The Global Importance of Increasing Design Rainstorms under Specific Return Periods in China

Abstract

Evaluating the correlation between changes in design rainstorms and extreme rainstorm events under climate change facilitates flood control and disaster reduction. Based on the daily rainfall data of 609 stations during 1958–2017, the Pearson-III curve, least square method, F-test, and other methods were adopted to study the changes in design rainstorms and the response to extreme rainstorms over nearly 60 years in China, and to explore the flood control capacity of engineering during climate change. The conclusions are as follows. (1) There is a general increasing trend in extreme rainstorm events in China. Most extreme rainstorm indices present upward trends in East China (EC), South China (SC), Central China (CC), Northwest China (NW), and Southwest China (SW) and downward trends in Northeast China (NE) and North China (NC). (2) The temporal series of design rainstorms shows general growth under each return period in China. In EC, SC, and SW, the design rainstorms increase significantly with rates of 3.0, 3.1, and 1.3 mm/10a, respectively, in the 20-year return period, while they decrease significantly by −2.0 mm/10a in NC (p < 0.05). (3) Design rainstorms have a positive response to extreme rainstorm events, which resulted in increasing rainstorms in a specific return period in EC, SC, NW, and SW. These results can promote the revision of engineering design standards and improve the flood control capability of engineering.

1. Introduction

Rainstorms and floods pose severe challenges to the natural environment and human society worldwide [1,2]. Approximately 6.8 million humans were killed by floods in the 20th century [3]. Large direct losses were observed in China, the United States, Canada, India, and various European countries [4]. For instance, 4713 people were killed by floods from 1959 to 2008, which caused an average of USD 8.2 bn in damages each year in the United States [5,6]. In Europe, 812 floods killed 2466 people locally from 1980 to 2018, and a flood caused approximately 200 thousand properties to lose power in July 2021. [7,8]. Due to the East Asian monsoon and atmospheric circulation, flood disaster losses, which account for approximately 10% of the global losses from 1990 to 2017, are particularly serious in China [9,10]. In recent years, most regions have experienced enormous rainstorms, such as the Shandong, Guangdong, and Henan Provinces, which have caused severe economic losses and disasters including mountain torrents, landslides, and urban waterlogging [11,12,13]. On 20 July 2021, the daily extreme rainstorm reached 552.5 mm at the Zhengzhou Station in the Henan Province, breaking the historical record and affecting 14.79 million people, with economic loss as high as CNY 120.1 billion [14]. To combat rainstorms and floods, numerous water conservancy projects, such as reservoirs, embankments, and flood gates, have been built in China, constituting a relatively complete flood control system. The constructed dams of various types involve nearly 100,000 sites, with total storage greater than 800 billion m³ [15].

With global warming, the intensity and frequency of extreme rainstorms are growing in most countries, and the changing stability of meteorological series may weaken the flood control capacity of engineering under specific precipitation standards [16,17]. The design rainstorm is a critical parameter for determining flood control standards during the design and operation period of engineering [18]. A design rainstorm refers to the maximum possible rainstorm at a specific design frequency, which can be calculated from measured long-time series precipitation through frequency analysis. Thus, the nonstationary change in extreme rainstorms may affect the design rainstorm in the same return period. Exploring whether extreme rainstorm events change the design rainstorm in a specific return period and weaken the flood control capability of engineering, has great significance for preventing flood disasters and for evaluating the reliability of flood-control engineering. Additionally, this has not only local (China) but global importance.

In recent decades, there has been growing concern on the characteristics, mechanisms, and trends of extreme rainstorm events in climate change conditions due to the increasing frequency of flood disasters [19,20]. With the intensification of climate warming, extreme rainstorm events will increase by 3~15% for every 1 °C rise in most countries, mainly in Southeast Asia, southern South America, North America, and western Europe [21,22]. The max 1-day precipitation amount (RX1day) and max 5-day precipitation amount (RX5day) show a significant upward trend in Malaysia, while there is a decrease in consecutive wet days (CWD). In China, the temporal variation in extreme rainstorms is similar to the global trend, but there are large spatial differences due to the vast territory [23]. Extreme rainstorm events increased significantly in Southeast and Northwest China but decreased in Northeast and North China [24]. Most extreme rainstorm indices have shown an increasing trend in Southeast China; for example, RX5day and very wet-day precipitation (R95p) increased at rates of 1.02 and 11.68 mm/10a, respectively [25]. In contrast, most extreme rainstorm indices show a significant decreasing trend in the arid areas of northern China [26,27].

Therefore, flood control design standards calculated using historical hydrological data may not meet expectations, resulting in threats to the reliability and safety of existing projects [28]. During an extreme precipitation event in 2017, several major reservoirs exceeded 85% of their total storage capacity in California, and the Oroville Dam experienced severe socioeconomic losses in the downstream reach due to spillway failures [29]. Similarly, this phenomenon is also very serious in China. For daily design rainstorms of 50-, 100-, and 200-year return periods, most drainage projects are overloaded in urban districts, leading to an increasing risk of waterlogging in Changzhou City, China [30]. The changes in design rainstorms influenced by increasing extreme rainstorm events have been widely investigated. Because more peak values of data points have been added to the daily maximum precipitation series, design rainstorms present an increasing trend in certain return periods [21,31]. For instance, the growth rate of daily design rainstorms was approximately 0.08~0.13 mm/a in the 100-year return period in China, some areas of which exceeded 0.3~0.8 mm/a [32]. During 1961–1990, the design rainstorm of the 100-year return period was 364.7 mm, and it increased by 11.56% during 2011–2020 in the Han River Basin [33].

Previous studies have yielded fruitful results on changes in extreme rainstorms and design rainstorms in China. However, it is unclear whether changing extreme rainstorm events cause a change in design standards so that existing projects fail to achieve their expected flood control objectives. Further investigation should be conducted on the response relationship between design rainstorms and the change in extreme precipitation. In this study, we aim to address the following questions: (1) How have extreme rainstorm indices and design rainstorms changed in recent years? (2) Have changing extreme rainstorm events impacted design rainstorms under specific precipitation standards? The study is based on the daily rainfall data of 609 stations during 1958–2017, combining the Pearson-III curve, least square method, F-test, and other methods. Our framework and resulting propositions can promote the reliability and safety of existing projects, provide technical support for the planning and safe operation of projects in China, and the optimization of human strategies to respond to extreme rainstorms.

2. Materials and Methods

2.1. Study Area

China is in East Asia and on the west coast of the Pacific Ocean (Figure 1). It is the world’s most populous country and a major emitter of greenhouse gases. It has a complex and diverse climate including a monsoon climate in the east, a temperate continental climate in the northwest, and an alpine climate on the Qinghai-Tibet Plateau. The topography exhibits a stepped distribution, with high elevation in the west and low elevation in the east. The regional diversity of natural conditions determines the spatial characteristics of meteorological elements [34]. Along the Qinling-Huaihe River line, precipitation is abundant in the south and scarce in the north.

Recently, natural disasters caused by extreme rainstorm events have led to tremendous ecological, social, and economic losses in China, especially flash floods, landslides, and urban waterlogging. To cope with these disasters, a large number of water conservancy projects have been conducted in China since 1951. According to the statistics of the International Commission on Large Dams (ICOLD), there were 23,841 dams in 2020 (height more than 15 m, or height more than 5 m and more than 3 million m³ in capacity), accounting for 40.6% of the 58,713 dams worldwide. Dam building developed most rapidly from 1951 to 1990, with a total of 21,500 dams constructed that accounted for 90% of the existing dams in China [35]. Based on geographical regionalization, the study area is divided into seven regions: Northeast China (NE), North China (NC), East China [36], South China (SC), Central China (CC), Northwest China (NW), and Southwest China (SW).

2.2. Data Sources

The precipitation data were extracted from the Daily Value Dataset of China Surface Climate Data (V3.0) provided by The National Meteorological Information Centre (http://data.cma.cn, accessed on 7 September 2021) with a time series from 1951 to 2017. Because precipitation data before 1960 are scarce, the study period was determined to be from 1960 to 2017. To reduce data bias, we removed 116 meteorological stations with missing data and selected 609 of the 756 stations as the research objects. The distribution of the meteorological stations is shown in Figure 1.

2.3. Design Rainstorm Calculation Method

The design rainstorms of hydrological engineering under specific precipitation standards are generally calculated using long-series historical precipitation data. To detect changes in the design rainstorms of projects constructed in different years, this study adopted cumulative segmentation to establish the design rainstorm sequence. First, the annual maximum daily precipitation samples of 609 stations from 1960 to 2017 were taken. In terms of the time series, the previous 30 years (1960–1989) represents the base period, which guarantees that the hydrological samples are long series. After 1989, the maximum daily precipitation samples were cumulatively segmented into 28 subseries (1960–1990, 1960–1991, 1960–1992…1960–2017) in one-year increments. The specific segmentation method is shown in Figure 2. The cumulative segmentation is similar to the dynamic hydrological sequence method. However, the maximum daily precipitation samples of 29 subseries provide a data basis for design rainstorm calculations and cannot be averaged directly.

Subsequently, based on the maximum daily precipitation sample, the Pearson-III method was selected to calculate the design rainstorm amount for return periods of 20, 50, 100, 200, 500, and 1000 years, forming the time series of the design rainstorm amount in different return periods. This method is suitable for describing statistical precipitation and other left-biased distribution variables and is widely used in hydrological design worldwide [37,38]. The probability density function of P-Ⅲ is shown as Equation (1).

$$f\left(x\right)=\frac{{\beta }^{\alpha }}{\mathsf{\Gamma }\left(\alpha \right)}{\left(x-a_0\right)}^{\alpha -1}{e}^{-\beta \left(x-a_0\right)}$$

(1)

where Γ(α) is the gamma function of α. α, β, and a₀ are statistical parameters that can be obtained using the coefficient of variation Cv, the coefficient of skewness Cs, and the mean x. The above parameters are evaluated using Equations (2) and (3).

$$Cv=\frac{1}{x}\sqrt{\sum _{i=1}^n{\left(x_i-1\right)}^2/n-1}$$

(2)

$$\alpha =\frac{4}{{{\displaystyle Cs}}^2} ; \beta =\frac{2}{\overline{x}CvCs} ; a0=\overline{x}(1-\frac{2Cv}{Cs})$$

(3)

According to the characteristics of heavy rain in China and practical experience, the ratio of Cs to Cv is usually considered a constant and is 3.5 [39].

2.4. Extreme Rainstorm Index Calculation Method

The bulleted lists look like this: Based on the definition of extreme rainstorm indices recommended by the Expert Group on Climate Change and Monitoring Indicators (ETCCDI) and the literature, nine indices were selected to represent extreme rainstorm events [40] (

Table 1. Definition of nine extreme rainstorm indices.

Category	Index	Indicator Name	Description	Unit
Amount Index	RX1day	Max 1-day precipitation amount	Monthly maximum 1-day precipitation	mm
	RX5day	Max 5-day precipitation amount	Monthly maximum consecutive 5-day precipitation	mm
	R95p	Very-wet-day precipitation	Annual total PRCP when RR > 95th percentile	mm
	R99p	Extreme-wet-day precipitation	Annual total PRCP when RR > 99th percentile	mm
	PRCPTOT	Annual total wet-day precipitation	Annual total PRCP in wet days (RR ≥ 1 mm)	mm
Intensity Index	SDII	Simple daily intensity index	Annual total precipitation divided by the number of wet days	mm·d−1
Days Index	CWD	Consecutive wet days	Maximum number of consecutive days with RR ≥ 1 mm	d
	R20	Number of heavy precipitation days	Annual count of days when PRCP ≥ 20 mm	d
	R50	Number of heavy precipitation days	Annual count of days when PRCP ≥ 50 mm	d

). The nine indices can be divided into three categories: the amount index (RX1day, RX5day, R95p, R99p, PRCPTOT), the intensity index (SDII), and the days index (R20, R50, CWD) [41]. The dynamic hydrological sequence method was used to calculate the extreme rainstorm indices, and is consistent with the long time sequence of design rainstorms. In terms of the time series, the study period of the extreme rainstorm indices was divided into multiple periods in one-year increments (1960, 1960–1961, 1960–1962…1960–2017). Then, the extreme rainstorm indices were converted to averages using the arithmetic mean method during these periods, forming a sequence of nine index averages. The difference between it and cumulative segmentation is that it does not set a base period for long series when dividing the time series, and directly averages the sample during data processing.

2.5. Mann–Kendall Trend Test Method

As a nonparametric trend detection method, the rank-based Mann–Kendall test has been widely used to detect the change trend of meteorological and hydrological sequences, and to verify the significance of the results [16]. The change trend is quantified using Z-statistics in this study. The positive (negative) values of Z indicate increasing (decreasing) trends of the variable; Z is equal to 0 and is stationary. If the absolute value of Z is more than 1.64, 1.96, and 2.56, the changing trend will reach significance levels of 10%, 5%, and 1%, respectively [42]. In this article, the 5% significance level was used to evaluate the significance of the trends.

To quantify the degree of change in extreme rainstorm events, the Sen’s slope estimator method was used to calculate the change rate of the precipitation index [43]. This method reduces the influence of outliers by calculating the median value (β-value) of the slope, usually used together with the M-K test.

2.6. Response Relationship Test Method

Principal component analysis (PCA) is an orthogonal statistical technique that can reduce the dimensionality of certain datasets. It can remove correlated and redundant features in high-dimensional data, replacing them with tractable low-dimensional data. Because the high-dimensional input data are processed using PCA to output new covariates with a high explanatory rate, the important information contained in the original data is not missing [44,45].

To explore the overall impact of the changes in the extreme rainstorm events on the design rainstorm, PCA was used to reduce the dimensionality of the change trend datasets of nine rainstorm indices and design rainstorms with six return periods. The change trend datasets of the nine indices were then converted to one-dimensional data, which represent the total change trend of extreme rainstorms. Likewise, the change trend datasets of the design rainstorms with different return periods of 20, 50, 100, 200, 500, and 1000 years were converted to one-dimensional data, which represent the overall trend of the design rainstorms.

Both extreme rainstorm events and design rainstorms have time-varying characteristics after dimensionality reduction. The Pearson’s correlation coefficient (r-value) was used to explore the correlation between them [37]. If r > 0, the two variables are positively correlated, and the correlation increases with the r value. In addition, the confidence ellipse and confidence band show the confidence level of the correlation (confidence level is 95%).

3. Results

3.1. Changes in Extreme Rainstorm Events

3.1.1. Temporal Change Characteristics

Figure 3 shows the interannual variation in extreme rainstorm indices in China from 1960 to 2017. In addition to extreme-wet-day precipitation (R99p), the amount index (RX1day, RX5day, R95p, R99p, and PRCPTOT) presented increasing trends during the study period. RX1day showed a significant increasing trend (p < 0.05), and the change rates were 0.52 mm/10a. In China, 342 sites (55.3%) had an enhancement in RX1day, of which 75 sites failed the 5% significance test. RX5day, R95p, and PRCPTOT showed a nonsignificant increase, with change rates of 0.10, 0.64, and 6.00 mm/10a, respectively (p > 0.05).

The intensity index (SDII) increased significantly at a rate of 0.13 (mm d⁻¹)/10a (p < 0.05). There were 362 sites nationwide with a significant increase (p < 0.05) in SDII. In the index of extreme rainstorm days (R20, R50, and CWD), only R20 had a nonsignificant increase with a rate of 0.13 d/10a (p > 0.05). In addition, R50 and CWD significantly increased and decreased with change rates of 0.06 d/10a and −0.07 d/10a, respectively (p < 0.05).

In the past 60 years in China, the amount indices of extreme rainstorms showed a nonsignificant increase (p > 0.05), the intensity index significantly increased (p < 0.05), and the days index showed a slight increase. Among the nine extreme rainstorm indices, only R99p and CWD had decreasing trends, while the other indices presented increasing trends. Therefore, extreme rainstorms showed upward trends in China in the past 60 years.

3.1.2. Spatial Variation Characteristics

The spatial distribution of extreme rainstorm events in China from 1960 to 2017 generally follows the pattern of more in the south, less in the north, and more in the east, less in the west (Figure 4). The stations with the highest values of mostly extreme rainstorm indices were located in SC, and the lowest were located in NW. For example, the average value of RX1day was 139.3 mm in SC, while it was only 27.8 mm in NW.

Figure 5 and Figure 6 reveal the M-K trend spatial distribution of the extreme precipitation indices. The spatial variation was similar to those shown in the increase in the west and east and the decrease in the north in the nine indices. RX1day increased significantly in SC, CC, NW, and SW (p < 0.05) with rates of 0.62, 0.52, 0.21, and 0.49 mm/10a, respectively. However, there were significant decreasing trends in NE and NC, with rates of −0.47 and −0.63 mm/10a, respectively, similar to the spatial change laws of R95p and PRCPTOT.

Except for the NE region, the intensity index showed a significant increase (p < 0.05), with the highest rate of 0.13 (mm d⁻¹)/10a in SC. In the days index, R20 increased significantly at rates of 0.19, 0.07, and 0.16 d/10a in SC, CC, and EC (p < 0.05), respectively, and decreased significantly at rates of −0.11 and −0.02 d/10a in NE and NC, respectively. Compared with R20, R50 also showed a significant increase in NW and SW (p < 0.05), while CWD decreased significantly with a rate below −0.05d/10a in regions other than SC (p < 0.05).

Overall, the amount indices showed significant increasing trends in SC, CC, NW, and SW (p < 0.05), and decreased significantly in NE and NC. The intensity index showed a significant increase outside of NE (p < 0.05). In the days index, CWD decreased significantly outside of SC, and the other indices were similar to the amount index.

Figure 7 demonstrates the spatial distribution of the changes in the extreme rainstorm events using PCA dimensionality reduction. The average number of extreme rainstorm events increased at a rate of 2.0/10a in China, mainly showing an increasing trend in EC, SC, CC, NW, and SW. There were larger changes in EC and SC with rates of 6.2 and 8.5/10a, respectively, and lower changes in CC, NW, and SW with rates of 2.1, 0.8, and 0.2/10a, respectively. However, there were stations with negative and larger changes in the extreme rainstorm events in the north of EC, east of NW, and northeast of SW. The remaining areas showed a decrease, among which NE and NC had larger rates of −7.9 and −3.6/10a, respectively. In general, the extreme rainstorm events mainly increased in EC and SC, and decreased in NE and NC.

3.2. Changes in Design Rainstorms

According to the temporal changes in the design rainstorms in 20-, 50-, 100-, 200-, 500-, and 1000-year return periods from 1960 to 2017 (Figure 8), the increasing rates of the design rainstorms were 1.3, 1.6, 1.9, 2.1, 2.5, and 2.7 mm/10a, respectively (p < 0.05). The rates of change increased with the return period.

Figure 9 illustrates that the spatial distribution of the design rainstorms in different return periods had a stepped decrease from southeast to northwest, similar to the extreme precipitation. The design rainstorm in SC was largest (an average of 244.7 mm in the 20-year return period), followed by EC and CC; it was smaller in NE, SW, and NC and was smallest in NW (an average of 50 mm in the 20-year return period).

The spatial distribution of changes in the design rainstorm with different return periods were identical (Figure 10 and Figure 11), showing a significant increase in EC, SC, and SW and a significant decrease in NE and NC (p < 0.05).

In the 20-year return period, a significant increasing trend (p < 0.05) of the design rainstorm series appeared in EC (3.0 mm/10a), SC (3.1 mm/10a), and SW (1.3 mm/10a), and over 30% of stations in these regions showed a significant upward trend. There were nonsignificant growth trends in CC (0.3 mm/10a) and NW (0.59 mm/10a). A significant downward trend was exhibited in NC (p < 0.05) with a rate of −2.0 mm/10a, 58% of stations having decreased significantly, and a nonsignificant decrease (−0.7 mm/10a) observed in NE. With the return period, the rate of change in most regions was characterized by the rate of increase and decrease both gradually increasing. With the return period reaching 1000 years, the increasing rates of the design rainstorms reached 5.5, 6.2, and 3.2 mm/10a in EC, SC, and SW, respectively, and the reduction rates reached −2.3 and −4.2 mm/10a in NE and NC, respectively.

After the dimensionality reduction using PCA, the comprehensive design rainstorm indicated an increasing trend with a rate of 0.8/10a in China; however, diverse trends appeared in different regions (Figure 12). The design rainstorm showed significant upward trends in EC, SC, NW, and SW, and the average rates of change were 1.7, 5.9, 0.8, and 1.8/10a, respectively. In addition, the design rainstorm in other regions had a downward trend. NC had the highest rate of −4.2/10a, followed by NE and CC at rates of −3.9 and −1.2/10a, respectively. In general, the design rainstorm increased in EC, SC, NW, and SW, and decreased in NE, NC, and CC.

3.3. Response of Design Rainstorms to Extreme Rainstorm Events

The correlation between the comprehensive change in design rainstorms and that of the extreme rainstorm events was calculated using the Pearson correlation coefficient, as shown in Figure 13. There was a strong positive correlation between them nationwide (p < 0.05), with 315 stations showing the same trend of design rainstorms and extreme rainstorm events (52%). Except for the CC regions, there was a significant positive trend between the change in the design rainstorms and the change in extreme precipitation over other regions (p < 0.05). The strongest correlation coefficient was greater than 0.5 in SC, followed by NE, NC, EC, and SW (0.2 < r < 0.3), while the weakest was in NW (r = 0.2).

Based on the change in the extreme rainstorm events and design rainstorms, the increasing design rainstorm was mainly influenced by increasing extreme rainstorm events in China. The significant increase in the extreme rainstorm events was the major factor leading to design rainstorm increases in EC, SC, SW, and NW.

4. Discussion

4.1. Comparison of Existing Research Results

In this study, an upward trend was detected in China during 1960–2017 for extreme rainstorm indices, including RX1day, RX5day, R95p, R99p, PRCPTOT, SDII, R20, and R50, which was consistent with the results of previous studies [46,47]. Regional differences were also found in the trends in extreme rainstorm indices in China. There was a significant increase in the amount and intensity index (p < 0.05), which was confirmed by [48,49]. In contrast, the regions in NE and NC are more likely to show significant decreases in amount and intensity, similar to previous studies of Central and North China [24,50]. In addition, a downward trend of CWD in the days index was demonstrated nationwide in this study (p < 0.05). Other studies have also found that the CWD of most stations in China showed a significant downward trend [51,52].

Previous studies have shown that design rainstorms have increased in most countries around the world [53]. In Canada, the design rainstorm with return periods of 10 years were shortened to one return period of 5.7 years [54]. Similarly, compared with the reference period, the design rainstorm with a 100-year return period increased by 15~40% in most European countries [8]. This study indicates that the design rainstorm in different return periods indicates significant growth and that the change rate increases with the return period in China. Regionally, significant increases were present in SC, EC, and SW, with significant decreases in NC. The design rainstorm in SC increased by 32.45% and 33.15% for the 10- and 20-year return periods, respectively, which is consistent with our results [55]. Ref. [32] found that design rainstorms increased in SC, EC, and NE but decreased in SW and NC, which is different from our results. In NE, the downward trend referred to in the study is the overall trend of the region, while their previous study found that the growth change existed only at some grid points. Our study also shows that approximately 30% of the stations increased in the middle of NE. Similarly, in SW, although the overall region presented a significant growth in this study, the trend mainly existed in the eastern SW, while the middle of SW decreased. These results are consistent with [32], which confirms the correctness of our results.

4.2. Causes of the Changes in Extreme Precipitation

The results show that extreme rainstorm events increased at a rate of 2.0/10a in China, which may be a consequence of global warming [56]. With rising temperature, the intensity and maximum value of precipitation, water vapor, and runoff show a significant upward trend worldwide [57]. According to the Intergovernmental Panel on Climate Change (IPCC) report, the growth rate of extreme rainstorms is 7% K⁻¹, which is larger than the average precipitation (1~3% K⁻¹) [36]. As moisture convergence is the dominant factor in extreme rainstorms, this phenomenon may be related to increasing water vapor resulting in climate warming [58]. With climate warming, the weakening of the large-scale monsoon circulation results in significant changes in extreme rainstorms [59,60].

In China, the intensity of extreme rainstorm events increases by 6.52% for each 1 °C increase in warming, which will become more severe with an increased population density [61]. At different warming degrees, the extreme rainstorm events in the monsoon regions in China are mainly driven by the ENSO, IOD, NAO, AMO, and PDO [62]. In SC, CC, EC, and SW, increasing extreme rainstorms are caused by the growing water vapor transport in summer [63]. As the East Asian summer monsoon moves westward, the water vapors are weakened and increased over NC and NW, respectively, leading to extreme rainstorm changes [27,64]. The increase in the WPSH is beneficial to water vapor transport over the tropical Pacific, causing heavy rainfall in EC and eastern SW, and the reducing extreme rainstorms in eastern NE and southern SC [25].

4.3. Impact on Flood Control Capability and Engineering Safety

According to the correlation analysis, there is a significant positive correlation between the change in design rainstorms and extreme rainstorm events [65]. This suggests that, on a global scale, extreme rainstorm events caused by climate change could affect the flood protection capacity of projects and increase flood risk. Hosseinzadehtalaei et al. [66] investigated the impact of climate change on short-duration extreme precipitation and estimated the future intensity–duration–frequency (IDF) curves over Europe. They found future IDF curves will be uplifted and steepened. Additionally, François et al. [20] discussed the challenges in designing floods under climate change and proposed that dam design standards may be affected by flood risk. Additionally, in China, with the changes in local extreme rainstorm events, the design rainstorms in a specific return period will grow in EC, SC, SW, and NW, and decrease in NE and NC. Huang et al. [15] believe that extreme rainstorm events have an influence on design rainstorms in engineering. Therefore, in conditions of increasing extreme rainstorm events, the flood control capability and the reliability of engineering may be weakened under specific design standards [43].

In California, the design rainstorm with the 100-year return period corresponds to the 30-year return period after precipitation changes, for 13 dams built according to the design standard in the 1980s [29]. Thus, the design standard of numerous projects has been unable to meet the anticipation of flood control. For buildings in the 1980s and unfortified Chinese projects, flood control capacity and safety may be threatened when they experience the specific return period of rainstorms. The probability of project failure will increase (such as leakage, landslide, and dam breaks), and the risk of flood further downstream will grow. In particular, engineering flood control capability is more likely to be weakened in areas where there is a significant increase in design rainstorms in the same return period. In EC and SC, the change rates of design rainstorms are the largest, increasing by 4.0 and 4.4 mm/10a, respectively (in the 100-year return period), and the decline in engineering flood control capacity and reliability are the riskiest. In NW and SW, the design rainstorm increases at rates of 0.8 and 2.1 mm/10a (in the 100-year return period), respectively. Their flood control capacity and the reliability of projects are riskier, while the other areas are less risky.

As global climate warming continues, meteorological series will involve more and higher precipitation maxima [67]. The design rainstorm with the 100-year return period will be reduced by at least twice in 40% of the regions in the world [68]. Based on the prediction of CMIP6, the possible increase in extreme rainstorm events in China will reach 20–30% at the end of the 21st century [69]. In the future, the design rainstorm in the specific return period may continue to grow with extreme rainstorms. The threat of flood control engineering and its reliability will be further intensified, especially in regions of significantly increasing design rainstorms, such as EC, SC, NW, and SW. Therefore, it is urgent to check and revise the flood control standard of the major projects to ensure flood control capacity and normal operation. Of course, in the course of the study, there are many issues have not been fully discussed for further in-depth inquiry, such as exploring the potential causes and drivers of the observed trends in extreme rainstorm events and design rainstorms and assessing the potential impacts of these trends on different sectors and regions.

5. Conclusions

(1)

Extreme rainstorm events in China generally indicate an increasing trend with regional differences, showing an upward trend in EC, SC, CC, NW, and SW, and a downward trend in NE and NC. Among the nine indices, most of the amount and days indexes showed increase trends. Rx1day, SDII, and R50 increased significantly, and the rates are 0.52 mm/10a, 0.13 (mm·d⁻¹)/10a, and 0.06 d/10a, respectively (p < 0.05).

(2)

The design rainstorm with different return periods showed a significant increase in China, increasing in EC, SC, NW, and SW and decreasing in NE, NC, and CC. In the 20-year return period, the design rainstorm had an increase at the rate of 1.3 mm/10a, which was mainly distributed in EC, SC, and SW, while NC showed a significant decrease (p < 0.05).

(3)

The design rainstorm had a significant positive response to the changes in the extreme rainstorm events. In the specific return periods, the design rainstorm will change with the extreme rainstorm events, such as the increase in the design rainstorm in EC, SC, NW, and SW and the decrease in NE and NC.

(4)

Against the backdrop of climate warming, design rainstorms will increase, resulting in the aggravation of the project accident probability and flood risk severity increasing in the future. The research is intends to minimize the damage caused by floods at a global level.