This paper discusses the use of iron isotope ratios as a signal of climate change, based on simulations using a model of ocean biogeochemical cycles that includes iron isotope. The authors point out the possibility that changes in the iron cycle associated with ENSO and the global warming can be detected as changes in iron isotope ratios and that iron isotope ratios can be utilized as an indicator of iron limitation, and recommend that regular time series observations be conducted. In addition, the paper discusses possibilities of not being able to detect iron isotope signals when positive and negative anomalies are offset or when isotope ratios in the endmembers are not deviate significantly in the background values. The text is concise and clear, and the analysis is carefully presented. Overall, the paper provides useful insights; thus I recommend to accept this manuscript after minor revision. There are a few points that I have questions about, which are described below. It would be desirable to answer these points and revise the text if necessary.

In the equatorial Pacific, as the authors mensioned, the equatorial undercurrent (EUC), which is a sub-surface flow, transports iron from west to east (e.g. Slemons et al., 2010, GBC), and this iron is thought to be supplied to the surface layer as well. Does this model reproduce well the transport of iron by EUC? Although analysis of this study is limited to the surface layer down to 10 m, the iron cycle in the surface layer is also controlled by the distribution of iron in the sub-surface layer and may be particularly important in the tropical Pacific. Describing how the model can reproduce iron transport by EUC strengthens the reliability of the model results.

I understand that the positive iron isotope anomaly around 40S in the South Pacific (Fig. 1b) is due to strong uptake fractionation effects (Fig. S1d). I, however, do not understand why the uptake fractionation effects are stronger in this region. Adding supplementary explanations will help readers’ understanding.

For the hindcast simulation, there are no figures regarding the basic quantities (e.g., distribution of primary production, dissolved iron concentration and the isotope ratios, etc.). The standard deviation of the interannual variation of iron isotope ratios is suddenly shown as the first figure. I think that showing the figures for the basic quantities first will help the readers’ understanding of the model results and authors’ arguments. It should be noted that the most readers have no prior knowledge about the characterisity of this model.

If regular observations of iron isotope ratio were possible, the obtained data would be time series data at a certain station. The authors could strengthen their arguments if they select one station from the model results, as an example, and show how the use of iron isotopic ratio helps to explain interannual variations of dissolved iron concentration.

The following are really minor comments.

In Table 1, a case of "Hindcast neuSED" is mentioned, but the result is not presented as a figure in the main text. Since the description of this case in the text (L. 356) is enoughly understandable regarding the model setting, there is no need to list it as a case in Table 1. In addition, in the “d56Fe endmembers” column of the "Hindcast neuSED" case, there is description of "Sediments: -1 per mill,", which may be a mistake of “0 per mill”.

I think the abbreviation ONI in line 202 stands for Ocean Nino Index, so it should be written as (dark red, Ocean Nino Index; ONI).

In line 221, there is a description of "d56FeEM and d56FeUF (Fig. 1)". I think there is no need to refer “Fig. 1” here because d56FeEM and d56FeUF are clearly defined in Equations (1) and (3).

Observations of iron isotopic ratios are currently very restricted, especially with little known about interannual variability, and it will be a long time before the findings of this study are validated. Seasonal variations in dissolved iron have been shown in several studies, I felt that it would have been easier and more useful to use this isotope model to evaluate the factors that cause seasonal variations of dissolved iron concentrations. I am looking forward to seeing such a study in futre.

Reference

Slemons, L. O., J. W. Murray, J. Resing, B. Paul, and P. Dutrieux (2010), Western Pacific coastal sources of iron, manganese, and aluminum to the Equatorial Undercurrent, Global Biogeochemical Cycles, 24(3), doi:10.1029/2009GB003693.

The authors applied a global ocean model describing the Fe and ⁵⁶Fe cycling to study the impact of climate variability on surface distribution of iron concentration and its isotopic signals. The model considers different isotopic compositions of sources and fractionation by biological uptake and organic complexation of iron. Their previous publication (König et al. 2021) presented the modelled distribution of delta⁵⁶Fe and a thorough comparison with observations. In this study this model was driven by different climate forcing data and a series of sensitivity experiments were conducted to quantify the contribution of single factors to the inter annual variability of delta⁵⁶Fe. Strong responses of delta⁵⁶Fe to climate change were found in the model. I find the idea to study climate variability with Fe isotope fingerprints highly interesting and the article was well-written and easy to follow. However, I have some concerns about the analysis of model results and kindly ask the authors to give explanations for the following points:

1. Line 120-123: The effect of two single components, fractionation by biological uptake and organic complexation, on delta⁵⁶Fe is estimated from the difference between an experiment with all components switched on and another one with only one component switched off (Eq. (1) and (2)). But the effect of the third component, isotopic compositions of endmembers, is estimated in a different way (Eq. (3)) which assumes that the three components act independently on delta⁵⁶Fe in a linear relationship which is not true. An experiment with all endmembers set to 0 ‰ is to my opinion necessary to disentangle the effect of all single components, as the authors mentioned themselves as well (L. 125-126). If this experiment was already done I would like to see if the result is identical to the estimation presented now in the manuscript and why.

2. Line 135-138: If I understand it correctly, the authors calculated SD of each distribution of delta⁵⁶Fe resulted from Eq. (1) to Eq.(3) and then the fraction of each single SD in the sum of them. SD can demonstrate the variability around the mean state but tells nothing about the mean state itself. Responses of the three single components to the interannual climate variability can be reflected in SD but also in the mean state of delta⁵⁶Fe. So I don’t quite understand why just SD of different runs are used to examine the contribution of single components.

Furthermore, the sum of three SDs is not the same as SD delta⁵⁶Fe_diss of the experiment with all components switched on, due to the non-linear relationship between the single components and different signs of the effects. The authors only discussed about the latter in the manuscript. I have no doubt that the results of the three experiments are interesting and can help us to understand how the marine Fe isotope cycle responses to climate variability. Different SDs of the three distributions indicate that each component is differently sensitive to climate variability. However, the interpretation of the relative importance in percentage needs a justification. 

At this stage I would like to encourage the authors to revise the analysis and interpretation of the model results. After that, I would be happy to provide more detailed comments.