Summary: This manuscript presents and anlyzes land-surface data collected for multiple years at 4 locations in central Spain.  The authors investigate the relationship between air and subsurface temperatures, and then investigate the conductive and thermal properties of the subsurface using continuous temperature measurements as multiple depths.  The importance of this study ranges from interpretating paleoclimatic reconstructions derived from borehole temperature profiles to understanding of land-surface processes that control permafrost, soil carbon uptake/release, and surface extremes.

General Remarks: This is an interesting study that presents a kind of dataset that is rare, namely continuous measurements of subsurface temperatures over multiple depths exending meters to tens of meters into the ground.  The authors do a good job of contextualizing their work in the broader literature and provide some insightful analyses of the data.  I provide a detailed list of revisions below, but my general assessment is that the paper should be published after my comments are addressed.  While none of my comments require substantial changes, the number of small changes that are necessary amount to a necessary major revision.

Specific Comments

Ln 3: two shallow "profiles"

Ln 13: thermal diffusivity, not heat

Ln 17: "and ongoing anthropogenic"

Ln 20: has allowed "the attribution of"

Ln 23: The difference between land and ocean warming rates is not exclusively due to evaporative cooling.  The heat capacities are important, as is the circulation of the ocean.  An increase in air temperature is also not the principal means by which the ocean is warming, but rather changes in the energy balance at the land and ocean surfaces that is driving both air and ocean temperature change.

Ln 25: Rahmstorf and Coumou is a strange reference here.  There are many other papers on land-atmosphere interactions that would be more appropriate.

Ln 27: long-term scales should be changed to be specific (over decades to centuries).  Melo-Aguilar et al. as a singular reference is also an insufficient reference for this statement.  There is a long history of comparisons between GST reconstructions and SAT comparisons that make this case (see the work of Harris and Chapman, Huang and Pollack, Beltrami, etc.). 

Ln 29: SAT changes are not transmitted into the subsurface.  As stated above, the temperature changes in both SAT and GST are the product of changes in the energy balance at the land surface.

Ln 36: The polar changes do not have implications exclusively for climate in high latitude regions.  For instance, emissions of carbon from permafrost has global implications.

Ln 38: temperatures allow recovery of past SAT

Ln 43: sources is crucial.  It is also not obvious what "its" is refering to is in this sentence.  SAT-GST coupling?  If so, SAT-GST isn't such an important thing to understand in the context of climate change, rather the relationship is important for understanding other crucial processes.

Ln 45:  The list of references here is pretty limited.  Other important papers include the following:

Baker, D. G. and D. L. Ruschy: The recent warming in eastern Minnesota shown by ground temperatures, Geophys. Res. Lett., 20(5), 371–374, doi:10.1029/92GL02724, 1993.

Putnam, S. N. and Chapman, D. S.: A geothermal climate change observatory: first year results from Emigrant Pass in northwest Utah, J. Geophys. Res., 101, 21877–21890, 1996.

Smerdon, J. E., Pollack, H. N., Enz, J. W., and Lewis, M. J.: Conduction-dominated heat transport of the annual

temperature signal in soil, J. Geophys. Res., 108, 2431, doi:10.1029/2002JB002351, 2003.

Smerdon, J. E., Pollack, H. N., Cermak, V., Enz, J. W., Kresl, M., Safanda, J., and Wehmiller, J. F.: Air-ground temperature coupling and subsurface propagation of annual temperature signals, J. Geophys. Res., 109, D21107, doi:10.1029/2004JD005056, 2004.

Smerdon, J. E., Pollack, H. N., Cermak, V., Enz, J. W., Kresl, M., Safanda, J., and Wehmiller, J. F.: Daily, seasonal, and annual relationships between air and subsurface temperatures, J. Geophys. Res., 111, D07101, doi:10.1029/2004JD005578, 2006.

Smerdon, J. E., Beltrami, H., Creelman, C., and Stevens, M. B.: Characterizing land-surface processes: A quantitative analysis using air-ground thermal orbits, J. Geophys. Res., 14, D15102, doi:10.1029/2009JD011768, 2009.

Ln 53: This discussion is missing an acknowledgment that trends have also been investigated (not exclusively a harmonic assessment) in Lesperance et al. (2010).  The harmonics have also been tracked on an annual basis in thermal orbit applications, i.e. an approach that could theoretically be done for a single year of data (Sushama et al. 2007 and Smerdon et al. 2009)

Ln 56: subsurface temperature for at least

Ln 64: contribute to more accurate

Ln 65: The ending clause of this sentence is confusing.  I think the authors are making a reference to the amount of heat that has been stored in the terrestrial subsurface over the last 50-100 years, relative to other components of the climate system.  That should be clarified.  They should also cite Beltrami et al. (2002) in the context of their assessment.

Ln 66: compliance is not the right word in this sentence

Ln 73: This allows us to test

Ln 78: "multi-annual long temporal intervals" is overburdened.  Revise for brevity and clarity.

Ln 83: also allows diffusivity changes to be evaluated through time and connected to

Ln 85: This form of finger counting the following sections is unneccessary.  Outlining the questions addressed in the manuscript is useful at this stage, but enumerating the sections is not.  Either reframe as questions or remove this part of the intro.

Ln 89: mountan system that splits

Ln 133: Note the decrease in high-frequency variability with subsurface depth, which is consistent with

Ln 142: The authors detail the method of subsurface temperarture collection in great detail, but say nothing of how the SAT and snow depth measurements are collected.  A little detail is warranted here.  It should also be specifically mentioned at what height the SAT is measured.  Vegetation cover at the sites would also be worth noting.  If the authors have a picture of a representative sight, that would also be useful.

Ln 155: Smerdon and Steiglitz (2006) also discuss the harmonic solution in this context and provide the explicit expressions for the amplitude and phase evolution as a function of depth.

Ln 166: The authors use this parenthetical convention at multiple places within the manuscript.  It is a poor and confusing form that should be avoided.  This is not just my own hobby horse - there have been papers in our field about it:

Robock, A. (2010), Parentheses Are (Are Not) for References and Clarification (Saving Space), Eos Trans. AGU, 91( 45), 419– 419, doi:10.1029/2010EO450004.

Ln 176: each harmonic is also propagated

Ln 183: The conductive process is widely described as a lowpass filter in this literature, but the description is not accurate and should be avoided.  Higher frequencies are indeed damped more strongly with depth than lower freqencies, but as the authors have noted, each frequency is also phase shifted relative to each other.  This latter property is not characteristic of a lowpass filter, which would not shift the relative phases of the signal.  While this may seem pedantic, it has important implications and the community should be careful to avoid this misleading description.

Ln 200: A few points about the spectral method.  It does not appear to me that it is any different than the CA, but merely exploits multiple frequencies outside of the annual signal.  This is novel and useful, but it does not warrent a billing as a different approach, i.e. it is merely exploiting the properties of conduction across a wider frequency band than a method that exclusively tracks the annual signal.  On a very small note, it might be worth renaming the approach to yield an acronym that is not SM.  This is a widely used acronym for soil moisture, which makes it a bit confusing when used later in the manuscript.

Ln 211: deterministic is not the right word here

Ln 214: despite the fact that

Ln 215: The B.I.G. is notorious, but the damping of the annual signal is not.  Consider a different word.

Ln 217: Further should be Furthermore in several cases within this paragraph.

Ln 228: slightly weaker is vague.  Do you mean the amplitude of SAT is smaller?

Ln 229: "being SAT-GST anomalies negative" is incorrect.  Rephrase for clarity.

Ln 234: The general statement before noting that RPI and CTS do not follow the expected behavior with elevation is that the sites conform to the expectation.  But that leaves only two other sites!  This should be reframed to reflect the fact that two sites behave as expected and two sites don't.

Ln 241: that are worth mentioning

Ln 245: There are many studies that have investigated the impact of snow on GST.  In addition to some that have already been mentioned, there are also these:

Stieglitz, M., S. J. Dery, V. E. Romanovsky, and T. E. Osterkamp (2003), The role of snow cover in the warming of arctic permafrost, Geophys. Res. Lett., 30(13), 1721, doi:10.1029/2003GL017337.

Zhang, T. (2005), Influence of the seasonal snow cover on the ground thermal regime: An overview, Rev. Geophys., 43, RG4002, doi:10.1029/2004RG000157.

Ln 300: This is also discussed in the following:

Pollack, H. N., J. E. Smerdon, and P. E. van Keken (2005), Variable seasonal coupling between air and ground temperatures: A simple representation in terms of subsurface thermal diffusivity, Geophys. Res. Lett., 32, L15405, doi:10.1029/2005GL023869.

Ln 310: The authors presumably have the stratigraphy of the soil and bedrock horizons from the cores that the drilled.  Why not confirm whether these estimated transitions align with their observations?

Ln 315: This diffusivity estimate as a function of depth was done in Smerdon et al. (2003).

Ln 345: criterion

Ln 350: Again, this variable diffusivity due to surface processes was also discussed by Pollack et al. (2006).

Ln 359: In constrast to tother sites

Ln 382: is consistent with

Ln 386: quasi-steady, but it is not clear exactly what is meant here.

Ln 389: propagated to depth

Ln 393: The construction here implies that the diffusivities at other sites should compare to those determined for the Spanish sites.  Why should that be so?  It is distinctly possible that the diffusivities are simply reflective of different suburface materials, which is expected.

Ln 405: The authors use synoptical in multiple places in the manuscript.  The implied meaning is not obvious.

Figure 3: There are no temperature tick marks on this graph to vive a sense of the magnitude of the temperature variation.  Only the mean values of offset temperature are marked.  I also find the green shading very hard to see and understand what it is marking.  Moving the arrow to the beginning of the offest might improve things, as would providing a caption next to the arrow describing what it is pointing at.

Figure 5: There are some strange discontinuities in these figures.  For instance, do the authors think the feature between 1.5 and 2 m in the annual CTS BRH profile is real?

Figure 6: This figure would be a lot easier to read with a legend instead of a color description in the caption.

Figure 11e: This figure would be a lot easier to read with a legend instead of a color description in the caption.  The grey shaded area also looks more like a line, making the description of that feature in the graph confusing.

This is a very interesting study presenting data that are rare, linking shallow surface and deeper soil temperature regimes over time.  Near to the end of the Abstract the novelty is clear, but this is lost to some extent in the main body of the paper, where much of the presentation is limited to the data, rather than presenting what is new and exciting in the work.  There is emphasis placed on well-described phenomena, such as near surface temperature variability due to the drivers of soil formation, but less on the interesting aspects to link the shallow surface to deeper soil.

Abbreviations are used throughout.  Some of these are not necessary and they make following the paper difficult. 

The title would be better and more novel if instead of 'subsurface' it mentioned 'from surface soil to bedrock'.

The Abstract is good but try to bring some of the novelty presented at the end earlier on in the text.  This will convince readers that the paper is new and worth reading. The early part of the abstract reads like an observational study, but your paper is more than this.

The Introduction needs some restructuring. The text varies from assuming no knowledge to having expert knowledge, and the flow of arguments could be clearer.  In one sentence you mention permafrost, then many other parameters, and then permafrost again.  Make it easier for the reader to follow.

You need to be clearer on the research gap before describing your study. Remind readers of the ‘conductive hypothesis’.  Missing is some of the novelty. The coupling of shallow to deep thermal regimes is novel, so bring this out more in the last paragraphs.

Section 2 contains most of the necessary information and indicates an interesting study.  Where you describe the sensors at line 100 it would help to be clearer about depth intervals.  You miss describing the date range of the data-set.

Section 3 – give an indication of the extension of time in the text at line 148. The text is clear to follow.

Section 4 – please see comments on the marked version. You spend too much time writing about well known processes, such as shallow surface temperature variation, and not enough linking shallow and deeper soils.