Abstract

Disasters caused by extreme precipitation under global warming are anticipated to have a strong negative impact on urban construction and social security. In this study, daily grid precipitation datasets of the Guangdong-Hong Kong-Macao Greater Bay Area (GBA) for the period 1961–2018 were extracted to explore the temporal and spatial characteristics of extreme precipitation by using regression analysis, moving average and kriging interpolation. The frequency and intensity indices showed an increasing trend, whereas a decreasing trend was found for the persistence indices, which indicates that GBA tends to slowly become wetter. The mean values of extreme precipitation indices (EPIs) in GBA generally increased from west to east and from north to south. Except for the indices of consecutive wet days and consecutive dry days, other EPIs showed an upward trend in most regions, especially in coastal cities where floods are more likely to occur. Principal component analysis and regression analysis showed that the correlations between the EPIs mostly passed the 0.05 significance test, which suggests that they had a good indicator of extreme precipitation in GBA. This study provides a theoretical basis for extreme precipitation disaster prevention and control within the urban agglomerations of the GBA.

HIGHLIGHTS

  • Nine indices were applied to spatio-temporally analyze extreme precipitation trends.

  • Time-series analysis and spatial interpolation were used to examine the characteristics of extreme precipitation in the GBA.

  • Except for CDD and CWD, the other seven EPIs had an increasing tendency.

  • Coastal regions generally showed a stronger upward trend of extreme precipitation.

INTRODUCTION

Extreme precipitation usually refers to precipitation of high intensity but with a low occurrence probability. Disasters caused by extreme precipitation can have a serious impact on the natural environment and human life. According to the fifth Climate Change Assessment Report by the Intergovernmental Panel on Climate Change (IPCC), global mean surface temperature (GMST) increased by 0.85 °C between 1880 to 2012. Global warming has accelerated the hydrological cycle and increased atmospheric water vapor fluxes, leading to substantial changes in the intensity and frequency of extreme hydrological events (Papalexiou & Montanari 2019). Research on extreme precipitation is of great importance to social development and human health.

The past decade has seen many studies devoted to the temporal and spatial characteristics of extreme precipitation. Extreme precipitation events have occurred increasingly frequently around the globe during the 20th century. Sun et al. (2021) examined changes in extreme precipitation (measured by such indicators as the annual maxima of 1-day (RX1day) and 5-day (RX5day)) based on high-quality station data from 1900 to 2018, and found that extreme precipitation had increased at about two-thirds of the stations. It has been found that extreme precipitation has a statistically significant connection to the GMST, and extreme precipitation is anticipated to become dramatically stronger with further global warming (Myhre et al. 2019). There have been regional studies that yielded different patterns of extreme precipitation (Soltani et al. 2016; Li et al. 2020; Tian et al. 2020). In Alaska, the magnitude of the projected increase of RX1day was the largest in the southern area, although the percentage increase was largest in the north of Alaska (Bennett & Walsh 2015). However, the great majority of extreme precipitation indices (EPIs) in the Chichaoua-Mejjate region in Morocco have shown a non-significant downward trend, although temperature-based indices of all stations showed increasing trends (Hadri et al. 2020). In the West Rapti River basin of Nepal, EPIs are anticipated to increase significantly in future, especially RX1day and RX5day (Talchabhadel et al. 2020).

With the rising temperature in China, the characteristics of extreme precipitation of the country have also changed. Cao et al. (2020) analyzed the daily data of 554 meteorological stations throughout China from 1960 to 2017 and found that the magnitude and intensity indices of extreme precipitation were increasing, while the indices of sustainability were decreasing, and these indices are included in the nine indices in this paper. To date, there are few studies focusing on the temporal and spatial changes of extreme precipitation in the Guangdong-Hong Kong-Macao Greater Bay Area (GBA), and most of the related studies focused on areas either larger or smaller than it. Zeng et al. (2017) found that extreme precipitation in most parts of western and southeastern China showed an upward trend during 1961–2010, such as the northwest river basin, Huai River basin, eastern Yangtze River basin, southeast river basin, southwest river basin and eastern Pearl River basin. Yu et al. (2018) selected the summer daily precipitation data from 429 stations with no missing data over eastern China between 1961 and 2012, and found that the precipitation thresholds roughly increased from northwest to southeast. In addition, using the generalized Pareto distribution fitting, they found higher probabilities of extreme precipitation in South China. Wang et al. (2016) examined extreme precipitation in Guangdong Province, and discovered that in most regions the frequency and intensity indices continued to increase from 1960–2013. Liu et al. (2018) found that the intense short-term rainfall (above the rain rate of 20 mm/h) in Foshan stayed around a certain level in the April-May-June rainy season but showed an increasing trend in July-August-September. There have been several studies on extreme precipitation in relation to global warming and urbanization in recent years (Papalexiou & Montanari 2019; Jiang et al. 2020), focusing on the impact of urbanization on extreme precipitation.

Previous studies have implied that extreme precipitation is likely to increase in GBA, which is located in southern China and home to large cities, but the spatio-temporal characteristics of extreme precipitation in this region are yet to be explored. This study attempts to identify the spatio-temporal characteristics of extreme precipitation in GBA, with a vision to provide a theoretical basis for the water resources management in response to climate changes, as well as technical support for improving water resources regulation and flood risk management capabilities in the GBA.

DATA AND METHOD

Study area

The GBA (111.35°E–115.43°E, 21.56°N–24.40°N) is located in the central-southern part of Guangdong Province and the estuary area of the Pearl River Basin in China. This region is humid and the water cycle is extremely active, with dense river networks and abundant precipitation. The temporal and spatial distribution of rainfall is extremely uneven in the GBA. As one of the four largest bays in the world, GBA is home to nine cities (Shenzhen, Huizhou, Dongguan, Guangzhou, Jiangmen, Zhongshan, Zhuhai, Foshan and Zhaoqing) in the Pearl River Delta, Hong Kong and Macau. It covers an area of approximately 56,000 km2 and has a population of approximately 70 million. Figure 1 shows the location of GBA and the distribution of meteorological stations in this region. The GBA boasts high economic vitality and inclusiveness, and plays a very important role in China's social development. In 2019, The Outline of the Development Plan for the Guangdong-Hong Kong-Macao Greater Bay Area issued by China's State Council specified the goal of building this area into a world-class bay area and creating a high-quality development model. However, the rapid economic development of GBA requires greater urban flood and drainage control.

Figure 1

Distribution of the meteorological stations and precipitation grid in the GBA.

Figure 1

Distribution of the meteorological stations and precipitation grid in the GBA.

Precipitation data

The precipitation data were taken from the 0.5° × 0.5° daily grid precipitation datasets (DGPD) in China for the period 1961–2018, established by the National Meteorological Information Center, Meteorological Data Room. The locations of the grid points are shown in Figure 1. The datasets can be downloaded from The China Meteorological Data Service Center (https://data.cma.cn/). We extracted the 27 grids of the nine cities in the Pearl River Delta by using R language. Because the meteorological stations used to build the DGPD do not include Hong Kong and Macau, the Hong Kong precipitation data were taken from the Hong Kong Observatory (HKO) (https://www.hko.gov.hk) and the Macao data were from the Macao Meteorological and Geophysical Bureau (http://www.smg.gov.mo).

According to the 2,472 national meteorological stations across China, the National Meteorological Information Center established the DGPD by using the thin plate spline method and a digital elevation model for interpolation. The root mean square error (RMSE) between the grid data and the actual precipitation is 0.49 mm, and the relation coefficient amounts to 0.93. Especially in southeast China, the relative bias error is mostly between −10% and 10% (Zhao et al. 2014).

Research methods

Nine EPIs were selected in this paper and the annual average of each index was calculated. Regression analysis and the moving average method were used to analyze the temporal trend of each index, and kriging spatial interpolation was used to analyze the average value of each index.

EPIs

Due to the differences in climatic conditions and topographical conditions among regions, there is generally no uniform standard for extreme precipitation. Referring to most scholars, in this study, we selected nine EPIs (Table 1) from the extreme climate index system defined by the Expert Team on Climate Change Detection and Indices (ETCCDI), and divided them into persistence, frequency and intensity indices (Soltani et al. 2016; Talchabhadel et al. 2020).

Table 1

The EPIs applied within this paper

TypeIndicatorSymbolDefinitionUnit
Persistence Consecutive dry days CDD Maximum number of consecutive days with PRCP <1 mm 
Consecutive wet days CWD Maximum number of consecutive days PRCP ≥1 mm 
Annual total wet-day precipitation PRCPTOT Annual total precipitation in wet days (defined as PRCP ≥1 mm) mm 
Frequency Number of slightly heavy precipitation days R10 Annual count of days when PRCP ≥10 mm 
Number of heavy precipitation days R25 Annual count of days when PRCP ≥25 mm 
Intensity Simple daily intensity index SDII Annual total precipitation divided by the number of wet days in the year mm/d 
Very wet days R95p Annual total precipitation when PRCP ≥95th percentile mm 
Max 1-day precipitation amount RX1day Annual maxima of 1-day precipitation mm 
Max 5-day precipitation amount RX5day Annual maxima of 5-day precipitation mm 
TypeIndicatorSymbolDefinitionUnit
Persistence Consecutive dry days CDD Maximum number of consecutive days with PRCP <1 mm 
Consecutive wet days CWD Maximum number of consecutive days PRCP ≥1 mm 
Annual total wet-day precipitation PRCPTOT Annual total precipitation in wet days (defined as PRCP ≥1 mm) mm 
Frequency Number of slightly heavy precipitation days R10 Annual count of days when PRCP ≥10 mm 
Number of heavy precipitation days R25 Annual count of days when PRCP ≥25 mm 
Intensity Simple daily intensity index SDII Annual total precipitation divided by the number of wet days in the year mm/d 
Very wet days R95p Annual total precipitation when PRCP ≥95th percentile mm 
Max 1-day precipitation amount RX1day Annual maxima of 1-day precipitation mm 
Max 5-day precipitation amount RX5day Annual maxima of 5-day precipitation mm 

PRCP, daily precipitation.

Regression analysis method

In the regression analysis method, a mathematical model was established to describe the statistical relationship between the EPIs x and the year T as follows:
formula
(1)
where a is the constant value and b is the regression coefficient.

The sign of the b indicates the tendency of x. When b > 0, it means that with the increase of time T, x shows an upward trend, whereas when b < 0, it means that with the decrease of time T, x shows a downward trend. b × 10 is called the climate tendency rate, and its unit is mm/10a. The regression coefficient b and constant a can be estimated by the least square method.

Moving average method

The principle of the moving average method is to calculate the moving average by sequentially increasing and decreasing the new and old data, so as to eliminate accidental changes, find the development trends and make predictions. Given a random sequence of n samples, the moving average sequence can be expressed as:
formula
(2)
where k is the moving length, .

Kriging interpolation

In a random process under certain conditions, the optimal linear unbiased estimator can be obtained by kriging interpolation in geostatistics (Le & Zidek 2006). The spatial distribution changes of extreme precipitation were presented using the Kriging spatial interpolation method in the ArcGIS data processing platform.

RESULTS AND DISCUSSION

Temporal changes

Temporal changes of mean annual values of different EPIs and its mean 5-year moving values (M5YMV) in GBA for 1961–2018 are shown in Figure 2. The trends of all indices were tested for significance with a confidence level of 0.05.

Figure 2

Temporal variation of each EPI in GBA for 1961–2018.

Figure 2

Temporal variation of each EPI in GBA for 1961–2018.

In the persistence indices, CDD showed a downward trend (−0.2 d/10a), with the highest value in 2005 (61 d) and the lowest value in 2017 (21 d). There was no clear trend in the M5YMV. The CWD showed a downward trend (−0.1 d/10a), with the lowest value in 1980 (9 d) and the highest value in 2016 (26 d). The trend of the M5YMV was decreasing overall for CWD (−0.2 d/10a). CDD is slightly larger than that of CWD, indicating that the drought of the GBA has eased a little. PRCPTOT showed an upward trend (17.9 mm/10a), with the lowest value in 1963 (983 mm), and the highest value in 2016 (2,330 mm). The M5YMV PRCPTOT trend was increasing (12.4 mm/10a), indicating that the total precipitation in GBA has increased.

In the frequency indices, R10 and R25 both showed an upward trend (0.6 d/10a and 0.4 d/10a), as well as the M5YMV (0.5 d/10a and 0.3 d/10a). R10 had its lowest value in 1963 (31 d) and its highest value in 2016 (72 d), while R25 had its lowest value in 1963 (8 d) and its highest value in 2016 (27 d).

In the intensity indices, SDII, R95p, RX1day and RX5day all presented an upward trend, with trends of 0.1 mm/(d·10a), 10.5 mm/10a, 0.3 mm/10a and 3.2 mm/10a, respectively. The trends of the M5YMV were 0.1 mm/(d·10a), 8.6 mm/10a, −0.3 mm/10a and 2.4 mm/10a, respectively. RX5day had a larger amplitude than RX1day, and the frequency of extreme precipitation events has increased. SDII had the minimum value (8.8 mm/d) in 1963, and the maximum (15.6 mm/d) in 2008. R95p had the minimum value (102 mm) in 1963, and the maximum (762 mm) in 2008. The minimum RX1day (56 mm) occurred in 1963 and the maximum (141 mm) occurred in 2018. The minimum RX5day (120 mm) occurred in 1963 and the maximum (276 mm) occurred in 2008. Except for RX1day, the relative increase of intensity indices is larger than that of PRCPTOT, indicating that precipitation in the GBA is becoming more concentrated, which is not conducive to water resources management.

According to the temporal mean annual value of EPIs, the persistence, frequency and intensity of extreme precipitation in GBA generally increased. Except for CDD and CWD, the other seven EPIs showed a non-significant upward trend, indicating that the GBA slowly became wetter from 1961 to 2018 and its extreme precipitation was gradually increasing. The overall drought in the GBA had slowed down, which was consistent with the national trend, but slighter (Cao et al. 2020). The highest value of the high-intensity precipitation indices occurred after 2008, usually in 2008 and 2016, when climate anomalies happened, and Wang et al. (2016) also found that the highest value of extreme precipitation event intensity appeared in 2008. The increase of the long-lasting index RX5day was greater than that of the short-lasting index RX1day, indicating that the risk of extreme precipitation in GBA has increased in recent years, because multiple days of precipitation can easily cause floods. The ratio of the mean annual R95p to the mean annual PRCPTOT was 28.2%, indicating that nearly one third of the total annual precipitation was contributed by extreme precipitation, which was less than the corresponding value in the entire Guangdong Province.

Spatial changes

Spatial changes of mean values of the indices

The spatial distribution of mean annual values of different EPIs is shown in Figure 3. In most areas, CDD was within the range 36.5–45.2 d, with the highest values in the southeast, whereas in the north-central part of Zhaoqing, the value was below 36.4 d (Figure 3(a)). In general, CWD was within the range 13.9–18.2 d, and the values were the highest in the north and northeast GBA, whereas the values were less than 13.8 days in southeast Hong Kong and Macau (Figure 3(b)). In the majority of the GBA, PRCPTOT was around 1,310–1,810 mm, but in the southern coastal area, including Zhuhai, Hong Kong and Macau, it could exceed 1,970 mm (Figure 3(c)).

Figure 3

Spatial variation of the mean values of each EPI in the GBA for 1961–2018.

Figure 3

Spatial variation of the mean values of each EPI in the GBA for 1961–2018.

The R10 was irregularly distributed (Figure 3(d)). In general, the values in the northern and southeastern GBA were higher than 47.8 d, whereas the values in some parts of Zhaoqing were below 45.5 d. R25 (Figure 3(e)) was higher than 19.7 d in Zhuhai, Macau and Hong Kong, located in the southern coastal GBA, while the values in some parts of the west GBA were less than 14.4 d.

SDII in Hong Kong and Macau was above 18.2 mm/d, and in Zhaoqing was generally below 11.5 mm/d (Figure 3(f)). R95p was mostly within the range 407–544 mm, and in the south of Zhuhai, Hong Kong and Macau, it could exceed 614 mm, whereas it was less than 406 mm in the west of Zhaoqing (Figure 3(g)). RX1day (Figure 3(h)) was mostly within the range 72–144 mm, but it could exceed 169 mm in Hong Kong and Macau. RX5day (Figure 3(i)) was within the range 135–259 mm, but it could exceed 343 mm in Hong Kong.

Overall, the extreme precipitation in GBA decreased from the southeast coastal areas to the northwest inland areas, and CDD also showed similar changes, indicating that the precipitation concentration in the southeast coastal cities was relatively high, which could lead to more floods in the wet season and higher risk of drought in the dry season, especially in Hong Kong, Macau and Zhuhai.

Spatial changes in the EPIs trends

The spatial distribution of trends and tendencies of different EPIs is shown in Figure 4. The trends of all indices of each grid were tested for significance with a confidence level of 0.05. The variation range of CDD (Figure 4(a)) was mostly around −1.0 to 0.4 d/10a, and no changes passed the significance tests. Overall, 65.5% of the grids showed a downward trend, whereas the northwestern part of Zhaoqing, Zhuhai, and parts of Macau showed an upward trend. CWD (Figure 4(b)) varied from −1.2 to 0.6 d/10a, and 62.1% of the grids showed a downward trend. Among them, CWD in the southwestern part of Zhaoqing increased significantly, whereas it decreased significantly in the northern part. PRCPTOT varied from −6.1 to 57.8 mm/10a (Figure 4(c)), and 96.6% of the grids showed a non-significant upward trend. The areas with an obvious upward trend were mainly in Shenzhen and Hong Kong in the southeast GBA.

Figure 4

Spatial variation of the trends of each EPI in GBA for 1961–2018.

Figure 4

Spatial variation of the trends of each EPI in GBA for 1961–2018.

The variation in R10 was from −0.6 to 1.9 d/10a (Figure 4(d)), and 86.2% of the grids showed an upward trend, with significant increases in Shenzhen and Hong Kong. The variation in R25 was within −0.1 and 1.3 d/10a (Figure 4(e)), and 89.7% of the grids showed an upward trend, with significant increases in the central part of GBA, including Guangzhou and Dongguan, and in Shenzhen and Hong Kong.

The changes in SDII (Figure 4(f)) varied from 0.04 to 0.42 mm/(d·10a), and the entire study area showed an upward trend. The northern Zhaoqing, Guangzhou and Shenzhen increased significantly, and the upward trend in Hong Kong was also obvious. The changes in R95p ranged from −7.3 to 26.6 mm/10a (Figure 4(g)) and 96.6% of the grids showed a non-significant upward trend. The areas with an obvious upward trend were located in the northern part of Zhaoqing. RX1day (Figure 4(h)) was between −6.5 and 3.3 mm/10a with 62.1% of the grids showing an upward trend. The areas with an obvious upward trend were mainly in Guangzhou and other areas in the northern part of the GBA; the northwestern part of Guangzhou increased significantly, and the eastern part of Zhaoqing decreased significantly. The variation in RX5day was between −2.7 and 7.1 mm/10a (Figure 4(i)), and 79.3% of the grids showed an upward trend, with significant increases in Guangzhou and other areas of the northern part of the GBA.

The strengthening of the subtropical high pressure causes the East Asian summer monsoon to continue to enhance, and more water vapor is transported into China. Most of the EPIs in the GBA showed an upward trend, with a significant increase in some areas. The annual precipitation in northern and southeastern coastal areas showed an increase, as well as the precipitation concentration and the frequency of extreme precipitation events. According to the temporal changes, it could be seen that the increase in extreme precipitation in the GBA was mainly manifested in the southeast coastal area, where the task of urban disaster prevention and mitigation will be more difficult than in other areas. However, although the overall drought in the GBA declined, the CDD in the northwestern part of Zhaoqing showed an upward trend, while other indices mainly showed a downward trend, indicating that the drought increased in this area. In addition, except for RX1day and RX5day, the increase rate of the indices in coastal cities was greater than that of inland areas. The conclusions above are consistent with those achieved by Wang et al. (2016), Zeng et al. (2017), and Cao et al. (2020).

As urbanization continues, greenhouse gas emissions have increased. Moreover, the hardening of the underlying surface has reduced infiltration and increased evaporation. Many studies have shown that urbanization changes urban extreme precipitation (Chang et al. 2018; Paul et al. 2018; Papalexiou & Montanari 2019; Jiang et al. 2020), but the magnitude of this effect varies among cities. As a world-class urban agglomeration, the GBA is one of the most economically vigorous regions in China. As in Figure 5, from the statistical yearbook data of each city, the urbanization rate of the GBA was above 47.7% at the end of 2018, and some cities were even close to 100%. The high urbanization and intensive human activities will inevitably influence the temporal and spatial distribution of extreme precipitation within the region. According to this paper, the extreme precipitation in most parts of the GBA was slowly increasing, and the coastal areas such as Hong Kong, Macau, Zhuhai and the southern part of Shenzhen were increasing more rapidly. These areas, where urbanization rates are more than 90%, play an important role in the development of the GBA, therefore flood prevention should not be underestimated. For example, monitoring and early warning of extreme weather, analyzing the causes and mechanisms of extreme precipitation events, and taking effective countermeasures to reduce the losses caused by disasters. In addition, although the urbanization rate of Zhaoqing was the lowest, there was large room for development. According to the results of this paper, Zhaoqing may have a trend of increasing drought, so attention should be paid to the reasonable allocation of water resources, including promoting water saving measures and optimizing water utilization structure.

Figure 5

The spatial distribution of urbanization rate in the GBA.

Figure 5

The spatial distribution of urbanization rate in the GBA.

This paper lacks the further discussion of factors that affect extreme precipitation. In addition to climate warming and urbanization, there are other factors in connection with extreme precipitation, such as El Niño, and the relationship between them needs to be further studied, using data such as temperature, underlying surface, time of occurrence of El Niño, to obtain more substantial results by using cycle analysis, mutation analysis and global climate mode (Chang et al. 2018; Paul et al. 2018; Yu et al. 2018; Sun et al. 2021; Talchabhadel et al. 2020). Further discussions on Chinese anomalies in 2008 and 2016 are also necessary. Besides, when there is a small range of heavy precipitation, the precipitation intensity of grid data is likely to weaken (Zhao et al. 2014). Therefore, it is necessary to conduct comprehensive inter-comparison and analysis of observation data from the meteorological stations and grid datasets.

Correlation of EPIs

The principal component analysis method was used to analyze the EPIs. Two consistent components with an eigenvalue variance greater than 1 were selected as the principal component. The trend of PRCPTOT and each EPI was shown to have good consistency (Table 2).

Table 2

Principal component load and explanatory variance of EPIs in GBA for 1961–2018

ComponentPRCPTOTCDDCWDR10R25R95pSDIIRX1dayRX5dayVariance (%)
0.956 −0.185 0.432 0.875 0.930 0.944 0.877 0.675 0.781 61.016 
−0.207 0.595 −0.610 −0.353 −0.123 0.180 0.220 0.484 0.390 15.290 
ComponentPRCPTOTCDDCWDR10R25R95pSDIIRX1dayRX5dayVariance (%)
0.956 −0.185 0.432 0.875 0.930 0.944 0.877 0.675 0.781 61.016 
−0.207 0.595 −0.610 −0.353 −0.123 0.180 0.220 0.484 0.390 15.290 

Correlation analysis among the EPIs showed a good correlation between each EPI and PRCPTOT (Table 3), except for CDD which was significantly correlated at the 0.05 level, the others were significantly correlated at the 0.01 level. The correlation coefficients between the two persistence indices (CDD and CWD) and the four intensity indices were relatively small, indicating that there was a weak influence among them. Except for CDD and CWD, the correlations between the EPIs was relatively high, all passing the 0.01 significance test. Therefore, the nine EPIs selected in this paper have a good indicator effect for temporal and spatial variations of extreme precipitation in the GBA.

Table 3

Correlation analysis for EPIs in the GBA between 1961 and 2018

PRCPTOTCDDCWDR10R25R95pSDIIRX1dayRX5day
PRCPTOT 1.000         
CDD −0.239* 1.000        
CWD 0.497** −0.175 1.000       
R10 0.960** −0.203 0.573** 1.000      
R25 0.940** −0.185 0.345** 0.892** 1.000     
R95p 0.864** −0.179 0.213 0.717** 0.865** 1.000    
SDII 0.754** −0.057 0.168 0.668** 0.836** 0.887** 1.000   
RX1day 0.527** −0.017 0.108 0.351** 0.427** 0.736** 0.584** 1.000  
RX5day 0.627** 0.068 0.240* 0.514** 0.607** 0.746** 0.695** 0.724** 1.000 
PRCPTOTCDDCWDR10R25R95pSDIIRX1dayRX5day
PRCPTOT 1.000         
CDD −0.239* 1.000        
CWD 0.497** −0.175 1.000       
R10 0.960** −0.203 0.573** 1.000      
R25 0.940** −0.185 0.345** 0.892** 1.000     
R95p 0.864** −0.179 0.213 0.717** 0.865** 1.000    
SDII 0.754** −0.057 0.168 0.668** 0.836** 0.887** 1.000   
RX1day 0.527** −0.017 0.108 0.351** 0.427** 0.736** 0.584** 1.000  
RX5day 0.627** 0.068 0.240* 0.514** 0.607** 0.746** 0.695** 0.724** 1.000 

Note: * and ** represent a significant correlation at the 0.05 level and the 0.01 level, respectively.

CONCLUSIONS

Based on the daily grid precipitation datasets of GBA for 1961–2018 and the precipitation data for Hong Kong and Macau, we analyzed the temporal and spatial changes of extreme precipitation in GBA with the nine EPIs recommended by ETCCDI.

The temporal changes of EPIs in GBA were generally gentle, with a generally wetting tendency overall. Among the persistence indices, CDD and CWD showed a downward trend, whereas PRCPTOT showed an upward trend. The frequency indices (R10 and R25) and the intensity index (R95p, SDII, RX1day and RX5day) both showed an upward trend.

The mean multi-year spatial distribution of the EPIs in the GBA increased from west to east and north to south. Except for CWD, the EPIs showed an upward trend in many areas, and coastal regions increased significantly. There were clear upward trends of PRCPTOT, R10, R25, and SDII, which usually occurred in Guangzhou, Dongguan, Shenzhen and Hong Kong. SDII had an increasing tendency throughout the GBA. The increasing areas of R95p, RX1day and RX5day were mainly in the northern part of the GBA. The increasing trend of extreme precipitation was greater in regions with more developed economies and denser populations.

Principal component analysis and regression analysis showed that PRCPTOT had a good consistency with each EPI. The nine EPIs selected in this paper therefore were good indicators of precipitation in the GBA.

While vigorously developing the economy, attracting talent and investment, the GBA should devote a certain budget to early warning and forecasting of extreme precipitation as well as to responding to disasters that may be caused by extreme precipitation, such as heavy rains, floods and continuous droughts, so as to reduce the economic losses. Subsequently, we can truly create a high-quality life circle that is suitable for living, business and traveling. The findings of this study provide a theoretical basis for GBA to formulate a water resource management strategy to address climate change, and provide technical support for improving the regional water resources regulation and control capabilities as well as the ability to reduce the impacts of droughts and floods.

ACKNOWLEDGEMENTS

This study was supported by the Scientific Research Project approved by the Department of Education of Guangdong Province (No. 2020KQNCX125) and the Scientific Research Foundation of Xinhua College of Sun Yat-sen University (No. 2019KYYB08).

DATA AVAILABILITY STATEMENT

All relevant data are included in the paper or its Supplementary Information.

REFERENCES

Bennett
K. E.
&
Walsh
J. E.
2015
Spatial and temporal changes in indices of extreme precipitation and temperature for Alaska
.
International Journal of Climatology
35
(
7
),
1434
1452
.
doi:10.1002/joc.4067
.
Cao
Q.
,
Hao
Z.
,
Fu
X.
,
Hao
J.
&
Lu
C.
2020
Analysis of spatial-temporal changes of extreme climate elements in China from 1960 to 2017
.
Yellow River
42
(
2
),
11
17
.
doi:10.3969/j.issn.1000-1379.2020.02.003
.
Chang
X.
,
Xu
Z.
,
Zhao
G.
,
Cheng
T.
&
Song
S.
2018
Spatial and temporal variations of precipitation during 1979–2015 in Jinan City, China
.
Journal of Water and Climate Change
9
(
3
),
540
554
.
doi:10.2166/wcc.2017.029
.
Hadri
A.
,
Saidi
M. E. M.
,
Saouabe
T.
&
El Alaoui El Fels
A.
2020
Temporal trends in extreme temperature and precipitation events in an arid area: case of Chichaoua Mejjate region (Morocco)
.
Journal of Water and Climate Change
1
21
.
doi:10.2166/wcc.2020.234
.
Jiang
X.
,
Luo
Y.
,
Zhang
D. L.
&
Wu
M.
2020
Urbanization enhanced summertime extreme hourly precipitation over the Yangtze River Delta
.
Journal of Climate
33
(
13
),
5809
5826
.
doi:10.1175/jcli-d-19-0084.1
.
Le
N. D.
&
Zidek
J. V.
2006
Statistical Analysis of Environmental Space-Time Processes[M]
.
Springer Science & Business Media, Berlin
, pp.
101
134
.
Li
M.
,
Han
T.
&
Hao
X.
2020
Reginonal characteristics of extreme precipitation events in Eurasia
.
Transactions of Atmospheric Sciences
43
(
4
),
687
698
.
doi:10.13878/j.cnki.dqkkxxb.20191202001
.
Liu
S.
,
Lei
Y.
,
Li
Y.
&
Zhang
Y.
2018
Analysis of the climatological characteristics of short-term intense rainfall of Foshan City and Its causal factors
.
Guangdong Meteorology
40
(
4
),
24
27
.
doi:10.3969/j.issn.1007-6190.2018.04.006
.
Myhre
G.
,
Alterskjær
K.
,
Stjern
C. W.
,
Hodnebrog
Ø.
,
Marelle
L.
,
Samset
B. H.
,
Sillmann
J.
,
Schaller
N.
,
Fischer
E.
,
Schulz
M.
&
Stohl
A.
2019
Frequency of extreme precipitation increases extensively with event rareness under global warming
.
Scientific Reports
9
(
1
),
345
357
.
doi:10.1038/s41598-019-52277-4
.
Papalexiou
S. M.
&
Montanari
A.
2019
Global and regional increase of precipitation extremes under global warming
.
Water Resources Research
55
(
6
),
4901
4914
.
doi:10.1029/2018WR024067
.
Paul
S.
,
Ghosh
S.
,
Mathew
M.
,
Devanand
A.
,
Karmakar
S.
&
Niyogi
D.
2018
Increased spatial variability and intensification of extreme monsoon rainfall due to urbanization
.
Scientific Reports
8
(
1
),
1
10
.
doi:10.1038/S41598-018-22322-9
.
Soltani
M.
,
Laux
P.
,
Kunstmann
H.
,
Stan
K.
,
Sohrabi
M. M.
,
Molanejad
M.
,
Sabziparvar
A. A.
,
Ranjbar SaadatAbadi
A.
,
Ranjbar
F.
,
Rousta
I.
,
Zawar-Reza
P.
,
Khoshakhlagh
F.
,
Soltanzadeh
I.
,
Babu
C. A.
,
Azizi
G. H.
&
Martin
M. V.
2016
Assessment of climate variations in temperature and precipitation extreme events over Iran
.
Theoretical and Applied Climatology
126
(
3/4
),
775
795
.
doi:10.1007/s00704-015-1609-5
.
Sun
Q. H.
,
Zhang
X. B.
,
Zwiers
F.
,
Westra
S.
&
Alexander
L. V.
2021
A global, continental, and regional analysis of changes in extreme precipitation
.
Journal of Climate
34
(
1
),
243
258
.
doi:10.1175/JCLI-D-19-0892.1
.
Talchabhadel
R.
,
Aryal
A.
,
Kawaike
K.
,
Yamanoi
K.
&
Nakagawa
H.
2020
A comprehensive analysis of projected changes of extreme precipitation indices in West Rapti River basin, Nepal under changing climate
.
International Journal of Climatology
1
19
.
doi:10.1002/joc.6866
.
Tian
Q. Y.
,
Li
Z. L.
&
Sun
X. L.
2020
Frequency analysis of precipitation extremes under a changing climate: a case study in Heihe River basin, China
.
Journal of Water and Climate Change
1
15
.
doi:10.2166/wcc.2020.170
.
Wang
W.
,
Yan
J.
&
Liu
Y.
2016
Variation characteristics of extreme precipitation events in Guangdong province
.
Bulletin of Soil and Water Conservation
36
(
2
),
293
299
.
doi:10.13961/j.cnki.stbctb
.
Yu
J.
,
Qi
M.
,
Sun
Q.
&
Tao
L.
2018
Statistical characteristics of summer extreme rainfall over eastern China and its relation with EL Niño
.
Trans Atmos Sci
41
(
1
),
77
84
.
doi:10.13878/j.cnki.dqkxxh.20151027001
.
Zeng
Y.
,
Tan
Y.
,
Qiu
X.
&
Liu
C.
2017
Analysis on the tendency of different precipitation in Ten major watersheds in China
.
Jiangsu Agricultural Sciences
45
(
14
),
189
195
.
doi:10.15889/j.issn.1002-1302.2017.14.052
.
Zhao
Y.
,
Zhu
J.
&
Xu
Y.
2014
Establishment and assessment of the grid precipitation datasets in China for the past 50 years
.
Journal of the Meteorological Sciences
34
(
04
),
414
420
.
doi:10.3969/2013jms.0008
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).