Innovative Cloud based solutions to map national scale landscape dynamics

By Josh Heyer, 2/13/2026

At RedCastle Resources our contributions to national mapping programs, developed in tandem with our partners in the USDA Forest Service, are at the forefront of augmenting landscape monitoring efforts and improving management outcomes across our public lands. As a Geospatial Data Scientist and Technical Team Lead who joined RedCastle in 2019, my work focuses on leveraging powerful, large-scale computational platforms to tackle complex environmental monitoring and land management challenges.

Central to this effort are two high-impact national mapping applications and products: the annual, wall-to-wall Tree Canopy Cover (TCC) product, and the Landscape Change Monitoring System (LCMS). Both are vital tools for better understanding and managing our nation's dynamic landscapes.

The immense undertaking of national scale mapping efforts requires a workflow that can process petabytes of satellite imagery efficiently. One solution we've successfully implemented over the past decade is harnessing the cloud computing power of Google Earth Engine (GEE).

The workflow developed by the team at RedCastle demonstrates precisely how GEE can be leveraged for large-scale, high-resolution landscape monitoring. GEE provides the necessary platform to move beyond localized or regional scale studies to comprehensive, wall-to-wall mapping.

Key Innovations

Our work with TCC has involved developing several key methodological innovations:

Wall-to-Wall Tree Canopy Cover (TCC) Prediction: RedCastle’s refined methodology provides a robust method for using GEE to predict tree canopy cover across vast areas, offering a comprehensive canopy cover product across CONUS.
Model Temporal Transferability: A major challenge in long-term monitoring is working with training data that are sparse or temporally constrained. To address this problem, we use model temporal transferability, a method that allows machine learning models to train on temporally constrained training data and make accurate predictions in years with no training data.
Quantifying Uncertainty: Transparency and reliability are paramount in scientific products that agency land managers can rely on for decision making. The TCC workflows implement a critical approach to quantifying uncertainty, a methodology that can be adopted in similar landscape monitoring work to ensure data integrity.

Sharing the Knowledge

Our LCMS work has developed innovative visualization applications. web based tools for processing and viewing geospatial data:

Data Viz: We have developed web-based data explorers that offer users a way to query and summarize LCMS and TCC data. The data explorers use the GEE API to do large-spatial data summaries on the fly and time-series analyses. Further, we have developed a LCMS API that can be used to generate data summaries more efficiently and generate reports.
geeViz Python package: RedCastle has developed the geeViz Python package that can be used to visualize geospatial data that are stored in GEE or awesome GEE. Redcastle has developed a robust library of geospatial tools in geeViz that can be used to process remote sensing data.

The insights and experience gained from producing LCMS and TCC using machine learning and Google Earth Engine were recently shared in peer reviewed journal articles and with the next generation of data scientists. I recently published “Annual national tree canopy cover mapping: A novel workflow with temporal transferability and improved uncertainty quantification" in Science of Remote Sensing, and was a coauthor on “Coincident maps of changing land cover, land use, and forest condition in the United States“ in Scientific Data. In addition, I organized and presented a talk to the University of Utah School of Environment, Society, and Sustainability, detailing how to effectively use GEE in nationwide mapping efforts.

This experience highlights the commitment of the team at RedCastle Resources to not only deliver critical environmental and ecological mapping solutions to our clients, but also to share the advanced techniques with the larger geospatial community that make these national-scale projects possible.

https://www.sciencedirect.com/science/article/pii/S2666017225001075?via%3Dihub