Risk Assessment of Compound Dry–Hot Events for Maize in Liaoning Province

Abstract

Extreme climates can result in marked damage to crop yields and threaten regional and global food security. Maize is a major grain crop in Liaoning Province which is severely affected by dry and hot weather events. This study was based on the maize yield and daily meteorological data from various meteorological stations in Liaoning Province from 2000 to 2020. We calculated the standardized dry and hot index and constructed a method for estimating the maize yield reduction risk under compound dry–hot events (CDHE) in Liaoning Province by combining the coefficient of variation in maize yield reduction, yield loss risk index, and the frequency of CDHE during yield reduction. The results showed that the high-risk area for the occurrence of CDHE in maize was Chaoyang City, located in the western part of Liaoning Province. Cities in the low-risk area accounted for approximately 64.3% of the total number of cities in Liaoning Province, mainly in the central and northern parts of Liaoning Province. This study emphasizes the impact of CDHE on agricultural production and provides an index for assessing the risk of CDHE on maize production.

1. Introduction

Climate is the main factor influencing crop yield variability [1,2]. Nearly one-quarter of all the damages and losses from climate-related disasters have recently occurred in developing countries. Drought and high temperatures have significantly reduced national cereal production by 9–10% [3]. The estimated reduction in maize production caused by future droughts is approximately 36–39%. The maize yield loss rate has increased over time [4]. Heat and drought can reduce crop yield by limiting carbon assimilation through photosynthesis, increasing carbon loss from respiration while limiting transpiration [5]. The frequency of heat waves in Northeast China, such as the 2018 Northeast Asian heat wave, has significantly increased [6]. As the scale of future droughts and heat increases [7,8,9,10], compound dry and heat events (CDHE) have become more hazardous than single disasters. Compared to the mid-twentieth century, compound heat and drought events have increased during the maize-growing seasons [8,11]. Since the 1950s, the global frequency of such events has roughly doubled [9,12], with substantial increases in China [12]. In 2012, the combination of severe heat and drought enhanced the heat sensitivity of maize and wheat in the US Great Plains [13], where maize production declined by approximately 20% compared with the national average [14]. Extreme dry–hot events in Europe from May to July 2018 led to maize yield reduction and ecosystem degradation [15]. From 2017 to 2020, Liaoning Province experienced four consecutive drought years. Rainfall during the maize reproductive period decreased by 51%, and the average maximum temperature was more than 3 °C higher than that in normal years. Severe drought and sustained high temperatures occurred in many places, and the area of maize affected by drought was as large as 772,467 ha [16].

Therefore, many studies have focused on CDHEs [17,18,19,20]. Measuring CDHEs requires selecting or constructing a correlation index [21] by selecting events in which drought and a given threshold number of high-temperature days occur simultaneously. This method can quickly locate a CDHE’s onset but cannot comprehensively analyze its development. The impacts of CDHEs depend not only on the frequency of their occurrence but also on their severity. Therefore, studies have increasingly assessed and quantified the hazards of CDHEs by constructing composite indices [17,22]. Mukherjee et al. [23] built a compound heat and drought index based on the modified Palmer Drought Index (sc_PDSI) and the daily maximum temperature using a copula function to identify compound heat and drought events in 26 global climatic zones between 1982 and 2016. Hao et al. [24] developed a Standardized Compound Event Indicator (SCEI) to characterize the severity and spatial distribution of CDHEs in southern Africa and globally using a meta-Gaussian model based on the Standardized Precipitation Index (SPI) and Standardized Temperature Index (STI). Li et al. [25] constructed a Standardized Compound Dry and Heat Index (SCDHI) based on the joint construction of daily scale SPEI and STI indices. Previous studies have mainly used precipitation and temperature to construct CDHEs. However, vegetation evapotranspiration is vital in understanding the impact of CDHEs on crops. The SPEI can more effectively characterize drought conditions by considering evapotranspiration than the SPI; therefore, the joint construction of the SPEI and STI was selected to characterize CDHEs in this study.

Liaoning Province is located on the east coast of the Eurasian continent, with a temperate continental monsoon climate, rolling hills on the east and west sides, and plains in the center. There are four distinct seasons: short spring, autumn, rain and heat in the same season, and frequent climatic disasters [26]. The overall temperature in Liaoning Province has increased due to global warming. The amount of precipitation has decreased during normal rainfall periods, with heat and drought events. Maize is the primary grain crop in Liaoning Province, where the production amount is closely related to food security in the region [27]. The maize sown area is 2,424,900 ha, accounting for 56.65% of the province’s grain crop area [28]. The frequent occurrence of CDHEs has caused significant economic losses, seriously affecting social and economic development. This has become an essential factor that restricts the sustainable development of the resources and environment in Liaoning Province.

Assessing the crop yield loss risk caused by climatic and socioeconomic conditions is essential for sustainable agricultural production and for investigating the uncertainties and risks associated with climate change [29,30]. Preliminary studies were conducted on risks associated with single meteorological hazards [31,32,33]. However, there is a lack of research on the assessments of CDHEs in maize, especially as heat and drought significantly impact the maize yield.

This study was based on maize yield data from 14 cities in Liaoning Province from 2005 to 2020. We combined the coefficient of variation in maize yield reduction, yield loss risk index, and the frequency of CDHEs during yield reduction. We analyzed the yield damage caused by CDHEs of maize in Liaoning Province and the zoning of yield damage across the entire province. Therefore, this study emphasizes the importance of examining the maize yield loss caused by CDHEs, with the establishment of a scientific monitoring and early warning system. This can effectively reduce losses caused by CDHEs, thus achieving a stable income increase in Liaoning and promoting the sustainable development of agriculture.

2. Study Area and Data Sources

2.1. Study Area

Liaoning Province (38°43′–43°26′ N, 118°53′–125°46′ E) is located in the southern part of Northeast China and comprises 14 prefecture-level cities, with a land area of 148,600 km² [27]. The terrain is high in the north and low in the south, with hills and mountains descending from east to west towards the central plains. The average temperature is 8.8 °C [16], decreasing gradually from the coast inland, with a temperature difference of 5 °C between the north and south. The average annual precipitation is 648 mm, with high rainfall concentrated in the summer months. Figure 1 shows the distribution of the meteorological stations in the study area.

2.2. Data Sources

The meteorological observational data for this study were obtained from the China Meteorological Science Data Sharing Network (https://data.cma.cn/, accessed on 31 December 2020). Climatic data, such as daily maximum temperature, daily minimum temperature, sunshine hours, relative humidity, wind speed, and precipitation from 2005 to 2020, were obtained from meteorological stations in Liaoning Province. The data from meteorological stations in Liaoning Province, China, were analyzed and used to calculate the CDHEs. Maize sowing data were sourced from the Liaoning Provincial Bureau of Statistics (https://tjj.ln.gov.cn/tjj/tjxx/xxcx/tjnj/index.shtml, accessed on 31 December 2020), which included data on the maize sown area, area under cultivation, production per unit area, and total production in each city.

3. Methodology

3.1. CDHE Construction

Hao et al. [17] improved the CDHE index. The Standardized Precipitation Evapotranspiration Index (SPEI) and the Standardized Temperature Index (STI) represented dry and hot conditions for 2005–2020, respectively. The severity (X) of a CDHE was assessed as X = SPEI/STI. Therefore, we defined a matrix for precipitation (P), evapotranspiration (E), and temperature (T), where X = G1 (P), evapotranspiration (E)/G2 (T). Functions G1 (P-E) and G2 (T) are the marginal probability distribution functions of precipitation minus evapotranspiration and temperature, respectively. Subsequently, X was normalized based on the SPEI. First, we fitted the marginal cumulative distribution, F, and then performed normalization according to the standard normal distribution, Φ. We referred to this index as the standardized dry–hot index (SDHI), which can be expressed as follows:

$$SDHI={\Phi }^{-1}\left[\mathrm{F}\left(X\right)\right]$$

(1)

When calculating the SDHI, three marginal distribution functions (G1, G2, and F) must be fitted to calculate the marginal probability distribution using Gringorten’s empirical method:

$$P=\frac{\mathrm{i}-0.44}{\mathrm{n}+0.12},$$

(2)

where i is the rank and n is the total number of observations.

Lower SDHI values indicate more severe compound dry and heat events. With calculations based on a standardized method similar to the SPEI, we chose the same grading method for CDHEs. Therefore, referring to the SPEI, a threshold of –0.5 was selected to define the occurrence rate of CDHEs [23]. The ratio of the number of stations where a CDHE occurs to the total number of stations is called the station ratio.

3.2. Mann–Kendall Mutation Test

The Mann–Kendall trend test, recommended by the World Meteorological Organization as a rank-based nonparametric trend test, constructs a standard Z-statistic to represent the trend detection value. The samples do not have to follow a sample-specific distribution and are not disturbed by outliers. For a time series of n samples, the standardized test statistic Z is as follows:

$$Z=\left\{\begin{array}{cc}\frac{S-1}{\sqrt{var\left(S\right)}}\hfill & S>0\hfill \\ 0\hfill & S=0\hfill \\ \frac{S+1}{\sqrt{var\left(S\right)}}\hfill & S<0\hfill \end{array}\right..$$

(3)

Among them

$$S={\displaystyle \sum _{i=1}^{n-1}{\displaystyle \sum _{j=i+1}^n\mathrm{sgn}(x_j-x_i)}},$$

(4)

$$sgn\left(x_j-x_i\right)=\left\{\begin{array}{cc}1\hfill & x_j-x_i>0\hfill \\ 0\hfill & x_j-x_i=0,\mathrm{and}\hfill \\ -1\hfill & x_j-x_i<0\hfill \end{array}\right.$$

(5)

$$\mathrm{Var}(S)=\frac{n\left(n-1\right)(2n+5)-{\displaystyle \sum _{p=1}^q{t}_p(t_p-1)(2t_p+5)}}{18},$$

(6)

where x_j is the number of j in sequence x, q is the number of arrays, and t_p is the number of data points in array p.

The UF statistic measures the difference between the numbers of increasing and decreasing extremes in a data series. Based on this, the variable UF(S) is defined as follows:

$$UF(S)=\frac{\left[S-E(S)\right]}{\sqrt{\mathrm{Var}(S)}}.$$

(7)

The UB statistic was used to measure the time interval between the extremes. Defining the reverse sequence X′ = {X_n, X_n−1, …, X₁}, the above calculation process was repeated to obtain the trend sequence UF′(S), where UB(S) is

$$UB(S)=-{UF}^{\prime }(S),$$

(8)

When UF intersects with UB, the intersection point is the mutation point of the time series.

3.3. Maize Yield Loss

The segregation of the maize meteorological yield is currently the most commonly used method for analyzing the relationship between meteorological factors and maize yield. The maize yield is generally divided into three components: trend yield, meteorological yield, and random error (generally negligible). Therefore, actual maize yields are often decomposed into two components, trend and meteorological yields, expressed as follows:

$$Y=Y_t+Y_w,$$

(9)

where Y is the actual yield, Y_t is the trend yield, and Y_w is the meteorological yield. A sliding average method for yield separation was used to obtain trend yield values with a window size of five years.

The relative meteorological yield was obtained from the ratio of the meteorological yield to the trend yield:

$$Y_{\mathrm{a}}=\frac{Y_w}{Y_t}\times 100\%.$$

(10)

Relative meteorological yields indicate the magnitude of food fluctuations and are comparable regardless of time and space. A positive Y_a indicates that meteorological conditions are generally favorable for maize production, with an increase in the maize yield; a negative value indicates that meteorological conditions are usually unfavorable for maize production, with a decrease in the maize yield. Years with relative meteorological yields of less than −5% are usually defined as disaster years. When Y_a is negative, its absolute value is defined as the yield reduction rate:

$${{\mathrm{Y}}_W}^{\prime }=\left|Y_{\mathrm{a}}\right|=\left|\frac{Y_w}{Y_t}\right|\times 100\%.$$

(11)

3.4. Coefficient of Variation in Maize Yield Reduction

The coefficient of variation in maize yield reduction (C_v) [33] refers to fluctuations in the maize yield during the year of reduction. Larger values indicate greater interannual fluctuations in the impact of meteorological hazards and more unstable yield losses; conversely, a smaller dispersion indicates more stable yield losses. This value partially reflects the fluctuation of meteorological hazards in a certain area and its ability to prevent and mitigate disasters:

$$C_{\mathrm{v}}=\frac{1}{\overline{{\mathrm{y}}_{\mathrm{i}}}}\sqrt{\frac{{\displaystyle \sum _{\mathrm{i}=1}^{\mathrm{n}}({\mathrm{y}}_i-\overline{y_i}{)}^2}}{n-1}},$$

(12)

where y_i is the yield in the reduction year (average yield in the reduction year).

3.5. Yield Loss Risk Index

To verify whether Y_a follows a normal distribution, the Lilliefors goodness-of-fit test was selected owing to the small sample size. The logarithmic method was used for a few samples that did not conform to a normal distribution. We established a probability density function for a normal distribution based on probability theory:

$$f(x)=\frac{1}{\sigma \sqrt{2\pi }}{e}^{-\frac{{(x-\mu )}^2}{2{\sigma }^2}},$$

(13)

where μ is the mean and σ is the standard deviation.

The probability that random variable x occurs in the interval (x₁, x₂) was defined as follows:

$$P(x_1\le x\le x_2)={\displaystyle {\int }_{{\mathrm{x}}_1}^{x_2}f(\mathrm{x})\mathrm{d}\mathrm{x}},$$

(14)

where P is the risk probability, and x₁ and x₂ are the yield reduction rates. Based on the actual reduction in the maize yield, the reduction rate was divided into four intervals [5%, 10%], [10%, 20%], [20%, 30%], and 30%. The risk index was calculated as follows:

$$Q_{\mathrm{t}}={\displaystyle \sum _{i=1}^n{E}_i{P}_i} \mathrm{and}$$

(15)

$$Q=\frac{Q_{\mathrm{t}}}{\max },$$

(16)

where Q_t is the intermediate variable, Q is the risk index on the interval (0, 1], P_i is the interval probability (Equation (14)), E_i is the median value of the interval, and max is the maximum value of the spatial distribution of Q_t. Larger Q values indicate greater yield loss risks.

3.6. Frequency of CDHEs at Reduced Yields

The yield reduction years due to CDHEs were obtained based on the yield reduction sequences. We determined the probability of the occurrence of compound heat and drought yield reduction years as follows:

$$F=\frac{d}{\mathrm{D}},$$

(17)

where d is the year of yield reduction due to compound heat and drought and D is the total number of years counted (16 years from 2005 to 2020).

3.7. Maize CDHE Loss Risk Index

The coefficient of variation of the maize yield reduction, yield loss risk index, and frequency of CDHEs that occurred in the selected disaster years were used to construct the maize compound dry and heat event loss risk index as follows:

$$\mathrm{Z}=C_{\mathrm{v}}FQ.$$

(18)

The risk type classification indicators were standardized using the extreme difference method:

$$M=\frac{Z_{\mathrm{i}}-Z_{\min }}{Z_{\max }-Z_{\min }}$$

(19)

where M ranges between 0 and 1. The degree of risk increased with an increase in the M value.

The indicators were automatically divided into four levels using the natural break classification (Jenks;

Table 1. Risk level classification.

Risk Level	Low-Risk	Medium-Risk	Second-Highest-Risk	High-Risk
M	(0, 0.23]	(0.23, 0.4]	(0.4, 0.65]	(0.65, 1]

).

4. Results and Analysis

4.1. Spatial and Temporal Variation Characterizations of CDHEs during Maize Growth

4.1.1. Time Variation Characteristics

Figure 2 shows the station ratio of CDHEs (SDHI ≤ 0.5) occurring in different maize growth periods in Liaoning Province. The ratio of stations with CDHEs during the sowing–seedling emergence stage showed a decreasing trend (Figure 2a). In 2005, 50% of the stations experienced CDHEs. None of the stations experienced CDHEs after 2013. The UF values were negative from 2005 to 2020, indicating a “downward” trend in the number of stations experiencing CDHEs. The UF and UB curves did not intersect, suggesting that there was no sudden change in the number of stations.

Figure 2b shows that there was a decrease in the ratio of stations experiencing CDHEs during the seedling emergence–jointing stage. In 2008 and 2009, 28.57% of the stations simultaneously experienced CDHEs. The ratio of stations that experienced CDHEs fluctuated by approximately 5% from 2012 to 2018. The UF and UB curves intersected in 2007 and 2011, and the UF value was less than zero after 2011, which implies that the station ratio of CDHEs showed a significant downward trend. In other words, the number of stations with CDHEs decreased during this period.

As shown in Figure 2c, the ratio of stations with CDHEs during the twelfth leaf showed an increasing trend. In 2020, 92.86% of the stations experienced CDHEs. Only 2008, 2012, and 2013 did not have stations with CDHEs. The UF value was negative from 2010 to 2017, indicating that the number of stations with CDHEs showed a “downward” trend during this period. The intersection of the UF and UB curves in 2019 indicates that the compound dry–hot station ratio changed abruptly during that year. The UF value was greater than zero, with an increasing trend after 2019, indicating that the station ratio showed a significant upward trend after this period.

As shown in Figure 2d, the ratio of stations with CDHEs during the tasseling stage showed a decreasing trend. The number of stations experiencing CDHEs in 2009 and 2014 was significantly higher than that in other years, accounting for 92.86%. The UF values were positive from 2005 to 2017 (except for 2012 and 2013), indicating that the number of stations with CDHEs showed an “upward” trend during this period. The UF and UB curves intersected in 2007 and 2019, indicating that the compound dry–hot station ratio changed abruptly in these years. The UF value was greater than 0 with an increasing trend after the mutation in 2007, implying that the compound dry–hot station ratio showed a significant upward trend thereafter.

Figure 2e shows that the station ratio of CDHEs during the maturity stage had an increasing trend, with no stations experiencing CDHEs before 2010. In 2011, 50% of the stations experienced CDHEs. The UF and UB curves intersected in 2019, indicating that the compound dry–hot station ratio mutated that year. The UF value showed an upward trend after the mutation in 2019, indicating that the compound dry–hot station ratio had an upward trend thereafter.

Figure 2f shows the ratio of stations with CDHEs throughout the plantation stage. The number of stations with CDHEs was significantly higher in 2014 than in other years (51.79%). The UF and UB curves intersected in 2008, 2011, 2014, and 2015, indicating that the composite dry–hot station ratios mutated in these years and that the UF values were all positive after the mutation, implying that the compound dry–hot station ratios showed a significant upward trend after the mutation.

Overall, from 2005 to 2020, only 2012 did not experience a CDHE. The ratios of the CDHEs in 2009 and 2014 were the largest, at 46.43% and 51.79%, respectively. The proportion of stations experiencing CDHEs during the two maize reproductive stages showed an increasing trend, with the most significant increase during the trumpet stage. CDHEs occurred more frequently during the twelfth leaf and tasseling stages.

4.1.2. Spatial Characteristics

In this study, the frequency of CDHEs occurring in the maize growth stages in Liaoning Province from 2005 to 2020 was calculated based on Equation (17). Figure 3 shows the spatial distribution results. Significant differences were observed in the frequency of CDHEs at different fertility stages. Figure 3a shows that there was no notable regularity in the spatial distribution of the frequency of CDHEs during the sowing–seedling emergence stage. Shenyang exhibited the highest frequency of CDHEs (18.74%). Four cities did not experience CDHEs during this period and were concentrated in the central and southern parts of Liaoning Province.

Figure 3b shows that the frequency of CDHEs occurring in the seedling emergence–jointing stage had a spatial distribution of low in the middle and high in the surroundings, with the maximum value occurring in Shenyang City and Huludao City, where CDHEs occurred three times in total. No CDHEs occurred in Anshan City, Liaoyang City, Panjin City, or Yingkou City during this period.

As shown in Figure 3c, the frequency of CDHEs that occurred in the twelfth leaf ranged from 18.76% to 50%, which was the most frequent period for CDHEs during the maize reproductive period. The large-value area was concentrated in the eastern part of Liaoning Province, and the maximum value appeared in Benxi City and Dandong City, where a total of eight compound high-temperature drought disasters occurred. The low-value areas were concentrated in the central part of Liaoning Province, with the lowest number of compound high-temperature drought disasters occurring in Jinzhou City (three occurrences).

Figure 3d reveals that the frequency of CDHEs during the tasseling stage ranged from 12.50% to 37.49%, with a large area concentrated in the western part of Liaoning Province. The maximum occurred in Chaoyang City, with a total of six CDHEs. The low-value area was concentrated in the central part of Liaoning Province, with Yingkou City and Panjin City experiencing the smallest number of CDHEs.

As shown in Figure 3e, the frequency of CDHEs in the maturity stage ranged from 0 to 31.22%, with the high-value area concentrated in the western part of Liaoning Province and the maximum value in Jinzhou City. The low-value areas were distributed in the northern, southern, and central regions of Liaoning Province, where Tieling, Dalian, and Yingkou did not experience CDHEs.

As shown in Figure 3f, the frequency of CDHEs during the entire plantation stage ranged from 8.75% to 22.4%, with a spatial distribution trend of low in the middle and high in the surrounding areas. The high-value areas were mainly concentrated in the western and eastern parts of Liaoning Province, with the maximum values occurring in Dandong City and Chaoyang City. Of these stations, 64.28% had CDHEs higher than 15%. The low-value areas were concentrated in the central part of Liaoning Province, with the lowest number of CDHEs occurring in Yingkou City.

Generally, the frequency of CDHEs in central Liaoning Province was relatively low during all the maize reproductive periods. Except for the trumpet stage, western Liaoning Province was prone to CDHEs. A total of 18 CDHEs occurred in Shenyang City, followed by Chaoyang City and Dandong City (17 events) during all the maize reproductive stages. For the CDHE occurrence characteristics at different fertility periods, the average CDHE frequency was greatest at the twelfth leaf and tasseling stages, which is the flowering period for early-, medium-, and late-maturing maize varieties. Changes in sowing dates can be used to adjust the maize flowering period to avoid the high temperatures in late July and early August and reduce the impact of high temperatures on yield [34].

4.2. Maze Yield Variation Characteristics

4.2.1. Variation Characteristics of Maize Yield

Figure 4 shows the overall fluctuations in the maize yield in Liaoning Province over the past 15 years. Owing to the level of agricultural technology, variety replacement, and adjustments to fertilizer application, the total maize output has recently risen from 954,000 to 1,303,000 tons, showing an upward trend. From 2005 to 2020, the number of years of maize yield reduction in Liaoning Province was five, accounting for 31.25% of the total number of years. The average disaster was a one-in-three-year event, indicating that maize production in Liaoning Province was more significantly affected by natural disasters. The years 2009 and 2014 had higher maize yield reductions, averaging 8.38%, which may be related to years with more CDHE occurrences. In 2010, 2015, and 2016, the actual maize yields were high and, in some cases, even positive; however, the corresponding years experienced a reduction in yield. This trend suggests that meteorological hazards are essential factors affecting maize yield.

4.2.2. Spatial Characteristics of the Coefficient of Variation of Maize Yield Reduction

To analyze the inter-annual variability and stability of maize yields, Figure 5 shows the spatial distribution of the coefficient of variation of maize yield reduction from 2005 to 2020. The high coefficient of variation implies that maize yields fluctuate significantly between years, and yields are susceptible to climatic and socioeconomic conditions. Figure 5 shows that there was no clear regularity or regionality in the distribution of the coefficient of variation for maize yield reduction. The values of the coefficient of variation of maize yield reduction ranged from 0 to 1.16. The maximum value occurred in Anshan City (center of Liaoning Province), characterized by the worst maize yield reduction stability. The high-value area was mainly distributed in the eastern part of Liaoning Province, with a few areas concentrated in the central and western parts. The coefficients of variation of maize yield reduction in Jinzhou City, Fuxin City, Tieling City, Shenyang City, and Liaoyang City were all less than 0.23, indicating relatively stable fluctuations in the maize yields.

4.2.3. Spatial Distribution of the Yield Loss Risk

The yield loss risk index (YLRI) combines the rate of reduction in maize yield and its probability of occurrence at the corresponding level. The risk of maize yield loss increased with the YLRI, implying that maize yields are vulnerable to climatic conditions. Figure 6 shows the spatial distribution of the maize YLRI in Liaoning Province from 2005 to 2020, with high-value areas in the western and southern parts of Liaoning Province while northern areas had a lower yield loss risk. Comparing Figure 5 and Figure 6 shows the differences in the spatial distribution of the coefficient of variation for maize yield reduction and the YLRI. The YLRI is only related to climatic conditions, thus explaining any notable differences. In contrast, the coefficient of variation of maize yield reduction is related to local socioeconomic factors (i.e., technological development, infrastructure, and investment) and climatic conditions.

4.3. Spatial Distribution of CDHE Frequency during Yield Reduction

Figure 7 shows the calculated and interpolated results for the frequency of CDHE occurrence during the 2005–2020 maize yield reduction in Liaoning Province. The frequency of CDHEs at the time of yield reduction showed spatial distribution characteristics of decreasing from west to east, with the occurrence frequency ranging from 6.25% to 31.24%. The large-value area was concentrated in the western part of Liaoning Province. The maximum value appeared in Chaoyang City, with a total of five CDHEs. The low-value areas were concentrated in the eastern part of Liaoning Province, with an average of two occurrences. Fushun City and Shenyang City had the lowest number of CDHEs, with one occurrence each. In total, 35.71% of the stations had a low drought frequency of 15% or higher. Overall, western Liaoning Province had a higher frequency of CDHEs during maize yield reduction, suggesting that the maize yield reduction is vulnerable to CDHEs.

4.4. Risk Assessment of CDHEs for Maize

The risk values of CDHEs for maize were calculated separately for each city in Liaoning Province to obtain their spatial distribution. Figure 8 reveals a clear spatial variability in the risk of CDHEs for maize in Liaoning Province. The low-risk zone was concentrated in the central and northern parts of Liaoning Province, with nine cities accounting for 64.3% of the total number of cities in Liaoning Province. The medium-risk areas were concentrated in the southeastern part of Liaoning Province. Panjin City and Huludao City were the second-highest risk areas for CDHE occurrence in maize. Chaoyang City, in the western part of Liaoning Province, is a high-risk area for CDHEs affecting maize. This region is located in northwestern Liaoning, characterized by the lowest precipitation and insufficient water resources.

Overall, there were some differences in the spatial distribution characteristics of the maize compound dry and heat event loss risk index and the maize growth period of dry–hot events. Among these, Chaoyang City was the most consistent. At the same time, Shenyang City, with the highest number of CDHE occurrences, was only a low-risk area, indicating that CDHEs had a relatively small impact on maize yields in Shenyang City. Due to topography and other factors, Panjin City, Dalian City, and Dandong City have also become medium-risk regions and second-highest-risk regions.

5. Discussion

Recently, compound climate disasters have occurred frequently in many locations [34,35,36,37,38,39,40]. These events have impacted China’s agriculture, water resources, human health, infrastructure, and ecosystems [15,41,42], resulting in serious constraints on sustainable economic and social development [43]. Droughts occurring in conjunction with extreme heat are often referred to as CDHEs, which can have amplified impacts on agriculture, ecosystems, and water resources that may be greater than those of a single disaster [44,45,46]. The frequency, duration, and severity of CDHEs have increased significantly in Northeast China [47]. Numerous studies have shown that heat and drought lower the yields of major crops such as maize and wheat [48,49,50]. The decline in yield is most significant when heat coincides with dry conditions [51,52].

Based on the improved compound dry and heat indices, we identified CDHEs during the maize-growing period in Liaoning Province and analyzed their temporal and spatial distribution characteristics. Hao et al. [23] showed a general trend for the increasing frequency of CDHEs in the northeast region. In hindsight, by comparing the yield reduction rate with the number of CDHEs in the corresponding years, CDHEs have recently become the main meteorological disaster associated with maize yield reduction in Liaoning Province, consistent with the results of a study by Shi et al. [53]. The growth trend in CDHEs during the twelfth leaf stage of maize was significant, consistent with that in the study by Guo et al. [54]. From a spatial perspective, more CDHEs have occurred in the western part of Liaoning Province; this result is consistent with the study by Wu [55] et al. which may be related to decreased precipitation and increased temperatures [56]. The low CDHE occurrence in central Liaoning Province may be due to the influence of an abnormal northwesterly airflow in Liaoning in conjunction with an anomalously weak subtropical high pressure in the Pacific Ocean, which has led to a cooling process and low precipitation [57]. As evapotranspiration has a significant impact on maize growth and development, a combination of SPEI and STI was selected to characterize the composite dry–hot scenario. Zhang et al. [33] assessed CDHEs using multiple drought indicators and showed that SPEI is a better choice for compound dry–hot assessments in arid and semi-arid regions.

A complete understanding of the processes leading to compound heat and drought disasters is essential for reliable risk projections under climate change. Previous studies have applied various crop-risk assessment methods and indicators [58,59,60]. Zhang et al. [33] conducted an agricultural risk assessment of CDHEs, observing a significant increase in hazards in northeastern China. In previous studies, a single-hazard meteorological disaster risk assessment combined hazard, exposure, vulnerability, and disaster prevention and mitigation capacity to construct a risk assessment model. This combination; however, does not adequately consider the difficulty of quantifying the indicators in a multi-hazard situation. In this study, the occurrence probability of different yield reduction rates, coefficient of variation of maize yield reduction, and CDHE frequency in the year of yield reduction were selected as the three indicators to construct the maize CDHE risk assessment models. This index improves the spatial comparability of risks and effectively considers compound disasters. These results are crucial for agricultural decision support systems and climate change assessments in Liaoning Province [61].

The western region of Liaoning Province was the region with higher values for both the coefficient of variation for maize yield reduction and the YLRI, indicating more significant fluctuations in the maize yield variance and susceptibility to disasters and changes at socioeconomic levels, consistent with Li et al. [62]. Overall, the YLRI was higher in western Liaoning Province from 2005 to 2020, consistent with the findings of Zhang et al. [63]. Tang et al. [64] examined the future risk of CDHEs in China, suggesting that it is high in the northern part of the country. Irrigation currently increases maize yields by 20–30% [65]. This reduces the impact of weather extremes on maize yields [66,67]; however, only 11% of Liaoning Province is effectively irrigated. Therefore, we must rationalize the exploitation of water resources for maize irrigation.

Maize compound dry and heat event risk assessment low-risk region is large, whereas the level of urban development is rapid; therefore, more water resource extraction can be appropriately expanded to maize planting areas. The medium-risk region is located in the southeastern coastal area of Liaoning Province, where rainfall is abundant. However, urban land use is extensive and the soil moisture is poor; therefore, authorities must strengthen field management and improve agricultural disaster mitigation techniques. Panjin City, Huludao City, and Chaoyang City, the second-highest- and highest-risk regions, should use early maturing and drought-tolerant varieties when planting maize.

The comprehensive, accurate, and rapid analyses of compound hazard intensities and risks form the basis for effective agrometeorological hazard risk management. Considering the nutritional and economic value, as well as widespread cultivation of maize, our study provides instructive information on the distribution and risk assessment of CDHEs. It can also be extended to risk assessments for other maize, such as food maize. In summary, we analyzed risk from the perspective of maize yield reduction and loss due to CDHEs, providing scientific guidance for the prevention and mitigation of maize disasters in Liaoning Province.

6. Conclusions

In this study, based on data from 14 meteorological stations in Liaoning Province, we constructed a compound dry and heat index to analyze the spatial and temporal distribution characteristics of CDHEs for maize in Liaoning Province during each growing period from 2005 to 2020. The yield disaster loss division of maize was constructed based on the risk analysis of CDHEs. The main conclusions are as follows:

• We revealed the temporal characteristics of CDHEs over the maize reproductive period. From 2005 to 2020, only 2012 did not have a CDHE. The ratios of CDHEs in 2009 and 2014 were the largest, at 46.43% and 51.79%, respectively. CDHEs occurred most frequently during the twelfth leaf stage, with a significant increasing trend. The second-highest number of CDHEs occurred during the tasseling stage. The spatial distribution characteristics were as follows: The frequency of CDHEs in central Liaoning Province was low. Western Liaoning Province was generally prone to CDHEs.
• The areas with high coefficients of variation for maize yield reduction were mainly located in eastern Liaoning Province, with several concentrated in central and western Liaoning Province. There were differences in the spatial distribution of the YLRI and the coefficient of variation of maize yield reduction, with the high-value areas in western and southern Liaoning Province.
• There was some variability in the spatial distribution characteristics of the maize CDHE risk index and the maize-growing period compound dry and heat index. The low-risk region was large, mainly in the central and northern regions of Liaoning Province. Chaoyang City in western Liaoning Province was a high-risk region for maize CDHEs, and Panjin City and Huludao City were the second-highest-risk regions.