Projecting Changes in Rainfall Extremes for the Huai River Basin in the Context of 1.5 ◦ C and 2 ◦ C Global Warming

: It is indisputable that global warming has triggered more frequent extreme weather and in turn led to severe ﬂood disasters. To understand the trend of extreme rainfall under 1.5 ◦ C and 2 ◦ C warming, we investigated the historical variation and future trends in extreme rainfall for the Huai River basin, which has frequently been hit by ﬂoods, using recorded meteorological data and a projection of ﬁve General Circulation Models in the Coupled Model Intercomparison Project 6. We used the years 1995–2014 as the baseline period to study the temporal and spatial changes in extreme rainfall under 1.5 ◦ C and 2.0 ◦ C warming scenarios. The results indicated that (1) temperatures in the Huai River basin have risen signiﬁcantly from 1995 to 2014, but there are insigniﬁcant variation trends in annual precipitation (AP), intensive precipitation (R95P), maximum daily precipitation (Rx1d) and heavy rain days (Rr50) during the same time span. (2) From 2015 to 2100, both temperature and extreme rainfall indices show increase trends, with a higher rate of increase under a higher emission scenario. (3) Under the warming scenario of 1.5 ◦ C, AP, R95P, Rx1d and Rr50 in the basin will likely increase by 4.6%, 5.7%, 6.2% and 13.4%, respectively, compared with that in the baseline period. Under the warming scenario of 2.0 ◦ C, AP, R95P, Rx1d and Rr50 will probably increase by 7.3%, 7.4%, 10.9% and 19.0%, respectively. (4) Spatially, the changes in extreme rainfall indices under the warming scenarios of 1.5 ◦ C and 2.0 ◦ C generally tend to increase from north to south. Higher intensity extreme rainfall will likely extend to the whole of the Huai River basin. It is therefore essential to study adaptive measures to cope with ﬂooding in the Huai River basin induced by the increase in future rainfall extremes.


Introduction
To some extent, extreme weather has an effect on socioeconomics [1]. As the planet has warmed, the frequency and intensity of extreme climatic events have been largely on the rise since the middle of the 20th century [2][3][4]. Extreme rainfall has been one of the main hazard sources, usually resulting in severe basin floods; moreover, it has a major impact on socioeconomic development [5]. During 17-23 July 2021, an extremely heavy rainstorm occurred in the Henan province with the storm center in the city of Zhengzhou. Over 200 mm fell in one 24-h period at Zhengzhou station, leading to huge economic loss and many deaths [6]. According to the annual report issued in 2022 by the Ministry of Emergency Management (MEM) of China, high-intensity rainfall over China in 2021 was 6% above the multiple-year average from 1961 to 2021. The direct economic losses induced by extreme rainfall amounted to about 245.89-billion-yuan, accounting for more than 70% of the total direct economic losses induced by natural hazards [7]. Climatic warming have been attracted broad attentions from not only government officials but also scientific communities [30][31][32].
The Huai River basin is one of the seven major rivers in China, situated in the transitional climate zone from arid climate in North China to humid climate in South China, where it is highly influenced by the Eastern Asia Monsoon. The river basin is densely populated, with a population density of over 600 persons per square kilometers, which is approximately 5 times the China average. Meanwhile, The Huai River supports over 17% of crop productivity of China. In recent decades, floods induced by extreme rainfall have been increasing in the Huai River basin, severely threatening food security in China and the safety of the local people. Although there are numerous studies available in the literature on climate change impact to the Huai River basin [33][34][35], previous studies mainly focus on water resources, with the used climate scenarios being issued in IPCC-TAR (Third Assessment Report), IPCC-AR4 (Fourth Assessment Report), and IPCC-AR5 (Fifth Assessment Report). There are only a few studies on the change in rainfall extremes under the newly issued SSPs scenarios, which can directly link with local communities. It is therefore critical and essential to understand variation trends of rainfall extremes in the Huai River basin under the newly issued scenarios for the planning of effective flood control of the Huai River. With multiple CMIP6-GCM projections, the main objectives of the study are to investigate changes in extreme rainfall for the basin under 1.5 and 2 • C global warming. The expected results could provide scientific support for the planning of flood control and adaptation to climate change in the Huai River. The paper is presented as follows: Section 2 introduces the study catchment and the methods used; Section 3 presents the results in detail; Section 4 discusses the findings of the study, followed by a summary and conclusions in Section 5.

Study Catchment
The Huai River basin is located in eastern China, between the Yangtze River basin and the Yellow River basin, with a drainage area of approximately 270,000 km 2 . The Huai River flows from west to east and runs across Henan, Anhui, Jiangsu and Shandong provinces, with a temperate monsoon climate in the northern part and subtropical monsoon climate in the southern part. The annual mean temperature ranges from 11 • C to 16 • C, decreasing spatially from south to north. The monthly average temperature peaks in July at about 25 • C and reaches a low of below 0 • C in January. The multiple-year average annual precipitation is about 920 mm and decreases from north to south. The precipitation in summer accounts for more than half of the annual precipitation.

Data Sources
Daily precipitation and temperature readings from 1995 to 2014 gauged at 177 meteorological stations were collected to analyze the historical variation in climatic variables. The major river system and locations of rain gauges in the basin are shown in Figure 1.
There are tens of GCMs available in CIMP5 and CMIP6. Previous studies indicate that five GCMs (CanESM5, CNRM-ESM2-1, IPSL-CM6A-LR, MIROC6 and MRI-ESM2-0) are reasonably suitable for China, especially for the eastern part [28,32]. Thus, projections of the five GCMs from CMIP6 were employed to investigate future trends of temperature and rainfall. The time spans of historical simulation and future projection under the multiple future emission scenarios of the five GCMs are 1995-2014 and 2015-2100, respectively. The general information of the five GCMs is shown in Table 1. In this study, the GCMs projections under the four Shared Socioeconomic Pathways (SSPs) scenarios were used to investigate future changes in climatic variables. SSPs is a newly launched scenario for climate change research that combines pathways of future radiative forcing and their associated climate changes with alternative pathways of socioeconomic development. The four adopted scenarios of SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5 represent the low  There are tens of GCMs available in CIMP5 and CMIP6. Previous studies indicate that five GCMs (CanESM5, CNRM-ESM2-1, IPSL-CM6A-LR, MIROC6 and MRI-ESM2-0) are reasonably suitable for China, especially for the eastern part [28,32]. Thus, projections of the five GCMs from CMIP6 were employed to investigate future trends of temperature and rainfall. The time spans of historical simulation and future projection under the multiple future emission scenarios of the five GCMs are 1995-2014 and 2015-2100, respectively. The general information of the five GCMs is shown in Table 1. In this study, the GCMs projections under the four Shared Socioeconomic Pathways (SSPs) scenarios were used to investigate future changes in climatic variables. SSPs is a newly launched scenario for climate change research that combines pathways of future radiative forcing and their associated climate changes with alternative pathways of socioeconomic development. The four adopted scenarios of SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5 represent the low forcing scenario, medium radiative forcing scenario, medium and high radiative forcing scenario and high forcing scenario, respectively.

CanESM5
Canadian Center for Climate Modeling and Analysis, Canada Japan Agency for Marine-Earth Science and Technology, Japan

Mann-Kendall Test
The Mann-Kendall nonparametric trend test (M-K method) is a commonly used method to detect the variation trend of a time series [36,37]. The method is based on an assumption of stability of a time series and does not require the samples to follow a certain statistical distribution. It is therefore suitable for hydro-meteorological series and has been widely used in climate sciences. Suppose {x 1 , x 2 , x 3 , . . . , x n } is an independent random time series, then the M-K statistic UF can be defined and calculated as follows: where, n is the length of the time series, Sgn is a symbolic function, x i and x j are the measured values at the year i and j, respectively. The statistic UF approximately follows a standard normal distribution. A positive value indicates an increasing trend, and a negative value indicates a decreasing trend. Given a significance level α, if |UF| > |Uα|, then a significant variation trend is detectable. UF is ± 1.96 at the significance level of 0.05.

Theil Sen Slope Calculation
The Theil Sen Slope is a nonparametric estimation method based on median estimation of a series trend. This method calculates the Kendall slope value β to determine the variation trend of a time series [37]. The positive β indicates an increasing trend, while the negative β indicates a decreasing trend. A higher absolute value of β means a greater increasing or decreasing trend of a series.
where, Median represents median function, x i and x j represent series sample values at i and j, respectively. The i and j are the order number of the series with 1 < i < j < n.

Extreme Rainfall Index
Drawing on the extreme climate indices recommended by the World Meteorological Organization, four extreme rainfall indices were selected to measure the climatic extremes of the Huai River basin, covering annual precipitation (AP), intensive precipitation (R95P), maximum daily precipitation (Rx1d) and heavy rain days (Rr50). These four indices reflect extreme rainfall situations at different time scales and have been widely applied to China, Asia, and throughout the world [38][39][40][41]. The definition of the four extreme rainfall indices is given in Table 2. AP indicates the climatic situation at an annual scale, while the other three indices closely link to regional floods.

Name
Definition Unit AP Sum of annual daily precipitation mm R95P Total precipitation with daily precipitation ≥ 95% quantile value mm Rx1d Maximum daily precipitation in a year mm Rr50 Days with daily precipitation ≥ 50 mm in a year day

Historical Variation and Future Trend of Temperature
Using the recorded temperature in the Huai River basin and the projected temperatures under different future climate scenarios, the basin average annual temperatures from 1995 to 2100 are shown in Figure 2. Trend rates and M-K statistics of the recorded temperature series from 1995 to 2014 and the ensemble mean temperature series from 2015 to 2100 were estimated by the Theil Sen Slope estimation and Mann-Kendall test (Table 3).  Total precipitation with daily precipitation ≥95% quantile value mm Rx1d Maximum daily precipitation in a year mm Rr50 Days with daily precipitation ≥ 50 mm in a year day

Historical Variation and Future Trend of Temperature
Using the recorded temperature in the Huai River basin and the projected temperatures under different future climate scenarios, the basin average annual temperatures from 1995 to 2100 are shown in Figure 2. Trend rates and M-K statistics of the recorded temperature series from 1995 to 2014 and the ensemble mean temperature series from 2015 to 2100 were estimated by the Theil Sen Slope estimation and Mann-Kendall test (Table  3).

Scenarios
Time Span Tendency Rate (°C/10a) M-K Statistics higher than those in the baseline period of 1995-2014.

Historical Variation and Future Trends of Rainfall Extremes
The four extreme rainfall indices were calculated using the recorded daily rainfall and the downscaled daily rainfall based on the five GCM projections. Historical and future variations in extreme rainfall indices in the Huai River basin are shown in Figures 3-6. The trend diagnosis results are given in Table 4.   higher than those in the baseline period of 1995-2014.

Historical Variation and Future Trends of Rainfall Extremes
The four extreme rainfall indices were calculated using the recorded daily rainfall and the downscaled daily rainfall based on the five GCM projections. Historical and future variations in extreme rainfall indices in the Huai River basin are shown in Figures 3-6. The trend diagnosis results are given in Table 4.        (1) during 1995-2014, the four extreme rainfall indices derived from CMIP6 multi-GCM simulation data fit the recorded data well, while the observed values of Rx1d and Rr50 are slightly larger than the simulated values for some years. (2) The four extreme rainfall indices present an insignificant decreasing trend over 1995-2014. The trends of the four extreme rainfall indices derived from CMIP6 simulations are generally consistent with that derived from the observed data, with the exception of AP. The   Figure 2 and Table 3 show that (1) the recorded temperature falls in the range of CMIP6 multi-model simulations and matches well with the ensemble mean of multiple GCM simulations. Statistical results indicate that the average simulated and recorded temperatures are 14.77 • C and 14.70 • C, respectively, with a relative error of 2.7%. It shows that the CMIP6 multi-model has a good performance on the temperature simulation of the Huai River basin. (2) During the period of 1995-2014, the observed and simulated temperature presented a rising trend of 0.20 • C/10a and 0.26 • C/10a, respectively. The M-K values of the simulated and observed annual temperature series are 3.7 and 2.04, which are higher than the threshold value of 1.96 at the confidence level of 0.05, indicating a significant rising trend in annual temperature. (3) From 2015 to 2100, the annual temperature exhibited significant rising trends for all four scenarios, with linear increasing rates of 0.14 • C/10a, 0.30 • C/10a, 0.50 • C/10a and 0.67 • C/10a under the SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5 scenarios, respectively. It can also be found that the annual temperature under the SSP1-2.6 scenario will reach a peak of around 2060 and then stabilize around 15-17 • C. The temperature trends under the SSP3-7.0 and SSP5-8.5 scenarios are much higher than those in the baseline period of 1995-2014.

Historical Variation and Future Trends of Rainfall Extremes
The four extreme rainfall indices were calculated using the recorded daily rainfall and the downscaled daily rainfall based on the five GCM projections. Historical and future variations in extreme rainfall indices in the Huai River basin are shown in Figures 3-6. The trend diagnosis results are given in Table 4.  In general, the multiple CMIP6 GCM simulations match well with the recorded extreme rainfall in the Huai River basin. In the historical period of 1995-2014, the extreme rainfall indices have insignificant decreasing trends. In the period of 2015-2100, the extreme rainfall indices show a significant increasing trend with the exception of Rx1d under SSP1-2.6 and SSP3-7.0. The growth rate is higher in higher emission scenarios, e.g., SSP5-8.5 than that in the lower emission scenarios, e.g., SSP1-2.6. As rainfall extremes are main drivers of floods in the Huai River basin, increases in the projected rainfall extremes mean flood issues in the Huai River will likely be aggravated in the context of global warming, particularly under the higher emission scenarios. It is essential to find an adaptive solution for guarantying the flood security of the Huai River basin.  Table 5.   Table 5 shows that the detected target global warming spans are associated with GCMs and emission scenarios. The likely time spans of 1.5 • C and 2.0 • C global warming are in the 2030s and 2050s. Under the higher emission scenario, e.g., SSP5-8.5, the time span of target global warming will occur several years earlier than that under the lower emission scenarios. It is also found that global temperature rises were projected by CNRM-ESM2-1, MIROC6 and MRI-ESM2-0 to be under 2 • C if the SSP1-2.6 emission scenario occurred, which means the goal of limiting global warming in the Paris Agreement will probably be reached under the lower emission scenario of SSP1-2.6.       Figure 9 shows the spatial distributions of changes in extreme rainfall indices relative to the baseline of 1995-2014, based on the ensemble mean of four extreme rainfall indices calculated by five GCM projections under the scenarios of 1.5 °C and 2.0 °C global warming.

Spatial Patterns of Changes in Extreme Rainfall Indices
Atmosphere 2022, 13, x FOR PEER REVIEW 12 of 18 The AP over the Huai River basin decreases roughly from south to north. The AP in the south of Anhui is more than 1400 mm, while that in the north of Shandong is about 600 mm. Figure 9a shows that (1)   The AP over the Huai River basin decreases roughly from south to north. The AP in the south of Anhui is more than 1400 mm, while that in the north of Shandong is about 600 mm. Figure 9a shows that (1) compared with the baseline of 1995-2014, changes in AP under the 2.0 • C warming scenario are higher than those under the 1.5 • C warming scenario over the Huai River basin. The spatial patterns of changes in AP under the 1.5 • C and 2.0 • C warming scenarios are similar. (2) Under the 1.5 • C warming scenario, the eastern part of the Huai River basin shows the lowest change in AP of about 20 mm, and changes in AP in Shandong province and northern part of Anhui province are higher than those in other areas. (3) Under the 2.0 • C warming scenario, changes in AP tend to decrease from west to east, with the highest change of over 80 mm occurring in the west of Henan province and the lowest change in AP of 10-20 mm in the east of the Huai River basin.
Throughout the basin, the R95P decreases from southeast to northeast, with the largest value of about 500 mm in the east of Jiangsu and the south of Anhui province, and the lowest value of about 300 mm in the north of Henan and the west of Shandong. It can be seen from Figure 9b that (1) under the 1.5 • C warming scenario, the changes in R95P decrease from north to south and from west to east, with the largest changes of more than 30 mm occurring in the northern part of the Huai River basin and the lowest changes of less than 10 mm in the eastern part of the basin. During the past period, storm days in the south and southeast of the Huai River basin are higher than that in the other areas of the basin, indicating a higher flood pressure in this area. However, under global warming, the increase in storm days in the northern part of the Huai River basin is higher than the other areas, which indicates that north Jiangsu and south Shandong in the Huai River basin will probably face a higher pressure of flood threat than before. The flood threat will likely cover the whole Huai River basin under 1.5 and 2.0 • C global warming. Flood control will not only focus on the local scale, but also extend to the whole Huai River basin.

Extreme Rainfall Indices
The Huai River basin is situated in a transitional climatic zone, from the arid climate of North China to the humid climate of South China, in which the climate is highly influenced by Asian Monsoon. More than 70% of precipitation is concentrated in the flood season from June to September, with the most of high intensity rainstorms usually occurring in this period. As a result, the river basin is frequently hit by extraordinary floods induced by extreme rainstorms [42], thus leading to huge economic losses. Changes in extreme rainfall have therefore attracted attention from the central government and local communities.
According to the definition of extreme rainfall (https://public.wmo.int/en/resources/ standards-technical-regulations, accessed on 23 January 2022), there are many extreme rainfall indices used in the literature [43][44][45][46]. However, annual precipitation, total precipitation with daily precipitation ≥ 95% quantile value, maximum daily precipitation in a year, and rain days with daily precipitation ≥ 50 mm are the most effective indices linked with river basin floods directly. These indices have been widely applied to investigate changes in rainfall extremes in China, as well as in the other countries of the world [8][9][10][11]24,47]. We therefore employed the four extreme rainfall indices to detect the variation trend of climatic extremes in the Huai River basin.

GCMs Projections and Uncertainty
In order to investigate future climate change and its potential effects, GCMs are believed to be the most powerful tool to project the future climate [17,18]. However, not all GCMs can be used for a specific area because different GCMs have different regional suitability [19,48]. It is of critical importance to select suitable GCMs to study river basins.  and Su et al. (2020) evaluated the suitability of CMIP5-GCMs and CMIP6-GCMs, and ranked the performance of multiple GCMs that have been applied to China [28,32]. The evaluation results suggested that the GCMs used in this study are more suitable to east of China than other GCMs. Even though the selected GCMs have good suitability to the Huai River basin, Uncertainty is still a big issue in climate change studies as different GCMs project different results, particularly for rainfall [49][50][51].
IPCC issued three scenarios of SRES (Special Report on Emissions Scenarios), RCPs (Representative Concentration Pathways), and SSPs (Shared Socioeconomic Pathways) which have been used for climate projection [52,53]. Most of previous studies of climate change and its impact mainly focus on future variation trends of temperature, precipitation and water resources [54][55][56]. Both this study and previous studies indicate temperature over the Huai River will continue to rise significantly, while precipitation was projected to increase for most of GCMs. Moreover, there exists a difference in the magnitude of projections of temperature and precipitation, even if the same GCMs were adopted. The use of multiple GCMs projections rather a single one is encouraged for the purpose of eliminating uncertainty in climate change studies [40,57]. Quartile analysis and ensemble mean could provide much more useful information for decision making based on multiple projection members [58,59] and was therefore adopted in this study.

Historical and Future Trend in Extreme Rainfall Indices
IPCC-reported extreme events will likely increase with global warming [2,3,39]. Increases in extreme rainfall have also been observed in most areas of China [8,11,12]. The observation of rainfall and its extremes in the Huai River was also detected to increase in most previous studies [10,42,60]. However, variation in extreme rainfall indices was detected to decrease insignificantly in this study, although temperature maintained a constant rising trend in a same time span. Compared with previous studies, tendency detection of extreme rainfall indices in this study is based on a data series from 1995 to 2014, which is much shorter than that in previous studies. The short data series used is the reason for the difference in trend detection results. Moreover, we compared the mean value of extreme rainfall indices from 1995 to 2014 with those in previous studies, in which a longer data series starting from 1960 was used [60,61], and found the mean values in this study are higher than that in previous studies, which could illustrate a detectable increase in extreme rainfall indices if a longer data series is adopted, and imply consistency in this study and previous studies.
Most previous studies on future extreme rainfall projections were based on scenarios of SRES and RCPs [21,33,34,62], and the projected extreme rainfall will likely increase in the coming decades, which is consistent with this study, although there exists a difference in magnitude of rainfall change. Global warming by 1.5 • C and 2.0 • C are two targets of global communities to limit global temperature rise. Although there are numerous studies on 1.5 • C and 2.0 • C global warming and its impacts [28,[30][31][32], there are few studies that focus on the change in extreme rainfall over the Huai River basin under 1.5 • C and 2.0 • C warming. Our study shows that extreme rainfall would likely experience a certain increase, even if the increase in global surface temperature was limited to below 1.5 • C and 2.0 • C. Moreover, extreme rainfall will likely extend to a larger scale area, even to the whole Huai River basin. To cope with such a situation, particularly flooding induced by changes in extreme rainfall, both structural measures (e.g., reservoirs and detention zone construction) and non-structural measures (e.g., building early warning and flood forecasting systems) have been suggested for use in the planning of flood control and Huai River basin management.

Conclusions
This study analyzed the historical variation of four extreme rainfall indices during the baseline period of 1995-2014 and investigated the changes in these indices under 1.5 • C and 2.0 • C, relative to that in baseline. We concluded that: (1) During the baseline period, temperature over the Huai River basin presented a significant rising trend, while the four extreme rainfall indices exhibited insignificant decreasing trends. From 2015 to 2100, temperature will likely continue to rise significantly, and the extreme rainfall indices were projected to increase significantly under all four future climates scenarios, with the exception of Rx1d under scenarios SSP1-2.6 and SSP3-7.0. (2) The time spans of 1.5 • C and 2.0 • C global warming will likely occur in the 2030s and 2050s. In comparison with the baseline period, AP, R95P, Rx1d and Rr50 over the Huai River basin will likely increase by 4.6%, 5.7%, 6.2% and 13.4% under the warming scenario of 1.5 • C, while they will probably increase by 7.3%, 7.4%, 10.9% and 19.0%, respectively, under 2.0 • C global warming. (3) The spatial patterns of changes in extreme rainfall i under the 1.5 • C and 2.0 • C warming scenarios are generally similar. For all four extreme rainfall indices, changes tend to decrease from north to south and from west to east, which means flood risk induced by global warming will probably extend to the whole Huai River basin.
Global warming is evident during the last few decades. Although the great efforts made globally might mitigate climate change to some extent, global surface temperature, as well as climate extremes, will likely continue to increase in the coming decades. It is therefore essential to adapt to climate change, particularly for areas sensitive and vulnerable to climate change, such as the Huai River basin.