Opening Remarks
 

1:05 - 1:25 p.m.

Habitat and Periphyton Mapping in Small Forested Streams via UAS

Teresa Mathews, Oak Ridge National Laboratory (ORNL)

Remediation at mercury (Hg) contaminated sites is challenged by the fact that the metallic and dissolved Hg released into the environment can be microbially transformed into methylmercury (MeHg) which is a more bioaccumulative and more toxic form of Hg. In stream systems, evidence suggests that benthic algal biofilms known as periphyton are hotspots for MeHg production, making it critical to understand periphyton communities and their dynamics to guide mercury mitigation and remediation efforts in contaminated streams. Traditional methods for the quantification of periphyton distribution and abundance in the field are time-consuming, labor-intensive, and leave large spatial gaps, requiring tenuous interpolation and introducing significant uncertainty. One approach to address the scaling issue associated with periphyton field data collection is to use remote sensing approaches in combination with field data collection to map periphyton distribution and abundance. Here we present results from near-surface remote sensing via unoccupied aircraft systems (UAS) as a cost-effective and flexible alternative for targeted data collections in smaller study areas.  We focus on two streams on the Department of Energy’s Oak Ridge Reservation:  the Hg-contaminated East Fork Poplar Creek and Bear Creek, which has similar MeHg concentrations despite having very low aqueous Hg concentrations.  We compare results from the two creeks using lidar, spectral sensors, and subcanopy imagery to describe different components of these forested stream systems. We then overlay these data with instream sampling data with the ultimate goal of modeling and mapping periphyton distribution to inform management decisions on watershed, riparian corridor, and instream habitat that could serve to reduce MeHg production.

Coauthors: Chris DeRolph, Paul Matson, and Scott Brooks (ORNL)

1:25 - 1:45 p.m.

Remotely Sensed Data for the Assessment of Uranium Mining Rehabilitation in Northern Australia

Keith Tayler, Office of the Supervising Scientist, Australian Government

The Australian Government Office of the Supervising Scientist applies remotely sensed data gathered at a range of scales across a variety of platforms to monitor and assess the success of uranium mining rehabilitation in the Northern Territory of Australia.

The Supervising Scientist is an independent statutory function established in 1978 to ensure the protection of the environment of the Alligator Rivers Region, including the dual World Heritage listed Kakadu National Park, from the effects of uranium mining.

The Office of the Supervising Scientist, which includes the Environmental Research Institute of the Supervising Scientist, conducts independent environmental research into the impacts of uranium mining and undertakes a comprehensive environmental monitoring program to ensure environmental protection requirements are achieved.

The Ranger uranium mine is the largest project in the Region, having produced 132,000 tonnes of uranium oxide between 1980 and 2021, and is currently undergoing rehabilitation at an estimated cost of AUD$3 billion.

The Office of the Supervising Scientist is using a suite of data gathered by satellite, aircraft, UAV, telemetry and underwater videography across multiple spatiotemporal scales to monitor for environmental impacts, establish closure criteria and assess the success of rehabilitation at Ranger.

Mr Tayler, Supervising Scientist, will detail how his team applies these various platforms and data sources and discuss the innovative AI-driven data processing techniques which have been developed by the Environmental Research Institute of the Supervising Scientist to analyse the large amounts of data generated in a cost-effective and timely manner.

Coauthor: Josh Koh (Office of the Supervising Scientist, Australian Government)

1:45 - 2:05 p.m.

Next-Generation Drone Technologies for Radiation Sensing and Environmental Remediation

Jonathan Rogers, Georgia Institute of Technology

Autonomous drone technology is evolving at a rapid pace. In the context of environmental remediation, autonomous vehicles play an important role as remote sensing platforms that can be deployed without risking exposure of personnel to harmful levels of radiation or hazardous materials. This talk will highlight recent research conducted at Georgia Tech to develop both software algorithms and hardware technologies that enable drones to be used as valuable sensing platforms at nuclear facilities and in environmental remediation scenarios. On the algorithms side, the presentation will describe recent advancements in source term estimation that enable autonomous vehicles equipped with radiation sensors to locate, identify, and characterize point sources of radiation in the environment. These algorithms are specifically designed to operate in cluttered settings, and use advanced particle filter techniques and spectrum unfolding methods to locate multiple sources and identify their isotopic composition. Experimental results will be presented wherein a robot using a radiation sensor employs the algorithms to identify source locations, strengths, and isotopic composition in an environment with numerous obstacles.

A second aspect of the talk will focus on recent hardware advancements that allow drones to attach to large objects for the purpose of long-term sensing or deployment of non-destructive testing (NDT) sensors. This new technology pairs a standard drone platform with an attachment anchor that allows the drone to attach to the inspection target via suction. Once attached, contact sensors can be deployed for NDT purposes, or radiation detectors can be installed or serviced in hard-to-reach areas. The presentation will describe recent progress in the development of these self-attaching drones and their potential use in applications related to nuclear facilities and environmental remediation.

2:05 - 2:25 p.m.

IAEA Activities in Support of Radiological Site Characterization

Milan Matos, International Atomic Energy Agency

Development of gamma-ray spectrometers has resulted in numerous devices that provide a wide range of tools for radiological characterization of remediated sites. Advances in robotics and automation in parallel with the progress in mobile computational power allowed the deployment of autonomous portable and mobile radiation detection systems to become widely available.

The IAEA is active in supporting the Member States in several emerging technologies used for radiological characterization of the remediated sites, including ground and aerial detection systems. This presentation will include an overview of project and activities that have been carried out in support of the IAEA Member States.

Coauthors: P. Sladek (IAEA); S. Altfelder (Federal Institute for Geosciences and Natural Resources, Germany); K. Kanaki and D. Ridikas (IAEA)

Open Discussion
 

BREAK
 

3:15 - 3:35 p.m.

The Threat Engine™: A Multi-Agent System for Context-Aware Remote Sensing and Autonomous Threat Detection

Halina Harvey, Innovative Physics, Ltd.

Autonomous sensing and remote telemetry are critical to modern environmental and security intelligence. The Threat Engine™ by Innovative Physics represents a paradigm shift in how field- and lab-scale sensing platforms transform raw sensor data into strategic, real-time intelligence. Rather than simply collecting environmental data, this AI-powered system interprets information contextually—determining not just that a threat exists, but where it is, what it is, and why it matters.

At its core, the system fuses sensor data from ground, aerial, or spaceborne platforms with open-source intelligence (OSINT), proxy indicators, and historical context to build a rich, multi-dimensional operational picture. Whether mounted on an unmanned aerial vehicle (UAV) flying over a high-risk site, or integrated into fixed sensors monitoring infrastructure, the platform’s autonomous agents collaborate to recognize patterns, adapt to situational context, and deliver precise, time-critical insights. These agents excel at assessing spatial and temporal signals using location-aware logic, learning from each new data point, and dynamically refining their threat classifications in real time.

The system’s Data Collection Platform (DCP) acts as the command-and-control backbone—consolidating telemetry from diverse nodes, augmenting with historical data, and contextualizing each observation to detect subtle but significant deviations. This architecture enables true decision-support autonomy in the field, reducing the cognitive load on analysts and operators.

By combining machine learning, sensor integration, and situational awareness within a multi-agent framework, the Threat Engine™ is ideally suited for deployment in remote, hazardous, or data-denied environments. From environmental remediation and emergency response to critical infrastructure protection and geospatial reconnaissance, it allows for faster, safer, and more intelligent site monitoring—transforming Earth-to-Orbit data into actionable foresight.

Coauthors: Mike Anderson and David Prendergast (Innovative Physics, Ltd.)

3:35 - 3:55 p.m.

Autonomous Ground-Penetrating Radar Robot for Subsurface Mapping and Monitoring of Hazardous Waste Repositories

Pieter Hazenberg, Florida International University (FIU)

Ensuring the structural integrity and environmental safety of legacy radioactive waste disposal sites presents an evolving challenge for long-term stewardship agencies such as the U.S. Department of Energy's Office of Legacy Management. Traditional surface-based inspection methods are increasingly insufficient to detect early signs of subsurface instability, erosion, or leachate migration. To address this gap, we present a novel autonomous robotic platform equipped with integrated ground-penetrating radar (GPR) and lidar for comprehensive three-dimensional mapping and monitoring of disposal cell environments.

Our system combines solar-powered endurance, all-terrain mobility, and sensor fusion to autonomously traverse riprap and uneven terrains typical of engineered covers. The robot performs in-situ subsurface imaging without requiring prior baseline data, enabling real-time detection of structural anomalies and groundwater level shifts. Field trials at the Rifle and Mexican Hat Disposal Cell in Colorado validated the platform’s ability to navigate harsh terrain, map subsurface features, and contribute meaningful data for risk assessment and site management.

This work will present the mechanical, electrical, and software architecture of the system, deployment insights across multiple legacy waste sites, and feedback informing the development of next-generation designs. Furthermore, we will present some of the detailed information about the subsurface as measured by the platform. The proposed robotic solution represents a significant advancement in autonomous geophysical inspection, offering a scalable, low-maintenance approach to enhancing the safety and efficiency of long-term environmental surveillance at high-consequence infrastructure sites.

Coauthors: Anthony Abrahao (FIU); Shawn Cameron (Savannah River National Laboratory): Tyler Coy, Sahaira Paz, Mackenson Telusma, Ravi Gudavalli, and Leonel Lagos (FIU)

3:55 - 4:15 p.m.

Remote Sensing and Machine Learning Accurately Predict Soil Moisture Dynamics within an In-Service Evapotranspiration Disposal Cell Cover

Chris Jarchow, RSI EnTech, LLC

Conventional Uranium Mill Tailings Radiation Control Act disposal cell covers include a layer of compacted, clayey soil (called a “low-permeability radon barrier”) designed to limit the surface flux of radon and protect groundwater by controlling percolation through the cell and to do so for a period of 1000 years, to the extent reasonably achievable. However, natural ecological and soil-forming processes have already changed as-built engineering properties of conventional covers. Compacted soil layers no longer have a low permeability, and radon diffusion and flux have increased. The US Department of Energy Office of Legacy Management, its Legacy Management Support contractor, and other collaborators are investigating options for future management of selected conventional covers in the western US as evapotranspiration (ET) or water balance covers while still protecting human health and the environment and ensuring the remedy remains protective. To limit percolation, ET covers rely on a thick unsaturated soil “sponge” layer to store precipitation that is seasonally released back into the atmosphere by plant transpiration and soil evaporation. Percolation can occur if the water content exceeds the water storage capacity of the soil sponge, potentially mobilizing tailings contaminants, and radon flux rates can exceed the regulatory standard if the soil-water content is too low. This paper demonstrates the application of an innovative methodology to monitor soil moisture (SM) using remote sensing and machine learning (ML) technologies. We incorporated 8 years of in situ SM data and freely available optical and synthetic aperture radar satellite images and gridded precipitation datasets into a multisource ML framework to model SM across three-dimensional space using a field-scale drainage lysimeter embedded within the in-service ET cover at the Monticello, Utah, Disposal Site. The model was trained using in situ SM measurements from 2019 and validated using data from 2014–2018 and 2020–2021. The approach produced accurate SM estimates for all six soil layers to a depth of 2 meters (m), with correlations between modeled and in situ SM ranging from 94% in the topsoil to 75% at the bottom of the cover. Error was also low, with minimal bias and root mean square error ranging from 0.003 to 0.017 cubic centimeters (cm3)/cm3. The approach also captured seasonal SM variability and spatial heterogeneity at 30 m pixel resolution. In situ SM measurements in the top half of the cover profile differed considerably from the spatially aggregated ML results, suggesting the existing SM sensor network may not adequately capture SM heterogeneity in this system. Our results suggest that vegetation is the primary driver of SM dynamics in this ET cover, and remote sensing and ML may be considered useful tools in maintaining regulatory compliance at ET cover disposal cells.

Coauthors: John S. Kimball (Numerical Terradynamic Simulation Group, University of Montana), Alison Kuhlman (DOE-LM)

4:15 - 4:35 p.m.

Soil Moisture Profile Monitoring and Forecasting Using In-Situ Sensors, Satellite Remote Sensing, and Machine Learning

Jinyang Du, University of Montana

Soil moisture (SM) is an essential climate variable that governs land-atmosphere interactions, runoff generation, and vegetation productivity. Timely monitoring and forecasting of SM spatial distribution and vertical profiles are critical for applications such as drought early warning and assessing the performance of engineered covers, including those for uranium mill tailings disposal cells. Arid and semi-arid regions are often selected for waste storage due to low groundwater recharge, yet verifying cover efficacy requires non-invasive and spatially representative SM data. Traditional monitoring methods are invasive, costly, and limited in capturing spatial heterogeneity, particularly over vegetated disposal systems.

To address these challenges, this study developed a satellite-driven machine learning (ML) approach to model high-resolution (30-m) daily SM dynamics by integrating multisource remote sensing data with in situ multi-layer SM measurements from the Montana Mesonet. The ML model established robust relationships between diverse predictors and in situ measurements across 4-, 8-, and 20-inch soil depths, achieving strong accuracy (R > 0.91; RMSE ≤ 0.047 cm³/cm³) with 1- to 2-week forecast lead times. The forecast system successfully predicted the onset, progression, and termination of the 2017 Montana flash drought, which was not fully captured by operational monitoring systems. By leveraging non-invasive remote sensing and ML, this approach offers a novel solution for long-term disposal cell monitoring, drought prediction, and water cycle assessments, with potential benefits for water resource management, precision agriculture, and environmental risk mitigation.

Coauthors: John S. Kimball (University of Montana), Christopher J. Jarchow (RSI EnTech, LLC)