Earth Science Digital Twins are rapidly evolving to advance earth science understanding and societally relevant decision making. Coastal hazards pose complex and compound threats to dynamic human settlements and land use, inviting their analysis using digital twin technologies. A geospatial approach requires a framework for data assimilation that spans Earth observations, in situ sensor networks, environmental models, and characterization of human population and infrastructure assets. In our study, an Open Data Cube is developed as the core integration platform for a series of prototype case studies: 1) coastal mosquito-borne disease exposure assessment; 2) dasymetric mapping of vulnerable populations; and 3) integrated modeling and observational assessment of a King Tide flood event for Hampton Roads, Virginia (USA.) These prototypes are described with preliminary results. The paper also discusses benefits of the ODC as a Digital Twin as well as impediments and future developments for enhanced capabilities.

The increasing availability and accessibility of commercial satellite data presents a unique opportunity for the Earth observation (EO) community to leverage new sources of remote sensing imagery. However, despite improved access to data and computation, conducting analyses with data from multiple platforms remains challenging due to interoperability barriers. Calibration and Validation (Cal/Val) efforts are essential to overcoming these obstacles by ensuring the quality and reliability of EO data and quantifying necessary adjustments for seamless integration. The Committee on Earth Observation Satellites (CEOS) Systems Engineering Office (SEO) has launched the CEOS Analytics Lab (CAL), an Open Data Cube (ODC)-based tool, to facilitate data processing and collaboration among researchers. In this paper, we present a pilot study using tools developed for CAL to compare commercial imagery to traditional civil government Earth observations. We demonstrate the usefulness of collaborative environments such as CAL and provide the community with tools which assist with commonly identified calibration and validation tasks necessary to perform interoperability studies. This demonstration lays a foundation for operations around the use and evaluation of commercial satellite data, supporting further developments in the CEOS and Open EO communities. We also present a set of tools, CalValToolbox, to the EO community to perform Cal/Val tasks.

Generative artificial intelligence (AI), specifically Large Language Models (LLMS) such as generative pretrained transformers (GPTS) have emerged as a transformational technology in modern research and applications. This technology will revolutionize a wide variety of fields, including remote sensing and Earth observation. We explore the use of LLMS to improve and automate EO analysis, making these results more accessible to a broad range of decision makers. We develop a framework using this technology to support the Committee on Earth Observation Satellites (CEOS) and NASA’s Open Science initiative. This framework leverages current work on Earth observation tools, including the Open Data Cube (ODC) [1] and the OEA Algorithm Hub (AlgoHub). [2]In our work, we investigate the ability of large language models to orchestrate and reason with tools created to access geospatial data using the Open Data Cube and the OEA Algorithm Hub [2]. In this paper we propose a natural language analytical framework for orchestrating Earth observation (EO) analysis. The framework uses large language models to interpret an analysis request, devise a solution based on ODC and AlgoHub tools, and execute the analysis. Our experimental investigations demonstrate that an LLM, with its embedded contextual knowledge, can effectively discern and sequence the necessary processing steps required to fulfill Earth observation analysis requests. This research demonstrates the potential of Generative AI (LLMs) in streamlining and enhancing EO analyses and empowering the observation community.In summary, Generative AI (LLMs) can enhance the capabilities of Earth observation analysis by automating tasks, providing natural language interfaces, and facilitating the extraction of valuable insights from the vast and complex datasets generated by remote sensing and EO technologies.

A Digital Twin (DT) for coastal resilience in a low-lying coastal city necessarily entails characterizing the geophysical terrain, complex hydrologic flows, and infrastructure modifications that affect flooding. Coastal flooding in ports such as Hampton Roads, Virginia, arise from coastal storms, regular tidal estuarine circulation processes, seasonal to interannual external forcing from the ocean via the Gulf Stream and Atlantic Meridional Overturning Circulation (AMOC), and associated local influences on sea-level anomalies. We take a geospatial, hub-based approach to integrate digital linkages and threads among real-time Internet of Things (IoT) water level and flood sensors, an operational hydrodynamic model, localized probabilistic sea-level rise projections, hydro-corrected LiDAR Digital Elevation Models and GIS stormwater infrastructure, and an open data cube that ingests, hosts, and processes Application-Ready Datasets (ARDs.) Our paper describes the structure and cloud integration of a flood sensor network, StormSense, and its augmentation with end-user identified gap sites to improve network coverage and sensitivity to vulnerable communities. The enhancement of the multiple flood sensor networks incorporates multiple APIs and Amazon Web Services to normalize and homogenize data feeds among municipal sensors, private sector sensors, NOAA and USGS gauges, and new sensors sited in partnership with emergency managers, non-profits, and the National Weather Service. We describe the development and improvements of a Delft3D model for hydrodynamic simulation, including characterization and grid improvements of shallow estuarine creeks. The model includes sea-level rise parameters for the Hampton Roads study area. The paper also outlines Earth Observing data in the Virginia Open Data Cube, including products from NASA, NOAA, ESA, and newly incorporated drone data. The data cube hosts multi-purpose data products, such as satellite derived Land Use/Land Cover, MODIS, Landsat, and Sentinel datasets, and large high-resolution, hydro-corrected LiDAR DEMs. We link the data cube to an ArcGIS Hub to provide end-products, disseminate content, and apply GIS analyses for real-time and simulated future scenario dashboard. The hub provides an end-user interface for impact modeling, what-if scenarios and decision support. The paper builds DT infrastructure to support emergency management, public health, and resilience planning.

The Open Data Cube (ODC) [1], [2], [3] is an open-source geospatial data management and analysis software package growing in popularity. With increased popularity there is increased demand for the computational power and storage capabilities required to analyze large spatial and temporal datasets. Cloud resource providers have bolstered their computational and satellite data offerings, simplifying access and management so that average users are able to tailor systems to their specific needs. Consequently, there exists substantial community interest in the comparative capabilities and technological performance of different cloud providers. Making informed choices in cloud providers offerings is crucial for optimizing geospatial data management and processing to meet user-specific needs. In this paper we begin to evaluate the deployment and performance of the ODC, standard notebooks, and datasets, to better determine functional differences. Our work finds that datasets saved in the local environment execute operations significantly faster than from the cloud, Google Earth Engine (GEE) has longer execution times when the metadata for a given asset is indexed locally, and that more thorough benchmarking is required to better understand the end-to-end performance of EO processes on the cloud.

Riparian and shoreline buffers are vital landscape elements of coastal and riverine shorelines, providing significant ecosystem services, habitats, and aesthetic qualities. One critical role is the filtration and uptake of non-point source pollutants, particularly in estuaries that have seen extensive land cover change, loss of wetlands, and that face sea-level rise such as the Chesapeake Bay. Combining a remote sensing and coastal ecological approach, this project presents data science tools to investigate coastal riparian and shoreline vegetation relationships with in situ water quality sampling. This project provides tools and examples for quantifying riparian buffer vegetation, calculating Total Suspended Matter (TSM) from Earth Observation (EO) data, and demonstrates the use of data science methods to explore relationships between non-point source pollutants.

Collecting data via small Unmanned Aerial Systems (UASs) can fill a niche amongst geospatial data collection methods by providing recent, very high-resolution imagery, at a fraction of the cost of space-based providers. We detail an end-to-end system for acquiring, processing, and importing data into the Open Data Cube (ODC), a free and open-source geospatial data management and analysis platform. The process employs third party software that can be low-cost or free. The acquisition component of our pipeline consists of low-cost drones, either assembled with commodity components, or available from retail stores to enable the community at large to better take advantage of UASs as a reliable data collection platform. We discuss a selection of third-party photogrammetry suites which output assembled images into tiles that can be indexed by the ODC. Finally, we demonstrate the indexing of orthographic drone data into the Open Data Cube.