Sun-induced chlorophyll fluorescence and photochemical reflectance index improve remote-sensing gross primary production estimates under varying nutrient availability in a typical Mediterranean savanna ecosystem

摘要

This study investigates theperformances of different optical indices to estimate gross primaryproduction (GPP) of herbaceous stratum in a Mediterranean savanna withdifferent nitrogen (N) and phosphorous (P) availability. Sun-inducedchlorophyll fluorescence yield computed at 760 nm (Fy760),scaled photochemical reflectance index (sPRI), MERIS terrestrial-chlorophyllindex (MTCI) and normalized difference vegetation index (NDVI) were computedfrom near-surface field spectroscopy measurements collected using highspectral resolution spectrometers covering the visible near-infrared regions.GPP was measured using canopy chambers on the same locations sampled by thespectrometers. We tested whether light-use efficiency (LUE) models driven byremote-sensing quantities (RSMs) can better track changes in GPP caused bynutrient supplies compared to those driven exclusively by meteorological data(MM). Particularly, we compared the performances of different RSMformulations – relying on the use of Fy760 or sPRI as a proxy for LUE and NDVIor MTCI as a fraction of absorbed photosynthetically active radiation (fAPAR)– with those of classical MM.Results showed higher GPP in the N-fertilized experimental plots during thegrowing period. These differences in GPP disappeared in the drying periodwhen senescence effects masked out potential differences due to plant Ncontent. Consequently, although MTCI was closely related to the mean of plantN content across treatments (r2 = 0.86, p < 0.01), it was poorly relatedto GPP (r2 = 0.45, p < 0.05). On the contrary sPRI and Fy760 correlatedwell with GPP during the whole measurement period. Results revealed that therelationship between GPP and Fy760 is not unique across treatments, but it isaffected by N availability. Results from a cross-validation analysis showedthat MM (AIC_cv = 127, ME_cv = 0.879) outperformed RSM(AIC_cv =140, ME_cv = 0.8737) when soil moisture was usedto constrain the seasonal dynamic of LUE. However, residual analysesdemonstrated that GPP predictions with MM are inaccurate whenever no climaticvariable explicitly reveals nutrient-related changes in the LUE parameter.These results suggest that RSM is a valuable means to diagnosenutrient-induced effects on the photosynthetic activity.