Recent advances in AI-based weather modeling have given rise to large-scale “weather foundation” models pretrained on large amounts of weather data. In this study, we investigate whether the Prithvi weather foundation model pretraining on 40 years of MERRA2 data, characterized by a resolution of 0.5° × 0.625° facilitates adaptation to ERA5, which has a finer resolution of 0.25° × 0.25°. We fine‑tune the Prithvi model on three years of ERA5 data, and investigate if there is any advantage of MERRA pretraining We compare three architectures during fine‑tuning: Encoder–Decoder (ED): the original Prithvi backbone. Low‑rank Attention Side‑Tuning (LAST): augmenting the ED with low‑rank adapters in the attention layers. Convolutional‑Only (CONV): replacing the decoder with additional convolutional modules and adopting an encoder‑only transformer. Our experiments reveal: Resolution gap: The twofold increase in spatial resolution and the different variable set in ERA5 introduce significant domain shift, leading to longer training times even for the model pretrained on MERRA2. One‑month vs. three‑year fine‑tuning: Short‑term fine‑tuning accelerates initial adaptation but fails to close the convergence gap introduced by resolution and variable differences. Three‑year training still has considerable computational training cost. Architecture ablations: Among them, the CONV model achieves the fastest convergence and highest predictive skill. We quantify the ablation effects to isolate the benefit of MERRA2 pretraining versus architecture choice. Our results suggest that, while MERRA2 pretraining provides a useful initialization, substantial fine‑tuning (in terms of both data volume and compute) remains necessary when transferring to higher‑resolution datasets.

Drought is intensifying at an accelerating rate across India, posing serious challenges to human populations and socio-economic development. This underscores the critical need for effective drought monitoring systems. Various indices have been developed to assess drought conditions. In this study, we evaluate the application of several such indices; the Standardized Precipitation Index (SPI), the Standardized Precipitation Evapotranspiration Index (SPEI), and the satellite-derived Vegetation Condition Index (VCI), to analyze historical drought patterns in the Bundelkhand region of Uttar Pradesh, India. We used meteorological data from the India Meteorological Department (IMD) and remote sensing NDVI data from PKU GIMMS to examine the relationship between VCI and commonly used meteorological drought indices. The temporal variations in SPI and SPEI at 1 month, 3 month, 6 month, and 9 month timescales were compared against monthly VCI data to identify drought periods in the region. To assess drought severity, frequency, and duration, we applied Run Theory. Our analysis revealed that the correlation between VCI and the meteorological indices was weakest at the 1-month timescale, suggesting that short-term drought conditions are less effectively captured by VCI. The results showed that all districts in the region experienced similar drought events, particularly during common drought years, such as 2002, 2007, and 2015, over a 32-year period from 1990 to 2021. While conventional methods continue to perform reliably, integrating them with remote sensing-based indices provides a more robust and comprehensive framework for drought assessment.

Water and solute transport in the vadose and saturated zone of crystalline hard rock formations are driven by complex flow phenomena, and modeling these processes at the hillslope scale is challenging. Additionally, sub-seasonal dynamics in climatic parameters can affect the hydraulic properties (hydraulic conductivity and specific yield/ porosity) of the vadose zone media. This necessitates characterizing the permeability architecture while modeling these aquifers’ unsaturated and saturated flow. The present study attempts to understand the permeability architecture of a crystalline hard rock aquifer system with agriculture as the dominant land use in India’s Eastern Ghat mountain range at the hillslope scale. The hydrogeology was investigated using 70 surface geophysical tests, 20 borehole loggings, groundwater level monitoring, and delineation of rock outcrops in a 27.38 km2 study area. The investigation improved the understanding of stratigraphy, weathered thickness, and particle size distribution in the underlying hard rock formations. Groundwater head simulations indicate that water table fluctuations in the stratiform fractured layer result in a significant fall in the water table, in contrast to fluctuations in the saprolite layer. The permeability architecture derived from this description of aquifer response distinctly illustrates groundwater reserves and their susceptibility to exploitation. The framework of this analysis can be applied to examine the implications of climatic effects on weathered hard rock aquifers with similar hydrogeological settings.

Aerosol-cloud interactions exert a strong but highly uncertain negative radiative forcing, masking current greenhouse gas-driven warming and obscuring our ability to estimate future warming as well. ”Natural experiments” like ship tracks have been identified as a promising means of constraining aerosol effects on clouds; however, questions remain about the extent to which cloud adjustments observed in ship tracks generalize to other pollution effects. At the same time, inadvertent aerosol perturbations like ship tracks have inspired proposals to offset global warming impacts with deliberate aerosol cooling [e.g., marine cloud brightening (MCB)]. Ship track aerosol are a mixture of carbonaceous material and sulfate, but changes in fuel composition due to international regulations have shifted aerosol composition from larger sulfate-coated to smaller soot-dominated particles. There is currently a debate in the literature about whether reduced cloudiness observed in some ship tracks is due to cloud adjustments to aerosol-driven microphysical perturbations or to aerosol absorption, which is negligible for sea salt (as would be used for MCB). Through satellite observations both before and after these regulations went into effect and large eddy simulation modeling of cloud adjustments under sulfur-dominated, soot-dominated, and salt ”tracks”, we test the hypothesis that ship tracks observed in the historical record are good analogues for what salt tracks would behave like under MCB. If there are large differences between sulfur and salt tracks in particular, then current knowledge based on satellite imagery and field measurements of ship tracks would need to be interpreted cautiously for the case of MCB and may indeed underestimate efficacy, with implications for the merit of small-scale outdoor perturbative studies focusing on salt tracks in particular.

Serpentinization occurs due the hydration of ultramafic rocks in various tectonic settings. This hydration has been proposed as a source of geological hydrogen that could be engineered as a carbon-free energy carrier. However, the details of the hydrogen generation are complex and remain poorly understood. Water availability often limits reaction progress as the fluid-rock interactions modify permeability and generate grain-scale stresses due to the volumetric increase associated with the reaction. Most laboratory studies to date rely on powdered samples instead of solid rock cores and hence do not capture the full dynamics of fluid-rock interactions in natural settings. This study investigates how water availability and temperature influence serpentinization in intact dunite cores. Stress relaxation experiments were conducted in a Griggs-type solid medium deformation apparatus at 1.3 GPa confining pressure and 300–400 °C temperature for durations of 0 to 11 days with water contents ranging from 2 to 4 wt% relative to rock mass. The rocks were first loaded to ~200 MPa differential stress and left to creep. Microstructural analysis of post-experiment thin sections confirmed significantly increased serpentine content compared to the starting material and intricate fracture networks likely caused by both the volumetric increase during serpentinization as well as creep deformation. Continuous ultrasonic monitoring tracked in-situ P-wave velocity evolution (Vp). These results connect observed seismic velocity changes to fluid–rock reactions in the deep lithosphere and provide constraints on serpentinization kinetics in intact ultramafic rocks.

For over 30 years, satellite altimetry has been a vital tool for monitoring ocean and inland water surfaces. Since 1992, at least two missions have operated concurrently, beginning with TOPEX/Poseidon and ERS-1, and now including around 10 active satellites such as ICESat-2, Sentinel-6A, and SWOT. While most mission data are freely available, they differ in format, processing level, and reference systems (e.g., ellipsoid or time), complicating multi-mission applications. Additionally, generating high-quality, ready-to-use scientific products often requires specialized expertise.   To address these challenges, DGFI-TUM has developed the Open Altimeter Database (OpenADB), which enables consistent data management and integration. OpenADB comprises the internal Multi-Version Altimetry (MVA) repository and a user-friendly web portal. It offers access to along-track products such as sea surface heights and ocean tides, along with detailed information on satellite missions, configurations, and data characteristics. All products are freely accessible via https://openadb.dgfi.tum.de after registration.   A key focus of OpenADB is the coastal zone. Over two decades of data have been reprocessed using algorithms and geophysical corrections tailored to enhance sea level retrievals near coasts without compromising open-ocean standards. The coastal dataset, based on ALES and ALES+SAR retrackers, is the largest of its kind. It also underpins the EOT20 tidal model, which delivers superior performance in shelf and coastal regions, which are typically challenging for tidal estimation. The EOT20 model is also available as an independent product. Furthermore, all missions are cross-calibrated, allowing users to seamlessly combine data from different satellites for regional-scale sea level analysis.

Ecosystem recovery is intricately linked to human wellbeing. Human wellbeing in this context is an interdisciplinary perspective on what allows humans to thrive in relation to the natural environment and entails values, physical and psychological health, governance, and social, cultural, and economic wellbeing. In the Puget Sound region in Washington state, ecosystem health and recovery include both biophysical and human wellbeing goals, grounded in a socio-ecological systems model. Over the past 20 years, the Puget Sound has developed an existing set of “Human Wellbeing Vital Signs”, which measure progress towards recovery goals through subjective and objective indicators of wellbeing. The existing Human Wellbeing Vital Signs have informed the integration of health and wellbeing, Tribal and cultural knowledge, and environmental justice into ecosystem recovery projects. Despite investments and advances towards integrating human wellbeing into ecosystem restoration in the Puget Sound, there remain crucial gaps and opportunities. These include unequal expertise, funding, and personnel capacity compared to the natural sciences; gaps between social science research, management decisions, and recovery program implementation; challenges to accurately monitor and report on human wellbeing outcomes; and limited involvement of overburdened and vulnerable communities in planning, monitoring, and implementation. To address this last challenge, the Puget Sound Partnership has engaged with various community-based organizations that represent various overburdened and vulnerable communities across the Puget Sound region – including youth, Black communities, Asian American and Pacific Island communities, and immigrant communities – using community-based participatory methods to better understand (1) resonance of existing Human Wellbeing Vital Signs, (2) additional community dimensions of human wellbeing not reflected by the existing Human Wellbeing Vital Signs, and (3) opportunities to revise the Human Wellbeing Vital Signs to improve its resonance with overburdened and vulnerable communities. This presentation will share the results of this community-based participatory research project and its applications for natural resource management decision-making processes.

Mountainous areas in South Asia are generally prone to natural hazards like floods, landslides, and earthquakes; yet conventional geospatial hazard analysis does not always make the best use of the rich cultural memory embedded in traditional lore. This study proposes an interdisciplinary approach that synthesizes geomythology—consisting of myths, legends, and oral tradition that contain environmental information—into geospatial decision support systems (GDSS) for hazard mitigation. Through the analysis of ancient texts (e.g., the Nilamata Purana), regional ballads, and folklore, we georeferenced and plotted narratives of major hazard-prone events, e.g., the Kedarnath Floods (2013), the Kangra Earthquake (1905), and the Great Assam Earthquake (1897).By superimposing oral tradition on geospatial information—such as fault lines, landslide data sets, and river catchment basins—we generated GIS layers that reveal significant correlations between physical hazards and cultural topics. Sample case studies suggest that temples, sacred landscapes, and mythic warnings (such as Dhari Devi’s wrath or extinguishing Jwala Mukhi’s flame) often coincide with geologically active zones. These markers are de facto ”seismic sentinels” in local folklore, retaining knowledge about hazards over generations. Our methodological integration provides a concept test for the incorporation of traditional ecological knowledge (TEK) into scientific models. The combined framework increases public engagement, facilitates culturally appropriate disaster communication, and broadens the temporal horizons of hazard perception—connecting deep time and proximal risk. This study contributes to emerging work in disaster anthropology and resilience research by demonstrating how myths function as cultural expressions as well as historically grounded sources of empirical insight into environmental risk.

The mosquito-borne disease dengue is dependent on climate, in part because of the climate sensitivities of its Aedes mosquito vectors. These climate–dengue associations are leveraged in dengue risk assessment models, which incorporate statistical relationships between meteorological variables (e.g., rainfall, air temperature) and dengue cases. However, the mechanistic pathways that result in different statistical relationships in different locations remains understudied. A mechanistic understanding of how interannual variability in climate translates to interannual variability in dengue incidence may aid in predicting dengue dynamics across climatically distinct locations and beyond dengue’s well-known seasonal cycles. In this work we investigate interannual variability within a key pathway in the climate-disease process chain: climate impacts on the Aedes vectors’ breeding habitats. We have implemented a process-based modeling pipeline that simulates climatic influence on habitat water dynamics (using the energy balance container model WHATCH’EM) and thereby on the relative population size of the vector (using models of environment-dependent vector survival and development rates for eggs, larvae, and pupae). We use this modeling pipeline, driven by meteorological data from 2001 to 2020, to simulate monthly relative population sizes of the vector for three climatically distinct locations in Sri Lanka (Negombo, Jaffna, Nuwara Eliya). We find that modeled vector population sizes are plausibly associated with climate conditions (e.g., in the cold, mountainous city of Nuwara Eliya, years of high 2-m air temperature tend to yield high populations) and that the climate drivers of population vary among locations. Moreover, tercile-tercile comparisons of dengue against model variables support the idea that dengue’s connection to temperature is strongly affected by the effect of temperature on vector population size. These results justify future work on rigorously assessing this modeling pipeline’s utility within a dengue risk assessment framework, where its process-based structure may be leveraged to inform intervention strategies in high-risk years and simulate impacts of future climate conditions.

We present results from deploying machine learning models into production for detecting methane leaks in data from imaging spectroscopy satellites. We use and publicly release one of the largest global and most diverse datasets of methane plumes from three different imaging spectroscopy missions. We compute three different variants of methane enhancement products and quantitatively compare their suitability for using with deep learning models (showing an improvement of more than 47% in the F1 score). Using the data from different sensors, we demonstrate generalisation capabilities and transfer knowledge learned on the largest dataset to the other smaller sets. With attention to the capabilities needed for model deployment, we extend prior evaluation methodologies from small tiled datasets to full granule evaluation. Using these, we identify that most typically used methods create a large number of false alerts. We propose model ensembles, which reduce the number of false alarms by over 57%, while further improving the F1 score by additional 3.2% over the non ensembled variant. Finally, we deploy the proposed models in a real methane leak detection pipeline used at the United Nations Environment Programme’s Methane Alert and Response System. Detections of our model are used as proposals for a team of analysts, which effectively speeds up the detection process and reduces the time needed to make these detections on each capture. In 9 months our system helped detect over 1465 methane leak events in data from EMIT, PRISMA and EnMAP. The created datasets are part of a living system and ready to be used for model adaptation and improvement over time. Our work is an important step towards a global AI assisted methane leak detection system, which will be necessary in the light of a large number of new imaging spectroscopy missions being launched by NASA, ESA and private actors. Relying solely on manual data detection is no longer feasible.

Dams are crucial structures for efficient water resource management. Previous studies have presented the interlinkages between seasonal dam deformation and hydrometeorological factors. However, the spatiotemporal changes in this relationship have not been widely quantified. We combined InSAR-based dam deformation, rainfall (RF), reservoir water levels (RWL) time series, and quantified Colwell’s seasonality across six dams in California. We also applied continuous wavelet (CWT) and cross-wavelet (XWT) transform to examine seasonal periods and coupling mechanisms between dam deformation and hydrometeorological variables. We find that the studied dams undergo seasonal deformations that are closely linked to RF and RWL. The Colwell indices (range 0-1), Predictability (P: 0.16-0.42), and Contingency (M: 0.15-0.42) both fall within the significance threshold (p<0.05), indicating strong predictability and seasonal dam deformation patterns. In contrast, the index Constancy (C: 0-0.08) shows no statistical significance across any of the dams, confirming these structures do not exhibit consistent deformation patterns over time but rather respond dynamically to seasonal variations. The Pardee (PAR) Gravity Dam (38.25N, -120.85W) has the highest index values for P and M, as it also indicates consistent annual deformation patterns over time. The Melones (NML) Embankment Dam (37.95N, -120.52W) has the lowest index values for P and M, as its yearly pattern varied during 2016-23. The seasonal period of deformation is around 1 year in CWT spectra for all the dams except the NML dam. It has CWT significant spectra dominated within a 6-month to 1-year range. XWT indicates that RF and RWL exhibit different interlinking with dam deformation for each dam. The dam deformation exhibits a lag response time of up to ¼ of a year, characterized by a 90° phase relation with the RWL as it rises and falls. RF has an anti-phase relation and no direct influence on the dam deformation. There is a strong possibility of having an influence of other factors like atmospheric loading and geological influence. This highlights the need for further analysis to examine the influence of these factors on seasonal dam deformation. Our findings enable improved dam safety monitoring protocols and support climate-resilient infrastructure management.

Fusion power has the potential to provide energy with less environmental impact than other energy technologies. However, inventories of early-stage designs indicate wide variability in material inputs. These differences may result in significant scale-up constraints and impacts, especially since many material inputs are critical to other energy technologies. We use a consequential life-cycle assessment (LCA) framework to evaluate environmental and social impacts of fusion as it scales up. We compare these impacts across different fusion system designs and energy technologies, including nuclear fission, solar, and natural gas power generation. We follow three steps: Quantify material inputs and outputs. We collect inventories of early-stage fusion designs and estimate material requirements for a 1 GW capacity plant. Model the scale-up of fusion power generation. We use growth rates of historical analogs to fusion, such as nuclear fission, to project near-term capacity milestones. We model multiple scenarios to characterize the feasibility space of fusion scale-up. Select impact categories. We synthesize literature and expert interviews to select impact categories such as global warming potential, water usage, toxicity of material inputs, supply chain stability, and worker hours. We combine the outputs of these steps to compare impacts of fusion systems at scale with other energy technologies. Preliminary results for a hypothetical ITER-style plant suggest low global warming potential compared to other energy technologies but higher water usage from mining and refining of critical minerals like lithium. This LCA will highlight other tradeoffs such as the social impacts of developing the critical materials supply chains. This LCA is a crucial step for macro-energy systems (MES) models that incorporate fusion power. It uses granular data about fusion designs and technology scale-up to improve its characterization in integrated assessment and other systems models. It also highlights dynamics of fusion scale-up, leveraging MES modeling to explore the role of fusion in decarbonization pathways and tradeoffs across competing technologies. By evaluating the impacts of fusion at an early stage, innovators can address scale-up constraints before they become locked in to the technology design.

Solar energetic particles (SEPs) are high-energy particles consisting of protons, electrons and heavy ions emitted by the Sun during solar flares and coronal mass ejections. Energetic particles emitted from the Sun range in energy from suprathermal (several keV) up to relativistic (several GeV), and can pose significant radiation risks to spacecraft, satellites, and human astronauts. SEP events occur relatively infrequently, and their time series profiles can often exhibit complex variability. Most SEP catalogs are still curated manually by experts, but these are frequently affected by inconsistencies and varying data quality or classification standards. As a result, SEP detection and classification remain time-consuming and prone to inconsistency. While the recent rise of machine learning in the space weather field has been primarily focused on SEP forecasting, machine learning also offers a promising tool for automating and standardizing SEP detection and classification. In this study, we present a first step toward such a system by developing a binary classification convolutional neural network (CNN) trained on almost 20 years of proton time series data from seven energy channels (13.6–33.4 MeV) recorded by the High Energy Telescope (HET) onboard STEREO-A between December 2006 and May 2025. Using five-fold cross-validation, the model correctly classifies about 99% of non-SEP samples and 94% of SEP samples. Our results provide a foundation for further application of machine learning in SEP detection and classification, contribute to a more quantifiable and consistent definition of SEP events, and, once extended to additional instruments, could help pave the way toward a fully automated and standardized SEP event catalog.

Airborne maritime lidars are widely used for measuring coastal bathymetry. However, these lidars require manned aircraft for operation and are expensive to build and field. Additionally, Alaska’s remoteness, inclement weather, and limited personnel capacity provide unique challenges to data gathering with traditional maritime lidar not seen in the contiguous United States. Droneborne maritime lidar is a promising solution to address these challenges and enhance the data gathering capability in the Arctic.Here, we report on the ongoing development of a droneborne maritime lidar at the University of Alaska-Fairbanks (UAF) Geophysical Institute. This lidar is designed to fly on a fixed-wing unmanned aerial vehicle (UAV) owned by the Alaska Center for Unmanned Aerial System Integration. This UAV has a 25 m/s cruise speed and a flight endurance of 900 km. The current status of the design will be reported, as well as modeling results showing the signal-to-noise ratio and expected depth performance. We also determine the optimal spot diameter for bathymetric profiling in Alaskan waters where this system will be fielded.

A document by Cecilia Johansson. Click on the document to view its contents.

Understanding sea ice evolution in Antarctica is critical for deciphering climate dynamics in polar regions and icy planets over multiple spatio-temporal scales. Satellite-derived ice condition models, primarily ice concentration, are frequently assimilated into coupled sea ice-ocean modelling systems. Accurate ice concentration measurements are crucial to the accuracy of climate models, as the percentage of ice coverage and form of ice greatly affect ocean surface albedo and heat exchange. However, the gap between the ice condition model and dynamics of the sea ice landscape in-situ is known to be significant. We present our in-situ ice observation data gathered during the NBP2501 expedition that provides high temporal and spatial resolution categorization of ice thickness, concentration, form and stage of development through observational analysis of the transition from open ocean to close pack ice in the Ross Sea, Antarctica from Feb. 18 to Apr. 6 2025. We document relative ice concentration based on analyzing the complexity in the variability of ice thicknesses, forms, and stages of development. These ice categorizations are used to train a machine learning algorithm to assess the ice condition, and compared to existing satellite imagery, and microwave satellite-based ice concentration. This analysis is used to establish an unprejudicial ice condition index that considers relative thickness, form, concentration and stage of development. In collaboration with ice forecasting company Drift+Noise, these ice condition index data could augment the ARTIST Sea Ice Concentration algorithm developed by Kaleschke et al (2001) and later refined in Spreen et al (2008), with the goal of providing a more comprehensive ice concentration that includes ice form, thickness, and stage of development. We also demonstrate the application of this data to inform operational decision making for scientific expeditions. Each operation conducted on the NBP2501 cruise is analyzed in the context of this index with a description of the operational challenges and adaptive strategies. Our high-resolution categorization of ice conditions aims to lessen the temporal and spatial gap between satellite-derived models and ground truth reality and provide a more robust framework for scientific operational decision making.

Spatial terrain parameters (TPs), such as slope, aspect, curvature, and catchment area, are crucial for understanding natural phenomena, including runoff generation, erosion, soil moisture distribution, and wildfire propagation. These TPs are derived from Digital Elevation Models (DEMs); however, computing them at high resolutions over large geographic regions remains computationally intensive and memory-constrained. GEOtiled-SG addresses these challenges with a scalable, open-source framework that partitions DEMs into spatially buffered tiles, enabling efficient parallel computation of terrain parameters. GEOtiled-SG builds on both the GDAL and SAGA GIS libraries, extending their capabilities to scale over large regions and offering users tools to compute more than fifteen parameters, including wetness indices and hydrological flow metrics. Its modular structure includes: optimized concurrent cropping of buffered tiles into smaller, manageable units for parallel processing; efficient mosaicking that reassembles the cropped tiles into a seamless surface while removing buffers; and unified concurrency across cropping, computation, and merging to maximize concurrent computing and minimize process overhead. We benchmark GEOtiled-SG on 10 m and 30 m DEMs covering Tennessee and the Continental U.S., leveraging up to 64 vCPUs. With these optimizations, GEOtiled-SG not only achieves up to 59× speedup in cropping and 18× in mosaicking (10× end-to-end runtime reduction) but also provides a robust and reproducible platform for high-resolution terrain analysis in hydrology, agriculture, and wildfire modeling. All software and data are openly available at https://github.com/TauferLab/GEOtiled.

Land subsidence in the greater Houston region was first identified at Goose Creek Oil Field in Baytown, Texas, with three feet of measured subsidence over eight years beginning in 1918. With increased industrialization and population growth, the region relied on groundwater to support and sustain development. This excessive groundwater use caused aquifer compaction which led to land subsidence measured at over nine feet in some areas. Some subsidence impacts were increased flooding and infrastructure damage that ultimately led to the creation of the Harris-Galveston Subsidence District (HGSD) by the Texas Legislature in 1975. HGSD’s mission is to regulate groundwater withdrawal to prevent further subsidence.