Related Experiment Video
Updated: Jun 7, 2026

Use of Principal Components for Scaling Up Topographic Models to Map Soil Redistribution and Soil Organic Carbon
Published on: October 16, 2018
A 0.6-meter resolution canopy height model for the contiguous United States.
Scott L Morford1, Brady W Allred2, Shea P Coons2
1Numerical Terradynamic Simulation Group, University of Montana, Missoula, MT, USA. scott.morford@umontana.edu.
We created NAIP-CHM, a detailed 3D model of U.S. landscapes using aerial imagery. This tool accurately maps vertical structures like trees and buildings, aiding ecosystem monitoring and land management.
Area of Science:
- Geospatial science
- Remote sensing
- Ecology
Background:
- Above-ground vertical structure is vital for ecosystem monitoring, carbon accounting, and land management.
- Airborne lidar, while accurate, is costly and has limited coverage, restricting its widespread use.
- Existing models often exclude human-made structures, limiting comprehensive landscape analysis.
Purpose of the Study:
- To develop a high-resolution (0.6m) canopy height and structure model (CHM) for the contiguous U.S. using publicly available aerial imagery.
- To create a model that characterizes the full vertical structure of landscapes, including vegetation, buildings, and infrastructure.
- To provide accessible tools and data for broad application in land management and ecosystem monitoring.
Main Methods:
- Utilized the National Agriculture Imagery Program (NAIP) aerial imagery to derive the NAIP-CHM.
- Employed a U-Net convolutional neural network with attention mechanisms and environmental conditioning.
- Trained and validated the model using a large dataset of 22.8 million co-registered NAIP imagery and lidar-derived CHM pairs, with stratified sampling for diverse ecosystems.
Main Results:
- The NAIP-CHM model achieved a pixel-wise root mean square error (RMSE) of 2.28 meters and an r² of 0.87 across the contiguous U.S.
- For forested sites specifically, the model yielded an r² of 0.82 and an RMSE of 3.82 meters.
- The developed model successfully characterizes both natural and human-made vertical structures.
Conclusions:
- NAIP-CHM offers a cost-effective and comprehensive solution for mapping landscape vertical structure across the U.S.
- The model's accuracy and broad coverage support improved ecosystem monitoring, carbon accounting, and land management.
- The provision of dataset, source code, and cloud tools facilitates widespread adoption and application without specialized computational needs.
Related Concept Videos
Topographic Surveying and Contours
Precipitation Gravimetry
In determining nickel by gravimetric analysis, a precipitant of ethanolic dimethylglyoxime is added to a hot nickel salt solution. This is quickly followed by the dropwise addition of dilute ammonia solution until precipitation occurs. A...
Methods of Obtaining Topography
Light Acquisition
Plotting of Topographic Maps
Precipitation and Co-precipitation
