Robust retrieval of biophysical traits under shade gradients in young Hopea hainanensis using visible and near-infrared hyperspectral hybrid modeling
Frontiers in Plant Science · 18.06.2026 · 17
DOI: 10.3389/fpls.2026.1788941 ↗Kurz gefasst
Fachartikel in der Zeitschrift Frontiers in Plant Science, online erschienen am 18.06.2026. Verfasst von Lin Chen und 9 weiteren Autor:innen. Beteiligte Einrichtung: Hainan Academy of Forestry (Hainan Academy of Mangrove), Haikou, China. Die Arbeit nennt 66 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Introduction
Shading regulation is crucial for cultivating and restoring endangered Hopea hainanensis saplings. However, the high variability in light conditions introduced by shading gradients can increase noise in visible and near-infrared (VNIR) canopy hyperspectral reflectance and exacerbate the domain shift between measured spectra and those simulated by radiative transfer models (RTMs). In addition, dry leaf density (DLD) retrieval from VNIR spectra remains challenging. These limitations restrict the robust nondestructive monitoring of physiological and structural responses of saplings to shading.
Methods
We developed an inversion framework that integrates multiple spectral scale-normalization methods with a hybrid RTM–deep learning modeling approach under controlled shading gradients. This framework was designed to improve cross-domain spectral consistency under shading-induced variability and to introduce a new carbon-based constituent index, DLD CBC , as a VNIR-compatible proxy for DLD.
Results
Inter-group shading differences in original reflectance were easily obscured by noise, whereas scale normalization substantially enhanced the shading response in the green peak and red-edge regions and better-preserved physical consistency between RTM-simulated and field-measured data. The hybrid modeling achieved the highest accuracy for chlorophyll (R 2 = 0.820; RMSE = 3.390 μg·cm -2 ), while nitrogen retrieval remained limited under VNIR (R 2 = 0.404; RMSE = 0.207). DLD CBC exhibited a consistent trend with measured DLD (r 2 = 0.747; RMSE = 4.867 mg·cm -2 ). Both the inverted chlorophyll and DLD CBC can effectively reproduce the directional response of measured traits along the shading gradient.
Discussion
This study provides a practical VNIR-based pathway for nondestructive and robust retrieval of key biophysical traits of H. hainanensis saplings under highly variable shading conditions.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Hainan Academy of Forestry (Hainan Academy of Mangrove), Haikou, China.
- The Innovation Platform for Academicians of Hainan Province, Haikou, China.
- Research Institute of Forest Resource Information Techniques, Chinese Academy of Forestry, Beijing, China.
- Precision Agriculture Lab, School of Life Sciences, Technical University of Munich, Freising, Germany.
Zitieren
APA
Chen, L., Yang, X., Dong, X., Lin, L., Shi, M., Chen, F., Huang, C., Yu, H., Yuan, Y., & Han, M. (2026). Robust retrieval of biophysical traits under shade gradients in young Hopea hainanensis using visible and near-infrared hyperspectral hybrid modeling. Frontiers in Plant Science, 17. https://doi.org/10.3389/fpls.2026.1788941
Vancouver
Chen L, Yang X, Dong X, Lin L, Shi M, Chen F, et al. Robust retrieval of biophysical traits under shade gradients in young Hopea hainanensis using visible and near-infrared hyperspectral hybrid modeling. Frontiers in Plant Science. 2026;17. doi:10.3389/fpls.2026.1788941
BibTeX
@article{Chen2026,
author = {Chen, Lin and Yang, Xiaoli and Dong, Xiaona and Lin, Ling and Shi, Mengmeng and Chen, Feifei and Huang, Chuanteng and Yu, Huilin and Yuan, Ying and Han, Miaoyi},
title = {Robust retrieval of biophysical traits under shade gradients in young Hopea hainanensis using visible and near-infrared hyperspectral hybrid modeling},
journal = {Frontiers in Plant Science},
year = {2026},
volume = {17},
doi = {10.3389/fpls.2026.1788941}
}
EUVIMED