Detection of solar-induced fluorescence from lower resolution field-measured spectra over the past decade in deciduous broad-leaf forest in Japan

Abstract:

In this study 1) we determined a calibration factor for SIF of low spectral resolution (SR) data by an analysis of calculated SIFs correlations using ultrafine- and low-SR spectroradiometer data in deciduous forest on a mountainous landscape in Takayama (TKY), central Japan, 2) we calibrated the SIF retrieved from spectrum data over the past decade observed in Phenological Eyes Network (PEN), which installed low-SR spectroradiometer, and compared the calibrated SIF with satellite SIF products to evaluate the extent to which the satellite product represented the ground observation.

Both low and high SR spectra have been recorded by MS-700 (Eko limited inc: FWHM = 10 nm) and by QEPro (Ocean Optics, Inc., FWHM = 0.24 nm) at 10 minutes interval, respectively, in TKY. MS700 has been accumulating data since 2003. QEPro began acquiring quality data since 2019. SIF values were calculated by aFLD for low SR, and by spectral fitting model (SFM) for high SR data at 30 min interval from August to December 2019. Analyses were conducted only in stable sky condition and true reflectance (not contaminated by fluorescence) of less than 0.5. To investigate the influence of SR on SIF value, virtual spectra in same SR as MS-700 were also generated from QEPro data (=mimic MS-700). We used the global OCO-2 SIF data set (GOSIF) for 2008-2017 from the website (Li, X. & Xiao J. 2019).

The correlation between SIF from QEPro and those from mimic MS-700 showed a high correlation with r ² = 0.667 and a slope of 0.037. The SIF values obtained from MS-700 by aFLD showed a strong correlation with the reference SIF obtained from QEPro by SFM (r ² = 0.504). The linear slope (= 0.0428) was within the relatively same range of that of mimic MS-700 and QEPro. This result suggested that we can easily obtain accurate SIF values from low SR using a calibration factor. We calibrated SIF over past decade (2008-2018) of PEN data using the calibration factor. The monthly correlation coefficient in 2008-2017 between GOSIF and SIF was 0.858, with the satellite product overestimating the SIF. Satellites generally acquire passive data only at little cloudy day, which caused satellite SIF products to overestimate actual SIF value. In conclusion, the technique to calculate SIF from such low SR would allow for low-cost and long-term monitoring of SIF.