Comparative analysis of field and laboratory spectral reflectances of benthic diatoms with a modified Gaussian model approach

Benthic diatom biomass on an intertidal mudflat was estimated by field spectrometry, a non-intrusive optical method operating in the visible-infrared wavelength range. Spectral reflectance (400-900 nm) of natural assemblages was related to the amount of principal photosynthetic and accessory pigments measured by high-performance liquid chromatography (HPLC) in the top 2 mm of sediment. Relationships established in situ were compared with those obtained in the laboratory with monospecific cultures. However, to compare both types of reflectance spectra differing in their overall shape (continuum) and to isolate the pigment absorption features from other effects, we applied hyperspectral data processing, the Modified Gaussian Model (MGM), to remove the continuum and model the main absorption bands by a succession of Gaussian curves. The Gaussian band depths associated with the absorption by chlorophyll a and the pigments specific to diatoms (chlorophyll c, fucoxanthine, diadinoxanthin) displayed linear relationships with the logarithm of chlorophyll a. The most relevant spectral feature to quantify the microalgal biomass in situ was the 632 nm absorption band associated with chlorophyll c. A significant but equivocal statistical relationship was obtained at 675 nm, due to the overlapping absorption by chlorophyll breakdown products (pheophytine a and pheophorbide a) present in all the samples analyzed in the field. Fucoxanthin absorption at 550 nm can become an indicator of chlorophyll a biomass for benthic diatoms, but the effect of seasonal photophysiological adaptation should be considered. The comparative analysis of field and laboratory reflectance showed that the chlorophyll a present in the top 2 mm was not a good estimator of the photosynthetic active biomass (PAB) and that a shallower depth should be sampled in muddy sediments colonized by microphytobenthos. The approach of MGM band depth retrieval offers the possibility of quantifying PAB and has the potential to characterize (by accessory pigments) microalgae present in the photic zone of different sediment types, removing albedo variability due to grain-size scattering effects or sediment moisture content. Its accuracy relies, however, on an appropriate ground-truth sampling, excluding the contribution of redundant pigments not detected by the sensor. (c) 2006 Elsevier B.V. All rights reserved.