plasma and shield out the external magnetic field. By measuring magnetic field perturbations, we can calculate the current densities necessary to create the observed perturbations (with some restrictive assumptions). We determine horizontal currents in the ionosphere of Titan from magnetic field perturbations during Cassini flybys closer than 1500 km, i.e., within the collisional ionosphere. Data from individual flybys with similar trajectories can show significantly different ionospheric current configurations. By grouping the flybys by external magnetic field and plasma conditions and Titan orbital position, we attempt to determine how ionospheric currents change with respect to external conditions. These currents induced in Titan’s ionosphere are but one way Saturn’s magnetospheric environment impacts ionospheric processes and dynamics at Titan.

Acknowledgements: funded by the Ghent University Special Research fund and Scientific Research Fund (FWO; K208225N).

Soil erosion poses a significant threat to agricultural productivity and water resource sustainability, particularly in monsoon-dominated semi-arid basins. This study estimates soil erosion in a sub-basin of the Upper Krishna River (UKR) using the Revised Universal Soil Loss Equation (RUSLE) within the Soil and Water Assessment Tool (SWAT) model. The Indian Meteorological Department (IMD) historical climate data from 1991 to 2021 was used for the calibration and validation of the SWAT model, and the annual average soil loss was found to be 105 tons/ha. Future projections were evaluated using climate data from two General Circulation Models (GCMs), CanESM5 and EC-Earth3, under SSP2 and SSP5 scenarios, to assess changes in annual average soil erosion until mid-century (2049). The results of the study indicate an increased average annual precipitation, which affects the soil erosivity and results in increased soil erosion. Under SSP2-4.5, projected soil erosion is 127.45 tons/ha for CanESM5 and 93.22 tons/ha for EC-Earth3, whereas under SSP5-8.5, the values rise to 208.69 tons/ha and 122.58 tons/ha, respectively. The results of the study highlight the impact of climate change on the soil erosion dynamics and identify the erosion hotspots in the study area to help policymakers and planners for sustainable land management practices.

Monitoring smoke from prescribed fires offers critical insights to stakeholders—from fire practitioners evaluating burn tactics to community leaders seeking to protect sensitive populations.  Prescribed burns often occur in complex, coastal terrain, making it difficult to predict their smoke behavior. Whereas meteorological forecasts lack the resolution necessary to capture the terrain complexities, hyperlocal sensing networks can provide a toolset to understand an environment before, during and after a burn. The key factor limiting these sensing networks is their placement around burn perimeters. This study presents a methodology for optimizing the placement of smoke monitoring sensors around prescribed fire perimeters to maximize data capture. In collaboration with Stanford University's SMesh Project—a student-led initiative focused on deployable smoke sensor networks—we apply failure-case analysis to assess sensor network performance under varying operational scenarios. Since sensors are often deployed at least 24 hours before ignition, small deviations in forecasted wind conditions can drastically reduce monitoring effectiveness.Our approach integrates a terrain-aware Gaussian plume dispersion model with a fuzzing-based technique, introducing randomized perturbations to simulate wind uncertainty and evaluate plume behavior. Gaussian plume models are used as they are lighter-weight computationally than other approaches, particularly when modeling across hundreds or thousands of possible scenarios. To demonstrate our methods, we conducted simulations across three ecologically distinct sites in California—Malibu, Shasta, and Henry Coe State Park—to identify optimal placement strategies and failure modes. These findings support more resilient sensor network designs for prescribed fire operations. Upcoming field deployments during the 2025 Fall prescribed burn season will help validate and refine this approach.

Accurate measurement of posture traits, such as curvature and droop, can provide critical insight into the structural phenotype and stress response of tall grasses, including key perennial bioenergy crops. These posture traits vary across genotypes and growth stages and can influence both productivity and harvestability. This study builds on previous efforts to quantify whole-plant posture, extending the analysis to incorporate deep learning and multi-view UAV imagery. High-resolution oblique, sideways, and nadir images were collected over experimental plots comprising multiple grass genotypes with diverse morphologies. A deep learning framework, incorporating instance segmentation models, was developed to classify plants based on posture metrics extracted from multiple viewing angles. By training on oblique and nadir views together, this work explores the degree to which plant posture can be inferred from each perspective, helping evaluate the contribution of each data acquisition approach and exploring the possibility of combined data acquisitions. Validation against manually annotated datasets evaluates the ability of the approach to differentiate posture traits at high resolution, enabling non-destructive, plant-specific phenotyping at scale. The work supports longitudinal monitoring of canopy structure and facilitates the integration of posture descriptors into breeding programs. It also provides a framework for future fusion with LiDAR and Structure-from-Motion (SfM) data to enhance 3D modeling of genotype-specific canopy architectures in switchgrass or other tall grasses.

The Pacific-Japan (PJ) pattern, a major mode of East Asian summer monsoon variability, features a tri-polar or dipolar structure, centered over southern Japan with opposing anomalies in the western North Pacific (WNP) and Sea of Okhotsk. Since it exhibits both interannual and decadal variability, capturing its phase transitions—especially the decadal shift from tripole to dipole—is crucial for understanding East Asian summer climate. This study evaluated the performance of three large ensemble models—CESM2-LE, MIROC6, and CanESM5—in reproducing the PJ pattern’s spatiotemporal characteristics, using 20CRv3 reanalysis data as a reference, with particular attention to their ability to capture its decadal variability. To robustly assess this, a multi-model and multi-ensemble framework was employed, allowing for comparisons across different models and ensemble members. Initially, performance of the climatology of three key East Asian summer monsoon (EASM) components was assessed. It was found that precipitation was substantially underestimated over the WNP, a key convective region driving meridional teleconnections essential for the PJ pattern development. This bias motivated the use of both precipitation and the zonal-wind–Coriolis parameter product (U*f) to more robustly evaluate the PJ pattern. Based on these variables, the PJ pattern was defined as the leading mode from empirical orthogonal function (EOF) analysis. The analysis showed that climate models tend to simulate excessively strong variance in the first mode of EOF, implying potentially weaker representation of lower modes of EASM and its decadal variability. This feature was further confirmed by power spectrum analysis of PC1, which revealed that most ensemble members were dominated by interannual variability. These results highlight the models’ capacity to simulate the PJ pattern yet limitations in simulating its complex feature and further decadal variability.

The application of scientific visualization techniques focused on identifying locations where backscatter exceeds a threshold revealed unexpected locations of potential or actual venting (as well as the anticipated plumes). The plume dataset using the Cabled Observatory Vent Imaging System (COVIS) attached to the Ocean Observatories Initiative’s Regional Cabled Array in the ASHES vent field in the caldera of Axial Seamount on the Juan de Fuca Ridge. The dataset contains backscatter intensity measurements from July 2020 to September 2023. For each month, plume-imaging data — which record plume shape — were acquired on five different days (weekly). On each of those days, twelve measurements were taken at two-hour intervals, generating twelve grid files. The dataset was processed using MATLAB. First, morphological operations were used to extract the six largest clusters, and the center of mass of each was calculated from the mean of their x, y, and z coordinates. A nearest-neighbor tracking algorithm then matched each cluster with its counterpart in previous time steps. Results showed persistent cluster formation at certain locations. To validate this finding, the matched clusters over time were plotted on the bathymetry map where the original data were collected. In addition to constant cluster formation at the Inferno and Mushroom vents, significant clusters appeared at locations that are not expected to host large plumes. Unexpectedly, we identify recurring clusters 16 meters north and 18 meters west of known Inferno vent orifices (see yellow and green clusters in figure). Future work is required to investigate possible causes of the clusters at these unexpected sites. Nevertheless, this study provides valuable information for the analysis of acoustic data intended to capture the incidence of hydrothermal plumes at the ASHES vent field.

Accurately predicting Active Layer Thickness (ALT) is essential for assessing permafrost stability and mitigating risks to Arctic infrastructure, ecosystems, and communities. Traditional physics-based models struggle with generalization due to sparse in-situ observations and complex environmental drivers. This study introduces a data-driven framework for ALT prediction in the Circum-Arctic combining field measurements, satellite and reanalysis data, and machine learning (ML) techniques. Maximum ALT for a given year is paired with geospatial inputs from earlier years, enabling prediction at four horizons: Same year, +1, +2, and +5 years. A robust multi-stage feature selection strategy is then applied to identify the most informative and temporally stable ALT drivers. To predict ALT, a weighted ensemble of three ML models (Categorical Boosting, Extremely Randomized Trees, and Bootstrap Aggregating) is trained separately for each horizon. The framework is evaluated using data from 115 sites across the Circumpolar Active Layer Monitoring network. Results show that the proposed approach consistently outperforms multiple baseline models across all horizons, with R^2 above 0.8 and RMSE under 25 cm. Future work will be focused on conducting ALT trend analyses to develop a multiclass index of permafrost degradation and regeneration at the Circum-Arctic scale.

For scientists, policymakers, decision makers, local communities and other stakeholders working in Alaska’s rapidly changing permafrost regions, accessible and timely insights derived from geospatial data are critical. This paper introduces a web-based Analysis and decision Support system for Alaska Permafrost (ASAP), designed to predict and visualize various permafrost-related phenomena, including wildfire risks and soil temperatures in Alaska. Utilizing a three-tier Service-Oriented Architecture, ASAP integrates real-time data acquisition, sophisticated high-speed data processing, and multiple machine learning models to deliver accurate and actionable insights. The system’s modular architecture enhances scalability, security, and user experience by seamlessly incorporating geospatial and meteorological data from both external and internal sources. The webapp’s automated processes streamline data selection, providing users with an intuitive interface that offers detailed visualizations and predictions. By bridging data science and environmental monitoring, ASAP offers a powerful tool for proactive decision-making in the face of accelerating Arctic transformation.

The Himalayas are globally recognized as a biodiversity hotspot, harbouring a rich array of species. However, this ecological wealth of the region is under significant threat from anthropogenic activities and climate change. Key challenges to Himalayan biodiversity include deforestation, habitat fragmentation, overgrazing, and unsustainable tourism practices. Conservation efforts in the Himalayas are multifaceted and involve collaboration among local communities, governments, and international organizations. A fundamental step in any conservation initiative is species mapping, which helps prioritize limited conservation resources by identifying areas of high conservation value. The Himalayan landscape presents heterogeneity within species due to variations in slope, elevation, and aspect, further complicating conservation efforts. This study employs a semi-hypertemporal (SH) approach to map Pinus roxburghii, Quercus leucotricophora, and Alnus nepalensis in the Himalayan landscape surrounding the Dudhatoli forest area of Uttarakhand. Sixteen SH Sentinel-2 images between January and December 2022 were utilized. The normalized difference vegetation index (NDVI) was computed for all images to reduce the dimensionality of the dataset, creating an SH NDVI database. This database was then processed using various classification algorithms, including Maximum Likelihood Classification (MLC), Support Vector Machine (SVM), and fuzzy machine learning (ML) modified possibilistic c-means (MPCM) models with both conventional mean and the novel individual sample as mean (ISM) approaches. While ML and SVM are well-established methods known for their robustness and accuracy, they often struggle with the complexity and heterogeneity of mountainous landscapes. Our results indicate that the fuzzy MPCM ISM training approach outperforms ML and SVM classifiers, as well as the conventional mean approach. This is reflected in the overall accuracy values: 51% with MLC, 65.5% with SVM, 66.7% with conventional MPCM, with 82% for MPCM ISM training approach. The ISM approach effectively addresses the heterogeneity within the same class in the landscape. This study highlights the superiority of the MPCM ISM method contributing to more precise species mapping and informed conservation strategies.

Science Gateways offer web based environments that enable distributed research communities to access advanced computational resources, including high performance computing (HPC), specialized workflows, curated datasets, and domain specific analytical tools. These platforms transcend geographic and institutional boundaries, facilitating remote job submission, workload orchestration, and collaborative data analytics and visualization. In this work, we present the Science Gateway integrated components within the A-KBS (Arctic Knowledge Based System), a next generation cyberinfrastructure framework designed to support multi user collaboration, AI enhanced predictive modeling, and seamless interaction with supercomputing resources for global-scale Arctic geospatial and extreme weather data processing. The A-KBS Science Gateway serves as the central interface through which researchers configure, launch, and monitor machine learning driven simulations across heterogeneous computing environments. In this paper, we detail the underlying system architecture, including the deployment of the PARSL semantic layer and Globus Compute services, highlighting their roles in workflow management, data mobility, and scalable execution. This integration empowers Arctic focused research teams to perform real time, data intensive modeling of permafrost, weather variance, and infrastructure resilience. The paper discusses implementation strategies, cross institutional coordination mechanisms, and anticipated impacts on collaborative Arctic research, offering a scalable model for domain specific Science Gateways in other remote and resource constrained environments.

Root zone soil moisture (RZSM) and subsurface fluxes such as capillary rise, root water uptake, and deep percolation are critical for effective irrigation planning and groundwater recharge assessment. However, direct measurement of these variables is spatially limited and costly to implement at scale. This study presents a computationally efficient, reduced-order variational data assimilation (RA-VDA) framework that combines satellite-based land surface temperature (LST) and surface soil moisture (SSM) observations with a physically-based soil moisture–temperature model (HYDRUS-1D). The approach leverages the complementary strengths of LST—sensitive to surface energy fluxes—and SSM—directly linked to near-surface soil water content—to improve RZSM and flux retrievals. Synthetic experiments are conducted under three scenarios: SSM-only, LST-only, and joint assimilation. Results show that SSM constrains moisture in deeper layers, while LST significantly enhances the estimation of water and energy fluxes, including evaporation and infiltration. Joint assimilation yields the best overall performance, reducing root-mean-square error in both state variables and fluxes. Importantly, the framework enables accurate estimation of capillary rise and deep percolation, which are essential for evaluating recharge potential and optimizing irrigation timing. The RA-VDA method uses Proper Orthogonal Decomposition (POD) to reduce the dimensionality of the optimization space, enabling scalable implementation without compromising accuracy. This integration of satellite remote sensing and physical modeling offers a practical, low-cost solution for soil moisture monitoring in data-scarce regions. The framework provides a strong foundation for advancing operational water management tools that support agricultural sustainability and groundwater conservation.

This paper describes a high-performance, distributed AI-based approach for Arctic-focused multivariate geospatial feature prediction, enhancing model accuracy and computational efficiency through parallel processing and distributed dataset preprocessing. The proposed approach was tested using an Alaska wide time series air temperature dataset spanning 25 years. The use of parallelized high performance computing architecture including, but not limited to Kubernetes (K8s) clusters, have the capacity, when properly configured with suitable cyberinfrastructure such as in the authors’ Arctic Knowledge Based System (A-KBS), to significantly accelerate large scale data analysis. The preprocessing and training pipelines were implemented in the JupyterLab environment utilizing Ray.io and TensorFlow. By leveraging the Ray.io library and employing a data chunking technique, the dataset was distributed across multiple worker nodes, enabling parallelization of data preparation tasks and reducing pre-processing time. We investigated eight feature selection techniques across three categories: filter, wrapper, and embedded methods, and compared their computational performance to identify the most relevant features for accurate and efficient air temperature prediction. Four generalized AI models trained using databases from Alaska were selected from linear, nonlinear, time series machine learning, and non time-series deep learning categories to predict air temperature. The impact of the eight different feature selection techniques on model performance was evaluated using MSE, sMAPE, KGE, MPE (%), and prediction time. The results showed that the proposed workflow significantly reduces processing time for most feature selection techniques and prediction models, particularly for those with high pre-processing demands. Notably, Kendall+DT and Pearson+MI methods exhibit better scalability as the number of allocated resources and dataset size increases, proving their effectiveness in handling large size datasets.