A document by Adrija Datta. Click on the document to view its contents.

Coastal communities, economy, and biodiversity are increasingly threatened by escalating environmental challenges. Addressing these necessitates multi-scale data-driven solutions. Digital Twins of the Ocean are emerging as a transformative approach, evolving into regional and local applications that integrate Earth observation (EO), in situ monitoring, numerical models, and Machine Learning (ML) into science-based and user-oriented solutions.  In this context, we present a portfolio of developments targeting key coastal challenges: (1) A ML method for rip current detection from beach imagery in support of emergency services. (2) An automated tool for seagrass wrack identification from coastal cameras, to assist local authorities in balancing tourism and conservation needs. (3) A multi-scale coastline monitoring, exploiting the synergies of fixed stations, crowd-sourced imagery, and satellite-derived shorelines, enabling high-resolution assessment of coastal state. (4) A Marine Heatwaves monitoring and forecasting system that integrates EU Copernicus Marine data for their operational assessment on  Marine Protected Areas. (5) A satellite-derived bathymetry mapping approach, lying on ML models trained on echosounder measurements, in support of coastal assessment and coastal interventions. (6) Benthic habitat mapping products, for sustainable marine planning, using EO, ground truth data, bio-optical models and ML to map benthic vegetation and density.  Through modular and problem-specific pipelines, this research integrates multi-platform data into contextualized, evidence-based intelligence for coastal planning, risk management, and environmental conservation. By ensuring citizen engagement and co-designing with stakeholders, this framework showcases the potential for coastal-to-ocean monitoring and forecasting with a user-oriented focus, prototyping a holistic information system.This research is part of the European Union (EU) funded projects FOCCUS (Grant Agreement 101133911) and iMagine (Grant Agreement 101058625). Views and opinions expressed are those of the authors only and do not necessarily reflect those of the EU or the European Health and Digital Executive Agency (HaDEA). Neither the EU nor the granting authority can be held responsible for them.

The Atmospheric Radiation Measurement (ARM) Data Center (ADC), at Oak Ridge National Laboratory (ORNL), is a central hub of the U.S. Department of Energy’s ARM user facility and supports global climate and atmospheric science by managing one of the most comprehensive and openly accessible archives of atmospheric observations in the world. With over seven petabytes of curated data spanning more than 30 years, the ADC processes and distributes over 50 terabytes of new data each month from fixed, mobile, and aerial observatories operating in climate-critical regions worldwide.As data grows in volume, complexity, and value, the ADC is advancing new AI- and ML-enabled capabilities to transform how users discover, access, and interact with ARM data. These efforts align with broader DOE and ORNL initiatives in building scalable, FAIR-aligned (Findable, Accessible, Interoperable, and Reusable) data ecosystems supporting the next generation of open science.Central to this transformation are AI-powered, user-centric tools that assist researchers through natural language interfaces, simplifying the process of identifying relevant data products across ARM’s complex metadata hierarchies. These tools reduce the entry barrier for new users while enabling personalized discovery pathways for more experienced users. By bridging and translating across diverse metadata structures, these tools enhance data usability and streamline cross-domain discovery. These tools also leverage a scalable vector-based infrastructure that enables contextual understanding, user-aware recommendations, and integration into future AI-ready APIs.The ADC’s broader portfolio includes AI-powered metadata and abstract generation, anomaly and epoch detection models, a Primary Measurement Type classifier, and a data recommendation engine – each advancing automation, knowledge extraction, and scalable ML pipeline readiness. Strategic collaborations with the USGS, the Earth Science Information Partners (ESIP), and ORNL’s AI Program further strengthen these efforts to accelerate AI-driven workflows. Together, these efforts position the ADC at the forefront of intelligent Earth science data platforms and open science innovation, enabling intuitive, efficient and more impactful scientific discovery.

Urban road networks underpin human mobility, economic activity, and timely access to essential services. Yet these lifelines are increasingly susceptible to climate‑driven compound flooding, when fluvial, pluvial and coastal processes interact to create complex urban inundation scenarios. While resilience assessments have begun considering equity, many fail to account for the nuanced impacts of hazard driven infrastructure disruptions on sub-populations' access to critical facilities. Here we examine how compound flooding events reshape access to hospitals, shelter houses, and transportation centers across a representative coastal metropolis of Boston by combining a high-resolution 1D–2D hydrodynamic flood model with dynamic mesoscopic traffic simulation. This approach captures both street-level inundation and evolving traffic conditions, including rerouting and congestion during extreme events. We evaluate changes in travel time to critical facilities under both normal and flood conditions. Results reveal that even localized flooding can lead to widespread travel delays, significantly reducing or even eliminating access to emergency services. Further, these disruptions are spatially unequal, with socially vulnerable neighborhoods experiencing significantly longer delays. Our findings highlight the need for resilience strategies that not only maintain system functionality but also ensure equitable access. This includes distributed siting of critical services, adaptive traffic management, and nature-based flood mitigation to reduce future systemic risks.

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

Land subsidence has been occurring in the Western U.S. since the early 1900s. In southwestern Utah, evidence for land subsidence and earth fissures related to excessive groundwater withdrawal dates back to the early 2000s. However, there has yet to be a survey to identify or monitor land subsidence throughout the state of Utah. In this study, we process InSAR time series across the state, using freely available datasets from ALOS-1 (2006-2011), Sentinel-1 (2015-2025), and ALOS-2 ScanSAR (2015-2025). We identify land subsidence with rates ranging from 0.5 cm/year (Snake Valley) to 4 cm/year (Escalante Desert) over the past 20 years, as well as highlight seasonal signals with amplitudes of up to 2 cm/year (Salt Lake Valley). We demonstrate the feasibility of combining these results with groundwater level trends to estimate aquifer properties and storage change on a regional scale. Finally, we discuss how novel regional datasets, such as those produced by the OPERA North America DISP Product, may be able to inform future groundwater-flow models.  

A document by Lauren Chilutti. Click on the document to view its contents.

A document by Matthew Penn. Click on the document to view its contents.

This abstract was presented as Poster at AGU Fall Meeting 2025 :Abstract ID 1919756 Final Abstract Number and Title: GP33B-0321*: Integrated and Efficient 3D EM Inversion for CSEM and MT    Session Number and Title: GP33B: Frontiers in Electromagnetic (EM) Geophysics II Poster Session Date and Time: Wednesday, 17 December 2025; 14:15 - 17:45 CST Location: New Orleans Ernest N. Morial Convention Center (NOLA CC),  Hall EFG (Poster Hall) 

Green plants play a vital role in mitigating urban air pollution, and selecting pollution-tolerant species is essential for sustainable greenbelt development. This study evaluates 12 dominant roadside plant species from National Highway 16 (NH-16) at Kharagpur, India, using two indices: the Air Pollution Tolerance Index (APTI) and the Anticipated Performance Index (API). Four biochemical parameters such as relative water content (RWC), leaf extract pH, ascorbic acid (AA), and total chlorophyll (TCH) were used to calculate APTI. The IIT Kharagpur campus was considered the control site to compare exposure levels and corresponding changes in plant species between the control (campus site) and polluted (roadside) environments. Duranta erecta and Neolamarckia cadamba demonstrated strong performance in APTI and API on both environments, whereas Cascabela thevetia, Albizia lebbeck etc. scored poorly. Interestingly, most species exhibited higher APTI values at the road site (mean=23.71) compared to the control site (mean=18.11), suggesting enhanced internal defence mechanisms in response to pollution stress. A multiple linear regression (MLR) model identified AA as the most influential significant predictor of APTI (β = 1.002, p<0.001), followed by pH, TCH, and RWC (β = 0.114, 0.100 and 0.077, respectively). There was no multicollinearity among the selected parameters, as all VIF values were below 2. In addition to biochemical traits, key morphological features such as stomatal density and cuticle thickness were examined as adaptive indicators that showed increased values under polluted conditions, suggesting physiological adaptation. Moreover, Duranta erecta exhibited the highest increase in stomatal density (58.95%), while Sasmanea saman showed the greatest increase in cuticle thickness (59.35%), reinforcing their suitability as roadside species for long-term air pollution abatement. Therefore, the integrated assessment of physiological, biochemical, and morphological traits of the selected dominant species highlight that the study outcomes can be used to select plant species for greening initiatives along National Highway corridors.Keywords: APTI; API; Multiple linear regression; Stomatal density

A document by Mohammad Zamehrian. Click on the document to view its contents.

This research expands on work in the field of Distributed Acoustic Sensing (DAS), a new technology that exploits unused optical fiber infrastructure for seismic strain sensing. When deployed on telecommunication cables, particularly offshore, the method has the potential to significantly enhance the coverage of Earthquake Early Warning (EEW) systems. However, while current EEW systems are designed to receive measurements of peak ground acceleration (PGA) from conventional seismic stations, the correlation between DAS amplitudes and ground motion is still not fully understood. There exists a need for more observations to validate this relationship before DAS technology can be included in EEW operations.In this study, DAS peak strain-rate (PSR) values are converted into ground motion predictions by estimating the apparent wave speed using a slant-stack approach. These predictions are compared with real measurements of PGA from co-located seismic stations from the Community Seismic Network to observe correlation and evaluate the accuracy of the method. Results provide new insight into the relationship between DAS amplitudes and ground motion and demonstrate the potential for DAS measurements to augment existing EEW systems.

A document by David M. Tratt. Click on the document to view its contents.

This project introduces Cognitive-Based Emulation (CBE), a new theoretical field that reframes trauma, cognition, and identity through symbolic projection and recursive epistemic feedback. From this field we created a cosmological framework, unique from other models. From this epistemic methodology we create an interdisciplinary baseline equation for the theory of everything based on first principles.  \cite{Newell_2026}   This model is then analyzed through a rigorous connection to Mikhail Medvedev’s (2023) QED-plasma strong field equation for theoretical validation. We then test this efficient, self-consistent, to obtain results parallel to recent findings from data, cosmological simulations, and similar work inspired through Mikhail’s extensive life production. We have explored several applications of the equation, including relativistic Weibel shocks, kinetic instabilities, dark matter dynamics, and potentially QED self-generated fields near bow shock regions. While traditionally studied individually we highlight the physical interwoven complexity of turbulence, magnetic reconnection, and shocks, phenomena traditionally studied in isolation, through a single analytical framework. This enables a consistent interpretation of energy transfer and particle energization without relying on subgrid models or artificial dissipation. In conclusion, we effectively bridge theory, simulation, and observation to enable fast, high-resolution analytical modeling. Therefore, this model naturally provides multiscale temporal and spatial resolution internally, preserving global regularity. Overall, this equation reduces to classical Relativity, magnetohydrodynamic, and quantum electrodynamics, suggesting that these traditional frameworks emerge as special cases of the framework.

Scientific communities are straining under mounting challenges—knowledge fragmentation, outdated publication processes, institutional erosion and funding cuts—that threaten our capacity to address urgent global problems. Traditional scientific processes and infrastructure are driven by misaligned incentives that reward competitive publications over collaborative knowledge-building.

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.

Precipitation downscaling is essential for applications such as climate modeling, hydrological simulations, and renewable energy planning. Stochastic downscaling methods, which aim to generate realistic fine-scale precipitation patterns from coarse atmospheric predictors, are particularly important due to the high spatial and temporal variability of rainfall. Despite significant advances in modeling techniques and computational power, precipitation downscaling still remain less accurate and reliable compared to most other meteorological variables, highlighting the need for continued research. In this work, we explore how AI Foundation Models can improve downscaling under the hypothesis that these very large models can better encode physical aspects of weather phenomena. We investigate precipitation downscaling at 0.1° resolution using diffusion models conditioned on ERA5 atmospheric variables and ERA5-Land static fields via embeddings from the Prithvi Weather Foundation Model. The conditioning data include surface and vertical (pressure-level) variables from ERA5 aggregated to 1° daily resolution, combined with ERA5-Land static fields. The target dataset is ERA5-Land daily precipitation at 0.1°. Unlike other approaches, we do not use ERA5 precipitation at coarse resolution; instead, we generate precipitation from noise guided by context variables, so called, channel-synthesis. We explore multiple conditioning strategies, including concatenation, cross-attention with a convolutional encoder, and feature embeddings derived from the Prithvi model to enhance performance. Experimental results show that diffusion models conditioned on Prithvi embeddings improve high-resolution precipitation generation compared to baseline conditioning models.

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.