Impact Assessment of Ozone Absorbed through Stomata on Photosynthetic Carbon Dioxide Uptake by Japanese Deciduous Forest Trees: Implications for Ozone Mitigation Policies

: Photosynthesis by trees is expected to contribute to preventing climate change including global warming. However, the current levels of tropospheric ozone (O 3 ) reduce the uptake of photosynthetic carbon dioxide (CO 2 ) by forest trees in Japan, which is a concern. Furthermore, it is unknown how much O 3 should be reduced to prevent plants’ O 3 -induced damage. The objective of the present study was to assess the negative e ﬀ ects of the current levels of O 3 absorbed via stomata and the impact of its mitigation on the CO 2 uptake by trees in Japanese forests. Impact assessment, targeted in 2011–2015, was performed for four deciduous broad-leaved trees: Fagus crenata , Quercus serrata , Q . mongolica var. crispula , and Betula platyphylla var. japonica . The assessment was based on species-speciﬁc cumulative stomatal O 3 uptake (COU) and species-speciﬁc responses of annual photosynthetic CO 2 uptake to COU. Annual COU di ﬀ ered between the four trees; the average COU of F . crenata , Q . serrata , Q . mongolica var. crispula , and B . platyphylla var. japonica across Japan was 41.7, 26.5, 33.0, and 29.1 mmol m − 2 , respectively, and the reductions in CO 2 uptake by the four trees were 14.0%, 10.6%, 8.6%, and 15.4%, respectively. Further analysis revealed that reducing the atmospheric O 3 concentration by approximately 28%, 20%, 17%, and 49% decreased the O 3 -induced reductions in photosynthetic CO 2 uptake to 5% in F . crenata , Q . serrata , Q . mongolica var. crispula , and B . platyphylla var. japonica , respectively. In the near future, implementing mitigation measures for the O 3 damage in plants is expected to enhance the photosynthetic capacity of Japanese forest tree species.


Introduction
Climate change, including rising air temperatures and changes in precipitation, is recognized globally as a serious environmental problem [1]. Atmospheric carbon dioxide (CO 2 ) is one of the major factors causing climate change [1]. The CO 2 concentration has increased to approximately 400 the trees were distributed, respectively. Third, COU in the tree distribution areas (Ministry of the Environment, http://www.biodic.go.jp/index.html) was estimated by interpolating the COU at the monitoring stations of O 3 using ordinary kriging. Fourth, according to species-specific responses of annual CO 2 uptake to COU reported by Yamaguchi et al. [18], O 3 -induced reductions in CO 2 uptake for the four trees were calculated. Finally, the effects of O 3 mitigation, i.e., impact of reducing the surface O 3 concentration, on CO 2 uptake were calculated for the four tree species by simulating the input O 3 concentrations. A summary of information on simulation settings, required data, and data sources is provided in Table 1.
where F st is the stomatal O 3 uptake, [O 3 ] is the atmospheric O 3 concentration, R s_O3 and R b_O3 are the stomatal resistance and leaf boundary layer resistance to O 3 , respectively, g s is the stomatal conductance to H 2 O, L d is the crosswind leaf dimension, and u is the wind speed. The g s was estimated according to the g s model provided by Kinose et al. [19], originally developed by Jarvis [21]: where g max is the maximum g s ; f min is the ratio of minimum g s to g max ; and f phen , f light , f temp , f VPD , f SWC , and f O3conc represent responses of g s /g max to leaf phenology, photosynthetic photon flux density (PPFD), air temperature, vapor pressure deficit, volumetric soil water content, and atmospheric O 3 concentration, respectively. These parameters are species-specific and were essentially based on our previous results (see details in Kinose et al. [19]). The leaf emergence date and leaf senescence date were provided by MODIS Land Cover Dynamics Product, MCD12Q2 v. 006, in 500-m grids in the world. Seasonal changes in enhanced vegetation index (EVI) of the MCD12Q2 were substituted for the f phen . These phenological data were obtained considering tree distribution using vegetation raster data (45 × 30 mesh) of the National Survey on the Natural Environment (Ministry of the Environment), indicating the distribution of dominant species, although uncertainties exist in the dataset because the forest is not necessarily pure. In the present study, g s was considered to be unchanged according to soil water content, i.e., f SWC = 1, because the amount of precipitation was relatively high in Japan (annual mean precipitation: approximately 1700 mm). The concentrations of photochemical oxidants are officially monitored at approximately 1200 monitoring stations across Japan (National Institute for Environmental Studies). In the present study, the concentration of photochemical oxidants was considered as the concentration of O 3 [10]. The meteorological data such as air temperature and wind velocity officially monitored at approximately 1300 monitoring stations throughout Japan (Japan Meteorological Agency) were used for the calculation of COU. PPFD was estimated based on the study by Masaki et al. [22]. The vapor pressure deficit (VPD) was estimated regarding the vapor pressure during the daytime as the saturated vapor pressure at the minimum temperature on the day, as the relative air humidity is almost 100% at the minimum temperature. The distribution of monitoring stations of photochemical oxidants and meteorological factors and regional partition in Japan is shown in Figure 1. changes in enhanced vegetation index (EVI) of the MCD12Q2 were substituted for the fphen. These phenological data were obtained considering tree distribution using vegetation raster data (45′′ × 30′′ mesh) of the National Survey on the Natural Environment (Ministry of the Environment), indicating the distribution of dominant species, although uncertainties exist in the dataset because the forest is not necessarily pure. In the present study, gs was considered to be unchanged according to soil water content, i.e., fSWC = 1, because the amount of precipitation was relatively high in Japan (annual mean precipitation: approximately 1700 mm). The concentrations of photochemical oxidants are officially monitored at approximately 1200 monitoring stations across Japan (National Institute for Environmental Studies). In the present study, the concentration of photochemical oxidants was considered as the concentration of O3 [10]. The meteorological data such as air temperature and wind velocity officially monitored at approximately 1300 monitoring stations throughout Japan (Japan Meteorological Agency) were used for the calculation of COU. PPFD was estimated based on the study by Masaki et al. [22]. The vapor pressure deficit (VPD) was estimated regarding the vapor pressure during the daytime as the saturated vapor pressure at the minimum temperature on the day, as the relative air humidity is almost 100% at the minimum temperature. The distribution of monitoring stations of photochemical oxidants and meteorological factors and regional partition in Japan is shown in Figure 1. The gaussian model was used as a semi-variogram model in the ordinary kriging to interpolate the annual COU at the O3 monitoring stations, which was clipped by the tree distribution. As in a previous study [23], we did not interpolate the input data for calculating COU, i.e., meteorological data and O3 data, because the interpolation is extremely difficult due to enormous calculation time. Kriging interpolation was performed using SAGA GIS v. 6.4.0 [24].

Impact Assessment of O3-induced Reduction in Annual Photosynthetic CO2 Uptake
The responses of annual photosynthetic CO2 uptake to COU were fitted by the Weibull function, and the O3-induced reduction in annual CO2 uptake was calculated as follows: O3-induced reduction in annual photosynthetic CO2 uptake = 100 − 100 × exp[−(COU/ω) λ ] where ω and λ are species-specific parameters of the Weibull function. ω and λ were originally proposed by Yamaguchi et al. [18] and were modified to make the intercept 100 ( Figure 2). The gaussian model was used as a semi-variogram model in the ordinary kriging to interpolate the annual COU at the O 3 monitoring stations, which was clipped by the tree distribution. As in a previous study [23], we did not interpolate the input data for calculating COU, i.e., meteorological data and O 3 data, because the interpolation is extremely difficult due to enormous calculation time. Kriging interpolation was performed using SAGA GIS v. 6.4.0 [24].

Impact Assessment of O 3 -induced Reduction in Annual Photosynthetic CO 2 Uptake
The responses of annual photosynthetic CO 2 uptake to COU were fitted by the Weibull function, and the O 3 -induced reduction in annual CO 2 uptake was calculated as follows: (6) where ω and λ are species-specific parameters of the Weibull function. ω and λ were originally proposed by Yamaguchi et al. [18] and were modified to make the intercept 100 ( Figure 2). Since previous studies indicated that leaf morphological and physiological characteristics, such as leaf mass per area and gross photosynthetic rate, are related to the sensitivity to O3 [7,25], the leaflevel photosynthetic sensitivity to O3 might be different between seedlings and mature trees due to the differences in their leaf characteristics. However, this has not been clarified to date. In the present study, therefore, we assumed that the responses of CO2 uptake to COU obtained in the seedlings in Yamaguchi et al. [18] were the same as those in the mature trees, although this might cause estimation errors in the impact of O3 on CO2 uptake to some extent.

Figure 2.
Responses of annual CO2 uptake per unit leaf area to cumulative stomatal O3 uptake (COU) of F. crenata, Q. serrata, Q. mongolica var. crispula, and B. platyphylla var. japonica seedlings. These responses were obtained from the study by Yamaguchi et al. [18] with slight modifications. The Yintercept was modified to 100 in the present study. Plots indicate the first-flush or early leaves (○) or second-flush or late leaves (×). The Weibull function was fitted to the response, and its parameters, ω and λ, and the mean absolute error (MAE, %) are indicated in each figure.  Responses of annual CO 2 uptake per unit leaf area to cumulative stomatal O 3 uptake (COU) of F. crenata, Q. serrata, Q. mongolica var. crispula, and B. platyphylla var. japonica seedlings. These responses were obtained from the study by Yamaguchi et al. [18] with slight modifications. The Y-intercept was modified to 100 in the present study. Plots indicate the first-flush or early leaves ( ) or second-flush or late leaves (×). The Weibull function was fitted to the response, and its parameters, ω and λ, and the mean absolute error (MAE, %) are indicated in each figure.

Results
Since previous studies indicated that leaf morphological and physiological characteristics, such as leaf mass per area and gross photosynthetic rate, are related to the sensitivity to O 3 [7,25], the leaf-level photosynthetic sensitivity to O 3 might be different between seedlings and mature trees due to the differences in their leaf characteristics. However, this has not been clarified to date. In the present study, therefore, we assumed that the responses of CO 2 uptake to COU obtained in the seedlings in Yamaguchi et al. [18] were the same as those in the mature trees, although this might cause estimation errors in the impact of O 3 on CO 2 uptake to some extent.

Discussion
The present study firstly evaluated the impact of the current levels of O 3 absorbed via stomata on the annual cumulative photosynthetic CO 2 uptake per unit leaf area of four Japanese forest tree species. The O 3 -induced reduction in annual CO 2 uptake averaged across Japan was 14.0%, 10.6%, 8.6%, and 15.4% in F. crenata, Q. serrata, Q. mongolica var. crispula, and B. platyphylla var. japonica, respectively (Figures 5 and 6). This result indicated that the current atmospheric O 3 concentrations have the potential to reduce the capacity of CO 2 uptake by Japanese forest tree species.
Kinose et al. [19] indicated that when the four tree species were grown in the same environmental conditions, the COU was ranked as follows: B. platyphylla var. japonica > F. crenata > Q. mongolica var. crispula > Q. serrata. However, the tree distributions and environmental growth conditions greatly differ between the four tree species. Betula platyphylla var. japonica is mainly distributed in the Hokkaido region, where the stomatal O 3 uptake was relatively low (Figure 3). As a result, the COU of B. platyphylla var. japonica averaged across Japan was lower than that of F. crenata and Q. mongolica var. crispula (Figures 3 and 4). Nevertheless, the impact of O 3 on the CO 2 uptake of B. platyphylla var. japonica was more remarkable than that of the other tree species (Figures 5 and 6). This is due to the relatively higher sensitivity of photosynthetic CO 2 uptake to COU in B. platyphylla var. japonica (Figure 2).
Several studies have assessed the effects of O 3 absorbed via stomata on gross or net primary production (GPP or NPP, respectively) in forest ecosystems [9,26]. Lombardozzi et al. [26] estimated the O 3 -induced reduction in GPP targeted at global forest based on the relationships between COU and maximum carboxylation rate and between COU and stomatal conductance derived in a specific tree species, the tulip poplar. Sitch et al. [9] evaluated the impact of O 3 on the GPP of five plant functional types, namely broad-leaved trees, conifers, C 3 grasses, C 4 grasses, and shrubs, considering different sensitivities of the net photosynthetic rate to O 3 among the plant functional types. However, the sensitivity of the photosynthetic capacity to O 3 is extremely different between tree species possibly because of the differences in the defensive ability to O 3 [18]. The advantage of our impact assessment is that we considered the species-specific responses of photosynthetic CO 2 uptake to COU, even though there was uncertainty in the evaluation of O 3 -induced reduction rates by 3.1%, 3.6%, 3.8%, and 6.2% in F. crenata, Q. serrata, Q. mongolica var. crispula, and B. platyphylla var. japonica, respectively ( Figure 2).
In the present study, COU and O 3 -induced reduction rates of CO 2 uptake were relatively high in Kanto, Chubu, Kinki, and Chugoku regions (Figures 3-6). Akimoto et al. [15] reported that the contribution of the transboundary transport of O 3 from the Asian continent to its concentration in Japan seemed to be relatively small except for that in the Kyushu region. According to Nagashima et al. [27], the contribution to the O 3 concentrations in Japan from China and the Korean Peninsula were estimated to be approximately 10% and 4%, respectively, although the domestic contribution was approximately 42% in the Japanese summer. Furthermore, O 3 is recognized as a short-lived climate pollutant (SLCP). These results indicate the necessity and effectiveness of domestic countermeasures against O 3 formulation for avoiding the O 3 -induced reduction in CO 2 uptake by Japanese forest tree species.
The concept of critical level of O 3 has been discussed in Europe to prevent its damage to plants. The critical levels of O 3 for European beech and European birch were defined as 5 ppm h of daylight AOT40 to diminish the growth reduction to a value of 5% [28]. In Japan, a provisional critical level based on AOT40 is suggested for the tree species sensitive to O 3 , namely 8-15 ppm h (6 months) corresponding to a 10% growth reduction [29]. However, the critical level based on stomatal O 3 uptake has been developed because the O 3 absorbed through stomata is more physiologically related to the O 3 damage to tree growth [30]. Critical levels of the two European trees were set at 5.2 mmol m −2 of POD 1 (phytotoxic O 3 dose above the threshold of 1 nmol m −2 s −1 ) corresponding to a 4% growth reduction [28]. Although this concept of the critical level of O 3 is more scientifically plausible, how much O 3 should be reduced is unclear. While the stomatal O 3 uptake concept is significantly useful to assess the impact of O 3 on plants, to create and implement mitigation measures against O 3 damage to plants, determining the atmospheric O 3 concentration is needed. To the best of our knowledge, the present study is the first to evaluate the impact of the mitigation of O 3 absorbed via stomata on the annual photosynthetic CO 2 uptake. As a result, we clarified that O 3 concentrations should be reduced by 28% (2011-2015 range: 19-34%), 20% (11-23%), 17% (8-21%), and 49% (46-51%) for F. crenata, Q. serrata, Q. mongolica var. crispula, and B. platyphylla var. japonica, respectively, to diminish the O 3 -induced reduction to a value of 5% (Figure 7). Because the domestic contribution of O 3 formation is 42% in the Japanese summer [27], it would be impossible to prevent the O 3 damage by 5% for B. platyphylla var. japonica by domestic control of the emission of O 3 precursors. In contrast, such a mitigation measure may be effective for the other tree species: F. crenata, Q. serrata, and Q. mongolica var. crispula. We provisionally propose that the O 3 concentration should be reduced by approximately 30% to decrease its damage on CO 2 uptake below 5% in some Japanese forest tree species. We hope our results will contribute to creating and implementing mitigation measures against the detrimental impact of O 3 on forest tree species in Japan and to enhancing their photosynthetic CO 2 uptake.
There are some uncertainties and future issues in the assessment on the impact of O 3 on the photosynthetic CO 2 uptake by forest tree species, including estimation accuracies of the stomatal conductance and errors of COU induced by the interpolation or extrapolation by kriging. One of the major issues is whether the parameterization results of the stomatal conductance model and the responses of CO 2 uptake to COU obtained in the particular environment of the experimental studies by Kinose et al. [19] and Yamaguchi et al. [18] could be applicable to the whole Japan in 2011-2015. Alonso et al. [31] reported different parameter values of the stomatal conductance model of Quercus ilex between marine-influenced sites and continental sites, indicating that the values of the model parameters depend on the environmental growth conditions of the trees. Actually, the optimum air temperature for stomatal opening depends on the growth conditions of F. crenata seedlings in Japan possibly because of an acclimation to the air temperature [32]. Therefore, the application of the stomatal conductance model of Kinose et al. [19] to the whole Japan may cause estimation errors of COU. Furthermore, Kinose et al. [7] indicated a higher sensitivity in the high-fertilized F. crenata seedlings due to the higher COU. In a forest area with a higher soil nutrient status, stomatal conductance may be higher via photosynthetic enhancement and be limited to some extent with shorter leaf longevity [33], indicating nutrient-dependent g max and f phen . In the present study, nutrient-dependent f phen may be partly considered because f phen was determined from the seasonal changes in EVI provided by satellite data. However, g max was regarded as a constant in the present study, which could cause estimation errors of COU. The nutrient-induced modification of g max in the calculation of COU is an issue to be considered in the future. On the other hand, the responses of CO 2 uptake to COU can differ depending on the growth conditions because the environmental conditions, such as atmospheric CO 2 concentration and soil drought, affect the leaf detoxification capacity to O 3 [34,35]. However, information on the effects of tree growth environment on the responses of CO 2 uptake to COU is insufficient to date. In order to perform a more accurate evaluation of the impact of O 3 on the photosynthetic CO 2 uptake by forest tree species, the effects of environmental growth conditions on COU and responses of CO 2 uptake to COU should be considered.

Conclusions
We evaluated the O 3 -induced reduction in annual CO 2 uptake per unit leaf area of Fagus crenata, Quercus serrata, Quercus mongolica var. crispula, and Betula platyphylla var. japonica in 2011-2015 based on the COU per unit leaf area and species-specific responses of the annual CO 2 uptake to COU. The current levels of ambient O 3 reduced the annual CO 2 uptake of F. crenata, Q. serrata, Q. mongolica var. crispula, and B. platyphylla var. japonica by 14.0%, 10.6%, 8.6%, and 15.4%, respectively. Further analysis revealed that, to diminish the O 3 damage to a value of 5%, O 3 concentrations should be reduced by 28%, 20%, 17%, and 49% for F. crenata, Q. serrata, Q. mongolica var. crispula, and B. platyphylla var. japonica, respectively. We hope that implementing mitigation measures against O 3 damage will help increase the photosynthetic CO 2 uptake by Japanese forest tree species.