Abstract The Compact Reconnaissance Imaging Spectrometer for Mars (CRISM) Visible‐to‐near‐infrared Hyperspectral Reduced Data Record (VRDR) data set provides near‐global (∼99%) coverage of the Martian surface between 0.371 and 1.017 μm across 92 spectral channels, at a spatial resolution of 90 m/pixel. This wavelength interval overlaps with that of high spatial resolution multispectral imagers, such as the Colour and Stereo Surface Imaging System (CaSSIS) and High Resolution Imaging Science Experiment (HiRISE), while providing substantially greater spectral detail. Previous work has shown that CaSSIS spectra are capable of distinguishing between key rock‐forming minerals using only four spectral bands. We have resampled the complete CRISM VRDR mapping data set to the four CaSSIS bands (Near‐Infrared “NIR,” Red “RED,” Panchromatic “PAN,” and Blue “BLU”). The result is the CaSSIS Resampled Data Record (CaRDR) data set: a new, simulated four‐band CaSSIS view of the Martian surface at 90 m/pixel available as a global set of 5° × 5° tiles, covering Mars quadrangles mc02–mc30. CaRDR products include GIS‐ready four‐band cubes, 8‐bit GeoTIFF browse products of various band combinations, and Color Band Ratio Composites. Together, these products enable (a) identification of new high‐priority imaging targets and improved spectral–contextual planning for high‐resolution cameras such as CaSSIS, HiRISE, and future orbital imagers, and (b) new opportunities to explore visible–near‐infrared heterogeneity at the sub‐hectometer scale, including spatial trends in mineralogy, surface composition, and regional geologic processes across Mars.
Abstract The North American‐Pacific Geopotential Datum of 2022 (NAPGD2022) is anticipated to become the official height system of the United States in 2027. This new datum is realized through a suite of models, including GEOID2022 (geoid) and DEFLEC2022 (deflection of the vertical). GEOID2022 targets 1–2 cm height accuracy in the contiguous United States, but its performance in areas containing terrestrial gravity data gaps and varying geology has not been well studied. In this study, we evaluate the 2025 beta versions of the GEOID2022 and DEFLEC2022 models using an established GNSS network to assess their performance in Oregon's Willamette Valley, a populated region with relatively sparse terrestrial gravity data coverage. A total of 23 selected marks in the GNSS network were occupied with the QDaedalus astrogeodetic system to measure geoid slopes with sub‐microradian accuracy. Geodetic coordinates were derived from long‐duration static GNSS observations. Eleven of the selected marks also have leveled orthometric heights. These three observation types enable validation of GEOID2022 using both GNSS/leveling and astronomical leveling. Thin‐plate splines were used to fit residual geoid undulation grids to observed astronomical geoid slopes. Despite limited terrestrial gravity data coverage, the astrogeodetic and GNSS/leveling observations are broadly consistent with both GEOID2022 and DEFLEC2022, within 2 cm and 0.7 arcseconds RMS. However, higher‐elevation sites exhibit pronounced multi‐centimeter and multi‐arcsecond systematic errors, particularly near Oregon's geologic surface density contrast boundaries. These results highlight the robustness of NAPGD2022 in regions with sparse gravity data, while underscoring the need for more complete and consistent terrestrial gravity coverage.
Abstract Snow water equivalent (SWE), which quantifies the amount of water in a snowpack, is critical for water supply forecasting. This study explores machine learning (ML) models' capability of representing the spatial distribution of SWE across the western United States. Nine graph convolution network (GCN) models using different loss functions, including a series of physics‐informed loss functions, are developed and compared to a baseline extra tree (ET) model. Different approaches for physics‐informed loss function are examined according to the correlation among geographic and meteorological conditions and are expected to help ML models better capture SWE variability. Terrain characteristics, meteorological variables, and snow cover products are used as input features for regional SWE estimation. All ML models are trained on data sets at selected in situ sites during 2018–2021 and evaluated with 9 days in January–March 2025. Results show that GCN models generally outperform the ET model, capturing SWE spatial variability more effectively and achieving higher statistical metrics. The use of the SWE‐elevation‐precipitation‐based physics‐informed loss function provides the best results among all models with an improved coefficient of determination ( R 2 ) of 0.28 and the lowest root mean square error (RMSE) of 0.22 m, while underestimation is observed for high SWE. A site‐by‐site comparison reveals that model estimates from the GCN models generally agree with observations with R 2 of 0.5–0.9, showing a reliable capability of estimating SWE at a single given site. However, the GCN models still have limited spatiotemporal generalization affected by the restricted representation of site measurements.
Abstract This study describes a machine learning model cascade for automating large‐scale shoreline mapping from time‐series of coastal satellite imagery. Our model cascade consists of independent supervised deep learning models for image suitability, image segmentation, and image segmentation suitability. The segmentation model determines shoreline locations, while the other models in this cascade filter imagery not suitable for segmentation or segmentations not suitable for shoreline detection. The SegFormer‐based segmentation model is trained with almost 2,300 labeled satellite images from a large variety of globally distributed coastal environments. The image segmentation model operates on imagery that has been deemed suitable by an image classification model that classifies whole satellite images as either suitable or unsuitable for segmentation. Another supervised whole image classification model is used to determine the suitability of each segmentation output. Finally, a spatial kernel density filters removes outlier shoreline points, with further timeseries analysis techniques to remove any remaining outliers. This study demonstrates the effectiveness of this system at two sites that are challenging to map: one at a sandy spit near Moses Point Fishing Village, Alaska, USA, and one at the mouth of the Elwha River, Washington, USA. This study demonstrates that these automated techniques produce shorelines with similar or better accuracies when compared to semi‐automated techniques. Ensembling model outputs better constrains the error and uncertainty of shoreline positions. Together, the set of models presented in this paper constitute an operational workflow for the generation and quality control of large‐scale records of satellite‐derived shoreline change.
Abstract Radial diffusion, driven by ultra‐low frequency (ULF) waves, is an important process for the transport of electrons in the outer radiation belt. Ground magnetometers give us constant observations of such ULF waves but their usefulness is limited by the models used to transform ground measurements into their progenitor fields in space. A typical assumption is that of a dipole field, and so measurements can only be reliably performed during the day. We have chosen a series of 17 magnetometer stations at different sets of geomagnetic latitudes, with stations in each set separated by several hours in longitude, and compared their Pc5 (1.67–6.67 mHz) ULF power measurements, and the resulting calculated radial diffusion coefficients, from 11 years of their data. By comparing the mismatch of their results when they were in different time sectors, and for different values of Kp, we were able to assess the median deviation of measurements conducted before dawn or after dusk with those on the day side. This will be useful for projects requiring a longitudinally limited array of magnetometers, such as FARBES (Forecast of Actionable Radiation Belt Scenarios), or for parts of the world where such coverage is limited.
