Author replies in italics.

The Hieronymus et al., have done through approach toward modeling phytoplankton and N2 fixation in Baltic Sea. The model seems to be well constructed developed upon the accumulated bodies of modeling and certainly provides new aspects of the regional modeling targeting this area. One challenge I had was to see how the cyanobacterial life cycle is simulated. A schematic and narrative would be useful. That said, I like how explicitly the filamentous bacteria is simulated, which is the unique feature of Baltic Sea. The manuscript is suitable Biogeosciences. The following are my comments hoping to improve the manuscript.

Authors: Thank you for your thorough review and for pointing out that we need to be more clear in explaining the life cycle part of the model. We will include a detailed schematic figure of the CLC model in parallel to current figure 2 in the revised version of the manuscript and clarify the description of it where needed (see the enclosed New Figure 2 for a first version). 

Main text:

L20: in timing of -> in the timing of

Authors: This will be corrected in the revised version of the manuscript. 

L21: runs we -> runs, we

Authors: This will be corrected in the revised version of the manuscript.

L28: of its -> of their

Authors: This will be corrected in the revised version of the manuscript. 

L30: nitrogen fixing -> nitrogen-fixing (there are other cases below, which I would not mention)

Authors: This will be corrected in the revised version of the manuscript. 

L46: in bloom formation -> in the bloom formation

Authors: This will be corrected in the revised version of the manuscript. 

L46: e.g. -> e.g.,   (there are other cases below, which I would not mention)

Authors: This will be corrected in the revised version of the manuscript. 

L54: In the model growing -> in the model, growing (for clarity)

Authors: This will be corrected in the revised version of the manuscript. 

L66: in abundance -> in the abundance

Authors: This will be corrected in the revised version of the manuscript. 

L73: gain understanding -> gain an understanding

Authors: This will be corrected in the revised version of the manuscript.  

L81: which has -> that has

Authors: This will be corrected in the revised version of the manuscript. 

L83: three dimensional -> three-dimensional

Authors: This will be corrected in the revised version of the manuscript. 

L88: the northern -> northern

Authors: This will be corrected in the revised version of the manuscript. 

L101: tense should be consistent

Authors: This will be corrected in the revised version of the manuscript. 

L112: remove ‘a’ or make it to ‘the’

Authors: This will be corrected in the revised version of the manuscript. 

L144: for the entire period -> for the entire period of (or put 1850 – 2008 in parenthesis)

Authors: This will be corrected in the revised version of the manuscript.  

L150: by very large burial -> by a very large burial (or ‘the’ very ...)

Authors: This will be corrected in the revised version of the manuscript. 

L175: For this work -> For this work,

Authors: This will be corrected in the revised version of the manuscript. 

L176: post processed -> post-processed

Authors: This will be corrected in the revised version of the manuscript. 

L186: includes also -> also includes

Authors: This will be corrected in the revised version of the manuscript. 

L189: true also -> also true

Authors: This will be corrected in the revised version of the manuscript. 

L193: A -> The (or remove ‘A’)

Authors: This will be corrected in the revised version of the manuscript. 

L197-210: There are different modeling experiment. I wonder which one is considered as default. Is there a model run that includes all the factors, which could be considered as default? I think it was done in the previous study? In any case, it would be useful to compare these sensitivity analysis to be compared with the default, so I suggest putting the results from the default along with these simulations.

 Authors: There is no default in the original model. Since there has not been any previous 3D modeling efforts of the Baltic sea where the CLC model has been included, and the original model had unlimited P availability, we needed to evaluate P dependency before estimating nitrogen fixation and cyanobacterial biomass. Therefore, several runs using different settings were performed with the aim of producing the best fit to observations and in order to understand the CLC in the Baltic Sea and its effect, and depence on, the nutrient composition. From these sensitivity runs, one was chosen as the best fit (weak P limitation) and used in the estimates. This will be clarified in the revised version of the manuscript, starting with background about P limitation in the Baltic Sea and the lack of P dependency as a setting in the model already in the introduction.  

L220: which -> ,which

Authors: This will be corrected in the revised version of the manuscript. 

L225: in this case -> in this case,

Authors: This will be corrected in the revised version of the manuscript. 

L227: generating -> , generating

Authors: This will be corrected in the revised version of the manuscript. 

L228: faster growing -> faster-growing

Authors: This will be corrected in the revised version of the manuscript. 

L240: we -> , we (for clarity and improving readability)

Authors: This will be corrected in the revised version of the manuscript. 

L245: I am having a hard time understanding what is meant by the life cycle model. Is that the diurnal cycle or is that longer cycle? A schematic (in addition to figure 2) and additional explanation (model summary with a few sentences) would be useful. I am suspecting  it is a seasonal cycle, so it would be nice if it is clearly defined here.

Authors: The life cycle model is seasonal. This will be clarified in the method section as well in the new schematic image that will be included in the revised version of the manuscript (please see reply above). 

L268: release -> the release

Authors: This will be corrected in the revised version of the manuscript. 

L301: we -> , we (for clarity and readability)

Authors: This will be corrected in the revised version of the manuscript. 

L301: seasonality of -> the seasonality of

Authors: This will be corrected in the revised version of the manuscript. 

L314: bloom forming -> bloom-forming

Authors: This will be corrected in the revised version of the manuscript. 

L320: however -> however,

Authors: This will be corrected in the revised version of the manuscript. 

Figure 2: Many different shapes are used. I wish to have a list of explanations for different shapes. Also, It is less clear where and how phytoplankton are represented. There seem to be multiple functional types of phytolankton but it is hard to see from the figure. I suggest another figure or panel to focus on phytoplankton functional types, as well as the life cycle of them since they seem to be key in the paper.

Authors: We agree that this figure is a bit difficult to follow and will revise it to become more clear. In a revised version of the figure, the colours of the parts belonging to the CLC-model will be changed (see first version enclosed, New Figure 3). These colours will in the revised manuscript match the new figure of the CLC (described above). It is also described in the figure legend that the red lines are indicating the flow between the CLC components. We will also, along with the figure, have a table with all the abbreviations (e.g., AKIW, AKIB) so it is easier for the reader to look back when needed, as well as explain the most important ones in the figure legend. We will however not describe the life cycles of the other phytoplankton groups in more detail since this is not the focus of the paper. There is a description in the figure legend that A1 and A2 stand for the functional groups “diatoms” and “flagellates and others”, respectively. 

Figure 3: Model seem to show much higher values than observations. I wonder what are the reasons. 

Authors: This difference will be further discussed in the revised manuscript. Figure 3 shows all different sensitivity runs and therefore also the runs with very high biomass values. However, the wPlim that we chose for the model results is not very far from the observations. We will provide a more detailed comparison in the revised version of the manuscript, where we provide mean values of the summer biomass peak as well as the seasonal span in a table so that it is easier to evaluate the runs to the observations. Also, observations have a maximum frequency of once every two weeks. This means that much information about peak biomass and variance is lost, and the highest values can sometimes be missed. 

Figure 4: I am personally curious about how the population of N2 fixers change.

Authors: We are a bit uncertain about this comment. Is it regarding how the community composition changes? Because the current figure, the upper panel, demonstrates how the whole population changes over time. Since the model only includes one group of cyanobacteria we can unfortunately not demonstrate how each taxa changes over time. 

Figure 5: The rate of nitrogen fixation seems to match despite the difference in biomass shown in Figure 3. I wonder what explains this. Also, I wish to see discussion on how Heterotrophic N2 fixation may alter the result.

Authors: The strong coherence between model results and observed nitrogen fixation is somewhat surprising given the larger cyanobacteria biomass displayed by all model experiments compared to observations (Fig. 3). There are several potential causes for the deviation in carbon biomass estimations from the model and observations. The cyanobacteria biovolume from observations was used with different presumptions to estimate the nitrogen fixation rates and to calculate carbon concentrations, respectively. The modelled nitrogen fixation is calculated during the run from the growth of HET in the CLC model while the carbon content of cyanobacteria is calculated by the Redfield ratio between nitrogen and carbon in HET with a minor contribution also from REC. Hence, there are uncertainties in the calculations of carbon biomass from both observations and from model results. It is not easy to change the Redfield C:N:P ratio that is used in the model since the results from the entire biogeochemical cycle including the oxygen consumption in the model is dependent on this ratio. There are other biogeochemical models with variable C:N:P ratios that might be used to analyze the impact from these processes further. Uncertainties in the comparison of models and observations stem also from the fact that observations are done on small water samples from an area that is covered by an average value from a 3.7 km x 3.7 km grid in the model.

Heterotrophic N2 fixation is extremely low in the Baltic Sea (e.g., Farnelid et al. 2013) and is therefore not included here. It has been demonstrated in the Baltic Sea that its three taxa that dominate the N2 fixation (Klawonn et al. 2016). Heterotrophic N2 fixation is neither included in the observations nor the model, and would probably not make any notable difference since it is so small. 

We will enhance the discussion of this in the revised manuscript. 

Figure 6: How do these compare to the model simulation?

Authors: These are from model simulations.

Figure 8: Could this be compared with observation?

Authors: The data in red is from the model simulation wPlim and in black from observations. This will be better clarified in the revised version of the manuscript both in the text and the figure legend. 

Supplementary material:

I wish to get the explanations behind (8). Why is it power of 4? Is that based on some previous studies?

Authors: The equations for light and temperature limitation are adapted to RCO-Scobi from the original model by Beckmann and Hense (2004) and Hense and Beckmann (2006).

I wish to get the reasoning behind (9). What is it formed with the addition of square termed in square root instead of the simple additions? Is that based on some previous studies?

Authors: See previous answer.

I wish to get some explanations behind equation (10) and (11), especially the reasoning of the mathematical formulas and qualitative interpretation of them.

Authors: Eq. (10) describes the transition from the recruiting and vegetative state (REC) to the diazotrophic state (HET). The maximum growth rate (s-1) of REC is larger than that of HET but the growth (mmol m-3 s-1) is, in the previous state, also dependent on nitrogen. When the growth of HET is larger than that of REC a transition to HET occurs. 

Eq. (11) describes the transition of HET to pelagic akinetes (AKIW). If the growth of HET is below a critical value, a transition to AKIW occurs.

We will deepen the model description and include a schematic of the CLC model in the revised manuscript. 

Other points for discussion: 

There are studies suggesting that heterotrophic bacteria may contribute to N2 fixation. I suggest considering discussing their effect on the overall N2 fixation in the Baltic Sea. The following papers may be useful: (Bentzon-tilia et al., 2015; Farnelid et al., 2013; Bentzon-Tilia et al., 2014; Chakraborty et al.; Pedersen et al., 2018).

Authors: We agree that heterotrophic N2 fixation should be mentioned and it will be included in the discussion in the revised version of the manuscript. However, since the N2 fixation rates by heterotrophic bacteria are extremely low in the Baltic Sea it would not affect the overall input of N2 fixation in the studied region (heterotrophic bacteria: 0.44 nmol l-1 d-1 in Farnelid et al. 2013 as compared to up to 800 nmol l-1 d-1 by filamentous cyanobacteria in Klawonn et al. 2016). 

N2 fixers (or nitrogenase) are known to be sensitive to O2. However, heterocysts have glycolipid layer which may protect them from O2. I think the hidden assumption in the model is that O2 does not matter to heterocysts. To support the assumption, the authors may consider citing (Inomura et al., 2017), as it shows that respiratory protection is not required for heterocysts; otherwise the rate of N2 fixation would be O2 dependent.

Authors: We agree that heterocysts are designed to protect them from O2 and therefore is this model assumption correct. The suggested paper includes a totally different species (soil bacterium) so if needed we prefer to cite something closer to our study organisms. However, since the filamentous cyanobacteria in our study are photosynthetic they produce their own oxygen, and therefore always have O2 present around their cells, and why changes in O2 concentration does not affect the nitrogen fixation rates.

References:

Bentzon-Tilia M, Farnelid H, Jürgens K, Riemann L. (2014). Cultivation and isolation of N2-fixing bacteria from suboxic waters in the Baltic Sea. FEMS Microbiology Ecology 88: 358–371.

Bentzon-tilia M, Severin I, Hansen LH, Riemann L. (2015). Bacteria Isolated from Estuarine Surface Water. Mbio 6: 1–11.

Chakraborty S, Andersen K, Visser A, Inomura K, Follows MJ, Riemann L. Quantifying nitrogen fixation by heterotrophic bacteria in sinking marine particles. Nature Communications Accepted.

Farnelid H, Bentzon-Tilia M, Andersson AF, Bertilsson S, Jost G, Labrenz M, et al. (2013). Active nitrogen-fixing heterotrophic bacteria at and below the chemocline of the central Baltic Sea. The ISME Journal 7: 1413–1423.

Inomura K, Bragg J, Follows MJ. (2017). A quantitative analysis of the direct and indirect costs of nitrogen fixation: a model based on Azotobacter vinelandii. The ISME Journal 11: 166–175.

Pedersen JN, Bombar D, Paerl RW, Riemann L. (2018). Diazotrophs and N2-Fixation associated with particles in coastal estuarine waters. Frontiers in Microbiology 9: 1–11.

Author reply references:

Beckmann, A., & Hense, I. (2004). Torn between extremes: The ups and downs of phytopiankton. Ocean Dynamics, 54(6), 581–592. https://doi.org/10.1007/s10236-004-0103-x

Hense, I., & Beckmann, A. (2006). Towards a model of cyanobacteria life cycle-effects of growing and resting stages on bloom formation of N2-fixing species. Ecological Modelling, 195(3–4), 205–218. https://doi.org/10.1016/j.ecolmodel.2005.11.018

Schneider B., Eilola K., Lukkari K., Muller-Karulis B., Neumann T. (2015) Environmental Impacts—Marine Biogeochemistry. In: The BACC II Author Team (eds) Second Assessment of Climate Change for the Baltic Sea Basin. Regional Climate Studies. Springer, Cham. https://doi.org/10.1007/978-3-319-16006-1_18

Meier, H.E.M., Eilola, K., Almroth-Rosell, E. et al. Disentangling the impact of nutrient load and climate changes on Baltic Sea hypoxia and eutrophication since 1850. Clim Dyn 53, 1145–1166 (2019). https://doi.org/10.1007/s00382-018-4296-y

Author comments in italics.

Major comments:

The paper conducts a 3-D modeling of Baltic Proper to simulate filamentous diazotrophic cyanobacteria and associated N2 fixation. The two major points that the authors appear to emphasize is that, the incorporation of cyanobacteria life cycle (CLC) dynamics and phosphorus dependence in the model greatly improve model performance in correctly simulating seasonality of the cyanobacteria biomass and N2 fixation. However, I found that the paper lacks clear focus, introduction, and thorough analyses. I was also puzzled by some results particularly from P dependence schemes.   

Authors: Thank you for this observation, we will revise the manuscript to become more focused and streamlined to the aims of the paper (to include CLC to a 3D model and perform phosphorus sensitivity runs to optimize specific to the Baltic Sea). Please see more detailed replies to the comments regarding the mentioned issues here below. 

(1) CLC appears to be one of the key issues the paper aims to resolve. Although the lead author and others have published a series of papers of CLC simulations, which making me not quite clear if CLC is still one of the key schemes that this paper would focus, at least both in the abstract (line 17-18) and conclusion (Line 301-304) the CLC is described as the main point of the paper. However, scientific background of life cycle of filamentous cyanobacteria is not sufficiently introduced in the paper, making it difficult to understand the different stages set in Method Section 2.3. Even in this section, CLC model part is not clearly described. There is Fig. 2 included in the paper which appears to be the structure of CLC and the model, but this figure is not referred and described. More importantly, model results of CLC (of different stages) are not shown. How the CLC improve the model is not analyzed and compared (such as to a model version without CLC or to the previous studies), except only two conclusing sentences (line 244-245).

Authors: It is true that CLC is the main focus of the paper since this is the first 3D modelling effort that includes the CLC in the Baltic Sea. Although the CLC model has been published before by two of the co-authors (Hense and Beckman 2006; 2010, Hense et al. 2013) this was neither combined with a 3D model nor specifically validated to observations from the Baltic Sea. In these previous publications there was free P availability, which is not true for the Baltic Sea, where P is very limiting for the filamentous cyanobacteria during the summer blooms together with weather conditions (e.g., Klawonn et al. 2016; Olofsson et al. 2016; Degerholm et al. 2006). This will be included in the revised version of the manuscript starting with the background of cyanobacteria life stages and P limitation in the introduction. We will clarify both that this is the first time CLC and 3D model is used as a combination in the Baltic Sea and why we need to evaluate P dependency before estimating biomass and nitrogen fixation in the revised version of the manuscript. 

We will also more thoroughly compare the model results of biomass and timing of bloom with a run that excludes CLC. 

(2) P dependence and weak P limitation (wPlim). I have problem to follow the key scheme (wPlim) that chosen as the main scheme. Half saturation concentration is very low (10^-6 nM), how it can be effective? Naturally there should be no complete absence of phosphate. The current half saturation of 10^-6 nM is extremely low, several order of magnitude lower below the detection limit. Indeed, Fig. 6. shows the lowest phosphate concentration is still much higher than 10^-6. How it can make substantial different results from noP as shown in such as Figs. 3 and 4? Even in these two Figures, Fig.3 and 4, wPlim results are not consistent (compared to other setups). For example, in Fig. 3 seasonal experiments, wPlim produce lower biomass for three of four sites (BY2\5\15). How in the Fig. 4 interannual variation, wPlim gives higher biomass?

Authors: The difference between noP and wPlim is that noP requires no P at all to bloom while wPlim blooms as long as P is present, even in tiny (~10^-6) concentrations. This means that when all P is consumed, the cyanobacteria can no longer grow. 

In the noP case, the end of bloom is completely dependent on the temperature and light availability (cf. Eq. (1), (2) and (6) in Table S3 and Table S5). Furthermore, there is no uptake or release of phosphorus by the cyanobacteria in this case which means that they do not affect the nutrient composition of the water column.

The relationship between the biomass and the P limitation scheme is not straight forward. It is not possible to say that wPlim will always generate the highest biomass compared to the other setups. The choice of P limitation not only affects the biomass but also the nutrient composition of ambient water which in turn affects the biomass of all functional types. Fig. 6 shows that during the early period, the phosphate concentrations are higher and the DIN lower compared to the other experiments generating higher cyanobacteria biomass. During the later period, DIN is completely depleted after the spring bloom in all experiments, while the phosphate is lowest in wPlim generating the lowest cyanobacteria biomass. Note that Fig 6 shows the mean seasonal cycle of phosphate over two different time periods at monitoring station BY15 only and says nothing about the overall minimum concentration. 

We will deepen this discussion in the revised manuscript.

Overall, the authors emphasize both CLC and wPlim schemes give better bloom timing (or seasonal pattern). However, even the “timing” of bloom is not quantitatively defined, and therefore the comparison and the conclusion they give better bloom timing is not supportive. For example, in Fig. 3, I cannot directly identify the difference of the start and end time of the bloom in each experiment. The authors should quantitatively define and show in numbers the start and termination of the blooms both in observations and the four experiments.

Authors: In the revised version of the manuscript we will include a table with mean peak Cyanobacteria biomass values and timing (date of the peak) of both model runs and observations and also the seasonal span of the blooms (between which dates) of the different model runs and observations, in order to more easily compare their differences. This comparison will also be more thoroughly discussed in the revised version. 

We will also compare the model results of biomass, date of bloom peak, dates between the bloom span with a run that excludes the CLC. 

Specific comments:

Section 2.2. Model structure Fig 2 should be referred and basic structure of the model sufficiently described.

Authors: We will include a modified version of Fig 2 in the revised manuscript that is clearer and also include a table to explain abbreviations. It will also be more clearly referred to in the text. Furthermore, we will include a schematic that more readily describes CLC.

Section 2.3. The text of this section should be reorganized to logically describe the stages and transitions.

Authors: We will include a new schematic of the CLC (see enclosed New Figure 2 for a fist version) parallel to the current figure 2 in the revised version of the manuscript. We will also revise the text of this section using the chronological order of the new CLC figure to help the reader understand the text of this section better and in a more logical way. 

Line 117-119: Two versions? Or is it the “simplified version” of the "modified Version"?

Authors: It is a mix of the original model of Hense and Beckmann (2006) and the simplified model of Hense and Beckmann (2010). We agree that this is confusing and need to explain and revise this sentence.

Line 125-126, logically it is not justified why the differences among the species cannot influence the main patterns?

Authors: Since we are using an average salinity and temperature preference range we can not look at specific regions where for example one of the taxa dominates because they might be outside the mean range. We will clarify this in the revised version of the manuscript. It might for example be difficult to apply our settings in a low salinity region of the Baltic Sea where we may have cyanobacteria that can grow in salinities down to 0, while the model has the lower range set to salinity 3. 

Line 132, the difference between AKIW and AKIB is not described.

Authors: Akinetes are pelagic (AKIW) or benthic (AKIB) and can be transferred between these reservoirs through sinking and resuspension. This is described in lines 132-133, but all abbreviations will be included in a new table in the revised version of the manuscript. 

Line 139-140: “For the transition between AKI (AKIB and AKIW) and REC we prescribe a fixed germination window - from April 20 to the end of April”: It is unclear. How the germination window defined? So, there is only one full life cycle each year? Before April 20, it is HET to AKI; and after end of April, it is always REC?

Authors: The germination is defined in Eq. (30) in Table S.3. Between April 20 and the end of April, germination occurs at a constant rate times the AKI concentration. The transition from HET to AKIW is defined in Eq. (11) in Table S.3. It is dependent on the temperature and occurs when the growth of HET has fallen below a critical value. There is thus only one full cycle each year. This will be clarified in the revised version of the manuscript and more easy to follow when referred to the new figure which will show the seasonal CLC. 

Line 146, “Growth of HET and REC are inhibited under anoxic conditions.” Why? Normally diazotrophs prefer anoxic conditions, right?

Authors: We do not understand why they would prefer anoxic conditions. As they spend their active life stage in surface waters where there is enough light for them to perform photosynthesis there is plenty of oxygen around all the time, as they also produce oxygen themselves. Heterotrophic diazotrophs might prefer other conditions but the organisms in this paper are all photoautotrophs. 

Line 147, What is difference between this salinity dependent window and the above-mentioned time window (April 20-end)?

Authors: The salinity dependence has little to do with the seasonality but is an effect of the observation that the optimum growth conditions of cyanobacteria occur in salinities between approximately 3 and 10 PSU. This span is taken to approximately represent the optimum span of N. spumigena, Aphanizomenon sp. and Dolichospermum spp. (Rakko and Seppälä, 2014). Its effect is more to limit the growth spatially than seasonally. 

Line 150, The range already described two sentences above.

Authors: The sentence above line 150 is referred to AKI and REC and the one on line 150 is AKIB. We will revise these sentences so this is more clear in the new version of the manuscript. We will also include a table with all abbreviations so the reader can easily look at them when needed. It is easy to miss that there are different versions of AKI for example. 

Line 151-153, Between 11C and 28C, it increases linearly from 10% to 100%?

Authors: The temperature limitation is defined by Eq. (8) in Table S.3 and is shown in the enclosed Figure R1. It is an adaptation to RCO-Scobi from the original model by Beckmann and Hense (2004).

Line 181-182, Nitrogen fixation rates appear to be an important observation parameter. How exactly calculated (estimated) from biomass in this paper?

Authors: Detailed calculations can be found in Olofsson et al. 2021 as we also refer to in the method section, but we will also add a few more explaining sentences on the calculations in the revised version of the manuscript. Biovolume (mm3 L-1) of the three different species were obtained from the SMHI database and mean values of volume-specific measurements of nitrogen fixation were obtained from in situ measurements of thousands of cells of each of the three taxa across two summer seasons (From Klawonn et al. 2016, as referred to in the manuscript). Observed taxa-specific biovolume (mm3 L-1) were multiplied with the taxa-specific nitrogen fixation measurements per day (umol N mm3-1 d-1) to obtain nitrogen fixation rates per volume water per day (mmol N L-1 d-1), and further depth-integrated over 0-10 m (mmol N m-2 d-1) to obtain area-specific nitrogen fixation rates. These rates could then be summarized for the whole year and multiplied with the size of the Baltic Prover (200 000 km2) to provide nitrogen loads via nitrogen fixation by filamentous cyanobacteria (kton N yr-1). 

Fig. 3. Unclear the seasonal cycle is for earlier period (1960-1979), later period (1999-2008) or full period (1850-2010)?

 Authors: It is for the period 1999-2008. This will be clarified in the caption.

Line 195: Diazotrophs tend to have much higher C:P or N:P ratio than normal phytoplankton.

Authors: It is true that they have a flexible ratio which can be both above and below Redfield, but the difference is not huge. Ploug et al. 2010 and 2011 show a fixation ratio of 6.6 for Baltic Sea filamentous cyanobacteria for example, this is only slightly above Redfield of 6.5. There is a difference between diatoms and dinoflagellates as well (Menden-Deuer and Lessard 2000), but we can not include all differences in this model and have to make some simplifications. 

Line 252-253, How the N2 fixation is simulated? That may indicate wrong simulation of biomass-specific N2 fixation rate.

Authors: The N2 fixation is a function of the temperature, light availability, N/P ratio and P. It is fully described in the appendix of Eilola et al. (2009). This section will be clarified with these details in the revised version of the manuscript since nitrogen fixation estimates is a main focus of the paper. 

Also, the observed N2 fixation is derived from observed biomass (still unclear to me the method); is that unreliable?

Authors: It is based on many measurements (thousands of cells across two seasons; From Klawonn et al. 2016) and observations over a long period of time (ca. 10 years of monitoring data from biweekly sampling in the Baltic Proper) so we would say it is fairly reliable. We will describe how it was estimated in more detail in the methods of the revised version of the paper (please see a more detailed reply to this comment above).

Some format issue: such as some incorrect parentheses (line 47, 107, 148), missing unit (line 201), incorrect subscripts and superscripts (line 178, 182), inconsistent color codes of experiments across figures 3, 4, 6, 7. “Baltic proper” or “Baltic Proper”?

Authors: Thank you for these observations. We will correct these errors in the revised version of the manuscript. 

Author reply references:

Beckmann, A., & Hense, I. (2004). Torn between extremes: The ups and downs of phytopiankton. Ocean Dynamics, 54(6), 581–592. https://doi.org/10.1007/s10236-004-0103-x

Eilola, K., Meier, H. E. M., and Almroth, E. (2009). On the dynamics of oxygen, phosphorus and cyanobacteria in the baltic sea; a model study. Journal of Marine Systems 75, 163 – 184. doi:https://doi.org/10.1016/j.jmarsys.2008.08.009

Hense, I., & Beckmann, A. (2006). Towards a model of cyanobacteria life cycle-effects of growing and resting stages on bloom formation of N2-fixing species. Ecological Modelling, 195(3–4), 205–218. https://doi.org/10.1016/j.ecolmodel.2005.11.018

Hense, I. and Beckmann, A. (2010). The representation of cyanobacteria life cycle processes in aquatic ecosystem models. Ecological Modelling 221, 2330 – 2338. doi:https://doi.org/10.1016/j.ecolmodel.

Hense, I., Meier, H. E. M., & Sonntag, S. (2013). Projected climate change impact on Baltic Sea cyanobacteria: Climate change impact on cyanobacteria. Climatic Change, 119(2), 391–406. https://doi.org/10.1007/s10584-013-0702-y

Klawonn, I., Nahar, N., Walve, J., Andersson, B., Olofsson M., Svedén, J.B., Littmann, S., Whitehouse, M.J., et al. 2016. Cell-specific nitrogen- and carbon-fixation of cyanobacteria in a temperate marine system (Baltic Sea). Environmental Microbiology 18: 4596–4609.

Menden-Deuer, Susanne, Lessard, Evelyn J., (2000), Carbon to volume relationships for dinoflagellates, diatoms, and other protist plankton, Limnology and Oceanography, 3, doi: 10.4319/lo.2000.45.3.0569.

Olofsson, M., Klawonn, I., and Karlson, B. (2021). Nitrogen fixation estimates for the Baltic Sea indicate high rates for the previously overlooked Bothnian Sea. AMBIO 50(1): 203-214.

Ploug, H., Adam, B., Musat, N., Kalvelage, T., Lavik, G., Wolf-Gladrow, D., and Kuypers, M.M.M.( 2011). Carbon, nitrogen and O2 fluxes associated with the cyanobacterium Nodularia spumigena in the Baltic Sea. ISME Journal 5: 1549– 1558. 

Ploug, H., Musat, N., Adam, B., Moraru, C.L., Lavik, G., Vagner, T., Bergman, B., and Kuypers, M.M.M. (2010). Carbon and nitrogen fluxes associated with the cyanobacterium Aphanizomenon sp. in the Baltic Sea. ISME Journal 4: 1215–1223.

Rakko, A. and Seppälä, J. (2014). Effect of salinity on the growth rate and nutrient stoichiometry of two Baltic Sea filamentous cyanobacterial species. Estonian Journal of Ecology doi:10.3176/eco.2014.2.01

Author replies in italics

Jenny Hieronymus et al. Modeling cyanobacteria life cycle dynamics and historical nitrogen fixation in the Baltic Sea

This study incorporated a cyanobacterial life cycle model with phosphorus dependency, which improved the prediction of diazotrophic cyanobacterial blooms in the Baltic Sea. The research is quite interesting and challenging; however, I found the whole manuscript lacks a clear hypothesis, clear clarification of why phosphorus is important, and the interpretation of results is not deep enough. I could see that the authors were trying to explain the methodology as it is a complicated study, however I got lost easily as there is not a clear approach or conceptual diagram to lead the readers. I have also got a few major concerns as listed below.

Authors: Thank you for this observation, we need to clarify our aims of the paper better in the introduction. We will revise the introduction to introduce the phosphorus dependency on an earlier stage, and also include a conceptual image of CLC (First version enclosed as New Figure 2) along with figure 2. 

To clarify, we have for the first time included the CLC model into a 3D model for the Baltic Sea. Previously the CLC model has been used by itself (Hense and Beckman 2006; 2010) and together with a 1d water column model (Hense et al. 2013). The P dependency of cyanobacteria has not been previously included, and since phosphate is limiting nitrogen fixation in the Baltic Sea during summer (Degerholm et al. 2006, Olofsson et al. 2016 etc.) the level of P dependency needed to be evaluated to not completely overestimate the biomass of cyanobacteria. As our study demonstrated in the model experiment noP (which is reflecting the settings in the previous studies), the biomass is far above observed levels and therefore discarded as a suitable setting for the Baltic Proper. Instead we found wPlim to be closest to observations in timing and magnitude of biomass and this was chosen for the estimates (as described in lines 240-241). We will make sure this is further clarified in the revised version of the manuscript. Please see more detailed replies to these issues below.

Introduction

From the manuscript it is not clear to me phosphorus utilization of the diazotrophic species is important in the Baltic sea, and what critical roles P plays in the dominance of the three species.

Authors: We will clarify this early on in the revised version of the manuscript. We have some details on the topic in lines 67-71 but will extend this section to also explain the background of the model settings. The noP run is what happens when applied to the Baltic Sea if P is not limiting and no uptake or release of P by cyanobacteria occurs, with cyanobacteria biomass far above observations. Please see the extended reply to the comment above.

Methods

Fig. 2 seems very complicated and busy to me, and I cannot tell what processes the authors have modelled and tried to test. What is your hypothesis? To someone who is not familiar with CLC model, I am suggesting the authors making Fig. 2 easier to follow, also by adding a conceptual diagram to illustrate what life cycle really means – what are the physiological processes, timescale, input conditions and output, etc.

Authors: We will include a CLC schematic image parallel to current figure 2 in the revised version of the manuscript (enclosed New Figure 2). We will also color code the CLC parts of the current version of Fig. 2, for example by red as the lines/arrows are in the current version as well as a new table with all the abbreviations (see enclosed New Figure 3 for a first version). 

Please also specify why CLC model needs to be modified to include P utilization. Did you mean by superior P uptake, P storage or DOP scavenging startegies?

Authors: Phosphorus dependency has not been considered in previous versions of the CLC model but must be considered in the Baltic Sea where P is often limiting the growth of filamentous cyanobacteria (e.g., Klawonn et al. 2016, Olofsson et al. 2016; Degerholm et al. 2006). This argument for the importance of this focus will be explained better in the introduction in the revised version of the manuscript. The sensitivity run “noP” demonstrates how the CLC model reacts when there is no phosphorus limitation as well as no uptake or release of phosphorus in cyanobacteria; biomass gets far above observations and is why limitations are clearly needed for the CLC model for the Baltic Sea. 

L100 – why some of the predicted temperature were much higher? Please kindly explain the reason behind it.

Authors: The temperature is well represented by the model. Slightly higher temperatures can be found in the upper parts of the halocline at BY15 (central Baltic sea, Fig. 10 in Meier et al., 2018). The reason is not clear but an exact reconstruction of the past is not to be expected by any model. For further details, please refer to Meier et al. (2018).

L110 – But you could already see cyanobacterial species vary in physiology from a great many publications. I wonder if it could be better to allocate a range or different C:N:P ratios for the modelled species, maybe a sensitivity analysis could help you find out whether this ratio really matters for the simulated outcomes.

Authors: We will think of this for the future, but for now it is too complicated. We will include a discussion around different C:N:P ratios in the revised manuscript as this may impact the ratio between biomass and nitrogen fixation.

L125 – I am sure the internal nutrient quotas also affect the growth and life cycle transitions; however, I cannot tell if you have included internal nutrient quotas impacts. A schematic including how processes involved in the model, alongside the methodology of this manuscript may help clarify the uncertainty here.

Authors: Internal nutrient and energy quotas were included in the original CLC model by Hense and Beckmann (2006). In a following publication (Hense and Beckmann, 2010), they constructed a simplified model, where the internal quotas were excluded, with the aim of obtaining a model efficient enough to be included in a 3d climate model. The model that we have used is an adaptation of their simplified model but where they separated the diazotrophic and non-diazotrophic stages into a two compartment model, we have instead summed up the recruiting and the vegetative (growing but without heterocysts) stage (REC) and obtained a four compartment CLC model including pelagic akinetes (AKIW), benthic akinetes (AKIB), recruiting and vegetative non-diazotrophic cells (REC) and cells with heterocysts (HET). We will include a clearer graphic of this in the revised manuscript.

Results and discussion

L565 – There are some extreme biomass values that were not predicted, why is that?

Authors: The colored dots in Fig. 3 show the two daily values of simulated biomass for the different experiments. The high frequency of the model output compared to the maximum sampling frequency of once every two weeks for the observations, generates a higher probability to capture extreme highs and lows. That being said, the simulated biomass is higher in the model compared to the observations even for the best fit simulation wPlim. There could be many reasons for that such as a too high or low C:N ratio, or non optimal choices of other constants. We will, in the revised manuscript, include a table or figure showing the monthly mean values of biomass in order to more clearly see the relation between biomass and the nitrogen fixation presented in Fig. 5.

Author reply references:

Degerholm, J., Gundersen, K., Bergman, B., & Söderbäck, E. (2006). Phosphorus-limited growth dynamics in two Baltic Sea cyanobacteria, Nodularia sp. and Aphanizomenon sp. FEMS Microbiology Ecology, 58(3), 323–332. https://doi.org/10.1111/j.1574-6941.2006.00180.x

Klawonn, I., Nahar, N., Walve, J., Andersson, B., Olofsson M., Svedén, J.B., Littmann, S., Whitehouse, M.J., et al. 2016. Cell-specific nitrogen- and carbon-fixation of cyanobacteria in a temperate marine system (Baltic Sea). Environmental Microbiology 18: 4596–4609.

Meier, H. E. M., Eilola, K., Almroth-Rosell, E., Schimanke, S., Kniebusch, M., Höglund, A., Pemberton, P., Liu, Y., Väli, G., & Saraiva, S. (2018). Disentangling the impact of nutrient load and climate changes on Baltic Sea hypoxia and eutrophication since 1850. Climate Dynamics. https://doi.org/https://doi.org/10.1007/s00382-018-4296-y

Olofsson, M., Egardt, J., Singh, A., and Ploug, H., (2016). Inorganic phosphorus enrichments in Baltic Sea water has large effects on growth, carbon fixation, and N2 fixation by Nodularia spumigena. Aquatic Microbial Ecology 77: 111–123.