Mid-Atlantic CHM and Digital Elevation Models Dataset

This dataset delivers high‑resolution (1‑meter per pixel) Canopy Height Models (CHMs) and Digital Elevation Models (DEMs) for the mid‑Atlantic United States. The products are generated by training Deep‑Learning (DL) models concurrently on high‑resolution multispectral imagery and high-resolution LiDAR samples from NSF’s NEON (National Ecological Observatory Network) project, which provide supervisory signals during training. This fusion yields a cost‑effective, scalable alternative and complement to wall‑to‑wall airborne LiDAR surveys, producing highly accurate canopy structure estimates that improve upon the quality of current DL‑based CHM products. Developed through innovative PhD research at the University of Delaware, the dataset supports applications across ecology, forestry, urban planning, hydrology, and climate science, from biodiversity conservation to disaster preparedness. The data are freely available to accelerate research and inform decision‑making across disciplines. The dataset is based on multispectral imagery acquired in 2023 and will be updated with 2025 imagery. Additional temporal layers can be produced whenever new high-resolution multispectral data become available, either captured or DL-super-resolved, enabling ongoing monitoring over time. The project website https://zaidud.github.io/midatlantic-elevationmap/ provides interactive maps for visualization of the dataset and easy download of subregions.