acp-2021-670     

Ozone deposition impact assessments for forest canopies require accurate ozone flux partitioning on diurnal timescales

Auke J. Visser et al.

General comments

This paper evaluates the potential added value of a multi-layer representation of vegetation canopies with respect to the traditional big leaf approach (e.g. Wesely, 1989) for simulating ozone deposition and ozone impact metrics for forest canopies. In the Appendix, It also compares for corrected version (vapor pressure and soil moisture of dry deposition such as Zhang et al (2003)).

This is an relevant paper making an interesting comparison between the proposed MLC-CHEM scheme and W89. Although the authors mention that the use of W89 mechanism is not valid for flux-based assessments (e.g from model simulation or observations), they do not discuss how their MLC-CHEM scheme could be implemented in a practical way in chemical transport models.  Since the proposed methodology is constrained by flux measurement observations, it cannot technically be implemented in regional transport models since it would require a spatial density of ozone flux measurements which does not exist over the domain of such models.  As the authors point out:  “sites with long-term ozone flux measurements are scarce especially these sites with long-term ozone flux measurements”. Therefore, the authors should explain how MLC-CHEM is an added value for current air quality models providing near real-time simulations (and using modified versions of W89).

Similarly, the authors also wrote: “Observational studies indicate that ozone deposition exhibits substantial temporal variability that is not reproduced by atmospheric chemistry models due to a simplified representation of 5 vegetation uptake processes in these models”. The authors should elaborate a bit on that and mention that the  lack of observational routine data for dry deposition is an handicap to appropriately simulate processes described by MLC-CHEM. It would be nice if clear recommendations would be given to modelers in that respect (other than “W89 is invalid”).

Note that W89 is rather obsolete nowadays as many authors have shown its weaknesses through the years. For example,  LAI scaling to canopy was omitted in W89 and should be included. VPD weakness has been corrected by using Jarvis (1976), etc.

Specific questions/issues

1) The authors wrote: “This invalidates the use of the W89 mechanism for flux-based assessments…”. The  authors would be careful with such general statement. Moreover,  the comparison with Wesely (1989) could not be entirely appropriate since very few institutions and models use W89 as in its original form. Most of models use different corrected forms of W89 ( exchange of ozone, e.g.  accounting for stomatal closure based on the vapor pressure deficit (VPD) and soil moisture, LAI scaling correction in W89,  

2) Line 115-166. W89  can be deemed representative for the representation of dry deposition in other atmospheric chemistry model. Please be careful since several models have upgraded W89 to correct for some of its weakness (include VPD deficit, wetness on foliage, corrected form LAI missing in W89, etc.). e.g. (Jarvis, 1976; Ref. Valmartin, M, Heald, CL and Arnold, SR (2014) Geophysical Research Letters, 41 (8). 2988 - 2996.etc.).

3) Conclusion: How can MLC-CHEM  be included in a regional transport model. ?. Avoidance of Wesely type parametrization might not be possible unless CO2 is better evaluated by these models (which does not seems to be the case now) and routine flux and near-real time measurements are available.

4) In the Appendix it would be useful and informative to also show  the diurnal profile of CO2 concentration (not only CO2 fluxes). It would be also interesting to provide values of CO2 compensation point and diurnal variations of CO2 concentrations  for the benefit of the reader (in Supplementary material).

Technical corrections:

1) Line 77:  detailed below --> which are detailed below.

2) Typo: Line 117  uptak   --> uptake

3) Typo: Legend of Figure 2 shaed --> shade

4) Legend of Figure 1: “and their in- and output variables” --> and their input and output variables.

5) Line 131 and 175 is a repeat of the same information.

6) Please avoid starting a section or a paragraph with the word Figure (section 3.1.1 and elsewhere)

7) Figure 5. “The shaded area in panel d highlights the daytime period (defined as 8-20 h LT) over which the stomatal flux is calculated”. It looks that the gray (shaded) area rather corresponds to overnight time not daytime. Please review and correct.

This work compares a well know big leaf model with a multi-layer canopy model to demonstrate that the latter best represent ozone deposition in two forest canopies. As the author mention, there are not many field measurements globally, therefore testing models vs observation remains challenging. However, for future studies the authors are invited to use data from other field studies which have been collected in the last decades, especially those collected in dry sites.  Results are quite predictable in the sense that a multi-layer model constrained with observation return more realistic data especially in reproducing diurnal site. Unfortunately, without field data to constrain MLC models, it is hard to obtain realistic results and therefore global application of MCL model remains challenging. The work reads well, and shed light on the need to use multi-layer models to derive ozone fluxes, in particular CUO used as a metric for ozone damage to vegetation.

Introduction

“…these are limited to short timescales. For these reasons, quantifying temporal

variability in stomatal and non-stomatal ozone deposition solely based on observations remains challenging.”

>>Please define what is an appropriate timescale to derive ozone flux partitioning. In my opinion having year-long continuous measurements is timescale long enough.

Site description

Other than a temperature broadleaf forest and a coniferous forest, it would have been wise to select also a dry forest in order to test models in environments with rather different climate conditions and vegetation features (i.e. the capacity to emit different VOCs). Furthermore, the effect of high VPD and drought on Gs is acritical point for models, therefore you missed such representation.

Lines 150-154: please provide a definition and/or reference for the relaxation factor of 0.5.

Line 154: please provide the formula or the reference to estimate In-canopy resistance.

Line 158: unclear the way adopted to estimate Ag. The latter parameter is crucial to estimate properly Gs. Have you tried to compare Cumulative Ag with GPP derived from EC measurements?

Lines 220-230: il looks like a mid-day depression in ozone deposition most likely due to partial stomatal closure is not taken into account by the MLC model.

Line 326: why do you exclude relevant chemical removal of ozone which may contribute to non-stomatal deposition? Perhaps low NOx and low BVOC emitted by the two ecosystems? Sesquitrpenoids from pines and to a minor extent monoterpenoids may react with ozone at time scales even lower than those used in the study.

Line 354: have you tried gap-filling so to produce CUOst for all year?

Line 488: rather than using the term “plant’s interior” I would say “plant’s chloroplast” or more generally “plant’s carboxylation site”.