Large crop production losses induced by global ozone stress based on interval evaluation

Andong Cai Institute of Environment and Sustainable Development in Agriculture, Chinese Academy of Agricultural Sciences, Beijing 100081, P.R. China Bin Wang Institute of Environment and Sustainable Development in Agriculture, Chinese Academy of Agricultural Sciences, Beijing 100081, P.R. China Tianjing Ren Institute of Environment and Sustainable Development in Agriculture, Chinese Academy of Agricultural Sciences, Beijing 100081, P.R. China Wenju Zhang Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences Xiaoke Wang Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing 100081, PR China Hongmin Dong Chinese Academy of Agriculture Sciences Yue Li (  liyue@caas.cn ) Chinese Academy of Agriculture Sciences


36
Air pollution, including atmospheric (or ground-level) concentrations of ozone (O3), can significantly 37 damage plant growth and biomass accumulation in terrestrial ecosystems 1,2,3 . Atmospheric O3 enters 38 the plant body through leaf stomata and stimulates a series of biochemical reactions, which destroy the 39 cell structure and initiate physiological and metabolic disorders 4, 5 . Such adverse reactions can decrease 40 stomatal conductance and net photosynthetic rate and further result in losses of biomass and yield.

41
Since the late 1800s, atmospheric O3 has risen from approximately 10 ppb to 50 ppb today and will 42 increase by 40-60% until 2100 6 .

43
For agroecosystem, an accurate assessment of how elevated atmospheric O3 affects crop productivity,

52
As atmospheric O3 rises, crop O3 absorption does not increase proportionally because of stomatal 53 resistance 4 . Crops have specific adaptability and resistance to atmospheric O3 damage through their 54 natural defenses (e.g., antioxidants, detoxification, and nocturnal remediation capabilities) and the 55 specific triggering responses 13 . These differences in ozone sensitivity may lead to significant variation 56 of yield responses relative to low and high O3 concentrations and for a time of exposure to those 57 concentrations. However, this phenomenon to date is poorly noticed and understood when evaluating 58 O3 impact 4,9,14 . Here we propose a new approach with the O3 sensitivity of crop yield (Yo) by of observational O3 data, part of this challenge has been met using atmospheric models to predict 4 regional and global O3-induced crop yield losses 3,8,16 . However, with the establishment of numerous 66 O3 monitoring stations in recent years, more accurate global real-time O3 concentration data could be 67 obtained ( Supplementary Fig. 5). An interval evaluation of Yo based on observational real-time O3 data 68 is crucial to consistent predictions of crop yield response to changes in atmospheric O3.

69
Based on an evaluation of over 900 O3 fumigation experiments, we present a detailed analysis to 70 provide a dose-response of O3 exposure on yield of major crops (wheat, maize, rice, and soybean)

101
Generally, the stomatal conductance of soybean is more significant, and its response to O3 was 102 significantly weaker than wheat and rice ( Supplementary Fig. 10) 9, 21, 22 . In terms of non-stomatal 103 factors, the mesophyll of dicotyledonous (e.g., soybean and most vegetables) is differentiated into 104 palisade tissue and spongy tissue compared with monocotyledonous (e.g., wheat, maize, and rice).

105
Palisade tissue is close to the upper epidermis and contains more chlorophyll. High O3 can firstly 106 damage the palisade tissue and then cause cytoplasmic wall separation and cell content dispersion,

107
inhibiting crop growth and yield formation 23 . Therefore, soybean and vegetable were more sensitive 108 to elevated O3 than wheat, maize, and rice.

109
There were significant differences in the Yo for the same crop type among O3 fumigation

121
Simultaneously, the activities of superoxide dismutase, catalase, and peroxidase increase rapidly under 122 O3 stress, which is the first step to defend against reactive oxygen species damage caused by O3 4, 25 .

123
Other pathways enhance crop resistance and lead to the parabola relationship between the Yo and O3 124 concentration, i.e., high O3 stress can accelerate crop respiratory metabolisms and promote

126
Interestingly, when O3 concentration exceeded 100 ppb, the crop yield loss rate increased with an 127 elevated O3 concentration of 1 ppb h -1 (Fig. 1c). Long-term high O3 exposure can reduce stomatal 128 resistance and destroy the antioxidant system 26 , which would lead to irreversible damage to the crop.

129
Overall, the parabola relationship between the Yo and O3 concentrations is an advanced indicator of 130 how crop yield responds to different O3 concentrations.   intervals was robust by using the accurate Yo and hourly O3 data from more than 7,000 O3 monitoring 165 stations in our study ( Supplementary Fig. 4). The crop yield loss rate differed significantly among 166 divergent O3 intervals (Fig. 2, Supplementary Fig. 6-9). This was mainly because the occurrence 167 frequency of low O3 concentration was much higher than that of high concentration (Supplementary   168   Table 1). It is also affected by the relationship between the Yo and O3 concentration (Fig. 1c). Our      The mechanisms of crop production loss. The magnitudes of Yo varied greatly among experimental 206 sites, ranging from -44.9 ×10 -6 to -4.8×10 -6 ppb h -1 (Fig. 1). Path analysis showed a network of 207 inter-correlation of atmospheric O3 concentration, photosynthetic indexes, and agronomic indexes in 208 determining the Yo (Fig. 3a), implying that the effect size of Yo was regulated by multiple factors rather

268
To standardize the database, experimental data were only when the following criteria were met: (1) recorded and used to test the publication bias ( Supplementary Fig. 2). Overall, the sites of our global 306 study spanned from -1.26° to 57.92° and -123.23° to 140.21° in latitude and longitude, respectively.

307
The O3 sensitivity of crop yield. The Yo is crop yield loss rate (%) relative to elevated O3 308 concentration per 1 ppb h -1 . This is applied to normalize the effects of atmospheric O3 on crop 309 production. One primary objective of our study was to precisely define the Yo under different O3 310 intervals using a meta-analysis approach. Meta-analysis is a comprehensive statistical strategy to 311 systematically combine and quantitatively evaluate multiple independent research results with a 312 common research purpose, which is particularly suitable for the large-scale study 36 .

313
Before performing the meta-analysis, the quality of experimental data was using the "metainf" 314 package ( Supplementary Fig. 2) 37 . This is a method of combining publication bias and treatment to 315 explore any source of publication bias. Such discrimination could reduce the small-sample effects by 316 publication bias and ensure the credibility of the results 38 . If a control corresponds to more than one 317 experimental treatment at a study site, such treatments are considered non-independent of sampling.

318
Previous studies have shown that the non-independence of the sample can significantly affect the

361
The Yo (ppb h -1 ) was calculated as follows:

369
according to O3 concentration of fumigation treatment (Fig. 1c). The Yo of maize under the 30-40 ppb 370 interval was obtained using a fitting equation due to the missing data ( Supplementary Fig. 4). Based on 371 the above principles and methods, we also calculated the O3 sensitivity of photosynthetic indexes

395
where i, m, and g are the ith O3 intervals, the mth hour and gth day, respectively. ni is total O3 intervals 396 for an hour. nm is 12, meaning 12 hours per day (08:00-20:00). ng is the number of days of crop 397 growing season.

398
Although we have tried to obtain crop yield loss rate for more than 7,000 stations worldwide,         Table 1 Global and five areas (ranked top five in grain production) annual loss amount of wheat,

575
Financial supports from the National Natural Science Foundation of China (42007073) is gratefully 576 acknowledged. We thank all the researchers whose data were used in this global synthesis. We are also 577 grateful to some of the environmental sites for providing real-time ozone data.