Distinct contribution of the blue spectral region and far-red solar-induced fluorescence to needle nitrogen and phosphorus assessment in coniferous nutrient trials with hyperspectral imagery
Accurate monitoring of plant nutrient status, especially nitrogen (N) and phosphorus (P) content, via remote sensing can facilitate precision forestry, with environmental and management benefits. In previous studies, plant traits derived from hyperspectral data via radiative transfer models (RTMs) a...
| Autores: | , , , , |
|---|---|
| Tipo de recurso: | artículo |
| Estado: | Versión publicada |
| Fecha de publicación: | 2025 |
| País: | España |
| Institución: | Consejo Superior de Investigaciones Científicas (CSIC) |
| Repositorio: | DIGITAL.CSIC. Repositorio Institucional del CSIC |
| OAI Identifier: | oai:digital.csic.es:10261/397800 |
| Acceso en línea: | http://hdl.handle.net/10261/397800 https://api.elsevier.com/content/abstract/scopus_id/105010564220 |
| Access Level: | acceso abierto |
| Palabra clave: | SIF Blue region Coniferous Far-red solar-induced fluorescence Hyperspectral imaging Leaf nitrogen Leaf nutrients Leaf phosphorus Plant traits |
| Sumario: | Accurate monitoring of plant nutrient status, especially nitrogen (N) and phosphorus (P) content, via remote sensing can facilitate precision forestry, with environmental and management benefits. In previous studies, plant traits derived from hyperspectral data via radiative transfer models (RTMs) and solar-induced chlorophyll fluorescence (SIF) effectively explained the observed variability in leaf N concentrations in crops. However, their contribution to leaf P concentration is unknown. Furthermore, such an approach might not be transferrable to coniferous stands, which are structurally complex and heterogeneous. We evaluated the potential of using physiological plant traits derived from airborne hyperspectral imagery to explain the observed variability in needle N and P concentrations in Pinus radiata D. Don (radiata pine) with four datasets collected over three years in established nutrient trials. RTM-derived data on pigment content in needles, including chlorophyll a + b (C<inf>ab</inf>), carotenoid (C<inf>ar</inf>), and anthocyanin contents (A<inf>nth</inf>), as well as SIF quantified at the O<inf>2</inf>A absorption band (SIF<inf>760</inf>), explained variability in N (R<sup>2</sup> = 0.67–0.97 and NRMSE = 0.07–0.30) and P concentrations (R<sup>2</sup> = 0.60–0.95 and NRMSE = 0.09–0.27) in needles. Although C<inf>ab</inf> was the most important predictor of needle N concentration (ranking C<inf>ab</inf> > A<inf>nth</inf> > SIF<inf>760</inf> > C<inf>ar</inf>), SIF<inf>760</inf> contributed the most to explain the variability of needle P concentration (SIF<inf>760</inf> > A<inf>nth</inf> > C<inf>ab</inf> > C<inf>ar</inf>). Moreover, the blue spectral region was essential for assessing P but not for explaining N variability in needles. Among all reflectance-based indices and inverted traits evaluated, the blue indices best explained the variability in needle P concentration, followed by C<inf>ab</inf>, C<inf>ar</inf>, and A<inf>nth</inf>. The study revealed the distinct contribution of far-red SIF vs. the blue spectral region for needle P compared to needle N, describing new insights for the physiological assessment of nutrient levels in forest stands using hyperspectral imagery. |
|---|