Hyperspectral remote sensing of chlorophyll a and pheophorbide a pigments for macrobenthos mapping on an estuarine tidal flat

• Pheob-a/Chl-a is an indicator of the specific macrobenthic community. • Pheob-a/Chl-a can be mapped by hyperspectral remote sensing. • Microphytobenthos has a spatiotemporally dynamic bottom-up effect on macrobenthos. Macrobenthos play critical roles in estuarine tidal flats through bioturbation, biodeposition, and bioirrigation and serve as key elements in the food web, depending largely on microphytobenthos (MPB) and serving as prey for birds. As traditional field-based methods for determining the spatial distribution of macrobenthos are costly and time-consuming, in this study we investigated the potential of inferring macrobenthos from degradation products of chlorophyll-a, a proxy for MPB biomass. First, we identified the macrobenthic groups most closely related to the pheophorbide-a and chlorophyll-a (Pheob-a/Chl-a) ratios via in situ data, although relationships were typically not significant. We then characterized the spectral characteristics of this ratio through controlled indoor experiments, integrated them into a hyperspectral algorithm, and mapped the spatial distribution of macrobenthos in the Western Scheldt estuary with this algorithm via PRecursore IperSpettrale della Missione Applicativa (PRISMA) hyperspectral satellite data. Specifically, the ratio was significantly (but weakly) related to the macrobenthic group of carnivorous/omnivorous/scavenger-feeding Malacostraca and bivalves and subsurface deposit-feeding Gastropoda, as verified by independent data, without showing significant relationships with many other groups. The blue spectrum was sensitive to Pheob-a/Chl-a, and a stacking model leveraging this spectrum was used to estimate the ratio. The results indicated substantial spatial heterogeneity in the estimated Pheob-a/Chl-a from the PRISMA images. The spatiotemporal relationships between these macrobenthic groups and MPB varied. Thus, while the method is not suited for macrobenthos mapping, our proposed approach paves the way for further research using hyperspectral imagery for ecological assessment, with its potential and limitations discussed.