Estimating time series phytoplankton carbon biomass: Inter-lab comparison of species identification and comparison of volume-to-carbon scaling ratios

• Large variation among six taxonomists for carbon biomass estimation.
• Different volume-to-carbon conversion methods do not on average produce large differences in total phytoplankton biomass estimates.
• Uncertainty in community biomass was between 46–49% when fixed cell volumes was compare to cell volumes determined by sample.
• Diatom cell volume were the biggest source of uncertainty (114–109%.) because of the seasonal variations that typify of diatoms.
• Cell volume should be determined for diatoms by sample or alternatively by month to correct for the seasonal variation in their cell volume.

An inter-calibration exercise was conducted to assess the performance of six phytoplankton taxonomists working within the Danish National Aquatic Monitoring and Assessment Program (DNAMAP). For species abundance and cell volume, a 2-fold difference was found among different estimates for subsamples from the same sample, which in turn cascaded into large differences in the species-specific carbon biomass contribution. The mean total carbon biomass estimated showed high variability (CV 43%) among the six taxonomists, but large variations were present within results produced by individual taxonomists (CV 8–50%), and one of the taxonomists produced significantly lower estimates than the others. Using data from phytoplankton time series samples, we also assessed the effect using a table of species-specific cell volumes versus cell volume measurements from a sample on carbon biomass values. For an example, the older cell-volume-to-carbon conversion method with fixed carbon-conversion constants was compared to the more recent approach of scaling biovolume to carbon biomass based on established regressions. We found that the regression between community biomass estimated by the old method versus the more recent equation yielded a slope close to 1, thus indicating general similar community biomass estimated between the methods. Type II regression suggested a high degree of variability in the estimates (17%). The highest degree of uncertainty was found by type II linear regression, when we compared the community biomass of diatoms estimated by cell sizes measured by sample to diatom community biomass estimated from cell sizes from a table of fixed cell sizes. In this analysis variation among methods for carbon estimation of individual samples was as high as 114%. Therefore, we recommend that, particularly for diatoms, cell volumes should be determined from the sample, or that table values be based on monthly estimates for at least the dominant diatom species for each study area.