Meteorological Pre-warning Grade of High Temperature During Flowering Stage for Summer Maize in North China Plain

High temperatures during flowering stage affects the seed setting in maize and thereby result in significant yield penalties in recent changing climate era. Based on the daily maximum temperature data of the maize growing season from 1971 to 2019 at the surface meteorological observatory in Henan Province of China and maize observation data at the agricultural meteorological observatory, an early warning grade of high temperatures during the summer maize flowering stage was constructed. The daily maximum temperature of the summer maize flowering stage was ≥ 32 °C as the critical threshold of high temperature disaster, and ≥ 35 °C as the threshold of high temperature disaster. The number of high temperature days (HD) ≥ 32 °C and ≥ 35 °C during 10 days of the summer maize flowering stage were counted. The sequence of high temperature days of summer maize flowering stage was constructed and its normal distribution characteristics were verified. The quartile value of the normal distribution sequence was selected as the threshold value of high temperature in different grades, and the high temperature heat damage in the summer maize flowering stage was divided into three early warning meteorological grades. The index verification of the early warning meteorological grade was carried out by using the observation data of kernel number per ear in the agricultural meteorological station for maize. The results showed that the three grade indexes corresponding to the number of high temperature days with daily maximum temperature ≥ 32 °C were: 3 ≤ HD < 5 (Grade I), 5 ≤ HD < 7 (Grade II), and HD ≥ 7 (Grade III) and the three grade indexes corresponding to high temperature days ≥ 35 °C were: 2 ≤ HD < 3 (Grade I), 3 ≤ HD < 5 (Grade II), HD ≥ 5 (Grade III). The constructed indexes were verified as follows, the number of high temperature days ≥ 32 °C at about 90% stations was significantly negatively correlated with the number of maize grains per ear, and 74% stations which the number of high temperature days ≥ 35 °C was significantly negatively correlated with the number of maize grains per ear. Applying this index to evaluate the high temperature risk in the flowering period, we can indicate that the high temperature risk had trend of increasing gradually from north to south. In conclusion, the constructed high temperature warning meteorological grade index during the flowering stage could accurately reflect the damage grade of maize during the flowering stage.


Introduction
Maize is one of the largest grain crop in China, which plays an essential role in developing the national economy. Huang-Huai-Hai summer maize planting area is one of the dominant maize planting areas in China. Under the background of increasing shortage of cultivated land resources, it is more important to increase the yield of maize steadily.
The number of kernels per ear plays a vital role in the formation of maize yield, and pollen vitality is an important factor affecting the number of grains per spike (Lu et al., 2009). Flowering stage is the most sensitive period to high temperature in the whole growth period of maize. However, high temperatures during the flowering stage will cause serious decline in the pollen vitality of maize (Zhang, 2019). Abnormal high temperature have adverse effects on the structure and function of pollen grains, resulting in the decrease of pollen vitality, fertilization, ripening rate of the female spike, reducing the number of kernels per ear, and eventually reduce yield of the crop (Qiao et al., 2019;Waqas et al., 2021). Studies have shown that high temperature makes it difficult for the female ear to spin silk, delayed the female ear's ability of silking or making it uncoordinated, and also caused the female ear losing water after silking, and losing vitality, which hindered the germination of pollen. Hence, higher temperature resulted in reduction of kernels per spike and eventually grain yield through directly hinder the lifetime and quality of the silk (Waqas et al., 2021).
The suitable daily average temperature of maize from tasseling to silking is 25-28 °C, but the flowering and pollination period of summer maize in Henan is generally during the high temperature period from late July to early August, which is more apparent in the context of global warming (Yang et al., 2013). Summer maize often encounters extreme weather such as high temperature and drought during flowering and the crop yield drops sharply (He and Zhou, 2012). Therefore, the high temperature during flowering stage has become one of the main meteorological disasters that affect the yield of summer maize in Henan Province.
Disaster monitoring and early warning should be based on scientific, standardized, and simple meteorological indicators. Previous studies have carried out a lot of research on the high temperature index of the maize flowering stage. According to Xu et al. (1996), pollen vitality would be lost if temperatures exceeded 32 °C or even 35 °C, and insemination would not be possible. Xu (2002) pointed out that when the temperature was between 34.3 and 37.8 °C, the silk life was only 72 h. Microscopic studies revealed the effects of different degrees of high temperature on pollen vitality, silk physiological activities, yield, and quality (Sheng et al., 2020). Still, temperature accumulation in evaluating high-temperature impact and sets of easy to use and clear meteorological grade indicators for monitoring and early warning of high temperature in the summer maize flowering stage had not yet been considered and formed, thereby limiting the development of high-temperature early warning services in the maize flowering stage. In purview of given facts about negative impacts of temperature extremes on maize yield, the present study determined the different levels of hightemperature thresholds during the flowering stage of maize in Henan Province based on the occurrence law of high temperatures during the flowering stage. During the same period, by considering the cumulative effect of high temperatures, we constructed an early warning meteorological grade of high temperatures during the summer maize flowering stage, contributing to the improvement of the support capability of agricultural meteorological science and technology services.

Overview of Study Area
Henan Province of China belongs to the transition area from the northern subtropical zone to the warm temperate zone. The main planting system is the winter wheat and summer maize. Summer maize is usually sown in early June and harvested in late September. In the growing season, the total solar radiation is 1900-2400 MJ/m 2 , the precipitation is 400-600 mm, and the accumulated temperature ≥ 10 °C is 2600-3100 °C day −1 .

Data Source
Data in this study consist of daily maximum temperature data collected from 110 surface meteorological stations in Henan Province from 1971 to 2019, and observations from 19 agrometeorological stations in Henan Province during the summer growing season, with the observation of the tasseling and silking stage of maize as well as the number of kernels per ear in the observation section. The distribution of surface meteorological observation stations and maize agrometeorological observation stations are shown in Fig. 1.

Determination of Flowering Stage of Summer Maize
According to data pertaining maize cultivation in China, approximately 77.7-84.5% of tasseling occurs 2-5 days after flowering begins, with the highest concentration occurring in 3-4 days, and the spike usually lasts for 7-10 days. In order to take into account the climatic conditions in Henan Province and the growth and development characteristics of summer maize varieties currently planted, a period of 10 days was selected to reflect the meteorological grade index for early warning of high temperatures in the flowering stage of summer maize after the tasseling start. However, the flowering stage of summer maize is mainly concentrated from late July to early August. The annual value of the starting date of the maize tasseling period which was gradually delayed from south to north is shown in Table 1. The flowering stage of summer maize in southern Henan was the earliest and the tasseling period was from July 23rd to 25th. The flowering stage in central and eastern Henan was in the middle, with the starting date around July 30th, and the tasseling period was postponed to early August in western and northern Henan.

Selection of High Temperature Index for Summer Maize During the Flowering Stage
Numerous studies have been conducted on the influence of different degrees of high temperature on pollen vitality and yield formation in summer maize. Temperature exceeding optimal ranges affects the silking and pollen dispersal processes, and ultimately low seed setting thereafter under high temperatures (Chen et al., 2008;Jiang et al., 2016;Xu et al., 1996;Yang, 2005).
All the previous studies used the "daily maximum temperature" as a meteorological factor to evaluate the degree of heat damage caused by high temperature during the summer maize flowering stage. Pollen vitality and fertilization rate began to be adversely affected when the daily maximum temperature reached ≥ 32 °C and the degree of influence increased significantly when it reached ≥ 35 °C. However, to evaluate the degree of heat damage caused by high temperature during the flowering stage of summer maize, the cumulative "high temperature days" of ≥ 32 °C and ≥ 35 °C were selected in our case as an index to evaluate the level of heat damage caused by high temperature during the flowering stage.

Frequency of High Temperature Stress in Summer Maize at Flowering Stage
Based on the frequency of high temperature disasters occurring in different grades of summer maize during the flowering stage, the frequency of such disasters can be calculated as shown in Eq. (1): where i is the disaster grade caused by high temperature during the flowering stage of summer maize, and the values were 1, 2 and 3. P i is the frequency of high temperature disasters in different grades of summer maize at flowering stage, which could be expressed as P1, P2 or P3, n i is the frequency of high temperature disasters in different grades of flowering, and N is the number of years.

Stage of Maize
Climate risk index could be expressed as probability multiplied by intensity, and the comprehensive risk index of high temperature in summer maize flowering stage was calculated as follows: (1)  where I is the high temperature climate risk index of summer maize at flowering stage, ω i is the high temperature risk intensity of different grades, and i is divided into Grade I-III. The determination method of ω1, ω2, and ω3: According to the loss rate range of different grades of grains per spike determined in 1.5, the relative proportion of the loss rate of grains per spike of the three grades was calculated as the intensity coefficient of high temperature disaster stress at flowering stage. The calculation method is shown in Eq. (3): The risk index of high temperature in summer maize flowering stage was divided into three levels, such as light, medium and heavy. Firstly, the risk index was normalized, as shown in Eq. (4): where I an is the normalized high temperature risk index of the nth station; I max is the maximum value of the whole region's high temperature risk index; I min is the minimum value of the whole region's high temperature risk index; and I n s the high temperature risk index of the nth station, n = 1-110.
According to the expert scoring method, the threshold of grading was determined. When the normalized high temperature risk index of summer maize in the flowering stage was less than 0.4, it was a mild risk. However, when more than 0.7, it was a severe risk. While the intermediate transition value was a moderate risk.

Data Analysis and Processing
All statistical analyses, data processing, and mapping were performed using SPSS Statistics package 20 (SPSS Inc., Chicago, IL, USA) and Microsoft Excel 2013.

Correlation Analysis Between High Temperature Days and Grain Number Per Spike in Summer Maize Flowering Stage
The high temperature at the flowering stage of summer maize mainly affected the yield composition through the number of kernels per ear. Therefore, the correlation between the number of grains per spike and the number of high temperature days at flowering stage was analyzed by counting the crop observation data of 19 main summer maize producing areas. The results are shown in Table 2. In terms of the correlation coefficient between high temperature days ≥ 32 °C and grain number per spike (except Xiping and Zhumadian), more than 90% of stations passed the 0.05 significant level (Gongyi, Ruzhou, Xinxiang, and Zhengzhou stations achieved 0.01 significant levels), indicating a significant negative correlation. According to the correlation coefficients between high temperature days ≥ 35 °C and grain number per spike, Puyang, Qinyang, Sanmenxia, Yichuan, and Xiping stations had small correlation coefficients and failed the significance test, while nearly 74% of other stations passed this test. Most of the stations showed a good linear relationship between the number of high temperature days at the flowering stage and the number of grains per spike of summer maize. The number of days at flowering stage ≥ 32 °C and ≥ 35 °C of summer maize could well represented the damage degree of summer maize under high temperature stress, which also indicated that it was reasonable to choose the index of high temperature days at flowering stage.

Determination of Threshold of High Temperature Grade at Flowering Stage Based on Statistical Method
The frequency distribution of high temperature days in the summer maize flowering stage calculated in 3.1 using the number of days with the highest daily temperature of 32 °C and 35 °C at every station from 1971 to 2019 is shown in Fig. 2. The frequency distribution of high temperature days at ≥ 32 °C showed a single peak and the frequency was highest when the high temperature days lasted 4 days. While the frequency distribution of high temperature days at ≥ 35 °C showed a decreasing trend, and the frequency was highest when the high temperature days lasted 1 day, then decreased in turn. Figure 3 shows the distribution characteristics of the frequency of high temperature days in the flowering stage of summer maize as normal distribution test was performed on the data series. The P-P diagram showed that most of the points diverged little from the straight line which means that two series of days with high temperatures followed a normal distribution.
The cumulative percentage of high temperature days in the flowering stage of summer maize at ≥ 32 °C and ≥ 35 °C respectively, along with the quartile value of high temperature days series are shown in Fig. 4. As can be seen, the percentile values of 25%, 50%, and 75% of high temperatures days ≥ 32 °C were 3, 5, and 7 days respectively, and the three percentile values of high temperatures days ≥ 35 °C were 1, 3, and 5 days, respectively. The threshold values of three grades of high temperature days ≥ 32 °C were 3, 5, and 7 days. Percentile values for days with temperatures exceeding ≥ 35 °C were 1, 3, and 5 days, respectively. The overall distribution of high temperature days together with the professional opinion of cultivation experts suggested that the high temperature index be based on a 2 day limit since the peak period of flowering and pollination usually lasts for 2-5 days. Therefore, our recommendation was to limit the high temperature index to 2 days, so the three threshold levels of high temperature days ≥ 35 °C were 2, 3, and 5 days.

Loss Rate of Kernels Per Ear Corresponding to the Threshold of High Temperature Days in Summer Flowering Stage of Maize
The stations with ≥ 32 °C and ≥ 35 °C days that passed the significance test were screened out. The relationship model between the days and grains per spike for ≥ 32 °C and ≥ 35 °C was constructed in order to estimate the loss rate range of grains per spike corresponding to different early warning meteorological grades, as shown in Fig. 5. The calculated average loss rate corresponding to each grade is shown in Table 3.
A total of 17 significant sites were selected based on the index of high temperature days ≥ 32 °C. Among them, when the number of high temperature days were 3 and 5 days, the corresponding loss rate of grains per ear ranged from 6.8 to 13.3%, with an average of 9.4%. When the number of high temperature days were 5 and 7 days, the corresponding loss rate of grains per ear ranged from 11.5 to 21.5%, with an average of 15.4%. When the number of high temperature days were ≥ 7 days, the corresponding loss rate of grains per ear ranged from 15.4 to 25.8%. A total of 14 significant sites were selected from the index of high temperature days ≥ 35 °C. Among them, when the number of high temperature days were 2 and 3 days, the loss rate of grains per ear at each station ranged from 4.5 to 11.0%, with an average of 7.4%. When the number of high temperature days were 3 and 5 days, the loss rate of grains per ear ranged from 8.2 to 20.6%, with an average   of 13.5%. When the number of high temperature days was ≥ 5 days, the loss rate of grains per ear ranged from 15.1 to 36.1%, with an average of 24.3%. The classification standard of disaster impact grade used in agrometeorology indicates that the mild disaster rate is 5-10%, the moderate disaster rate is 10-20%, and the severe disaster rate is 20%. The loss rate of grains per spike corresponding for each high temperature early warning meteorological grade was approximately consistent with the classification range of disaster impact grade (Table 3).

Classification and Demarcation of Meteorological Indexes of High Temperature in Summer Flowering Stage of Maize
According to the correlation between high temperature days and grain number per ear and the loss rate of grain number per ear corresponding to different grades, the grades of high temperature days were divided into Grades I-III, as shown in Table 4.

Interannual Variation of Occurrence Regularity of High Temperature and Heat Damage in Summer Maize Flowering Stage
Station ratio with high temperature stress in different grades of florescence is shown in Fig. 6, however, there was no significant change trend. The number of stations with high temperature and heat damage varied greatly in the years. The number of stations with high temperatures and heat damage varied greatly over time. The number of stations with a high temperature in the third florescence varied the most with a coefficient of variation of 77.6%, while the number of stations with a high temperature in the second florescence varied the least with a coefficient of variation of 55.2%. There was no significant upward trend in the secondary ratio of total stations.

Spatial Variation of Occurrence Regularity of High Temperature Stress in Summer Maize Flowering Stage
As a result of the index of high temperature grade in the summer maize flowering stage, the occurrence frequency of high temperature heat damage in each region in the past 50 years from 1971 to 2019 was determined, and the frequencies of high temperature heat damage of different grades of damage were calculated as shown in Fig. 7. Areas with a high frequency of Grade I high temperatures were mainly situated in the north of Henan Province, including the majority of areas in Anyang, Puyang, Xinxiang, Zhengzhou, and Luoyang. Most of the high-value areas of Grade II were found in eastern Henan, including Kaifeng and Shangqiu, and the majority of other areas were below 30%. Gradually, the frequency of Grade III was increased  from north to south, and high-value areas with a frequency exceeding 40% were located in Nanyang, Zhumadian, and Zhoukou in southern Henan. Using the comprehensive index of high temperature risk zoning at the flowering stage, the light, medium, and heavy risk levels were divided and the results are shown in Fig. 8. During the flowering stage, the risk of high temperatures gradually increased from north to south, with Luoyang and Sanmenxia in the west of Anyang in the north being considered light-risk areas, Nanyang, Zhumadian, Xuchang, and Zhoukou in the south of Henan being considered heavy-risk areas, and other transitional areas as medium-risk areas. Xu et al. (2021) summarized the previous research results and concluded that the pollen germination rate of many genotypes decreased to near zero when exposed for long periods above 32 °C. Once the temperature increased beyond 35 °C, the impact on the pollen vitality of maize was fatal and when the temperature exceeded 38 °C, the tassel could not blossom. As for high temperature index of summer maize in flowering stage, Li et al. (2015) constructed a comprehensive climate index of high temperature and heat damage of summer maize in Huaibei Plain, based on extreme maximum temperature days with daily maximum temperature ≥ 35 °C, and average minimum relative humidity during this period. The risk of extreme high temperatures during the growth period of maize was evaluated by taking into account the accumulated days of high temperatures with daily maximum temperatures greater than 32 °C and the days with a maximum temperature exceeding 32 °C daily (Wang, 2018). According to the "GB/T 21985-2008 Main Crop High Temperature Hazard Temperature Index" (GB), when the daily maximum temperature was ≥ 30 °C and the relative humidity was ≤ 60%, the maize blossomed less; however, when the maximum daily temperature was ≥ 35 °C, pollen grains lost vitality, which was not conducive to flowering and was consistent with the two temperature indexes selected in this study. High temperatures and long durations result in more serious damage, which is in accordance with the incidence law and disaster mechanism of high temperature heat damage, and is in accordance with the research conclusions of high temperature index in maize flowering stage in various studies (Jiang et al., 2016;Wang, 2015Wang, , 2018Xiao & Ai, 1998).

Discussion
In the current study, only the temperature index was considered, and the air humidity index had not been taken into account in the meteorological classification of high temperatures during the summer flowering stage of maize. Previous studies had also found that there was an interaction effect between temperature and humidity on pollen tubes to a certain extent (Li et al., 2002). Low temperature and lower humidity could slow down the decline in pollen tube vitality and prolong the spinning time of pollen tubes. Li et al. (2015) pointed out that the high average minimum relative humidity during the high temperature process may weaken the influence of high temperature heat damage, whereas the low average minimum relative humidity may aggravate the effect of high temperature to a certain extent. However, the superposition mechanism of air humidity on high temperature and heat damage in the maize flowering stage has not been clearly explained, and the results of the research are limited. Therefore, only the most important temperature factor for high temperature stress in summer maize flowering stage was selected for this study.
In this study, 32 °C and 35 °C were considered as the meteorological grade indexes. A threshold of 32 °C could be used as the initial threshold for high temperatures, and 35 °C could be used as the threshold for a significant increase in high temperatures during the flowering stage. However, this study did not involve high temperature such as above 38 °C. A number of studies have demonstrated that 38 °C is the inflection point for most physiological characteristics of maize. For example, tassels would completely stop scattering powder after being stressed at 38 °C for 3 days (Ren et al., 2019). When the temperature was higher than 38 °C, the germination ability of maize pollen and the elongation ability of pollen tube would be significantly reduced, resulting in an average reduction of 21% in seed setting rate (Yang & Zhang, 2006). With global warming, the number of days with extremely high temperatures has increased (Guan et al., 2021;Yu et al., 2016). Therefore, in the future studies it is necessary to investigate the effect of extremely high temperatures on the number of grains per ear of summer maize at the flowering stage.
In addition, high temperatures at the flowering stage of summer maize have an obvious influence on its silk elongation, fertilization abilities, and the interval between male and female florets (Gao et al., 2020;Jia et al., 2020;Yang, 2005;Zhao et al., 2014). So, it is of great importance to clarify the occurrence characteristics and phenomenon of high temperatures during the flowering stage and to conduct technical research on variety selection, cultivation management, chemical control, and regulation of water and fertilizer in order to minimize the adverse effects of high temperatures on maize and ensure the high and stable yield of summer maize (He et al., 2020a, b, c;Wen et al., 2021). Different varieties of maize have different heat tolerance levels and screening high-temperature tolerant maize varieties is one of the most important measures that can be taken to combat climate change (Li et al., 2002;Sun et al., 2019;Wen et al., 2021). The results showed that the pollen viability of heatresistant maize genotypes was about 10% higher than conventional varieties under high temperature stress. However, the heat-sensitive genotype of maize was greatly affected by high temperature and its pollen vitality was about 28% lower than that of conventional varieties (Zhao et al., 2012). In order to strengthen the exploration of high temperature tolerant germplasm resources and to cultivate high temperature tolerant varieties, there is a need to improve the resistance of maize to high temperature and heat damage (Guan et al., 2021;Yu et al., 2016). Taking advantage of the difference in pollen quantity among different varieties and variety hybridization, and carrying out mixed cropping among varieties can also increase the fertilization and seed setting rate resulting in the increase of maize yield per unit area (Liu et al., 2008). The heat tolerance of maize varies with its growth stage (Waqas et al., 2021). By adjusting sowing date, the flowering stage of maize can be avoided, reducing the effect of high temperature heat damage on the critical growth period of maize (He et al., 2020a, b, c). Also, by optimizing water and fertilizer management, as well as creating a suitable microclimate environment in the field we can effectively prevent the harmful effects of high temperatures on the flowering stage of maize (Yuan et al., 2017).

Conclusion
The daily maximum temperature ≥ 32 °C at the flowering stage of summer maize was selected as the critical threshold for maize damage, and ≥ 35 °C was the threshold for significant aggravation of heat damage. The correlation between the number of high-temperature days and the number of grains per ear showed that the number of high temperature days could better reflect the high-temperature disaster situation at the flowering stage. Therefore, the harm degree of high temperature in summer maize flowering stage can be divided into three meteorological levels. The results showed that the three grade indexes corresponding to high temperature (≥ 32 °C) days were: 3 ≤ HD < 5 (Grade I), 5 ≤ HD < 7 (Grade II), and HD ≥ 7 (Grade III); the three grade indexes corresponding to high temperature (≥ 35 °C) days were: 2 ≤ HD < 3 (Grade I), 3 ≤ HD < 5 (Grade II), and HD ≥ 5 (Grade III). Correlation analysis revealed negative correlation between the number of high temperature days ≥ 32 °C and the number of grains per ear in more than 90% stations, and ≥ 35 °C at more than 74% stations. The more the high temperature during flowering stage, the more the grains per spike decreased.