| 01 Dec 2025
A potential explanation for the anomalously low nitrate to phosphate ratio in the well-oxygenated East/Japan Sea
Abstract. The East/Japan Sea (EJS), a well‑oxygenated marginal sea, exhibits an anomalously low nitrate (NO3⁻) to phosphate (PO43−) ratio (~ 2.6:1), diverging from the canonical Redfield ratio (16:1). To resolve this long‑standing biogeochemical enigma, we examined nitrogen (N) cycling genes and bacterial communities across depths and seasons. External phosphorus inputs – riverine, atmospheric, and crustal – were insufficient to explain the imbalance. Instead, high abundances of N‑reducing genes and affiliated taxa suggest a plausible role for bacterially mediated N loss throughout the water column. We propose that N removal may occur within particle-associated microenvironments (i.e., oxygen-depleted microzones inside sinking organic aggregates), despite oxygen-rich conditions. A dual-scale feedback – short-term anthropogenic N deposition enriching surface waters (i.e., increasing NO3⁻:PO43− ratio in the upper waters) and longer-term deoxygenation, driven by the weakening of deep-water formation, potentially favoring subsurface N loss (i.e., decreasing NO3⁻:PO4³⁻ ratio in the deep waters) – may promote a vertically stratified NO3⁻:PO43− regime in the future EJS. Our findings highlight the EJS as a sentinel system for how combined anthropogenic and climatic forces could reshape marine nutrient balances.
The paper «A potential explanation for the anomalously low nitrate to phosphate ratio in the well-oxygenated East/Japan Sea » by Kim et al. attempt to explore the contribution of bacterially mediated N loss for explaining the low N:P ratio encountered in the East/Japan Sea. The study is based on the hypothesis that N reducing pathways, i.e. nitrate reduction and denitrification, may occur at relevant magnitude in their study area (despite the well oxygenated conditions encountered), enough to explain low NO3-PO4 ratios. To test their hypothesis, the authors use sequencing data from samples collected between the surface and 1000 m depth during 5 cruises in the EJS, between February and October 2021. 16S sequencing was used to build up a taxonomic classification of bacterial communities at the order level. Furthermore, they applied a computational tool to infer functional properties of the obtained bacterial communities in an attempt to analyse the relative abundance of N-reducing genes.
I am not a specialist in genomics approaches but (or maybe because of that) I find that the Methods section lack literature references (with doi) for applied protocols, bioinformatics tools/pipelines and used databases. Hopefully, other reviewers more proficient in the field may assess whether the used methodology is sound.
This said, the main and critical caveat of this work is, in my opinion, that the conclusions are not supported by the data and the interpretation of the results is sometimes way too speculative. The authors base their conclusions in a series of numbers (percentages of N-reducing genes and associated bacterial communities) which do not give sufficient quantitative information. Throughout the manuscript, the authors state that they found « high » abundances of N-reducing genes ; but high compared to what ? And how can we know that the obtained percentages are « enough » to explain the observed N:P ratios ? Moreover, seasonal and vertical variability is presented without really assessing the reasons and consequences of this variability.
Nutrient data collected during the cruises are exploited only superficially although they should be at the center of their study. In the current version, all nutrient data have been merged to produce a unique value of NO3:PO4 ratio which is not presented in the results section but in the introduction. I strongly suggest to present the nutrient profiles for each cruise to explore seasonal and vertical variability in the same way as the genomics data. This may allow to compare the variability between nutrients and genomics data which may help assessing the role of N-reducing pathways in N loss in the water column.
The Results and Discussion section reads more like a literature review of the factors that may drive N:P ratios in the study area and I found that the results are not discussed enough. Moreover, there are a number of inconsistencies in the arguments given. For instance, the first part of the section claims that external phosphorus inputs are insufficient to explain low N:P ratios but later on, the authors state that N-enriched atmospheric deposition is responsible for high N:P ratios in surface waters. Again, presenting nutrient profiles will be very useful to understand whether N:P ratios are indeed higher in the surface that in deeper waters.
All the above are suggestions to help the authors derive clear findings from their hypotheses and data, and present substantial scientific insights. I suggest a thorough revision of the paper, particularly the Results and Discussion section, but I must say that I am not sure if such a thorough revision will finally make the manuscript suitable for publication in Biogeosciences.
Specific comments:
Lines 57-59. If low N:P ratios are a known feature of the EJA, I suggest the authors refer to literature data in the introduction and present their own data in the results section in a more detailed way.
Lines 64-66. Since genomics-based techniques are experiences rapid advances, the authors may want to cite more recent references.
Line 81. I guess it was filters and not filtered samples that were frozen and used for DNA extraction.
Line 83. Same comment as above.
As said before, the results and discussion section is awkward. Indeed, only 3.2. sub-section is properly results and a bit of light discussion. Sub-section 3,1 reads more like a literature review.
Lines 139-140. Please, add references for the selected biomarker genes.
Figure 2. Is there a possibility of adding error bars?
Line 169. Aren’t some of these orders (e.g. Alteromonadales, Pseudomonadales) ubiquitous in the ocean?
Line 178-179. The statement that proportions of 37.7% of N-reducing genes and 20.0% of taxa indicate the potential for bacterially mediated N loss is not very convincing. How these numbers compare with other regions?