Detecting plant physiological changes under stress using sun-induced chlorophyll fluorescence from mid-spectral resolution imagery

Abstract

Sun-induced chlorophyll fluorescence (SIF) serves as a valuable indicator for remote sensing of plant physiology. Recent studies have shown that far-red SIF at 761 nm (hereafter SIF) from mid spectral resolution (SR) imagers (3–7 nm full width half maximum (FWHM)) derived reliable fluorescence emission in relative levels. Such mid SR based SIF (SIF_MS; the subscript MS refers to mid SR basis) has been extensively utilized for assessing fluorescence variability in precision agriculture and plant stress detection across various platforms. However, the uncertainties regarding the ability of SIF_MS to detect physiological SIF yield (ΦF) remain inadequately investigated. Here, we evaluated various retrieval techniques (i.e., Fraunhofer line depth (FLD), its variant methods, and spectral fitting method (SFM)) under theoretical condition (SCOPE simulation), stress condition (DCMU experiment at field level and rice brown spot disease detection at drone level), and at airborne level employing mid SR (3–7 nm FWHM) and sub-nanometer SR (< 0.2 nm FWHM) sensors. Under theoretical conditions, the SIF_MS derived from aFLD, 3FLD and SFM showed strong linear relationships with the simulated fluorescence at 761 nm (R² > 0.9; RMSE <0.5 mW m⁻² nm⁻¹ sr⁻¹). With data collected under stress conditions, however, only the SFM method captured the stress-induced physiological changes in the field-ΦF_MS (p-value <0.05) and from the drone-ΦF_MS (p-value <0.05; R² = 0.51). SFM also produced a robust and clear map of airborne-SIF_MS (SIF_MS = 1.03 x SIF_SN + 0.48 where the subscript SN refers to sub-nanometer SR). Conversely, 3FLD and aFLD methods yielded a limited ability to adequately detect the physiological changes in field-ΦF_MS (p-value>0.05) and drone-ΦF_MS (p-value <0.05; R² ≤ 0.15) and exhibited a noticeable stripe noise in the airborne-SIF_MS map. Our finding demonstrated that mid SR imagery with SFM has the potential to track crop ΦF changes induced by stresses.