1 Introduction

Maize is a major cereal crop produced worldwide. As maize demand continues to increase for food, feed, and fuel, global production has increased to over 1100 million Mt in 2019 (FAO 2020). Meteorological disasters are the main factors that restrict stable maize production under climate change. Negative association between growing season temperature and maize production has been observed in many of the maize producing areas (Sakurai et al. 2012; Gabaldón-Leal et al. 2016). Extreme heat plays a significant role in this negative relationship (Lobell et al. 2011). Meanwhile, the frequency and severity of extreme heat have increased around the world during the past decades (Orlowsky and Seneviratne 2012; Horton et al. 2016; Lu et al. 2018). Disaster risk assessment has become a major research priority for managing extreme events in order to reduce the possible impacts of extreme events (IPCC 2012).

The negative association between extreme heat and maize production can mainly be explained by two mechanisms. The first mechanism includes the indirect effects of high temperatures that increase crop water stress. This can result from reductions in the soil water supply due to high transpiration, or because of high temperatures, increasing crop water demand for soil water (Lobell et al. 2013). The other mechanism is direct heat stress effects on maize organs, such as desiccation of pollen and reductions in net photosynthesis under extreme heat conditions (Bolanos and Edmeades 1996; Crafts-Brandner and Salvucci 2002). Although previous studies have confirmed the negative relationship between growing season extreme heat and production, it is still unclear as to how much of the impact is a result of the direct heat stress independent from the indirect water stress effects. It is hard to isolate the effect of direct heat stress from other environmental elements, especially with regard to statistical studies based on field conditions (Schlenker and Roberts 2009; Deryng et al. 2014) where precipitation shortages and extreme heat may occur concurrently during the growing season. Additionally, the frequency of concurrent drought and extreme heat events is increasing under climate change (Wei et al. 2020). It is critical to quantitively assess the direct impacts of extreme heat in order to accurately project crop yield and to develop and target management strategies under global warming.

Process-based crop growth simulation models provide a major approach for separating the effects of different factors (such as irrigation, soil, nutrients, crop management, and weather conditions) on crop growth and grain yield formation processes (Saseendran et al. 2005; Abedinpour et al. 2012; Liu et al. 2014; Ren et al. 2016). A large number of crop models have been built during the past decades, and crop models named Decision Support System for Agrotechnology Transfer (DSSAT), AquaCrop, Agricultural Production Systems Simulator (APSIM), World Food Studies (WOFOST) model, and Erosion Productivity Impact Calculator (EPIC) are most prominent models around the world (Camargo and Kemanian 2016). The fundamental structural may differ for different crop models. Before being used, models should be calibrated and validated for a certain region by using field experimental data. Many previous studies have proven that APSIM is an effective tool for estimating the potential effects of environmental and management elements on crop production. APSIM has been widely used for maize in the North China Plain and Northeast China (Liu et al. 2012; Chauhan et al. 2013; Liu et al. 2014; Ren et al. 2016; Sun et al. 2016). Liu et al. (2014) used APSIM to isolate the effects of changes in sowing dates and temperature on maize growth and yield. They demonstrated that although warming climate had a negative impact on maize, adaptation activities such as earlier sowing dates mitigated the negative impacts. Sun et al. (2016) observed that APSIM was able to quantify the positive effect that sunshine hours and the diurnal temperature range at the grain filling stage had on maize yield. Therefore, we selected APSIM to separate the direct impact of extreme heat on maize yield from the effects of other environment or management elements.

Extreme heat researches have mostly focused on the spatial–temporal distribution of extreme heat events and changes in extreme heat event characteristics (Horton et al. 2016; Yang et al. 2017), the relationship between extreme heat and crop growth (Schlenker and Roberts 2009; Lobell et al., 2011, 2013; Deryng et al. 2014), and public health (Johnson and Wilson 2009; Schwarz et al. 2020). Disaster risk assessment studies can estimate the likelihood of hazardous events and potential losses. Disaster risk assessment is necessary for the mitigation of and preparedness for natural disasters (Zhang 2004; Jia et al. 2012). There are several formulas used for disaster risk assessment, and the formula “risk = hazard × vulnerability” advised by the United Nations International Strategy for Disaster Reduction (UN/ISDR 2007) has been most commonly used in recent years (Shahid and Behrawan 2008; Jia et al. 2012; Zhang et al. 2017, 2019). Hazard is expressed by the frequency and severity of hazardous events with potentially large impacts (IPCC 2007). For extreme heat hazard, cumulation of degree days over the extreme heat threshold or the number of extreme heat days at crop growing season is mostly used as the severity of extreme heat (Deryng et al. 2014; Zhang et al. 2021). The vulnerability of a hazard-bearing body is the propensity or predisposition to be adversely affected (Zhang et al. 2019). There are several approaches to assess vulnerability, and the method of vulnerability curves has been used by a lot of researchers (Jia et al. 2012; Wang et al. 2013; Yin et al. 2014; Zhang et al. 2019). The curve is expressed as the function between potential losses change with intensities of hazardous events, and logistical function was most commonly used to fit the vulnerability curve in previous studies (Jia et al. 2012; Wang et al. 2013; Yin et al. 2014). Crop models can be used to estimate the potential losses under various environment conditions.

It is difficult to isolate the impacts of direct heat stress on crops, and previous agrometeorological disaster risk assessment research has seldom been concerned about extreme heat, especially the direct stress. In this study, we used summer maize as the hazard-bearing body to assess the risk of direct extreme heat stress by considering two aspects: hazard and vulnerability. Extreme temperature days during each maize growing stage were used to assess the hazard. The calibrated APSIM model was used to separate the direct impacts of extreme heat from other environment factors. Extreme heat vulnerability curves for several summer maize growing stages were constructed. Risk evaluation based on hazard and crop vulnerability can provide the theoretical and scientific bases for extreme heat risk reduction, mitigation, and management strategies.

2 Materials and methods

2.1 Study area

The Haihe Plain (34.93° N–41.06° N; 112.56° E–119.85° E) is the northern part of the North China Plain (Fig. 1). This region produces about 8.5% of China’s total maize output and accounts for 9.9% of China’s maize planting area (Xue et al. 2021). The cropping system is winter wheat and summer maize rotation. It is an area of China where maximum temperatures have been observed in the last few decades (Kang and Eltahir 2018). Compared with southern parts of the North China Plain, the Haihe Plain is with less precipitation and irrigation is required to supplement soil water to achieve much more crop yield (Wu et al. 2019). Consequently, the region suffers from some ecological and environmental problems, and irrigation may have aggravated the local extreme heat in recent years (Kang and Eltahir 2018).

Fig. 1
figure 1

Study area and meteorological and Agro-Eco-Experimental stations in the Haihe Plain, China

Full size image

2.2 Weather and phenological phase data

Daily maximum temperature, minimum temperature, precipitation, and sunshine duration of 26 meteorological stations at the Haihe Plain were used in this study (Fig. 1); the time range from 1981 to 2019. These weather data are maintained by the China Meteorological Administration and can be accessed at http://data.cma.cn/. Rigorous quality control procedures have been performed for the dataset by the maintainers, and it is one of the best daily meteorological datasets in China (Chen and Li 2017). Sunshine duration can be converted into solar radiation using the Ångström formula (Ångström 1924). These daily weather data were used to identify extreme heat and to drive APSIM.

The summer maize yearly phenological stage data included planting, six leave fully emerged (V6), tasseling (VT), and physiological maturity (R6) dates (stage designations are as given by Ritchie et al. 1986). These phenological dates were collected from the annual report of national agricultural meteorological observation stations, and the paper report forms are maintained by the China Meteorological Administration. These phenological stage dates were used to define three maize growing stages. Stage I: planting to V6 is the vegetative stage, lasting about 34 days; stage II: V6 to VT is the vegetative transitioning to reproductive stage, lasting about 20 days; and stage III: VT to R6 is the reproductive stage, lasting about 45 days. Because cropping calendars and seasonal temperatures can be different, the growing stages covered different time ranges in different years at different stations. This means that on the same day in different years, summer maize can be in different growing stages. Figure 2 shows that on any given date, the frequency represents the percentage of years that summer maize was within a particular growing stage on a given date. The station with the earliest planting dates is shown as the dashed lines, where planting occurred in late May or early June for most years. The station with the latest planting dates is shown as the dotted lines, where planting occurred in late June for most years (about a month later than the station with the earliest planting dates). The solid lines are the average of all of the stations in the Haihe Plain, where most stations were planted in middle June, V6 occurred in the middle of July, and VT occurred in early August. Variations in cropping calendars and seasonal temperatures result in substantial changes in the temperatures that summer maize may experience, so it is important to use the specific summer maize phenological stage dates.

Fig. 2
figure 2

Frequency of years with summer maize in a certain growing stage on each given date. Black lines show growing stage I, red lines show stage II, and blue lines show stage III. Dashed lines are for the station with the earliest planting dates (Tangyin station), the dotted lines are for the station with the latest planting dates (Huanghua station), and the solid lines show the mean developmental stage distribution averaged across all stations

Full size image

2.3 Experimental data

Field experiments were conducted at three Agro-Eco-Experimental Stations of the Chinese Academy of Sciences during 2004–2007 (Fig. 1), named Luancheng (LC), Yucheng (YC), and Fengqiu (FQ). All experimental data are supported by the Chinese Ecosystem Research Network (CERN). The multi-year average annual rainfall (AR), annual daily temperature (ADT), and total sunshine hours per year (TSH) are shown in Table 1. All stations belong to the warm temperate semi-humid monsoon climate, with rainfall mainly concentrated in summer. The soil is classified as loamy soil at LC and YC and sandy loam at FQ. The bulk density and hydraulic parameters in various soil layers at the experimental stations are shown in Table 2. These data are local experimental station data sets and from CERN.

Table 1 Climatic conditions at three Agro-Eco-Experimental Stations in the Haihe Plain, China, including annual daily temperature (ADT), total sunshine hours (TSH), and annual rainfall (AR)
Full size table
Table 2 Soil characteristics in different soil layers at three experimental stations in the Haihe Plain, China, including soil bulk density (BD), volumetric water content at saturation (SAT), drained upper limit (DUL), and lower limit of plant available soil water (LL)
Full size table

The field experimental data were used to calibrate APSIM. Experimental data at LC were recorded during 2005–2007 for maize variety “Jingyu 7.” Fertilizer was applied at V6 as 90 kg N ha−1, 90 kg ha−1 P2O5, and 45 kg ha−1 K2O. Experimental data at YC were recorded during 2005–2007 for maize variety “Ludan 9002.” Fertilizer was applied at V6 as 150 kg N ha−1, 75 kg ha−1 P2O5, and 75 kg ha−1 K2O. Experimental data at FQ were recorded during 2004–2007 for maize variety “Zhengdan 958.” Fertilizer was applied before planting as 90 kg N ha−1, 90 kg ha−1 P2O5, and 90 kg ha−1 K2O. Experimental data are all from non-water-stress experiments. Irrigation was applied when soil water content was low and the irrigation method was flooding. Phenological development, biomass, leaf-area index (LAI), and grain yield were observed at each growing season and were used to calibrate APSIM.

2.4 APSIM model and calibration

APSIM was developed by the Agricultural Production Systems Research Group (APSRU) under the Commonwealth Scientific and Industrial Organization of Australia (CSIRO) and the Queensland Government. The APSIM maize model has been well validated and can be downloaded from the APSIM website (https://www.apsim.info/). APSIM version 7.10 was used in this study to separate the direct impacts of extreme heat on maize yields from other climate or management elements and then to build the summer maize extreme heat vulnerability curves at different growing stages.

APSIM was calibrated by comparing simulated results with experimentally observed data that included aboveground biomass, leaf-area index (LAI), and grain yield at LC, YC, and FQ with a trial-and-error method (Seidel et al. 2018). The coefficients of determination (R2), slopes of regression lines, and root mean squared errors (RMSE) were used to quantify model performance.

2.5 APSIM model-based risk assessment process

2.5.1 Hazard assessment

The number of extreme heat days was used as the hazard intensity factor (H). For summer maize, daily maximum temperature over 35 °C can be considered an extreme heat day (Deryng et al. 2014). Because the sensitivities of summer maize to extreme heat vary at different growing stages, the hazard intensities were calculated for each stage separately. Based on the concept of hazard, a hazard assessment model was constructed from the perspectives of the probability and intensity (Zhang et al. 2017, 2019).

$$H={\sum }_{s=1}^{n}S\times f(S)$$
(1)

where \(H\) is the hazard assessment value for a certain station, \(S\) is the number of observed extreme heat days during a summer maize growing stage (denoting the intensity of extreme heat), and \(f(S)\) is the frequency of years with \(S\) days of extreme heat during this growing stage.

2.5.2 Vulnerability curves

In this study, we simulated two scenarios: standardized climate scenario with no heat stress (S1) and temperature change scenario (S2). Summer maize yields under the two scenarios were simulated by the calibrated APSIM model. S1 was conducted to simulate the yield under standardized climate conditions by driving the APISM model with multi-year average daily temperature, precipitation, and solar radiation data. S2 was conducted to examine the influence of changes in temperatures during a certain growing stage on final crop yield. We did this by driving the APSIM model with actual temperature data during a certain growing stage during the past 40 years, while the temperature data used during the other growing stages, as well as precipitation and solar radiation, were multi-year average daily data. For both of the scenarios, we simulated no water and nutrient stresses by setting automatic irrigation and fertilization in APSIM. Under such conditions, the simulated yield variations (YL) between S2 and S1 can denote the direct effects of temperature at the studied growing stage, excluding other elements, such as the water stress. The vulnerability curve of each growing stage was obtained by establishing the function between YLs and extreme heat hazard intensities (H). In this study, logistical functions were used to fit the vulnerability curves. YL values were calculated as

$${\mathrm{YL}}_{ij}= \frac{\mathrm{Yield}\_1-{\mathrm{Yield}\_2}_{ij}}{\mathrm{Yield}\_1}$$
(2)

where \({\mathrm{YL}}_{ij}\) is the yield variations at growing stage \(i\) and year \(j\). \(\mathrm{Yield}\_1\) is yield under S1. \({\mathrm{Yield}\_2}_{ij}\) is the yield under S2 for growing stage \(i\) and year \(j\).

2.5.3 Calculation of disaster risk

According to regional disaster system theory, disaster risk (R) is a function of hazard (H) and vulnerability (V) under the influence of environment. Equation (4) was used to express risk in this study (Shi 2002; UN/ISDR 2007; Yin et al. 2014; Zhang et al. 2017; Zhang et al. 2019).

$$R=f\left(H,V\right)={\mathrm{YL}}_{H}, V\{\mathrm{YL},H\}$$
(3)

where \({\mathrm{YL}}_{H}\) is the yield losses resulted by the direct impact of extreme heat; \({H}_{i}\) is the extreme heat hazard intensity with \(i\)-year return period, the return period is calculated by the exceeding probability of extreme heat intensity at each station; \(V\{\mathrm{YL},H\}\) is the function for \(\mathrm{YL}\) versus \(H\) (the vulnerability curve); \({\mathrm{YL}}_{{H}_{i}}\) can be calculated according to \(V\{\mathrm{YL},H\}\). The overall technological framework of risk assessment is described in Fig. 3.

Fig. 3
figure 3

Risk assessment process for direct extreme heat stress on summer maize

Full size image

3 Results

3.1 Performance of APSIM model

Comparisons of simulated maize LAI, aboveground biomass, and yield with measured values are shown in Fig. 4. The variations in observed LAI, aboveground biomass, and grain yield in different years and experimental stations are well captured by APSIM. Slope values (β) are near to 1.0, indicating that there are no obvious systematic over- or under-estimations. The R2 values indicate that the model-simulated values explained 97%, 98%, and 81% of the variation in measured LAI, biomass, and grain yield, respectively.

Fig. 4
figure 4

Comparisons between observed and simulated LAI, biomass, and grain yield from 2004 to 2007 at Luancheng, Yucheng, and Fengqiu. n = number of observations; R2 = coefficient of determination; β = regression slope; RMSE = root mean square error. The gray lines are the 1:1 lines; the black lines are the linear regression lines

Full size image

3.2 Extreme heat hazard assessment

During 1981–2019, the earliest date of extreme heat in the Haihe Plain appeared on April 17, 2019 at Shexian, and the latest date of extreme heat was on September 19. As shown in Fig. 5a, the frequencies of the occurrence of extreme heat days were highest from mid-June to early July, and the frequencies were mostly greater than 0.2. The frequency decreased as the day sequence increased, but a small spike appeared around July 23. Luancheng had the largest number of extreme heat days of all the stations (16.7 day per year). Changli had the least number of extreme heat days (2.5 day per year).

Fig. 5
figure 5

Distribution of extreme heat days in the Haihe Plain, China. a Frequency of extreme heat for a certain day during the past 39 years. The lines are the 5-day moving average frequency at different stations. The red solid line denotes Luancheng, the location with the most frequent extreme heat; the solid blue line denotes Changli, the location with the smallest number of extreme heat days; the solid black line is the average of all of the stations in the Haihe Plain; the gray lines are other stations. b Annual variation of the number of extreme heat days at each summer maize growing stage. The solid black line denotes the entire growing season; the solid red line denotes stage I; the solid blue line denotes stage II; the solid green line denotes stage III; the gray dashed lines are the multi-year average number of extreme heat days during the entire growing season before and since 1997

Full size image

Based on the summer maize yearly phenological stage dates at each station, we extract the number of extreme heat days at each growing stage, and the annual variations are shown in Fig. 5b. The year 1997 was very hot, and the average number of extreme heat days during summer maize growing season in the Haihe Plain reached 21.4 days (black solid line). Before 1997, the multi-year average number of extreme heat days was 7.0 days (gray dashed line). Since 1997, the number of extreme heat days increased to an average of 11.1 days. The multi-year average ratio of extreme heat days during the entire growing season to that during the year was 77.6%. The red, blue, and green solid lines in Fig. 5b show the annual variation of the number of extreme heat days during summer maize growing stages I–III, respectively. The percentage of the entire growing season with extreme heat days during the three growing stages were 74.0%, 19.0%, and 7.0%, respectively. This means that in the Haihe Plain, extreme heat mainly occurred during the summer maize growing season, especially during stage I (planting to V6).

Based on Eq. (1), extreme heat hazard values during each growing stage at each station are calculated, and the spatial distributions are shown in Fig. 6. The spatial distributions are consistent for the growing season and for growing stages I and II. The extreme heat hazards are the smallest at the northeastern coastal area for all stages and increased to the western area. For stage III, the extreme heat hazard for the entire region is very low. The spatial variation is larger at stage I, with a range of 9.7 days, while at stage II the range is only 3.8 days.

Fig. 6
figure 6

Summer maize extreme heat hazard distribution. a Entire growing season; b stage I; c stage II; d stage III

Full size image

3.3 Extreme heat vulnerability at each summer maize growing stage

Based on the calibrated APSIM model and the process introduced in Sect. 2.5, maize yields are simulated under different numbers of extreme heat days at each summer maize growing stage. Logistical functions are used to fit the yield losses against extreme heat intensity at each growing stage. The final vulnerability curves are shown in Fig. 7, and the function parameters and R2 values are shown in Table 3. The fitting goodness of vulnerability curve for stage I is the best, and its R2 is 0.44. The fitted curves of the three stages all at the 0.01 significance level explained most of the variability. By comparing the vulnerability curves of the three growing stages, it can be seen that summer maize is more vulnerable to extreme heat direct stress at stage III. At this stage, the number of extreme heat days over 1.6 days could result in yield losses, and yield losses increased rapidly with the number of extreme heat days. At stage I, the extreme heat days over 9.1 days may result in yield losses, and yield losses increased slowly with the number of extreme heat days. Stage I is the least sensitive stage among the three growing stages. Vulnerability of stage II is intermediate to stage I and stage 3, and the number of extreme heat days over 4.4 days could result in yield losses at this stage. In general, the yield losses resulting from direct heat stress are limited, as shown in Fig. 7. APSIM simulated yield losses of less than 10% under historical temperature conditions and with water demand fully met by irrigation (no water stress).

Fig. 7
figure 7

Physical vulnerablility curves for direct heat stress on summer maize at different growing stages. The yield loss and the extreme heat intensity of a site in a given year are indicated by a cross, black crosses represent growing stage I, blue crosses represent growing stage II, red crosses represent growing stage III. The curvers are fitted by using logistical function called vulnerability curves

Full size image
Table 3 Functions of vulnerability curves for summer maize different growing stages
Full size table

3.4 Risk assessment of direct extreme heat stress on summer maize

Risk of the direct extreme heat stress is assessed based on the vulnerability curves and extreme heat hazard. Annual extreme heat intensity for each summer maize growing stage at every station is recognized, and the exceeding probability of extreme heat intensity for each station is fitted. With the corresponding vulnerability curves, 5-year, 10-year, 20-year, and 50-year return period summer maize extreme heat risk maps are drawn (Fig. 8). As the return period increases, the risk values increase gradually, and the increased amplitude is quite large for stage I, especially in the western area. As shown in Fig. 6, growing stage I has the greatest extreme heat hazard, especially in the western region. Although summer maize is not very vulnerable to extreme heat direct stress at this stage, the extremely large numbers of extreme heat days make the risk values large. For stages II and III, extreme heat hazards are less, but stage III is the most vulnerable period to extreme heat direct stress, thereby making the risk values larger for stage III at the same hazard level.

Fig. 8
figure 8

Risk of direct extreme heat stress on summer maize at different growing stages in the Haihe Plain, China, with hazard levels of once in 5, 10, 20, and 50 years

Full size image

4 Discussion and conclusions

4.1 Discussion

Extreme heat and drought are often concurrent. Extreme heat can result in yield losses by increasing water stress or by direct heat stress effects. It is critical to analyze the direct temperature impacts for accurate yield prediction and targeted mitigation strategies (Carter et al. 2016). In our research, APSIM was used to separate the direct impacts of extreme heat from other parameters, including impacts associated with soil water stress, as we set APSIM to simulate with water requirement fully met. There is no doubt that extreme heat would reduce maize yield. Our results further quantify the direct impacts of extreme heat on summer maize yield during different growing stages. The number of these instances are relatively limited but cannot be alleviated by irrigation (Fig. 7). This is in agreement with earlier studies in the central USA (Lobell et al. 2013; Zhang et al. 2015) that reported that extreme heat effects resulted by reducing soil water supply played a greater role in yield reductions and irrigation can significantly reduce the sensitivity of maize to extreme heat. Over two-thirds of the negative effects can be reduced by irrigation. The area equipped for irrigation is large in the Haihe Plain (Kang and Eltahir 2018), and irrigation has relieved the impacts of extreme heat and insufficient precipitation. However, the large amount of irrigation water requirement has led to a declining groundwater table in this region and has exacerbated water shortages and environmental issues (van Oort et al. 2016). Risk adaptation strategies, such as breeding cultivars to ameliorate heat damage or shifting current cropping systems to use of heat tolerant crops, are necessary. Carter et al. (2016) assumed that improved genetics and management strategies have the potential to partially or completely eliminate the impacts of extreme heat on maize.

When studying extreme hazards during the different summer maize growing stages, some researchers have used fixed phenological dates, such as the average dates across years, or they have presumed the phenological dates are the same across regions (Jia et al. 2012; Zhang et al. 2017). These assumptions cannot represent year-to-year variations or differences between regions. Correctly resolving growing stages is quite important in considering yield sensitivity (Butler and Huybers 2015). In our study, specific summer maize stage dates were used to quantify extreme heat intensity and vulnerability curves for each growing stage. This method permits better understanding of extreme heat and crop growth and provides a more accurate risk assessment result for better guidance of heat stress mitigation strategies and yield prediction.

Many previous studies have confirmed that silking and grain filling are the maize stages most sensitive to extreme heat (Sánchez et al. 2014; Butler and Huybers 2015). Our results, as shown in Fig. 7, are consistent with these previous studies. Additionally, our results further determined that the extreme heat disaster risks are different at different growing stages. Even though growing stage III has the greatest vulnerability to direct extreme heat stress, the extreme heat hazards are relatively small, and the final disaster risks are not serious (Fig. 8). An effective method for reducing summer maize extreme heat disaster risk could be to adjust planting dates so that the highly temperature-sensitive stage does not occur during the time of year when high temperatures frequently occur.

Summer maize extreme heat vulnerability curves have the potential to predict yield losses according to the growing season extreme heat condition; this is quite valuable for extreme heat disaster early warning, risk management, agricultural insurance, etc. However, the fitting goodness of the vulnerability curve presented in this study still has the space to be further promoted. Moreover, it is built based on crop model. Crop model simulation results are related with the model calibration strategies; this may lead to some errors in regional scale application. Vulnerability curves built by the detailed and high-quality historical disaster records are closest to the production practice; however, such records are lacking in many areas. The more realistic vulnerability curve can greatly improve the accuracy of summer maize yield forecasting and disaster early warning, which should be further explored in future researches.

4.2 Conclusions

The present study assessed the risk of direct extreme heat stress on summer maize in Haihe Plain based on the regional disaster risk system theory. Hazard and vulnerability are considered in the risk assessment model. The major conclusions are as follows: (1) In Haihe Plain, extreme heat days most commonly occur in mid-June to early July. For most years and stations, this period coincides with summer maize growing stage I. This results in an extreme heat hazard at this stage that is significantly larger than observed for growing stages II and III. And the hazards appeared higher in the western part of the study area. (2) The fitted vulnerability curves for the three growing stages showed that summer maize is most vulnerable to direct heat stress at growing stage III, followed by stages II and I. However, yield losses that are a result of direct heat stress are relatively limited. (3) Summer maize extreme heat disaster risk increased as hazard levels rose, and the increased magnitude is larger in growing stage I, especially for the western area. This result is mainly due to serious extreme heat hazard. Growing stage III followed due to high vulnerability. The risk assessment results are significant because they provide the basis for a better understanding of how to reduce summer maize extreme heat disaster risk that may arise from a hotter future climate.