Comment on esd-2021-68

In the abstract there is no description of the differences in setups in the two ensembles of projections (ECOSUPPORT and BalticAPP / CLIMSEA simulations). It would be nice to list them in accordance with Table 3.

Despite the uncertainties related to global and regional climate and impact models, as well as the unknown pathways of GHG and nutrient emissions, which make it very difficult to compare these two ensembles of projections, the authors obtained a number of important new results. Among them, I note: 1) global mean SLR results in a more or less complete compensation for the Baltic Sea salinity reduction for the projected increasing river runoff, 2) the global SLR can be identified as a new driver that has a strong impact on bottom oxygen concentration, 3) most noticeable are the differences in projected biogeochemical variables between ECOSUPPORT and BalticAPP / CLIMSEA ensembles. The article is undoubtedly important and interesting for ESD readers not only with the presented results, but also with a discussion of knowledge gaps and prospects for further research. The main, but, incidentally, a small drawback of the manuscript is not always a sufficiently detailed explanation of the results obtained, which is discussed below in specific comments.

Abstract
In the abstract there is no description of the differences in setups in the two ensembles of projections (ECOSUPPORT and BalticAPP / CLIMSEA simulations). It would be nice to list them in accordance with Table 3.

L.114-115 "An overview was given by (Schrum et al., 2016)…"
Reference is missing from the bibliography L.193-208, 866-874 (Fig.3) Nutrient input scenarios in BalticAPP and CLIMSEA simulations are described using Fig. 3. The latter is of very poor quality and does not allow you to see the differences between the blue, green, orange and red curves (Fig.3, lower panels) corresponding different scenarios. The choice of colors appears to be unfortunate. In addition, plotting all the curves on one graph creates the impression of complete chaos, nothing more (although, perhaps, the authors wanted to show just that).
The ECOSUPPORT nutrient input scenarios (Gustafsson et al., 2011;their Fig. 3.1) are not displayed in the Fig. 3. At the same time, they are important (especially for the historical period) in order to better understand the differences in the projections of biogeochemical variables in the two Ñomparable ensembles, and it would be good to add them as a separate panel in Fig.3.

L.460-463
The good agreement between simulated annual mean surface phytoplankton concentrations and reanalysis data everywhere (with the exception of coastal regions) is somewhat surprising. As far as the author of the review is aware, the deviations of the simulated chlorophyll concentration from the individual observations can be very large. Perhaps the agreement is due to the small number of observations and their corresponding small contribution to the results of the reanalysis?

Bottom oxygen concentration"
This subsection discusses the differences between ECOSUPPORT and BalticAPP ensembles in bottom oxygen concentration. Discussion of differences ends with the phrase: "These results are explained by the historical nutrient input reductions and the slow response of the Baltic Sea". (L.658) This unsatisfactory, too general explanation makes the reader think, and due to what differences in historical nutrient input this situation could have happened, and he turns to Fig. 3 and does not find there the necessary information about historical nutrient input in ECOSUPPORT (see above remark about Fig. 3). Further, an inquisitive, but already somewhat irritated reader continues reading and discovers a detailed explanation of the differences between ECOSUPPORT and BalticAPP ensembles in the biogeochemical cycling (for some reason, in the Knowledge gaps section) in lines 751-763.
In my opinion, it would be better not to test the patience of the reader and move this explanation to section 3.2.6. It would also be good to expand this explanation by indicating the difference in the initial conditions for biogeochemical variables between ECOSUPPORT and BalticAPP ensembles and how this difference affected the final results.