This is an excellent paper form the highly reputed authors.

The manuscript is an easy read, it's logically structured, arguments and thought processes are well founded.

My only few recommendations would be:

1. Remove reference to the excess charge densities of rocks when discussing this property interpreted for trees. Given the expertise of some authors this is not surprising, but I don't find this comparison relevant. It would be better to compare the data with similar.

2. Add a brief description of the color scheme used in Figs. 4b-c. This could be done as a legend to the colored bar or a brief description in the caption.

3. Sections 5.2 seems a bit too speculative, so I'd recommend to shorten the discussion and append it to Section 4. However, I will leave the decision on whether to do this or not to the authors.

General comments

This article presents very interesting data and data analysis of a 1 year monitoring of self potential (SP) and sap flow (SF) measurements on 3 different mediterranean forests. 1 or 2 trees are measured in each site, with single SP measurement on each tree. The paper is well written, well organized (except some parts, cf specific comments). It necessitates some preliminary skills in SP, but should be understandable for a broad audience.  The strengths of the study are these long term measurements, the pertinent signal analysis used for understanding links (or no links) between SF and SP, which open way for understanding complicated signal from in situ experiments, and the quantifications with theoretical grounds for SP interpretation given by the authors.  Exemplarily, the authors shows that it is not so much evident to relate SF and SP despite various studies started in the 60’s and show the different processes that can affect the signal.  If the strength of the study is in the diversity of site and trees, in my opinion there is a lack of discussion in the interpretation of the results in the discussion according to the different sites and tree species. From experimental point of view, only 1 tree is measured per site (2 at Font Blanche, but 1 tree of each species).  If I understand that it could be a high workload, more replicated measured tree sampled would have been interesting, at least, to examine the range of variation for a same location.  From another point of view, most of the interpreted results are based on 2 weeks in April for a 1 year study. May be looking at other contrasted periods or seasons would have been interesting for deciphering the different processes acting on SP signal (ie in winter when transpiration is slowed down or halted), in autumn when trees are recovering from possible drought, in summer during the drought and limitation of transpiration by water availability).

I add below some more specific comments, which can also call for some discussion on some points.

Specific comments

13 Transpiration is => Plant transpiration is

14 Solely relying on sap flow measurements => Not really as at the stand scale (evapo) transpiration can also be assessed by flux towers’ measurements.

32-37 Again cite different methods for (evapo)transpiration measurement : flux towers, soil water balance.. With the specifity of sap floxw measurement being the direct method for trees for obtaining transpiration.

42 Gindl et al., 1999;) => suppress last ;

46 transpiration processes, which facilitate water and solute transport within the xylem and phloem of trees => Rather  : transpiration which relies on the transport of water and solutes in xylem….  Phloem transport is not generated by transpiration but by gradients of concentration of sugars which generate turgor pressure gradients and flow, and can occur without transpiration.   So add another sentence, for phloem, decoupled from transpiration  

46 , trigger electrokinetic and electro-diffusive effects, => a few words to  explain the origin of effects ?

132 low of sap induces a natural electric field in the opposite direction, as depicted in Fig. 1e => no clear indication  of opposite electric field in 1e ? Only in the transverse direction of the vessel with the subplot. Could you make the graph clearer for this electric potential difference generation.

134 no external currents (i.e., electrical flux equals 0). => what means external current here ? If there if a flow of ion charges what is the return current so that net current J=0 ?

137 of the capillary fluid => flowing fluid ?

140  an empirical relationship => for which media ?

147 where k (m2) denotes permeability => for a porous medium?

148 under the assumption of 1-D flow,  => and neglecting Jdiff  in (1)

149- 155  lead to an increase in solute concentration towards the crown => If available, give range of variation in concentration, that should be small as ions are absorbed by cells too..

Table 1 Soil depth is indicated in m (range 0-80 m !), it is rather cm I think  …   Evapotranspiration seems low, if potential. Is it actual or potential evapotranspiration ?

L240 the sap velocity in μm s => the sap velocity is obtained in µm s-1

L243  high-impedance multimeter controlled by a digital data logger => give the refernce of manufacturer and model. What is the sampling frequency of recorded voltages ?

L244 which length and diameter for screws ? Fig 2 it seems it is not screws for  wood works. True ?  Does screws penetrate sapwood or are just in contact with first outside layer of vessels ? 

L272 In addition to variational  mode decomposition, would Fourrier spectrum analysis give also similar or complementary results ?

Fig 3 The time axis is labelled as Jan.01, Feb.01.. May be Jan 23, Feb 23… Jan 24 would be better.  30minutes data are shown, daily values would be interting to show (cumulative as proxy of daily flow and average for SP probably)

 L306 coherence anomalies => Why anomalies ? not just the coherence,   the highest at 1 day, and meaning, I think, that the SP and SV are more linked in terms of variations to each at a 1 day period than other periods.

L307 negative correlation => ?, figure 4 is not correlation. If you refer to linear correlation in S1, put this sentence before to show the difference in data interpretation between wavelet and correlation.

Fig4 the meaning for arrows as lag phase is not very clear to me. It is not explained and used in the text in 4.1.2. If this data it is needed it should be explained and used.

L324 330 Can you comment not only on frequency differences but also on the “power” of signal associated to those frequencies?

Fig 5: x axis could be labelled Frequency (d-1), for better understanding. Indicate what means |u1(f)| … in the legend.

349 during a two-week period in the growing season => Which 2 weeks and why those 2 weeks ?

L352 – 354 Negative correlation, and flow direction. If OK for the 5^th mode of SV, looking at raw data on fig 3, SP of pine is negative, while for oak is positive … (and would lead to a positive and negative correlation), how to reconciliate with your statement?

L359 and others occurrence in the text : The fifth decomposed signals => rather the fifth decomposed mode of signals. In the text, and fig legend (eg in fig 6)  it could be clearer if you use “the fifth decomposed mode of” …

L361 applying linear regression => regression of SP vs SV

L360- 364 In Fig 6, the 5^th mode of SV shows negative values (almost half of data?), ie negative sap flow. What does that mean from a plant point of view ?

Fig 6 in legend: red dots indicating outliers => red dots indicating points (outliers) outside the 95% confidence level of the regression. In the graph => In blue Prec. could b replaced by rain. There is no rains at Font Blanche ? Why there is blue dots on graphs (b,d,f…). Indicate r² of correlations (a, c,…)

L364 383 : All of this etxt should be pput either at the end of the section, or a new result section created such as “excess charge density estimation”. Which u is used here ? from sap flow or 5^th mode of sap flow ? Why no calculation for other sites ?

L391 – 398 and fig 7 : All of this text is difficult to understand. What is actually done is a difficult to follow, and figure 7 is not well explained…. Could it be more clearly rewritten?

L413-415 analogy between unsaturated soil and plant hydraulic conductivity seems doubtful to me. In soil, variation in water content is linked to filling/emptying of pores? Water content oin plants is expressed generally on a (fresh) weight basis, not volumetric. When plant desiccates its volume change but not necessarily saturation.  Indeed the plant variation in water content is rather loss of water from cells, with a loss of turgor, and not from xylem which remains full of water except in very dry situation when embolism happens. If it could be the case in dry summer,  it might not be that for spring, autumn ? At least that should be discussed in such analogy.

L425-430 Xylem and wood is very different from mineral media and made of cellulose, hemi cellulose and lignin. Does any data exist for zeta potential of wood ? Isn’t  the comparison with mineral media is a bit limited by the fact you use eq 12 elaborated nfor mineral media. As zeta potential is an important coupling factor at least that points to the need to its experimental estimation..

L443-450 In Love eta l (20078) study the difference in pH is between the soil and the plant, when they show the correlation SP-pH. If a pH for soils of these is considered to be ~8 (calcareous soil) and xylem sap ~6, the SP related to pH would be ~ 120 mV, ie in the order of magnitude of measured SP signal, and electrokinetic effect. Could authors comment on that ?

L451 metal ions => rather alkaline earth cations

L478 481 to be discussed in discussion …