Reply on RC1

This manuscript investigates the underlying causes of extensive hypoxia in 2018 in the Gulf of Riga, using data from the regular monitoring program, a continuous profiling system at the deep part of the basin and a specific survey in September 2018. The authors conclude that the 2018 hypoxia was caused by a combination of several factors: 1) High freshwater and nutrient inputs in autumn 2017 and January 2018 promoting high productivity in the system, 2) inflow of saline waters from the EGB early in the year, due to unusual wind patterns, forming a deep located halocline (deeper than normal resulting in a smaller water volume in the NBL) that was maintained by north-easterly winds during summer, 3) rapid warming of the surface layer strengthening thermal stratification, and 4) reduced ventilation of the NBL during summer. These conclusions are not surprising and basically confirmatory to our present understanding of processes governing hypoxia.

As such, I concur with the authors in their conclusions, but it really surprised me that it takes 32 pages to underpin these conclusions. Unfortunately, the manuscript does not keep a clear stringent structure, as it is filled with repetitions and unnecessary details not used for supporting the conclusions. For example, the results section presents many and elaborate analyses (17 pages) without it being clear how these results support the conclusions. Reading the results section felt more like reading a WQ status report from an environmental agency, where all data have to be presented -relevant or not. The weak point is that the results section does not guide the reader towards the main conclusions! As I see it, the main problem is that the manuscript is not structured around clearly formulated hypotheses that are subsequently investigated in detail. The authors state that the objective is 'to evaluate the possible role of different forcing factors leading to the observed hypoxia'. I would strongly recommend that all the possible factors/explanations are outlined in detail with appropriate referencing to other literature studies in the introduction and that the M&M section describes how each of these hypotheses will be investigated with rigorous data analyses to address each of them separately. The results section should present only the analyses relevant to the hypotheses and finally, the discussion should centered around relevant scientific discussion points instead of repeating the results. This will require substantial rewriting, but I am also confident that the outcome will be more appealing to the readership of Biogeosciences. I estimate that the main text could be reduced by half.
Response: Thank you very much for the comments! We will revise the manuscript according to your comments. The main questions of the study were: 1) What was the reason for the observed extensive near-bottom hypoxia in the Gulf of Riga in 2018? and 2) Was it an exceptional event, or is it a feature that could occur in the Gulf of Riga and similar basins regularly and/or even more often in the future? We did not have clear hypotheses when analysing the measurement data. Of course, it could be suggested that the observed extended hypoxia was probably related to some specific meteorological/hydrographic conditions. The enhanced input of nutrients is just in the background, and it alone does not create sudden hypoxia/anoxia in late summer/autumn. Strong stratification is something supporting hypoxia development. However, for the year 2018, we found that the most important was haline stratification in the deep layer already in spring. The deep stratification was maintained by the inflow of saltier waters from the open Baltic. It should also transport oxygen to the deep layer. However, we estimated that the consumption is higher than advection and diffusion that resulted in extended hypoxia. By analysing historical data, we showed that hypoxia is occurring more often. Extrapolating this finding and analysing meteorological data in 2018 in relation to the longterm averages, we could conclude that probably such hypoxia will occur in the future even more often. This is the paper in short. We will revise the text in the Results section focusing on these two questions and only presenting the data and analysis supporting the conclusions. Accordingly, we will revise the Introduction and Discussion sections. The text will be shortened. We hope it makes the presentation clearer.
In my reading I also found several unclear sentences and sentences that could be sharpened. I have listed some of these under my technical comments, but I stopped commenting on the language after the introduction, realizing that a major rewrite would be required. I do recommend that the next version of the manuscript is proofread by a native English speaker.
Response: Thank you. We will follow your suggestions, and when the manuscript is revised, we will order proofreading by a native English speaker.

Specific comments:
The introduction is quite long and contains very general, and occasionally trivial, information about processes related to hypoxia (almost textbook like). This information could/should definitely be shortened to present only the most relevant information that leads to the formulation of the objectives and research questions. I suggest that the authors outline all the possible causes underlying hypoxia in the Gulf of Riga, leading to the formulation of specific testable hypotheses. Moreover, I do not think the introduction presents a stronger motivation for the study. I hope the author can present information that explains why 2018 is particularly interesting and why it is relevant to consider a single year. Are years like 2018 expected to be more frequent under current climate change scenarios -i.e. are we expecting more such events to occur in the future? Did 2018 have any ecological consequences such as fish kills, loss of benthic fauna, etc. In summary, the introduction needs to be terser.

Response:
We will revise and shorten the Introduction section. We will skip the trivial sentences and keep the most relevant text to two main scientific questions (mentioned above). We did not know about the presence of deep haline stratification to hypothesize its importance. However, with the present knowledge, the presentation of the study questions, suggestions, methods and results can be improved (as suggested).
There were no specific ecological consequences. However, the release of phosphates from the sediments is triggered by such hypoxia and counteracts the external nutrient load reduction.
It would improve the readability of the materials and methods section, if the different data analyses were more clearly linked with specific hypotheses stated in the introduction.
Response: We will revise this section and try to relate the methods to the questions.
The calculation of oxygen consumption rates are based on simple box model assumptions, but these calculations are also very sensitive to small differences in salinity bw stations 114 and G1 (the divisor in Eq. 1). The authors should comment on this and how a small potential bias in using values at stations 114 to characterise the inflow of saltier EGB water could influence the calculated rates for oxygen and phosphorus.
Response: Yes, we agree. We were able to estimate the uncertainties regarding the variability at station G1 (central Gulf of Riga). The same applies to profiles at station 114 (Irbe Strait). However, the measurements at station 114 could miss the inflowing water mass. Thus, an additional uncertainty factor is present. We will discuss this point in the revised text.
On page 8, it is described that a decline in oxygen concentration should be expected when physical processes are taken into consideration. Does this mean that the authors discarded observations that did not exhibit a decline in oxygen after adjusting for physical processes? If yes, this would bias oxygen consumption rates to higher values, as negative values can be expected by shear randomness. This needs to be clarified.
Response: Physical processes (advection and diffusion) should cause an increase in nearbottom layer oxygen concentration. We estimated the expected oxygen concentration and assigned the difference in the expected oxygen concentration and the observed oxygen concentration to oxygen consumption. We did not find negative consumption values based on the measurements (profiles) with a time step of 1-1.5 months.
On page 9, trends in oxygen and phosphate concentrations are investigated, but why are the authors interested in trends? What do they expect? This is one of many examples, where the formulation of a hypothesis would improve the storyline. Are the authors expecting that expanding hypoxia in the EGB will have an effect on the Gulf of Riga and increase the likelihood of spilling over?
Response: We will explain it better. The trends were analysed to determine whether the hypoxic conditions and high phosphate concentrations have occurred more often recently. It answers the question of whether hypoxia observed in 2018 agrees with the long-term changes. In addition, the finding is that the trends indicate worsening of conditions despite no increase in external nutrient load. Thus, we could conclude that meteorological/hydrographic conditions, e.g., prolonged stratified season, should be responsible.
The first paragraph of Section 3.1.2 (Page 10) presents changes over time in the physical parameters in the Ruhnu Deep. From reading, it is not clear why all this information (and with the high level of detail) is presented. Parts of the paragraph are trivial and the text could easily be reduced substantially (e.g. the two first sentences could be removed).
Response: We will shorten the text keeping only the relevant information (regarding the main questions).
On page 12 first paragraph, many numbers are presented, but why are these numbers relevant for the storyline.
Response: We will shorten the text keeping only the relevant information (regarding the main questions).
On page 12 second paragraph, the authors assess the uncertainty of the areal estimates of hypoxia by looking at the distribution of the depth of the hypoxia threshold value. Since the authors have many profiles that are spatially distributed, why didn't they investigate the spatial distribution of the threshold value to see if the depths are horizontally 'constant' over the domain? This would be a more meaningful analysis. Moreover, it is not clear how the authors will use their uncertainty estimate! What is the purpose of this calculation, if it is not used for substantiating the arguments later?
Response: We will skip this part. It is relevant to evaluate the uncertainty of the method (when just one profile in the central gulf is used to determine the hypoxic area extent). However, we agree that it is not important for this paper. Section 3.2.1: why are the authors presenting all this information on wind patterns? It would be easier to read if the authors formulated a hypothesis about which wind patterns promote hypoxia and then investigate these.

Response:
We will revise and shorten the text linking the analysis to the suggestion that north-easterly winds could create the inflows of saltier waters (and thus, deep layer stratification in the gulf) and low wind speed and high irradiance/air temperature could support vertical stratification and hinder vertical mixing. Section 3.2.2: Again, explain why these data are interesting! River discharge data should be presented in a more hypothesis-driven context. Describe the expectations for the data and underpin with analyses.
Response: We will revise and shorten the text linking the analysis to the suggestion that large river discharge could bring more nutrients and organic matter to the gulf. Section 3.2.3: Same comment.
Response: See answer regarding Section 3.2.1. Section 3.3.2: The authors use 1.5 page of text to describe different profiles. This section is longwinded and should be shortened. As an example on page 18 (L. 5-13), an entire paragraph is used to explain that stratification was stronger in 2018 than in 2017. This could be said with a single sentence. The whole section could easily be reduced to less than half size.

Response:
We agree and will revise this entire section.
The discussion is primarily a repetition of the introduction and results sections, and it doesn't read like a discussion section. It is important that the authors bring up pertinent research questions and treat these from the angle: -what do we know, what has this study shown and what can we learn? This approach to the discussion would also highlight the novelty of the study. I believe a rewriting of the discussion to follow the general style of a discussion would be needed.