the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 30 Aug 2024
Global flux-based ozone risk assessment for wheat up to 2100 under different climate scenarios
Abstract. The negative effects of tropospheric ozone (O3) on vegetation can lead to reduced photosynthesis, accelerated leaf senescence, and other negative outcomes which affect crop yields and biodiversity. This study presents a flux-based assessment of the global impact of O3 on bread wheat (Triticum aestivum) for the 21st century, under various climate scenarios (Shared Socioeconomic Pathways, SSPs). A dual-sink big-leaf dry deposition model is employed to estimate the phytotoxic ozone dose (POD) absorbed by wheat through stomata, integrating data from two Earth System Models (ESMs) from the Coupled Model Intercomparison Project 6 (CMIP6). The study explores spatial and temporal variations in O3 concentrations and the effects of climate variables on stomatal conductance, explaining changes in POD from the present time to the century’s end. The results indicate significant regional disparities in O3 dose for wheat, particularly under weak O3 precursor emissions control scenarios. The most vulnerable regions include Northern Europe, East China, and the Southern and Eastern edges of the Tibetan Plateau, where the POD increase by the end of the century is expected to be most pronounced. Conversely, POD decreases worldwide under stringent pollution emission control scenarios. However, in some regions, changes in POD may be driven more by climate variables and their interaction with O3, rather than by O3 concentrations alone. Therefore, this study emphasizes the need for effective emission mitigation policies of both O3 precursors and greenhouse gases to preserve global food security from O3 damages.
- Preprint
(5077 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-2573', Anonymous Referee #1, 04 Oct 2024
General comments
This manuscript describes a risk assessment exercise of ozone effects on wheat production at the global scale based on modelled phytotoxic ozone dose (POD6). POD6 values are calculated following the CLRTAP methodology, using as input data the ozone concentration and other meteorological variables produced by two Earth system models, run under different shared socioeconomic pathways with contrasting radiative forcing and atmospheric pollution control policies over the XXI century. Changes in POD6 under the different scenarios are described for different regions of the world. The influence of meteorological conditions and ozone concentration on the final POD6 estimate and relative wheat yield loss are also analysed. This kind of analysis is useful to understand how climatic changes may affect the negative effects of tropospheric ozone pollution on agricultural yield in the future rather than focusing solely on changes in ozone concentration. The analysis provides also insights on the relative weight of air pollution policies and radiative forcing scenarios on the risk of ozone effects on agricultural yields by the end of the XXI century. The relative influence of these two factors changes for different regions of the world, highlighting the co-benefits of controlling both ozone precursor and greenhouse gases emissions to the atmosphere.
The manuscript is well structured and written and make a relevant contribution to the field of risk assessment effects on agricultural production under future climatic conditions. The objectives, datasets, methods and results are clearly presented and easy to follow, although I have some comments that may need minor revisions of the manuscript:
Specific comments
Line 138 and Table 1. From the list of variables in Table 1, I am not able to find soil moisture. Thus, I understand that soil moisture was modelled based on GFDL and UKESM outputs. Please explain, or cite methods, on how soil moisture was modelled and how plant available water was computed from soil moisture data.
Line 169. Crop geometry refers to the development of LAI and SAI over the course of the growing season, as a function of thermal time, as described in Guaita et al 2023? If that is the case, I would specify this somewhere in the manuscript, since the term “crop geometry” seems confusing to me. Was plant height changed as well over the course of the growing season?
Line 176. The Mediterranean wheat parameterization for soil moisture refers to volumetric water content while in this modelling exercise, plant available water is used. Please describe how was this modified in your calculations.
Line 191. It seems that there is a typo in the reference Guaita et al 2023b, or the reference is missing in the bibliography.
Line 215. I think this sentence needs some clarification. The prescribed sowing dates come from Qiao et al. (2023), while thermal times describing the phenological development of wheat were taken from González-Fernández et al., 2013 and Grünhage et al., 2012). Thus, I understand that if the thermal time at leaf senescence (according to González-Fernández et al., 2013 and Grünhage et al., 2012) was not reached before the following prescribed sowing date (according to Qiao et al 2023), then the node was excluded from POD6 calculation.
Line 221. Unfortunately, I am not familiar with this sort of spatial analysis to assess its use, although the concept and results obtained look reasonable to me. However, I miss some comments about the tests conducted to assess the assumptions of the ANOVA analysis.
Line 239. How the accumulation period changed between the baseline and the 2100 scenarios? Was it earlier and shorter, or longer in different regions of the world?
Lines 244-247. This statement might be more appropriate for the discussion section.
Line 272. Desert regions or arid regions? In desert regions, it will be unlikely to find rainfed wheat crops?
Figure 1. Please explain in caption the meaning of areas in white.
Table 4. Means presented in Table 4 are averaged between the two CMIP6 models? Please clarify this.
Lines 366-369. The mean or median relative yield loss could be also a very helpful metric to describe the range of the expected risk to wheat production
Line 379. How should the very low Pearson correlation coefficients 0.15 > r > -0.15 covering relatively large areas in the map on Figure 6 be interpreted? Is this also reflecting a lack of agreement between models? Higher uncertainty in the expected changes?
Lines 403-406. The description of POD6 increases under non-limiting soil moisture conditions is confusing to me. Most extensive increases in POD6 happens in Southern Europe and South Asia, but the average value of increase seems lower than for other regions like South-East Asia or North-Africa. The values reported are averaged also across models?
Line 447. Does not
Line 455. The discussion should include some comments regarding the performance of the models chosen in this study to simulate ozone concentrations (as mentioned in line 245). Also, uncertainties stemming from the use of particular models: in this study the general pattern matches, but one model predicts bigger changes in POD6 and higher risks of O3 effects on yield compared with the other one, and large areas show small Pearson’s correlation coefficients between modes, as shown in figure 6. I wonder if multi-model ensembles could be a useful tool for future projections.
Also, there is one variable that will change depending on SSP scenarios compared with the current situation that also affects stomatal conductance and likely PO6 but is not taken into account with this methodology, such as the CO2 concentration. This should be considered an additional uncertainty and limitation of this approach.
On the positive side, it would be interesting to comment on the advantages of assessing tropospheric ozone effects on agricultural yield under future climatic conditions using the POD approach as compared with other assessments based only on changes in ozone concentration.
Finally, it could be stressed that the results presented here support the co-benefits of abating greenhouse gases and air pollution emissions jointly to help in the mitigation of air pollution effects in agriculture.
Line 479. Subscript missing in O3.
Citation: https://doi.org/10.5194/egusphere-2024-2573-RC1 -
RC2: 'Comment on egusphere-2024-2573', Anonymous Referee #2, 24 Oct 2024
This paper presents calculations of POD6 from current conditions to the year 2100. Although changes in crop yield due to ozone over this period are of interest, I find it difficult to know what to make of the results when I have not been given any impression about whether the models can actually predict ozone and especially POD6 to any satisfactory degree in the base run. The lack of comparison with measurements is even more surprising given that this manuscript was submitted to the TOAR-II Special issue!
I am afraid I find this omission to be too significant to be ignored, and therefore cannot recommend this manuscript for publication. More detailed comments follow.
Major comments
1. The major weaknesses of this paper are that the authors have chosen to model a very difficult ozone metric (POD6), and they present no evidence to show that the models used have any ability to model that metric (or indeed any other), even for present day conditions.2. First, about the metric itself. Why was POD6 chosen? It is well known that ozone metrics such as PODY can be very difficult to estimate, especially when the Y threshold is very high (e.g. Sofiev and Tuovinen et al, 2001, Tuovinen 2000, Touvinen et al, 2007). POD is also a difficult metric to obtain from observations because its calculation requires a large number of parameters, assumptions and auxiliary measurements that are usually not available. Such problems explain why the otherwise comprehensive TOAR database of vegetation-relevant ozone metrics (Lefohn et al, 2018, Mills et al, 2018a) did not include estimates of POD.
3. Further, the LRTAP mapping manual makes it clear that the so-called PODYSPEC metrics (including POD6SPEC) are intended for situations where ozone and meteorological variables can be accurately estimated at the flag leaf of a wheat plant. Global scale model simulations are not at all well suited to making accurate predictions of POD6. Indeed, the LRTAP manual suggests that large-scale simulations make use of the a lower Y threshold, and some simpler parameter settings, which they denoted POD3IAM.
4. For these reasons the global scale POD assessments of Mills et al. (2018b,c) made use of POD3IAM metric. And although neither of the Mills papers was able to evaluate even this POD3 metric globally, they did show that the EMEP chemical transport model that was used was able to satisfactorily reproduce some basic statistics, namely mean of daily maximum ozone and M7, at sites from around the globe (Mills et al., 2018b, SI), and that model had been extensively tested against field data relevant to ozone deposition and fluxes.
5. The usual problems of accurately modeling O3 and its metrics are exacerbated when climate models are used. In this case the meteorology is not constrained by reanalysis, and hence diverges more from the real-world than usually seen in current day chemical transport models. So, how well can your models predict O3, M7, and AOT40 for example (ie the metrics which can be derived from global observations), and indeed the hourly frequency distribution of O3?
6. A related issue is also that this paper seems to use quite short slices of meteorology. The base simulation is for 15 years (2000-2014), and the climate runs seem to be for 10 years (though I am a little confused by the 10 year slices given on L134 and the 15-year slice mentioned on L116). With short time-slices there is an increased risk that changes seen are due to random variations rather than to a true climate signal. Even with 20 year time-slices Langner et al. (2012) showed that the changes seen in summertime ozone were not significant at the 95% level over large parts of Europe.
7. In the manuscript here, these is no discussion of these key issues. Instead we are referred to Turnock et al. (2020) for information about model skill, but that paper states that "CMIP6 models consistently overestimate observed surface O3 concentrations across most regions and in most seasons by up to 16 ppb, with a large diversity in simulated values over Northern Hemisphere continental regions".
8. Given such issues with surface O3 and the modeling in general, I have no reason to believe that the POD6 values have satisfactory values, or that trends in this metric are any more reliable.
Other comments
p2, L48. Strange not to mention the Mills et al. papers here, or that of Van Dingenen et al., 2009; these papers both offer both global-scale assessments which involved a lot of work (including comparison with observations) and are widely cited.p4, Table 1. This table should also include the thickness of the lowest model, as this is the important for deriving crop-height O3 concentrations.
p5. on L116, we read that the baseline is calculated for 15 years, over 2000-2014. How many years are used for the 2100 simulations?
p5, L139. What does "Contextually" mean here? It sounds odd.
p5. Sect.2.2 The text suggests that wilting point and field capacity are needed, and derived as volumetric soil moisture (VSM) values. One issue is that the ESMs will have their own systems for dealing with soil water, and their calculations of near-surface ozone and resistances in general will presumably reflect their interpretation of soil-water effects. Possibly more serious is the use of volumetric soil water (VSM). The same VSM can represent very wet conditions in some soils, but very dry in others, but as far as I can tell the methods don't distinguish between different soils at all.
p6. The text here omits any mention of the difference between leaf and canopy scale resistances, but this is a key part of the DO3SE methodology (e.g. Tuovinen et al., 2009)
p7, L173. Again the word contextually is used. It fits better here than in the above example, but I think it is better to say "In the context of...".
p7, L174. The word parameterizations is a bit vague, and readers cannot be expected to know what this means. Please make a table with the parameter values.
p7, L184. In what way are the Feng et al. (2012) parameterizations incomplete? I would have thought that methods developed from China were more appropriate for global approaches than those from Spain.
p7, L195 "A well-established dose-response relationship...". Is this so well established? The mapping manual states "the percentage effect due to O3 impact on crop yield estimated in large-scale modeling should be calculated as follows:
(PODYIAM – Ref10 PODYIAM) * (% reduction per mmol/m2 PODYIAM.POD3)
And indeed, Mills et al (2018c) used:
RYL = (POD3IAM-0.1)*0.64
but this manuscript uses POD6 rather than the recommended POD3IAM for unexplained reasons, and makes no mention of the "Ref10" correction.
p7, L202--204. This sentence seems out of place compared to the preceding text.
p24, L456. Given all the uncertainties I mentioned in the major comments section, I wonder what the phrase "Our results may be associated with different degrees of confidence depending on the agreement between the two available CMIP6 models," means? The paper has barely mentioned the main sources of uncertainty I think.
p42, L890 UKESM - 20m. Is that cell depth, or cell-center?
References
Langner, J., et al., European summer surface ozone 1990–2100, Atmos. Chem. Physics, 12, 10097–10105, https://doi.org/10.5194/acp-12-10097-2012, 2012.
Lefohn, A. S., et al., Tropospheric ozone assessment report: Global ozone metrics for climate change, human health, and crop/ecosystem research., Elem Sci Anth., 6, 28, https://doi.org/http://doi.org/10.1525/elementa.279, 2018.
Mills, G., et al., Tropospheric Ozone Assessment Report: Present-day tropospheric ozone distribution and trends relevant to vegetation., Elem. Sci. Anth., 6, https://doi.org/10.1525/elementa.302, 2018a.
Mills, G., et al., Ozone pollution will compromise efforts to increase global wheat production, Global Change Biol., 24, 3560–3574, https://doi.org/10.1111/gcb.14157, 2018b.
Mills, G., et al., Closing the global ozone yield gap: Quantification and cobenefits for multistress tolerance, Global Change Biol., https://doi.org/10.1111/gcb.14381, 2018c.
Sofiev, M. and Tuovinen, J.-P.: Factors determining the robustness of AOT40 and other ozone exposure indices, Atmos. Environ., 35, 3521–3528, 2001.
Tuovinen, J.-P.: Assessing vegetation exposure to ozone: properties of the AOT40 index and modifications by deposition modelling, Environ. Poll., 109, 361–372, 2000.
Tuovinen, J.-P., et al., Robustness of modelled ozone exposures and doses, Environ. Poll., 146, 578–586, 2007.
Tuovinen, J.-P., et al., Modelling ozone fluxes to forests for risk assessment: status and prospects, Annals of Forest Science, 66, 401, 2009.
Turnock, S. T., et al., Historical and future changes in air pollutants from CMIP6 models, Atmos. Chem. Physics, 20, 14547–14579, https://doi.org/10.5194/acp-20-14547-2020, 2020.
Van Dingenen, R., et al, The global impact of ozone on agricultural crop yields under current and future air quality legislation, Atmos. Environ., 43, 604–618, https://doi.org/10.1016/j.atmosenv.2008.10.033, 2009.
Citation: https://doi.org/10.5194/egusphere-2024-2573-RC2 - CC1: 'Comment on egusphere-2024-2573', Owen Cooper, 29 Oct 2024
Data sets
Ozone risk assessment (model output) Pierluigi Renan Guaita and Giacomo Alessandro Gerosa https://doi.org/10.5281/zenodo.13485000
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 177 | 57 | 34 | 268 | 5 | 6 |
- HTML: 177
- PDF: 57
- XML: 34
- Total: 268
- BibTeX: 5
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1