Estimating the spatial distribution of zooplankton biomass by combining Video Plankton Recorder and single-frequency acoustic data

Estimation of zooplankton biological parameters such as biomass distributions from single-frequency acoustic surveys of zooplankton is generally not possible because of the diversity of taxa present in assemblages and resultant ambiguities in acoustic analysis. We present an approach that assumes that at a given, fixed depth, the relative proportions of sizes across species remains constant. Using this assumption, along with independently sampled Video Plankton Recorder (VPR) data on the species and size composition of the sound scattering assemblage, single-frequency acoustic data have been used to estimate the spatial distribution of zooplankton taxa and their biomass within a region of ocean without the use of acoustic scattering models. The data from a joint-ship study of the fine-scale physical and biological structure of a 3.5 km x 4.5 km x 75 m deep region of Georges Bank were used to demonstrate the utility of this approach. The spatial distributions of zooplankton biomass were estimated for pteropods, copepods, amphipods, euphausiids, chaetognaths and fish. Within each of the six parallel cruise legs that constituted the joint-ship survey, we estimated the relationship between mean zooplankton concentration from the VPR and mean acoustic volume backscatter within each 2 m depth bin. These relationships were used to scale the volume backscatter data throughout the transect into estimates of total zooplankton concentrations. The relative proportion of each taxon was estimated from the VPR for each 2 m depth bin within each leg and used to convert total zooplankton concentrations to taxon-specific concentrations. We then estimated biomass for each taxon using VPR estimates of mean sizes along the transect in conjunction with length-wet mass morphometric relationships. Several large acoustic patches were detected within the surveyed volume of water. Volume backscattering within the 10-20 m depth strata was largely associated with copepod biomass, while deeper scattering appeared to be a consequence of pteropod biomass. Amphipods contributed to scattering in the upper 10 m of the sixth leg of the grid. Euphausiids made minor contributions to the observed volume backscattering, while scattering due to chaetognaths and fish appeared to be negligible. Our results suggest that these patches were primarily due to the presence of copepods and pteropods. Our biomass estimates were highly correlated with data from a 1 m(2) MOCNESS tow in leg 6; however, they predicted only 19-29% of the measured silhouette biomass from the net tow. This discrepancy was explained by a systematic underestimation of zooplankton concentrations by the semi-automatic image analysis system used to process VPR data. The results described herein suggest that the combination of independent data from optical samplers, nets or pumps and single frequency acoustic surveys may provide a viable method for generating near real-time estimates of zooplankton biomass distributions. (C) 1998 Elsevier Science Ltd. All rights reserved.