Remote sensing-based estimation of gross primary production in a subalpine grassland

[摘要] This study investigates the performances in a terrestrial ecosystem ofgross primary production (GPP) estimation of a suite of spectralvegetation indexes (VIs) that can be computed from currently orbitingplatforms. Vegetation indexes were computed from near-surface fieldspectroscopy measurements collected using an automatic system designedfor high temporal frequency acquisition of spectral measurements inthe visible near-infrared region. Spectral observations were collectedfor two consecutive years in Italy in a subalpine grassland equippedwith an eddy covariance (EC) flux tower that provides continuousmeasurements of net ecosystem carbon dioxide (CO₂) exchange (NEE)and the derived GPP.

Different VIs were calculated based on ESA-MERIS and NASA-MODISspectral bands and correlated with biophysical (Leaf area index, LAI;fraction of photosynthetically active radiation intercepted by greenvegetation, fIPAR_g), biochemical (chlorophyllconcentration) and ecophysiological (green light-use efficiency,LUE_g) canopy variables. In this study, the normalized differencevegetation index (NDVI) was the index best correlated with LAI andfIPAR_g (r = 0.90 and 0.95, respectively), the MERISterrestrial chlorophyll index (MTCI) with leaf chlorophyll content (r = 0.91) and the photochemical reflectance index (PRI₅₅₁),computed as (R₅₃₁-R₅₅₁)/(R₅₃₁+R₅₅₁) with LUE_g(r = 0.64).

Subsequently, these VIs were used to estimate GPP using differentmodelling solutions based on Monteith's light-use efficiency model describingthe GPP as driven by the photosynthetically active radiation absorbedby green vegetation (APAR_g) and by the efficiency(ε) with which plants use the absorbed radiation to fixcarbon via photosynthesis. Results show that GPP can be successfullymodelled with a combination of VIs and meteorological data or VIsonly. Vegetation indexes designed to be more sensitive to chlorophyllcontent explained most of the variability in GPP in the ecosysteminvestigated, characterised by a strong seasonal dynamic ofGPP. Accuracy in GPP estimation slightly improves when taking intoaccount high frequency modulations of GPP driven by incident PAR ormodelling LUE_g with the PRI in model formulation. Similarresults were obtained for both measured daily VIs and VIs obtained as16-day composite time series and then downscaled from the compositingperiod to daily scale (resampled data). However, the use of resampleddata rather than measured daily input data decreases the accuracy ofthe total GPP estimation on an annual basis.

[学科分类] 地球化学与岩石