The manuscript by Tapia et al. describes P/Ca variations in several speleothems from Northern Iberia over two distinct climate transitions. The high increases in P/Ca at the onset of cool events are attributed to enhanced freeze-thaw cycling, facilitating the mobilization of P from the soil. The authors support their interpretation by a model sensitivity study. Overall, I think the data and the final conclusions are sound but I have at major aspects that I suggest to think about or include into the discussion, because these could influence the overall take-home messages a little bit. In addition, I suggest to reduce the number of Figures in the main text and part of it possibly moved to a supplement. Also, I suggest to use color-blind friendly palettes. Likewise, the main text, in particular the results section, could be streamlined and focused on the most important results. My impression is that a large part of the results section just deals with processes that are then shown to be not dominant, so in my opinion this could be largely shortened or moved to the supplement.

Main comments:

1. Soil pH. If I get the authors right, they focus their analyses on the contribution of organic P. However, adsorbed and precipitated P may be released to the soil solution under certain conditions as well, which are highly pH dependent. While I agree that vegetation itself may not have changed too much (L457ff), but soil pH may still have changed. If you have any idea about soil pH values at the sites, and how that could have changed in the past, I suggest to think about this pathway as a potential additional source of P to the drip water. and potentially include it into the discussion as well.
2. Sampling bias. I am not quite sure how the authors have taken into account that their stalagmites have varying growth rates, and how that might affect the observed P variability. Its hard to assess from the Figures also because age models have been published elsewhere, but it looks like some of the stalagmites change growth rate quite substantially, and that part of the increased variability in P/Ca could just be because of a sampling bias during periods of high growth rate? Therefore I suggest to down sample the stalagmite P/Ca records to equidistant , comparable resolution, to remove this effect.

Minor comments along the text

L35 and L49: I agree, but if I am not mistaken, none of these aspects is discussed further on of that may be part of the observed variability in the speleothems?

L116 repeated information

L153 It would be good to include some numbers of how much soil water pH may change and if that could influence the solubility of other P pools.

L210 In the distributions, is the temporal resolution due to different growth rates / sampling intervals taken into account? some peaks could have been missed/smoothed in slower growing stalagmites

Fig. 2 What to the blue and red colors mean? I wonder how the box plots would change if one would take into account the variations in growth rate.

L234 so, synchronously to the shift in d13C?

L236 How “low”? could you specify the magnitude of the shift also in d13C?

Fig. 4 would be nice to have similar lines indicating the “cooling” event(s)/spikes than in Fig 5

L?? section 3.2.2: I again ask myself if some of the observations of high/low variability may have to do with growth rate/sampling bias. I suggest to clarify, and possibly show / discuss records down sampled to equidistant temporal resolution

Fig. 7 the nomenclature is a bit confusing, in the Fig its “models A…D”, in the caption “panels A…D”, and then it is mixed up with the D from partitioning coefficient… maybe use lower case letters or numbers for the panels?!

L457 P export does not necessarily have to be related with a change of vegetation, but also a change in soil pH influencing the stability of the different P pools in the soil

General comments:

The manuscript by Tapia et al. presents P/Ca records from stalagmites of 2 caves in northwestern Spain. Their findings suggest accelerated phosphorus leaching during sudden climate transitions, as supported by sensitivity tests with a model. As a non-specialist I enjoyed reading this paper, although the number of figures and information presented in the paper felt a bit overwhelming at some points. To improve this I have listed some comments below. There are also smaller (technical) corrections that have to be made to improve the form and readability of the manuscript.

Specific comments:

Line 126: How was the temperature obtained?

Line 130: These values do not correspond to the values in Table 1. Are these values also presented in figure 2, results from new measurements? Should they not be presented after the method is introduced. Due to the large spread in the data (for LV especially), maybe also report the standard deviation with the average.

Line 140: first mention fCa, explain

Line 182: Not sure if you have defined D before.

Line 202: temperature

Line 206: I don’t count 12 stalagmites, and there are 2 bedrock samples in figure 2

Line 207: GAEL?

Line 207-213: Text is a bit repetitive, it feels like a list, due to the ‘In GAE’, ’PGM-LIG section of GAE’ two times in a row. Rewrite it more naturally.

Line 213: How many points or % data was removed by applying this filter? Is the uncorrected + corrected data available in the supplementary?

Line 329: what does A and B refer to? Panel or model?

Overall, there is some inconsistencies with the use of abbreviations in text and figures; Penultimate Glacial Maximum (PGM) to the Last Interglacial (LIG) are also called GL and DE in Figure 3. I assume you are talking about the same events, and since there are already a lot of abbreviations in the manuscripts, try to remain consistent. In the manuscript there is also GS 22 and GS22 (e.g. in text and SFigs)

Table 1 : Units are missing for P/Ca and Mg/Ca.

Figure 1:Description needs more details for panel C: ‘average precipitation per month (bars) and average temp…’

Figure 2: Penultimate Glacial Maximum. Maybe also note the abbreviations of e.g. PGM, LIG, GS 22 in the caption/figure. And what do the different colors mean?

Figure 3: Note in the caption that the P/Ca values presented in this figure are corrected using Al/Ca.

Figure 4: mention GS 22 or GS22 in the caption

Figure 5 and 6: why split up these figures, can be combined in 1

Figure 7: Again, the caption is a bit confusing, panels A, B, D, C and models A, B, C, D? I get a bit lost with the text and the reference to the panels. Maybe use letters for the individual panels and numbers for the models? Partitioning  of an element (E) is normally written as D_E and not just D.

Figure 8: Legend is not readable

Figure 10: I like this summary figure. Maybe add theoretical high or low stalagmite P/Ca for both scenarios (example p/ca with an arrow up or down next to it). Explain what red arrows mean (more or less intense P flux?)

Table S1: 2 of pCO2 in subscript, °C temperature

Figure S2: The rightmost column illustrates. 2 of pCO2 in subscript

Figures: In general, check if your figures are color-blind friendly, and if not, try to adjust them accordingly.

Data availability: Exactly how will the data be made available? Since the manuscript is already published on BGD and even citable, the data should be made available through an open repository as soon as possible.

Technical corrections:

In general, references in text have to be checked and corrected for form. Often a space is missing before the lit ref (example in line 37, 39, 42, 47…, 522 etc) or brackets have to be removed (examples line 164, 445, 465 etc.). Other examples of problems: Missing point in Kost et., al (line 129). Missing point after (Frisia et al., 2012) (line 57). Remove brackets on Kost et al., 2023 in line 125. Add closing bracket and point after ‘see Kost et al., 2023 for overview’ in line 117. There are numerous other small issues.

Also check the text for double spaces (line 39 – before ‘Current monitoring’ and line 47- before ‘Karst regions’), or lack of spaces (line 326: ‘Fig.7’).