The threshold for inclusion has been raised to 50% (days with less than five 30-min data points were removed), and all results updated based on this change. Only very slight changes to the results were observed, and we agree that our methodology is now more robust. 

Results 

Lines 277/278: The Rn values at which H is described as being greater than LE (for Kopuatai) or LE as greater than H (for Mer Bleue) are a bit ambiguous when the uncertainty intervals are taken into considerations in Figure 4. Are these thresholds mentioned in the text of Rn>=250W/m2 and >=350W/m2, respectively, based on statistical testing for difference between the two components considering the uncertainty estimates? E.g. for Mer Bleue at an Rn of 350W/m2 the plotted standard deviations still overlap considerably with the mean values for both LE and H, so it is a bit debatable whether one can be classified as being ‘greater’ than the other in a statistically significant context. It is obvious from the figure that the two sites have very different patterns of energy partitioning, but I would recommend to be more careful and clear with the language here. 

The threshold values mentioned in the text have been adjusted such that they reflect the points where error bars no longer overlap between LE and H. The new text is as follows: 

“At Kopuatai, H was a much larger component of the energy balance than LE (Fig. 5a); H was greater than LE at Rn ≥ 450 W m-2, which was reflected in a mean Bowen ratio (β) of 2.0 (β range based on binned H and LE values was 0.43–3.11). At Mer Bleue, however, LE was either similar to or greater than H at Rn ≥ 350 W m-2 (Fig. 5b). As a result, mean β was 0.80, with a range of 0.63–1.05.” 

Discussion 

Lines 394-401: I think this also supports the hypothesis mentioned earlier that evaporation from below canopy water is suppressed at Kopuatai, as partitioning towards LE increases significantly as soon as ‘open water’ is available in the form of canopy interception. If there was a significant contribution from evaporation of below canopy water the difference in energy partitioning between wet and dry conditions would likely not be as stark (which is exactly what is happening at Mer Bleue where there is very little difference in partitioning when comparing wet and dry). 

We thank the reviewer for this insight. We have added the following text: 

“The contrast in energy partitioning depending on wetness state indicates that the dense standing litter layer within the E. robustum canopy suppresses evaporation from the moist peat surface at Kopuatai.” 

Figure 7 and Appendix D 

In the Appendix Fig D1 the variability of EF for Kopuatai seems to be much larger at low VPD values, for both the unfiltered 30 min data as well as the growing season dry canopy MoD data points. This does not seem to be reflected in the shading of the 95% confidence interval in Figure 7 which barely changes with changing VPD. How was the 95% confidence interval determined? At Mer Bleue (Fig D2) the variability of the MoD dry canopy data does not seem to change as much with increasing VPD, so lines up better with the 95% CI depicted in Fig 7. 

We agree that it is unintuitive that the spread of the raw and MoD mean data points in Figure D1 was not reflected in the 95% confidence intervals. However, having modified the inclusion threshold to 50% (see above), the 95% CIs are now wider at low VPD compared to high VPD (perhaps due to improved data quality; see red dotted line in the image below). These 95% CIs are still not as large as expected at low VPD, however, this could be due to large bin counts from the fourth VPD bin (0.8 kPa) and higher. This trend is much clearer when plotting the same data with standard deviations (see blue dotted line in the image below), as the errors get considerably larger at low VPD; therefore, we do not consider this to be an issue. 

References: 

Aurela, M., Riutta, T., Laurila, T., Tuovinen, J.-P., Vesala, T., Tuittila, E.-S., Rinne, J., Haapanala, S., and Laine, J.: CO2 exchange of a sedge fen in southern Finland - the impact of a drought period, Tellus B, 59, https://doi.org/10.1111/j.1600-0889.2007.00309.x, 2007.   

Chang, J. and Root, B.: On the relationship between mean monthly global radiation and air temperature, Arch. Met. Geoph. Biokl., 23, 13-30, https://doi.org/10.1007/BF02247305, 1975.   

Grossiord, C., Buckley, T. N., Cernusak, L. A., Novick, K. A., Poulter, B., Siegwolf, R. T. W., Sperry, J. S., and McDowell, N. G.: Plant responses to rising vapor pressure deficit, New Phytol., 226, 1550-1566, https://doi.org/10.1111/nph.16485, 2020.   

Otieno, D., Lindner, S., Muhr, J., and Borken, W.: Sensitivity of peatland herbaceous vegetation to vapor pressure deficit influences net ecosystem CO2 exchange, Wetlands, 32, 895-905, https://doi.org/10.1007/s13157-012-0322-8, 2012.   

Takagi, K., Tsuboya, T., and Takahashi, H.: Diurnal hystereses of stomatal and bulk surface conductances in relation to vapor pressure deficit in a cool-temperate wetland, Agr. Forest Meteorol., 91, 177-191, https://doi.org/10.1016/S0168-1923(98)00078-1, 1998.   

Speranskaya et al. use two peatlands with different vegetation types (Kopuatai, NZ and Mer Bleue, Canada) to investigate evaporation (E) differences. Kopuatai has a vascular plant as its most dominant cover while Mer Bleue has more Sphagnum. This difference in the stomatal control of the vegetation is likely the reason for different responses of E to VPD and could then have implications on how the peatlands would respond to warming/drying changes. I find the paper to be well written with clear figures and very straight forward both in its presentation and in its findings. 

My principle complaint is the choice of bogs for the comparisons. It would have made for a more clear cut comparison if the non-vascular bog were to have been more similar to Kopuatai (no snow, less seasonal, etc.). I realize that might have been a tall order to arrange, but it would have made the interpretation and comparisons even more clear. Alternatively, it would have been interesting to see even more peatlands included (perhaps in a gradient of vascular plant coverage). I think the paper would have been much enriched if a look at more peatlands could have provided a bit stronger support to the findings about the importance of the vascular plants, or if by picking out drought periods in the record it could have been demonstrated that the vascular plants really would protect these sites better than those with more moss cover. Table 2 does get at this a little bit, but to a less extent than I am suggesting here. Otherwise, I have only minor comments so I think the paper can be accepted after they are addressed. 

We thank the reviewer for the positive comments on the quality of our paper, and for the suggested changes to its scope. In terms of selection of a comparison peatland, we chose Mer Bleue due to the length of its data record and its mid-latitude temperate climate status hence being a reasonable Sphagnum-dominated peat bog analog to Kopuatai. We also believe that a cross-hemisphere comparison is a novel approach to take, especially given the paucity of long-term EC datasets available from Southern Hemisphere bogs. We think that the inclusion of other peatlands might have been valuable, however, the use of only two sites enables a greater focus on describing the characteristics of their evaporation regimes, without this being lost within a wider synthesis. Nevertheless, a comparison of more sites (e.g., along a gradient of vascular plant coverage) could be an interesting avenue to explore in future research, now that the key distinctive features of the evaporation regime at Kopuatai have been described.  

Minor comments: 

Line 139 - Please add what is the difference in height between the hummocks and hollows. The WTD is used but one needs the heights to better understand how that might impact upon fluxes in the bog. 

The difference between hummock and hollow heights at Mer Bleue has been added (while Kopuatai does not have such topography). 

L147 - The window is 100 consecutive 'available' half-hours. This was then unclear to me how much this window might contract and expand due to gaps in the records. Did these windows ever get very large? 

The window of 100 available half-hours is fixed and centres on the data gap.  The observations used to develop the radiation relationship to fill the gap are not consecutive (and this wording has been corrected in the paper).  For periods where there is more missing data, the period of time covered by the 100 half hour dataset gets larger but the fixed number of data points ensures a statistically consistent gap-filling process and avoids the creation of large outliers that can occur with a fixed window size.  The window period is 3 to 5 days for 80% of the dataset gaps and less than 10 days for 98% of the dataset gaps.  For the rare occurrence of larger window periods, the maximum window period in summer (when LE is most variable) was 20 days. Longer window periods occurred (rarely) in winter but this was when LE was near zero and had limited temporal variability. 

L 223 - here is what I mean about knowing the hummock to hollow height. The 377 mm is only a bit below the hollow surface (I think they are 30 cm at Mer Bleue) so is actually closer to Kopuatai than it otherwise looks (at least over the part of the landscape that is hollows, which I think is roughly 50% at Mer Bleue). 

We agree that this may appear to be cause for concern, as the WTD is measured relative to the hummock surface, which is on average 0.25 m above the hollow surface at Mer Bleue. However, since there is a 70% cover of hummocks at Mer Bleue, the majority of the peat surface at Mer Bleue experiences a WTD similar to the measurements used in this paper. Therefore, our view is that, despite the height differences between hummocks and hollows, the actual WTD at Mer Bleue is still considerably lower than at Kopuatai (and less stable). We have added the following sentence to Section 2.1.2: “The hummock height at Mer Bleue is 0.25 m, with a 70% cover of hummocks.” 

A few places, I think you want to say 'radiation received' rather than 'receipts' which sounds weird to my ear. 

The wording of those phrases has been changed. 

Fig 2 - it would be good to add the site labels to the figures themselves. Yes, they are in the caption but it is easier to read if they are directly on the figure. 

Site labels have been added for Figures 3-6 (previously Figures 2-5). 

Fig 4 and 5 - Rn is in the caption but not in the figures 

This is not quite correct – Rn constitutes the independent variable (x-axis) of these figures. We have re-worded the figure captions to make this clearer. 

I would combine Fig 6 - 8 into a multipanel figure, it would make them easier to see and compare across and take up less pages. 

Thank you for this suggestion, we have combined these into the new Figure 7. 

I suggest calling gc, 'calculated gc' to make it clear that it is not directly measured but derived. 

As it is clearly stated in the Methods section that “gc was calculated...”, we feel that changing the name of gc to “calculated gc” throughout the document is unnecessary. This has been standard practice in many previous papers. 

Table 2: It may help people if (New Zealand) was added after Aotearoa as I note the authors even use New Zealand in their affiliation lines so Aotearoa is not yet fully adopted. (Also doesn't it only refer to the North Island?) 

Aotearoa has been added to all instances of New Zealand in the text (and vice versa), for consistency. This name refers to all of New Zealand. 

L 361 -It would be interesting to conduct an experiment where the litter layer is removed to see if the litter layer really does have much impact. Totally outside the scope of this paper, but a neat experiment nonetheless.  

We thank the reviewer for this suggestion – this will be noted as a potential future research topic. 

 A figure showing the actual study locations would be a nice touch.  

We have added photographs of the two study sites (Figure 1). We have not used a map because scale issues would render this meaningless. Lat/long coordinates are provided in the text to enable readers to locate and inspect the sites using (e.g.) Google Earth.