Picocyanobacteria (Prochlorococcus and Synechococcus) play an important role in primary production and biogeochemical cycles in the subtropical and tropical Pacific Ocean, but little biological information on them is currently available in the North Pacific Ocean (NPO). The present study aimed to determine the picocyanobacterial contributions to the total primary production in the regions in the NPO using a combination of a dual stable isotope method and metabolic inhibitor. In terms of cell abundance, Prochlorococcus were mostly dominant (95.7 ± 1.4%) in the tropical Pacific region (hereafter, TP), whereas Synechococcus accounted for 50.8%–93.5% in the subtropical and temperate Pacific region (hereafter, SP). Regionally, the averages of primary production and picocyanobacterial contributions were 11.66 mg C m⁻²·h⁻¹ and 45.2% (±4.8%) in the TP and 22.83 mg C m⁻²·h⁻¹ and 70.2% in the SP, respectively. In comparison to the carbon, the average total nitrogen uptake rates and picocyanobacterial contributions were 10.11 mg N m⁻²·h⁻¹ and 90.2% (±5.3%) in the TP and 4.12 mg N m⁻²·h⁻¹ and 63.5%, respectively. These results indicate that picocyanobacteria is responsible for a large portion of the total primary production in the region, with higher contribution to nitrogen uptake rate than carbon. A long-term monitoring on the picocyanobacterial variability and contributions to primary production should be implemented under the global warming scenario with increasing ecological roles of picocyanobacteria.

The CHEMTAX program has been widely used to estimate community composition based on major pigment concentrations in seawater. However, because CHEMTAX is an underdetermined optimization algorithm, underdetermined bias has remained an unsolved problem since its development in 1996. The risk of producing biased results increases when analyzing the picophytoplankton community; therefore, this study tested a new method for avoiding biased CHEMTAX results using the picophytoplankton community around the East Sea (Japan Sea). This method involves building a linear model between pigment concentration data and community composition data based on DNA sequencing to predict the pigment range for each operational taxonomic unit, based on the 95% prediction interval. Finally, the range data are transformed into an initial ratio and ratio limits for CHEMTAX analysis. Three combinations of initial ratios and ratio limits were tested to determine whether the modeled initial ratio and ratio limit could prevent underdetermined bias in the CHEMTAX estimates; these combinations were the modeled initial ratio and ratio limit, the modeled initial ratio with a default ratio limit of 500 s, and an initial ratio from previous research with the default ratio limit. The final ratio and composition data for each combination were compared with Bayesian compositional estimator-based final ratio and composition data, which are robust against underdetermined bias. Only CHEMTAX analysis using the modeled initial ratio and ratio limit was unbiased; all other combinations showed significant signs of bias. Therefore, the findings in this study indicate that ratio limits and the initial ratio are equally important in the CHEMTAX analysis of biased datasets. Moreover, we obtained statistically supported initial ratios and ratio limits through linear modeling of pigment concentrations and 16s rDNA composition data.