Preprints
Filtering by Subject: Artificial Intelligence and Robotics
Tracking Progress in General-Purpose VLMs and Agents for Earth Observation
Published: 2026-11-09
Subjects: Artificial Intelligence and Robotics, Remote Sensing, Sustainability
The rapid advancement of vision-language models (VLMs) and agent-based workflows is expanding their potential for Earth observation (EO). Existing studies primarily compare models at a single point in time, providing limited insight into how their capabilities evolve across generations. To address this gap, we conduct a longitudinal evaluation of 48 VLMs across successive generations using a [...]
What lies beneath: National aboveground carbon stock changes and under-canopy disturbance detection from 3D point clouds
Published: 2026-11-07
Subjects: Artificial Intelligence and Robotics, Environmental Monitoring, Environmental Sciences, Forest Management, Forest Sciences, Geography, Remote Sensing
Accurate and spatially explicit quantification of forest biomass and carbon stock changes over time is necessary for carbon accounting and verification at different scales. Here, we applied deep learning to full 3D point-cloud data to produce a high-resolution time series of forest attribute maps for Denmark. The maps include predictions for biomass and canopy height with estimated uncertainties, [...]
Station-trained air-quality forecasts without regional emission inventories
Published: 2026-10-31
Subjects: Artificial Intelligence and Robotics, Atmospheric Sciences, Computer Sciences, Environmental Monitoring, Environmental Sciences
Air-quality forecasts rely on regional chemical transport models (CTMs) and emission inventories, which are uncertain in many regions and costly to update. Here we present SOLACE, whose inner forecast is trained on monitoring-station observations alone and evolves gridded concentration fields through explicit transport, layer exchange, sources and sinks. With 140,000 trained parameters and no [...]
Explainable and Uncertainty-Aware Computational Geodetic Inversion: A Review and Framework for Multi-Source Earth Observation
Published: 2026-10-30
Subjects: Artificial Intelligence and Robotics, Geophysics and Seismology, Numerical Analysis and Scientific Computing
Geodetic inversion converts surface and remote sensing data into estimates of concealed Earth system parameters, including fault slip, magma-source geometry, subsurface mass changes, crustal strain, and time-dependent geohazard activities. This review explores geodetic inversion as a computational challenge that involves combining diverse data, optimizing and quantifying uncertainty, and [...]
Detecting Changes in Power Plant NOx Emissions Using Deep Learning and TEMPO Satellite Imagery
Published: 2026-10-24
Subjects: Artificial Intelligence and Robotics, Atmospheric Sciences, Climate
NASA's Tropospheric Emissions: Monitoring of Pollution (TEMPO) instrument provides hourly daytime observations of tropospheric nitrogen dioxide (NO2) across North America. These observations could support independent, facility-level monitoring of power plant nitrogen oxides (NOx) emissions, filling gaps when continuous emissions monitoring system (CEMS) measurements fail or facilities are not [...]
Three gaps for foundation models to address in satellite remote sensing of inland waters
Published: 2026-10-09
Subjects: Artificial Intelligence and Robotics, Computer Sciences, Earth Sciences, Environmental Sciences, Fresh Water Studies, Oceanography and Atmospheric Sciences and Meteorology
Long-term Earth observation of inland waters depends on continuity between satellite missions, adequate combinations of spectral, spatial and temporal resolution, and representative calibration and validation datasets. Discontinuities and trade-offs in these components constrain the records available for studying inland waters and their change. We examine three gaps affecting inland water remote [...]
Leveraging Ocean Data and Models: Building the Data Infrastructure, Workforce, and Modeling Capacity to Synthesize Earth’s Ocean Observations
Published: 2026-09-25
Subjects: Artificial Intelligence and Robotics, Atmospheric Sciences, Biogeochemistry, Climate, Computer Sciences, Databases and Information Systems, Earth Sciences, Environmental Sciences, Fresh Water Studies, Geology, Hydrology, Meteorology, Numerical Analysis and Scientific Computing, Oceanography, Other Earth Sciences, Other Oceanography and Atmospheric Sciences and Meteorology
NASA’s aquatic and marine research enterprise within the overarching NASA Biosphere Program is entering an era of unprecedented data abundance. MODIS and the Plankton, Aerosol, Cloud, ocean Ecosystem (PACE) mission already generate ocean color data at rates that strain legacy processing and archiving workflows, and forthcoming missions will add further volume and complexity. These satellite [...]
Seasonal and interannual dynamics of estuary connectivity along the California coast
Published: 2026-09-23
Subjects: Artificial Intelligence and Robotics, Environmental Sciences, Geography, Remote Sensing
Intermittently closed estuaries are globally widespread and ecologically important, yet their mouth-state dynamics are difficult to observe simultaneously at multiple estuaries, limiting understanding of how regional forcing and local controls shape estuary-ocean connectivity. We developed a multi-year record of estuary mouth connectivity for 66 intermittently closed estuaries along the [...]
Terrain-Dependent Generalization Failure in Multimodal Lunar Crater Classification: A Comparative Study Using Optical and Topographic Data
Published: 2026-09-08
Subjects: Artificial Intelligence and Robotics, Computer Sciences, Planetary Geology, Planetary Geomorphology, Planetary Sciences
Automated lunar crater detection is an important task in planetary mapping, geological analysis, and future lunar exploration. Recent studies have demonstrated the potential of combining optical imagery with topographic information; however, the influence of terrain variability on crater classification performance remains insufficiently explored. This study presents a comparative evaluation of [...]
Global Power Outage Detection at the Kilometer Scale from Satellite Nighttime Lights
Published: 2026-08-28
Subjects: Artificial Intelligence and Robotics, Geographic Information Sciences, Geography, Human Geography, Nature and Society Relations, Power and Energy, Public Health, Remote Sensing
Satellite nighttime light observations show great promise for remote sensing of power outages, yet have only been utilized in limited case studies to date. Here, we present generalized detection of power outages across the globe at ~1-kilometer resolution. Our method, the Outage frameWork for Luminosity Interruptions (OWL-I), reveals the power outages can be reliably detected from VIIRS/Black [...]
A Blueprint for Modern Floodplain Management: System Requirements and Architecture for AI-Ready Platforms
Published: 2026-07-29
Subjects: Artificial Intelligence and Robotics, Civil and Environmental Engineering, Computer Sciences, Engineering, Hydraulic Engineering, Physical Sciences and Mathematics, Water Resource Management
Floodplain managers play a critical role in mitigating flood risks and ensuring compliance with federal, state, and local regulations. Their daily work often depends on a patchwork of standalone tools, manual processes, and legacy systems that have evolved over time to meet regulatory and community needs. To establish a foundation for next-generation technological solutions, we synthesized [...]
Detecting illegal water abstractions using Earth Observation and artificial intelligence: challenges, research gaps, and future directions
Published: 2026-07-28
Subjects: Artificial Intelligence and Robotics, Environmental Monitoring, Hydrology, Natural Resources Management and Policy, Remote Sensing, Water Resource Management
Water scarcity and unsustainable withdrawals are major environmental challenges, particularly in semi-arid and Mediterranean regions where agriculture places strong pressure on limited hydrological resources. Illegal or unreported water abstractions intensify these pressures by undermining ecosystem stability, policy compliance, and long-term water security. Although Earth Observation (EO) and [...]
Cloud-Free Imaging Probability Forecasting for Optical Earth-Observation Tasking over Yerevan, Armenia
Published: 2026-07-17
Subjects: Artificial Intelligence and Robotics, Atmospheric Sciences, Environmental Monitoring, Numerical Analysis and Scientific Computing, Remote Sensing
Optical Earth-observation satellites such as Sentinel-2 cannot see through cloud, so a tasking attempt over a cloudy target wastes a limited imaging window along with onboard power and downlink bandwidth. We study whether next-day cloud conditions over Yerevan, Armenia can be forecast accurately enough to automate the binary Go/No-Go tasking decision. Daily cloud-free fractions are derived from [...]
Deep Neural Network-Based Inversion of Turbidites in Confined Basins
Published: 2026-07-11
Subjects: Artificial Intelligence and Robotics, Fluid Dynamics, Geology, Sedimentology
Turbidites generated by large earthquakes and other geological events are commonly preserved in small, topographically confined basins along active continental margins. Reconstructing flow conditions from these deposits is essential for assessing past hazards; however, existing inverse models have been validated only for unconfined settings, and their applicability to confined basins remains [...]
A machine learning approach for detecting biofouling in oceanographic data
Published: 2026-07-11
Subjects: Artificial Intelligence and Robotics, Fluid Dynamics, Numerical Analysis and Computation, Oceanography, Oceanography and Atmospheric Sciences and Meteorology, Physical Sciences and Mathematics
Autonomous ocean observing platforms collect long-term biogeochemical time series, but sensor degradation from biofouling introduces progressive biases that contaminate the climate record. This work focuses on the BGC-Argo fleet of profiling floats, where optical sensors measuring chlorophyll-a and backscatter are particularly susceptible to biofouling. Current detection relies on per-float [...]