Neural network guided interpolation for mapping canopy height of China's forests by integrating GEDI and ICESat-2 data

• Published in: Remote sensing of environment Vol. 269; p. 112844
• Main Authors: Liu, Xiaoqiang, Su, Yanjun, Hu, Tianyu, Yang, Qiuli, Liu, Bingbing, Deng, Yufei, Tang, Hao, Tang, Zhiyao, Fang, Jingyun, Guo, Qinghua
• Format: Journal Article
• Language: English
• Published: New York Elsevier Inc 01.02.2022; Elsevier BV
• Subjects: altimeters; Canopies; canopy height; Carbon; Carbon sequestration; China; Climate change; Deep neural network; Ecosystem dynamics; Elevation; environment; Estimates; forest canopy; Forest canopy height; Forest ecosystems; Forest management; Forests; GEDI; Global climate; ice; ICESat-2 ATLAS; Interpolation; Laser altimeters; Lasers; Lidar; Mapping; Monitoring methods; Neural networks; Performance evaluation; Saturation; Spaceborne lidar; Spatial interpolation; standard deviation; Terrestrial ecosystems; topography; China; Lidar; Spatial interpolation; GEDI; ICESat-2 ATLAS; Deep neural network; Forest canopy height
• DOI: 10.1016/j.rse.2021.112844
• ISSN: 0034-4257

Spatially continuous estimates of forest canopy height at national to global scales are critical for quantifying forest carbon storage, understanding forest ecosystem processes, and developing forest management and restoration policies to mitigate global climate change. Spaceborne light detection and ranging (lidar) platforms, especially the Global Ecosystem Dynamics Investigation (GEDI) and Ice, Cloud, and land Elevation Satellite-2 (ICESat-2) Advanced Topographic Laser Altimeter System (ATLAS), can measure forest canopy height in discrete footprints globally. Their coverage provides a promising data source for national to global-scale forest canopy height estimates. However, previous studies usually used a regression-based approach to develop spatially continuous forest canopy height distribution through the aid of optical images, which cannot take full advantage of the dense spaceborne lidar footprints and may still suffer from the saturation effect of optical images. In this study, we developed a novel neural network guided interpolation (NNGI) method to map forest canopy height by fusing GEDI, ICESat-2 ATLAS, and Sentinel-2 images. To evaluate the performance of the proposed NNGI method, we generated a 30-m forest canopy height product of China for the year 2019. More than 140 km2 drone-lidar data were collected across the country to train and validate the NNGI method. The results showed that the average forest canopy height of China is 15.90 m with a standard deviation of 5.77 m. We evaluated the interpolated forest canopy height product of China by over 1,100,000 GEDI validation footprints (R2 = 0.55, RMSE = 5.32 m), about 33 km2 drone-lidar validation data (R2 = 0.58, RMSE = 4.93 m), and over 59,000 field plot measurements (R2 = 0.60, RMSE = 4.88 m). Benefiting from the interpolation-based mapping strategy, the resulting product had almost no saturation effect in areas with tall forest canopies. The high mapping accuracy demonstrates the feasibility of the proposed NNGI method for monitoring spatially continuous forest canopy height at national to global scales by integrating multi-platform spaceborne lidar data and optical images, enabling opportunities to provide more accurate quantification of terrestrial carbon storage and better understanding of forest ecosystem processes. •Neural network guided interpolation (NNGI) was proposed to map forest canopy height.•Over 140 km2 drone lidar data were collected to train and validate the NNGI method.•A 30-m forest canopy height product of China was mapped with high accuracy by NNGI.•The NNGI method reduced the saturation effect of estimates in tall forests.•Fusion of GEDI and ICESat-2 alleviated the strip effect of each individual dataset.