1. Introduction
Floods are the most frequent and deadliest natural disasters in the world (Chen et al. 2021). They threaten human society in various ways, such as causing loss of life, destroying property, and impeding productive activities (Zaalberg et al. 2009; Tsakiris 2014; Bamberg et al. 2017). Historical statistics indicate that flood disasters in China are more severe than the world average (Zhang et al. 2002). Furthermore, flood risks are not evenly distributed in China. The lower reaches of the “seven big rivers,” which contain 50% of China’s population and contribute 70% to its industry and agriculture, are more prone to flooding (Zhang et al. 2002) than other regions. Among these areas, the Yangtze River valley (YRV) is the largest river basin (Chen et al. 2021) and has experienced many destructive flood events, such as those of 1870, 1931, 1954, 1998, and 2020.
The 1931 Yangtze River flood, also referred to as the 1931 China flood or the 1931 Yangtze–Huai flood, is regarded as one of the world’s deadliest natural disasters on record (Courtney 2018). According to historical records, a disastrous flood swept the Yangtze–Huai River basin from June to August 1931 (Li et al. 1994). The flood inundated approximately 180 000 km2, affected 25 million people, and claimed over 2 million lives according to the official report (National Flood Relief Commission 1933). Among the causes of deaths, approximately 150 000 people were drowned during the flood and many more failed to survive the subsequent famines and diseases (Buck 1932). The casualties of this individual event alone are nearly equivalent to the global mortality from natural disasters between 1974 and 2003 (Guha-Sapir et al. 2004). During this flood event, eight provinces in central-eastern China, including Hubei, Hunan, Jiangsu, Anhui, Jiangxi, Zhejiang, Henan, and Shandong, suffered most from heavy rainfall. Other regions, such as Northeast China in the north, Guangdong in the south, and Sichuan in the west, were also partly affected (Courtney 2018). All this evidence indicates that the 1931 Yangtze River flood covered a large area and was a local result of widespread flooding.
Despite its enormous social impacts, the causes of the 1931 Yangtze River flood have rarely been investigated, which is partly due to the limitation of historical records and meteorological data before the 1950s (Qian and Zhou 2014). Benefitting from some recently released historical datasets, such as the instrumental observations in China between 1912 and 1951 compiled by the National University of Singapore (NUS; Png et al. 2020), we are now able to expand the temporal coverage of continental China’s meteorological records back prior to the 1950s. Moreover, the release of the twentieth-century reanalysis products and historical sea surface temperature (SST)-driven climate model outputs from the sixth phase of the Coupled Model Intercomparison Project (CMIP6) also allow us to revisit the large-scale forcing of historical extreme events such as the 1931 Yangtze River flood.
Flooding is primarily caused by excessive rainfall (Sharma et al. 2018; Tabari 2020). The summer rainfall anomalies along the YRV are affected by the East Asian summer monsoon circulation system, including the western Pacific subtropical high (WPSH) and the subtropical westerly jet in the upper troposphere (Tao and Chen 1987; Wang and LinHo 2002; Ding and Chan 2005; Zhou and Yu 2005; Sampe and Xie 2010; Li et al. 2020). The abrupt northward advance of the WPSH axis to 25°N and the migration of the upper jet to the Tibetan Plateau’s north flank indicate the starting dates of the rainy season along the YRV (Tao and Chen 1987; Zhou et al. 2009a; Sampe and Xie 2010). When the WPSH extends southwestward and the westerly jet shifts southward, the YRV tends to experience increased pluvial events (Zhou and Yu 2005).
As part of the East Asian summer monsoon system, the summer rainfall along the YRV displays variability at multiple scales [see Zhou et al. (2009b) and Ding et al. (2020) for reviews]. On an intraseasonal time scale, the summer rainfall along the YRV is dominated by the intraseasonal oscillation of the East Asian summer monsoon (Yang and Li 2003; Li et al. 2015; Hsu et al. 2016; Ding et al. 2020), modulated by the Madden–Julian oscillation (Li et al. 2018; Liang et al. 2021) and summer North Atlantic Oscillation (Bollasina and Messori 2018; Liu et al. 2020). On the interannual scale, El Niño–Southern Oscillation (ENSO; Huang and Wu 1989; B. Wang et al. 2000, 2008; Yamaura and Tomita 2014) and extratropical wave trains (Ding and Wang 2005; Hsu and Lin 2007; Wang et al. 2018) are identified as drivers of summer rainfall variations.
Since the 1931 Yangtze River flood was preceded by an El Niño event (Rasmusson and Carpenter 1982; Yeh et al. 2009), the role of tropical oceans in driving flooding is our first concern. In summers during decaying El Niño phases, the YRV rainfall is usually above normal (Chang et al. 2000a; Xie et al. 2016; Ding et al. 2020). In response to an El Niño event, the WPSH tends to intensify (Xiang et al. 2013; Chen et al. 2019), and the westerly jet tends to shift southward (Liao et al. 2004; Zhu et al. 2013; Du et al. 2016). The effects of El Niño events on the East Asian summer climate are connected by the western North Pacific anomalous anticyclone (WNPAC; Chen et al. 2016; Zhu and Li 2016; Li et al. 2017). This anomalous anticyclone frequently forms during an El Niño mature winter and persists to the subsequent summer [see Li et al. (2017) for a comprehensive review]. The WNPAC increases the YRV summer precipitation in at least three ways: transporting more moisture, enhancing the pressure gradient to its northwest, and blocking the eastward propagation of synoptic-scale perturbances to produce a quasi-stationary front (Chang et al. 2000a).
The warming in the tropical Indian Ocean can modulate precipitation anomalies along the YRV (Liu and Duan 2017; Liang et al. 2021; Z.-Q. Zhou et al. 2021). The tropical Indian Ocean plays a delayed role in prolonging the climate effects of El Niño from boreal winter into summer. For instance, from the late boreal winter to the following summer of an El Niño event, the Indian Ocean displays a basinwide warming pattern (Yang et al. 2007), which is known as the Indian Ocean basin mode (IOBM). The IOBM enhances summer rainfall along the YRV by promoting southward displacement of the WPSH (Ding et al. 2021) and maintains the WNPAC (Yang et al. 2007; Wu et al. 2009; Xie et al. 2009).
The interannual variability of the East Asian summer monsoon system is also affected by extratropical wave trains, including a zonally oriented teleconnection pattern known as the Silk Road pattern (SRP), which is trapped by the upper-tropospheric subtropical jet (Lu et al. 2002; Enomoto et al. 2003), and a meridionally oriented teleconnection pattern known as the East Asia–Pacific (EAP) pattern (also referred to as the Pacific–Japan pattern) in the middle troposphere (Nitta 1987; Huang 1992). These teleconnection patterns, which are by their nature quasistationary Rossby wave trains, impact the interannual variability of East Asian summer rainfall by altering the location and strength of the monsoon circulation system (Wang et al. 2001; Lu 2004; Wakabayashi and Kawamura 2004; Tao and Wei 2006; Hsu and Lin 2007; Iwao and Takahashi 2008; Hu et al. 2023). These teleconnection patterns also exhibit subseasonal variability and modulate the subseasonal rainfall variability in East Asia (Huang 2004; Ding and Wang 2005; Hsu and Lin 2007; Ren et al. 2013; Guan et al. 2019).
In this study, we aim to address the following questions by combining historical records, reanalysis data, and climate modeling outputs: 1) Was the 1931 summer rainfall record-breaking along the YRV from a centennial perspective? 2) What was the underlying physical mechanism of this disastrous flood?
2. Data and methods
a. Observational and reanalysis data
The following rain gauge observations are used in our analysis:
-
1) 71 stations (1880–2007, hereafter S71) that contain seasonal and annual precipitation series over eastern China and were compiled by Wang et al. (2009).
-
2) NUS archived China Weather Database (1912–51, hereafter NUS) that contains daily instrumental records of temperature, precipitation, and sunshine, compiled by Png et al. (2020).
-
3) 756 stations (1951–present, hereafter S756) that contain daily ground observations provided by the China Meteorological Administration (CMA).
The NUS data, compiled from historical sources from the NUS Library, contains 319 stations in China (Png et al. 2020). This dataset provides high-frequency weather data dating back to 1912. However, these records are not temporally consistent at most stations. The number of records peaked in 1936 and stayed low from 1937 to 1950 (Png et al. 2020). No single station has a full range of records from 1912 to 1951 in this dataset, which means this dataset is currently not capable of presenting long-term trends. In the NUS dataset, 11 stations are located within the range of the YRV and its tributaries with available precipitation records during the summer of 1931. Unfortunately, some stations neither have latitude/longitude information nor have corresponding modern stations in the S756 dataset. After excluding these stations, we identified four stations to serve as the representative stations for the YRV, namely Yichang, Yuezhou (now Yueyang station), Hankou (now Wuhan station), and Nantong (Fig. 1a).
(a) Distribution of the four representative stations along the Yangtze River valley in the NUS database (red circle markers) and the corresponding modern stations in S756 (blue triangle markers). (b) Spatial distribution of the summer (June–August) mean precipitation (mm month−1) in 1931 based on the S71 rain gauge observations. (c) Monthly precipitation amount (mm month−1) and (d) precipitation days (days) along the Yangtze River valley based on the NUS and S756 rain gauge observations. In (a), the blue numbers correspond to the S756 station IDs. In (b), yellow dots show the locations of the S71 stations, and red stars show the locations of the four NUS stations. In (c) and (d), monthly precipitation characteristics of the 1931 and the 1971–2000 climatology are displayed as dark blue and light green bars, respectively.
Citation: Journal of Climate 36, 18; 10.1175/JCLI-D-22-0771.1
Three gridded monthly precipitation products are also used:
-
1) The Full Data Monthly Product at 1.0° × 1.0° horizontal resolution provided by the Global Precipitation Climatology Center (GPCC V2020, 1891–2019; Schneider et al. 2020).
-
2) The gridded monthly precipitation dataset at 0.5° × 0.5° horizontal resolution provided by the Climate Research Unit (CRU V4.05, 1901–2020; Harris et al. 2020).
-
3) The terrestrial monthly precipitation dataset at 0.5° × 0.5° horizontal resolution provided by the University of Delaware (UDel V5.01, 1900–2017; Willmott and Matsuura 2001).
All gridded observations are interpolated onto a 1.0° × 1.0° grid through bilinear interpolation for fair comparison.
In addition to precipitation datasets, several other variables are also used:
-
1) The observational SST derived from the Hadley Center Sea Ice and Sea Surface Temperature dataset (HadISST V1.1, 1871–present; Rayner et al. 2003).
-
2) The geopotential height, wind, specific humidity, vertical p velocity, and surface pressure fields were derived from the following three reanalysis datasets that cover the early twentieth century: the newest version of the Twentieth Century Reanalysis at 1.0° × 1.0° horizontal resolution and 28 vertical levels provided by the National Oceanic and Atmospheric Administration (20CRv3, 1836–2015; Slivinski et al. 2019); the European Centre for Medium-Range Weather Forecasts (ECMWF) Twentieth Century Reanalysis using surface observations only (ERA-20C, 1900–2010; Poli et al. 2016) at 1.0° × 1.0° horizontal resolution and 37 vertical levels; and the ECMWF 10-member ensemble of coupled climate reanalysis of the twentieth century (CERA-20C, 1901–2010; Laloyaux et al. 2018) at 1.0° × 1.0° horizontal resolution and 37 vertical levels. The 20CRv3 product provides daily and monthly average values from 1836 to 2015 and 3-hourly values from 1981 to 2015. Both the ERA-20C and CERA-20C products provide 3-hourly and monthly values from 1901 to 2010. Monthly and daily values are used in this study. Specifically, only pressure levels that are included by all three products, namely the 28 vertical levels of 20CRv3, are used for vertical integration.
b. Model experiments
The AMIP-hist simulation, an extended Atmospheric Model Intercomparison Project (AMIP) run from the CMIP6 archive, is used to verify the role of SST in this event. The AMIP-hist simulation is the tier-1 experiment in the Global Monsoons Model Intercomparison Project (GMMIP), which is one of the endorsed model intercomparison projects (MIPs) in the CMIP6 (Zhou et al. 2016). This experiment uses an atmosphere-only general circulation model (AGCM) configuration and is forced by the HadISST dataset. Unlike the standard AGCM experiments that begin in 1979 (Gates et al. 1999), AMIP-hist extends back to 1870 and covers the time period from 1870 to 2014.
In this study, 15 models participating in the GMMIP are used (see Table 1). The following variables are selected: geopotential height, wind, air temperature, specific humidity, vertical p velocity, and surface pressure fields. To ensure that each model occupies the same weight, only the first member (i.e., realization r1i1p1f1 if available; exceptions are CanESM5 with realization r1i1p2f1 and CNRM-CM6-1, CNRM-CM6-1-HR, and CNRM-ESM2-1 with realization r1i1p1f2) is used for models with more than one member. All the models are interpolated onto a 2.5° × 2.5° grid through bilinear interpolation before analysis.
Information on the CMIP6 models used in this study.
c. Analysis method
1) Regridding station data
For unbiased comparisons between station data and gridded products, the S71 data are interpolated onto a 1° × 1° grid using the iterative improvement objective analysis with influence radii of 5°, 3°, and 2° (using the function “obj_anal_ic_Wrap” in the NCAR Command Language; https://www.ncl.ucar.edu/) before analysis.
2) Extreme precipitation index
The monthly maximum consecutive N-day precipitation (RxNday) index, which is one of the extreme precipitation indices defined by the joint CCI/CLIVAR/JCOMM Expert Team (ET) on Climate Change Detection and Indices (ETCCDI), is used to indicate the extreme intensity and duration of rainfall events (http://etccdi.pacificclimate.org/list_27_indices.shtml). The RxNday index refers to the monthly maximum accumulated precipitation amount for the N-day interval. For example, Rx7day means the maximum accumulated rainfall over one week. This index measures the intensity of rainfall for small N (usually less than 7, such as Rx1day and Rx3day) and the duration of rainfall for large N (usually larger than 7, such as Rx14day and Rx28day). In this work, different N values (N = 1, 14, 28) are chosen to examine the extreme precipitation on different time scales, and the normalized RxNday indices are shown in the form of anomaly percentage (i.e., divide anomalies by climatological mean) instead of their absolute values as Zhang et al. (2020) and T.-J. Zhou et al. (2021).
3) Eddy geopotential height
The position and strength of the WPSH are conventionally measured by the 500-hPa geopotential height. However, the 500-hPa geopotential height field varies with climate background (Gong and Ho 2002; Zhou et al. 2009a; Huang and Li 2015). The eddy geopotential height is an ideal substitute for the measurement of the WPSH because it is comparable under changing climate conditions (Huang and Li 2015; He et al. 2018). In this work, the eddy geopotential height is obtained by subtracting the zonal belt mean between 0° and 40°N from the original field (Zhou et al. 2009a).
4) Climate indices
The following two SST indices are used: (i) the Niño-3.4 index, defined as the SST anomalies averaged in the region of 5°S–5°N, 120°–170°W; and (ii) the IOBM index, defined as the SST anomalies averaged in the region of 20°S–20°N, 40°–110°E, following Chambers et al. (1999). Before calculating the indices, a 9-yr high-pass filter is applied to the original time series to extract interannual variability signals.
5) Wave activity flux
3. The persistence and distribution of summer rainfall
We first examine the spatial coverage of the summer flooding and the seasonal distribution of 1931 rainfall along the YRV (Fig. 1). In the summer of 1931, there existed two major precipitation centers in eastern China: one meandered along the northern bank of the middle and lower reaches of the Yangtze River, which corresponds to the mei-yu rain belt, and the other was distributed along the upper reaches of the YRV and partly fell into the Pearl River basin; that is, the second rain belt centered on the intersection point of Hunan, Guizhou, and Guangxi Provinces (Fig. 1b). We further examine the monthly anomalies and find that rainfall in 1931 peaked in July when the YRV recorded monthly rainfall totals exceeding 450 mm, which exceeded the climate mean for 1971–2000 by a factor of 2.5 (Fig. 1c). The precipitation days in July 1931 added up to 21.3 days on average, which is nearly twice the climate mean (Fig. 1d).
The 1931 summertime rainfall displayed clear subseasonal variability. For example, while July experienced precipitation amounts and precipitation days that were both above normal, both June and August recorded normal conditions (Figs. 1c,d). A quantitative comparison finds that July contributed over 50% of the total summer rainfall in 1931 (Fig. 2b). Previous studies have shown that the subseasonal evolution of the mei-yu rainband based on data after the 1950s (for instance, see Hu et al. 2017; Wang et al. 2020; He et al. 2021; Zheng and Wang 2021, etc.), which indicates the necessity of investigating subseasonal-scale characteristics rather than seasonal means. Since the summer flooding in 1931 was dominated by anomalous July rainfall, we focus on the July condition in the following analysis.
(a)–(d) Spatial distribution of the anomalous precipitation percentage (%) in July 1931 based on three gridded precipitation datasets, (a) GPCC, (b) CRU, and (c) UDel, and their (d) arithmetic mean. (e) Time series of the regional mean July precipitation anomaly (mm month−1) averaged along the YRV (28°–34°N, 110°–122.5°E) against the climatology of 1971–2000. In (a)–(d), the purple rectangle represents the YRV, and the values in the top right corner denote the YRV-averaged precipitation amount in July 1931. In (e), bars represent the average of three datasets, and the yellow triangle, green square, and blue circles represent the results derived from the UDel, CRU, and GPCC datasets, respectively. The black dashed line in (e) indicates the year 1931.
Citation: Journal of Climate 36, 18; 10.1175/JCLI-D-22-0771.1
We further examine the spatial distribution of precipitation anomalies in July 1931 and their relative position over the past hundred years (Fig. 2). The spatial pattern of the July rainfall anomalies is consistent among the three observation-based datasets (Figs. 2a–d). The July rainfall resembles that of the seasonal mean, except that precipitation is greater along the YRV (figure not shown). The regional average of July precipitation along the YRV (28°–34°N, 110°–122.5°E) is more than 300 mm. During the past century (1901–2010), the 1931 rainfall ranked as the second strongest rainfall event, and the ranking sequence was consistent among different datasets (Fig. 2e). Hence, despite being previously acknowledged as “the world’s deadliest floods” on record (Courtney 2018), the summer rainfall in 1931 along the Yangtze River was not the largest event over the past century.
The newly released NUS dataset enables us to reveal the details of the 1931 summer rainfall along the Yangtze River at a daily frequency. We then examine the extreme nature of the heavy precipitation in 1931 as indicated by the RxNday (N = 1, 14, 28) indices for July and compare it to the index for the period from 1951 to 2010 (Fig. 3). The 1931 summer rainfall does not show pronounced short-period extreme precipitation events (Fig. 3a). However, it exceeds most extreme events during the last 50 years for intervals longer than two weeks (Figs. 3b,c). As shown on the 28-day interval, the rain that fell in the summer of 1931 was 160.36% greater than the climatological mean rainfall for July. Even after 1950, none of the heavy rainfall events were comparable to the 1931 summer rainfall event in terms of their duration. Hence, the 1931 summer rainfall event was caused by the accumulation of continuous precipitation instead of very extreme cases.
The maximum consecutive N-day precipitation (i.e., RxNday) anomaly percentage (%) for July from 1951 to 2010 (mm; N = 1, 14, 28). The horizontal dashed black lines and the values in the top-right corner denote the anomaly percentage of the extreme precipitation index for 1931. The calculation is based on the NUS data (for 1931) and S756 data (for 1951–2010) using the four stations marked in Fig. 1. The reference period is from 1971 to 2000.
Citation: Journal of Climate 36, 18; 10.1175/JCLI-D-22-0771.1
How many above-normal precipitation days occurred in July 1931? We use the 1951–2010 precipitation percentile threshold as a standard. For each day in July, we have 60 samples during the 60-yr period. This sample is not large enough to calculate the percentile threshold. Considering the influence of synoptic perturbations, it is reasonable to assume that weather conditions occurring 7 days before or after a date are also likely to be observed on that date. When including this 15-day window, the sample numbers increase to 900 (60 × 15) for each day (including both dry and wet days). In this case, we can count how many days in July 1931 exceed a precipitation percentile threshold, and the results are shown in Table 2. On the one hand, two-thirds of the days in July 1931 exceed the 75th percentile, and one-third of the days exceed the 90th percentile. On the other hand, there are 6 days exceeding the 95th percentile and 2 days exceeding the 99th percentile. Most of the days in July 1931 fall into the range from the 75th to 95th percentile, which again indicates that the 1931 summer rainfall is dominated by the accumulation of continuous processes rather than a few extreme cases. Hence, the 1931 summer rainfall features a long persistent characteristic.
Days in July 1931 exceeding the 1951–2010 precipitation percentile threshold (50th, 75th, 90th, 95th, and 99th percentile).
4. Forcing mechanisms for the 1931 Yangtze River flood
a. The atmospheric circulation patterns associated with the flood
To reveal the underlying physical mechanisms behind the 1931 Yangtze River flood, we examine the anomalous atmospheric circulation conditions in July 1931 based on the reanalysis data (Fig. 4). For climate mean conditions, the rainy season along the YRV starts in mid-June when the WPSH advances northward to approximately 25°N (Tao and Chen 1987; Zhou et al. 2009b). In the lower troposphere, the southwesterly anomalies transport abundant water vapor from the ocean to the YRV. In the upper troposphere, the subtropical westerly jet has two cores over Eurasia: one appears over the East Asian continent and the other over the western Pacific (Du et al. 2009). The mei-yu onset is associated with a northward shift of the eastern branch of the subtropical jet toward the northern flank of the Tibetan Plateau (Li et al. 2004). For the climatological state during July from 1916 to 1945, the ridge line of the WPSH is located at approximately 30°N, and the axis of the subtropical westerly jet is located at approximately 45°N (figure not shown).
Distribution of anomalous atmospheric circulation in July 1931 based on three reanalysis datasets. (a) Climatological mean 200-hPa zonal wind (contours; m s−1) and corresponding anomalies (shading), (b) climatological mean 500-hPa eddy geopotential height (contours; gpm) and corresponding anomalies (shading), (c) 850-hPa relative vorticity anomalies (shading; 10−6 s−1) and wind anomalies (vectors; m s−1, starting from 0.5 m s−1), (d) vertically integrated (from surface to 300 hPa) moisture convergence anomalies (shading; 10−2 g m−2 s−1) and moisture flux anomalies (vectors; 105 m s−1 g kg−1, starting from 0.5 × 105 m s−1 g kg−1). Positive values of moisture convergence denote net inflow of moisture. White dots indicate regions where all three datasets agree on the sign of the anomalies. Specifically, for vectors that have two components, if either component of a vector has the same sign, it is considered as the vector agreeing on the sign. In (c) and (d), only vectors that agreed between all three datasets are shown. The purple rectangle in (d) represents the YRV. The reference period is 1916–45, with the year 1931 centered on the climatology.
Citation: Journal of Climate 36, 18; 10.1175/JCLI-D-22-0771.1
In July 1931, the upper-level westerly jet core shifted southward by 5°–10° from its climatological position of 45°N (Fig. 4a). Following the meridional shift for the subtropical jet, the jet core over the western Pacific also shifted westward by 10° and was located to the north of the low-level southwesterlies (cf. Figs. 4a,c). In the middle and lower troposphere, the anomalous circulation in July 1931 mainly featured the existence of the WNPAC, which leads to a southwestward extension of the WPSH and strengthened low-level southwesterlies along its northern flank (Figs. 4b–d). The ridge line of the WPSH moved southward by 5° in latitude, and its westernmost point extended westward by 10° in longitude (Fig. 4b). The WNPAC transports more water vapor from the ocean via the southwesterlies along its northwestern edge (Figs. 4c,d). The above features are consistent among the three reanalysis datasets. Under these circulation anomalies, enhanced moisture transport from the South China Sea converged over the mei-yu rainband along the YRV and provided sufficient moisture supply for the summer rainfall there (Fig. 4d).
Previous studies have shown that when the WPSH extends southwestward, the low-level southwesterly jet along its northwestern flank strengthens, resulting in more water vapor being transported to the YRV from the Bay of Bengal and the South China Sea (Chang et al. 2000a; Zhou and Yu 2005). In July 1931, the enhanced low-level jet and moisture transport along the edge of the WPSH are evident in the reanalysis data (Figs. 4c,d). Additionally, the northeast wind anomalies to the north of the YRV prevent a farther northward intrusion of moisture transport to northern China, preserving more water vapor along the YRV. In the upper troposphere, the subtropical westerly jet shifts southward to approximately 35°N, which leads further toward a stronger upper-level divergence that is favorable for triggering intensified convection (Zhou and Yu 2005). Hence, both the change in the WPSH and the subtropical westerly jet were conducive for heavier summer precipitation along the YRV in July 1931.
b. Tropical sea surface temperature anomaly forcing
Why did the atmospheric circulation show anomalous patterns in July 1931? The atmospheric circulation patterns in 1931 resemble the typical pattern during an El Niño decaying summer (Liao et al. 2004; Xiang et al. 2013; Zhu et al. 2013; Du et al. 2016; Chen et al. 2019). To understand the effects of the tropical oceans in this event, we examine the temporal evolution of the SST anomalies and the corresponding 850-hPa wind anomalies based on the reanalysis and AGCM simulations (Figs. 5 and 6).
Temporal evolution of (a)–(c) sea surface temperature anomalies (shading; °C) and 850-hPa wind anomalies (vectors; m s−1, starting from 0.2 m s−1) and (d)–(f) 500-hPa vertical p-velocity anomalies (Pa s−1) from January to July 1931, based on three reanalysis datasets. Dots indicate regions where all three datasets agree on the sign of the anomalies. In (a)–(c), only vectors that had the same sign are shown. The black boxes represent the tropical Indian Ocean, and the white boxes represent the Niño-3.4 region. The reference period is 1916–45.
Citation: Journal of Climate 36, 18; 10.1175/JCLI-D-22-0771.1
As in Fig. 5, but for the multimodel ensemble mean of historical SST-driven AGCM simulations. Dots indicate regions where at least 80% of the models agree on the sign of the anomalies.
Citation: Journal of Climate 36, 18; 10.1175/JCLI-D-22-0771.1
In the reanalysis, during January of the preceding winter, the SST anomalies over the tropical eastern Pacific exhibit an El Niño warming pattern, and an anomalous anticyclone is formed over the Philippine Sea (Fig. 5a). This anomalous anticyclone forms in boreal winter and is maintained until the following spring because of El Niño remote forcing via the wind–evaporation–SST feedback mechanism (B. Wang et al. 2000) and anomalous wind–moist enthalpy advection mechanism (Wu et al. 2017a,b) [see Li et al. (2017) for a comprehensive review]. The historical SST-driven AGCM simulations display similar patterns as the reanalysis data during this period (Fig. 6a), which demonstrates that the SST anomaly pattern excites the WNPAC during El Niño mature winter.
After the El Niño mature winter, basinwide warming occurs in the tropical Indian Ocean due to El Niño–induced surface heat flux anomalies (Fig. 5b) (Nigam and Shen 1993; Klein et al. 1999; Liu and Alexander 2007). Meanwhile, the SST anomalies in the tropical Pacific Ocean decay rapidly (Fig. 5b).
In the following summer, the SST anomalies in the equatorial central-eastern Pacific almost return to a normal state (Fig. 5c). Meanwhile, the basinwide warm SST anomalies in the tropical Indian Ocean persist due to local air–sea interactions (Du et al. 2009). Ascending anomalies over wide areas near the equatorial Indian Ocean and the Maritime Continent are seen (Fig. 5f), which indicates that the Indian Ocean plays an active role in driving the atmosphere during El Niño decaying summers. In boreal summer, the anomalous warming patterns in the Indian Ocean contribute to the maintenance of the WNPAC through the Kelvin wave–induced Ekman divergence mechanism (Wu et al. 2009; Xie et al. 2009) and wind-induced moist enthalpy advection mechanism (Wang et al. 2022). In response to tropical Indian Ocean warming, an equatorial Kelvin wave is excited to the east and penetrates the western Pacific (Gill 1980). The Kelvin wave easterlies decay rapidly with latitude, producing anticyclonic shear on both sides of the equator. The anticyclonic shear drives Ekman divergence over the WNP (Wu et al. 2009; Xie et al. 2009). On the other hand, the easterly anomalies associated with the equatorial Kelvin waves transport dry (low moist enthalpy) air westward to the WNP (Wang et al. 2022). Both the Ekman divergence and the negative moist enthalpy advection suppress convection over the WNP and thus maintain the WNPAC. Hence, the Indian Ocean acts like a capacitor that prolongs the effects of El Niño into summer [see Xie et al. (2016) for a comprehensive review]. Similar patterns for the SST anomalies and circulation anomalies are seen in SST-driven AGCM simulations, demonstrating the ocean forcing to the WNPAC, although the intensity of the WNPAC is slightly weaker than the reanalysis due to the lack of air–sea interaction (Figs. 6c,f).
To confirm the effects of the tropical Indian Ocean, we further regress the 850-hPa wind anomalies and vertically integrated moisture flux anomalies onto the July Indian Ocean SST (i.e., regionally averaged SST over 20°S–20°N, 40°–110°E) from 1916 to 1945 with 1931 centered within the period (Fig. 7). For the reanalysis, similar to Fig. 4c, an anomalous anticyclone and negative relative vorticity anomalies stand out over the WNP (Fig. 7a), indicating that the basin-wide warming of the Indian Ocean in boreal summer can induce the formation of the WNPAC. The Indian Ocean SST forcing also enhances moisture convergence toward the mei-yu band (Fig. 7b). The AGCMs show similar patterns, again demonstrating the crucial role of tropical Indian Ocean SSTs in driving the WNPAC (Figs. 7c,d).
Distribution of anomalous atmospheric circulation associated with the summer Indian Ocean (20°S–20°N, 40°–110°E) SST anomalies during the time period of 1916–45 based on (a),(b) three reanalysis datasets and (c),(d) the multimodel ensemble mean of AGCM simulations. (top) The 850-hPa relative vorticity anomalies (shading; 10−6 s−1) and wind anomalies (vectors; m s−1, starting from 0.2 m s−1) regressed upon the July Indian Ocean SST anomalies and (bottom) vertically integrated (from surface to 300 hPa) moisture convergence anomalies (shading; 10−2 g m−2 s−1) and moisture flux anomalies (vectors; 105 m s−1 g kg−1, starting from 0.2 × 105 m s−1 g kg−1) regressed upon the July Indian Ocean SST. White dots indicate regions where all three datasets (or at least 80% of the models) agree on the sign of the anomalies, and only vectors shared by all three datasets (or at least 80% of the models) are shown. The purple rectangle in (b) and (d) represents the YRV. The reference period is 1916–45.
Citation: Journal of Climate 36, 18; 10.1175/JCLI-D-22-0771.1
We also quantitatively examine the intensity of oceanic forcing. The intensity of the Niño-3.4 index in the preceding winter of 1931 is 1.42 standard deviations above average and ranks 9th over the past century (Fig. 8a). Previous studies found that the amplitude of ENSO was relatively small before the 1960s and increased rapidly after the 1960s (Gu and Philander 1995). This kind of decadal variation is also evident in Fig. 8a, indicating that the climate regime might have changed after the 1960s. The IOBM index in the summer of 1931 is approximately 1.79 standard deviations and ranks as the fifth-highest value over the time period of 1900–2010 (Fig. 8b). The IOBM index also displays decadal variations and increases rapidly after the 1980s. Although both the intensity of the preceding wintertime El Niño measured by the Niño-3.4 index and the subsequent boreal summertime Indian Ocean warming measured by the IOBM index in 1931 are not the strongest in the century, we speculate that the atmospheric response to unit SST anomaly forcing is stronger in the small-amplitude periods than during the large-amplitude periods due to the change in climate regimes. This speculation warrants further investigation.
The 9-yr high-pass filtered time series of (a) the standardized detrended winter (December–February) Niño-3.4 and (b) the standardized detrended summer Indian Ocean Basin mode (IOBM) indices from 1900 to 2010, based on the HadISST dataset. Black dashed lines indicate the indices averaged from December 1930 to February 1931.
Citation: Journal of Climate 36, 18; 10.1175/JCLI-D-22-0771.1
c. Extratropical wave trains
We have demonstrated that El Niño–related SST forcing, especially tropical Indian Ocean basin warming, is conducive to heavy summer precipitation along the YRV by driving the WNPAC. Nonetheless, the slow-varying SST anomalies may be insufficient to explain the month-to-month variations in summer rainfall.
Previous studies have reported that extratropical wave activity can also influence precipitation anomalies along the YRV (Ding and Wang 2005; Hsu and Lin 2007; Wang et al. 2018). To examine the potential influence of extratropical processes, we first remove the ENSO signals by subtracting the linear regression for the Niño-3.4 time series from the original data during the period of 1916–45. We then examine the monthly change in mid- and high-latitude circulation anomalies and find that extratropical wave activity was able to affect the YRV in July 1931 (Figs. 9a,b). We note that such extratropical signals are much weaker in June and August (figures omitted).
(a) The 200-hPa wind anomalies (vectors; m s−1) and meridional wind anomalies (shading; m s−1), (b) 200-hPa geopotential height anomalies (contours; gpm; starting from 10 gpm every 20 gpm), wave-activity flux (vectors; m2 s−2) and SST anomalies (shading; °C), and (c) 200-hPa zonal wind anomalies (contours; m s−1, starting from 2 m s−1 every 2 m s−1) and precipitation anomalies (shading; mm month−1) in July 1931. The ENSO signals are linearly removed from atmospheric circulation fields by subtracting the linear regression for the Niño-3.4 time series from the original field during the period of 1916–45. The red solid contours in (b) and (c) represent positive values, while the blue dashed contours represent negative values. The black and purple lines in (b) represent the polar front jet axis and subtropical jet axis in July 1931, respectively. The purple contour in (c) represents the 18 m s−1 contour, indicating the location of the climatological mean westerly jet.
Citation: Journal of Climate 36, 18; 10.1175/JCLI-D-22-0771.1
In July 1931, there were two zonally oriented wavelike patterns in the upper troposphere over the Eurasian continent (Fig. 9a). The wavelike patterns resemble the coupling mode of two jet waveguides, namely the polar front jet waveguide and the subtropical jet waveguide (Iwao and Takahashi 2008; Xu et al. 2022). Along the polar front jet (whose climatological mean axis is located at approximately 60°N in July), the north branch of the wavelike pattern has four centers of action (i.e., the British Isles, western Russia, northern Siberia, and Okhotsk) that resembles the British–Okhotsk Corridor (BOC) pattern (Xu et al. 2022). Along the subtropical jet (whose climatological axis is located at approximately 45°N in July), the south branch of the wavelike pattern spanning from North Africa to East Asia resembles the SRP (Lu et al. 2002; Enomoto et al. 2003). The wave-activity flux clearly shows the bifurcated wavelike pattern along the polar front jet and subtropical jet (Fig. 9b).
The extratropical wavelike patterns influence the summer precipitation along the YRV by modulating the meridional shift of the subtropical westerly jet and the zonal shift of the WPSH (Tao and Wei 2006; Hong and Lu 2016; Li and Lu 2017; Guan et al. 2019). In July 1931, the negative phase of the SRP along the subtropical jet, characterized by the cyclonic anomaly over western Asia and East Asia and the weak anticyclonic circulation between them (Fig. 9a), moves the jet southward by strengthening (weakening) the westerlies to the south (north) of the jet axis (Fig. 9c). Meanwhile, the East Asian center of the negative SRP is enhanced by the positive phase of BOC along the polar front jet through strong wave activity energy transported southward from the positive Okhotsk anomalies (Fig. 9b). In addition to the zonal wavelike patterns, a meridional tripolar pattern of zonal wind anomalies is evident over East Asia and the WNP in July 1931 (Fig. 9c). This pattern is recognized as the EAP (also referred to as Pacific–Japan) teleconnection (Nitta 1987; Huang 1992). Hence, the coupling of BOC–SRP–EAP results in an equatorward shift of the subtropical westerly jet. In accordance with the equatorward displacement of the subtropical jet, the WPSH extends southwestward due to high pressure coupled with upper-level jet stream (Wang et al. 2001; Enomoto 2004; Lu 2004; Wang et al. 2016). This explains why the WPSH was relatively steady in July 1931 and could not move farther north as during normal years (Fig. 10). The stable WPSH maintained heavy rainfall by transporting abundant moisture to the YRV (see Fig. 4d shows above).
Daily evolution of western Pacific subtropical high (WPSH) in July 1931 based on three reanalysis datasets. (a) WPSH extent (denoted by the zero line of eddy geopotential height at 500 hPa) in July 1931 (red solid line) and 1916–45 climate mean condition (blue dashed line). (b) Longitude of the westernmost point of the WPSH in 1931 (red solid line) and 1916–45 climate mean conditions (blue dashed line) from 1 to 31 Jul. (c) Longitude–time section along 20°–25°N of daily geopotential height at 500 hPa in 1931 (blue shading) and 1916–45 mean condition (contours) from 1 to 31 Jul.
Citation: Journal of Climate 36, 18; 10.1175/JCLI-D-22-0771.1
To understand the persistence of the 1931 July rainfall (cf. Fig. 3 and Table 2), we further examine the daily evolution of the WPSH during this period. Previous studies have shown that persistent heavy precipitation along the YRV is accompanied by westward extension of the WPSH (W.-C. Wang et al. 2000; Liu et al. 2008; Mao et al. 2010; Lee et al. 2013; Ren et al. 2013; Chen and Zhai 2015). Hence, we examine the zonal position of the WPSH in July 1931 (Fig. 10). The monthly mean location of the WPSH extent (indicated by the zero line of eddy geopotential height) extends westward significantly compared to the climatology (Fig. 10a). On a daily time scale, the westernmost point of the WPSH body at 500 hPa is located to the west of its climatological mean position throughout July 1931 (Fig. 10b). In late July, the WPSH generally retreats eastward during climate mean conditions, but it remained west of 120°E in 1931 (Figs. 10b,c). Hence, the consistent westward extension of the WPSH anchored the rainband along the YRV and led to persistent rainfall in July 1931.
The above analyses demonstrate that the heavy rainfall in July 1931 was caused by the joint effects of El Niño–related tropical Indian Ocean warming and wavelike patterns resembling a coupling of BOC, SRP, and EAP in the extratropics. The Indian Ocean basinwide warming during the summer of an El Niño decaying year influences the rainfall along the YRV by maintaining the WNPAC and enhancing the moisture transported by low-level southwesterlies. The BOC–SRP–EAP-like wave activity is associated with a southward shift of the upper-level subtropical westerly jet, which couples with the low-level flows and is favorable for intensified convection along the YRV.
5. Discussion of springtime soil conditions
Although the persistent rainfall along the Yangtze River was extremely heavy in July 1931, the total precipitation amount averaged along the YRV during this period is not the heaviest from a centennial perspective. Hence, there might be other factors contributing to the deadliest aspects of this event. We note that the amount of precipitation in May 1931 was also above normal along the YRV (Fig. 1c). We further examine the extreme nature of the anomalous rainfall in May 1931 (Fig. 11). The consecutive extreme precipitation indices in May 1931 display similar features as those in the summer (see Fig. 3). The anomalous rainfall in May 1931 is 76.70% and 96.06% above average for the 14- and 28-day intervals, respectively (Figs. 11b,c). Hence, the springtime rainfall of 1931 is also extremely heavy. There is a possibility that the substantial May rainfall can further fuel the severity of summertime flooding. We examine the hypothesis by analyzing the persistence of soil moisture.
Previous studies have demonstrated the role of antecedent rainfall in flooding (Tramblay et al. 2012; Berghuijs et al. 2016; Grillakis et al. 2016; Woldemeskel and Sharma 2016; Bennett et al. 2018). Heavy antecedent rainfall can lead to elevated soil moisture and even soil saturation. The near-saturated soil moisture conditions make it easier for subsequent heavy precipitation to generate a flood. To test whether this mechanism applied to the 1931 flood event, we examine the springtime soil moisture conditions derived from reanalysis datasets (Fig. 12). In the 20CRv3 product, the soil moisture content in the upper 200-cm soil layer was slightly below normal in April 1931 (Fig. 12a). Later, the persistent rainfall in May significantly increased the soil moisture content along the YRV (Fig. 12b). The wet soil conditions were preserved throughout June when the rainband was located to the south of the YRV (Figs. 12c,e). In July, abundant precipitation fell into the YRV and increased the soil moisture (Fig. 12d). The coherent changes in precipitation and soil moisture indicate that the soil moisture content in the upper 200 cm has some kind of “memory,” which can persist for longer than one month. Therefore, the abundant antecedent springtime rainfall partly exacerbated the impacts of July heavy precipitation on the flood of 1931.
(a)–(d) Spatial distribution of soil moisture content (from 0 to 200 cm) anomalies (shading; kg m−2) and precipitation anomalies (contours; mm day−1; starting from 5 mm day−1 at 2.5 mm day−1 intervals) from April to July 1931 based on the 20CRv3 reanalysis data. (e) Monthly soil moisture content (kg m−2) averaged along the YRV [purple rectangle in (a)–(d)]; 1931 and the 1916–45 climatology are displayed as dark blue and light green bars, respectively.
Citation: Journal of Climate 36, 18; 10.1175/JCLI-D-22-0771.1
Additionally, the process described above resembles one category of compound events defined as “preconditioned events” by the Sixth Assessment Report (AR6) from the Intergovernmental Panel on Climate Change (IPCC) Working Group I, which means a climatic impact driver aggravated by weather-driven or climate-driven preconditions (Seneviratne et al. 2021). The compounding effects of the precondition and the climatic impact driver partly explain how a single event can trigger serious consequences that cannot be anticipated from the magnitude of the event alone.
The compound effects alone cannot fully explain the record-breaking hazards of 1931. The heaviest spring and summer rainfall over the past century occurred in 1954, and that event resulted in a loss of 32 000 lives (Luo 2006); however, the estimated death toll of the 1931 Yangtze River flood exceeded 2 000 000 (National Flood Relief Commission 1933). There is evidence that suggests that the inadequate flood disaster prevention measures in the 1930s were partly responsible for the disastrous consequences. In the first half of 1932, approximately 1.1 million workers were employed to reconstruct 2000 km of dikes (National Flood Relief Commission 1933). This record reveals that the 1931 flood destroyed nearly 2000 km of dikes. Even though the dikes were reconstructed, many of them collapsed later due to the subsequent flooding that occurred in 1935 (https://disasterhistory.org/central-China-flood-1931). In other words, the dikes in the 1930s were not sturdy enough to protect people and infrastructure against strong flooding.
Additionally, the official response to this disaster was inadequate. In the aftermath of flooding, many people failed to survive the subsequent famines and diseases. Of the 2 million deaths, only a small proportion could be directly attributed to drowning during the flood, while approximately 70% of the refugees died by disease (Buck 1932). In response to this disastrous flooding, the government established the National Flood Relief Commission, which employed many native and foreign experts across a range of fields, including hydrology and epidemiology (National Flood Relief Commission 1933). However, the mortality rate from disease in some relief camps was found to be much higher than that in rural communities (Buck 1932). All these documents indicate that in addition to the overwhelming climate-driven conditions, the poor response and capability to adapt to the 1931 flood was also responsible for the disastrous consequences. This historical event warrants deeper examination within the climate change adaptation community.
6. Summary and concluding remarks
The flooding that occurred during the summer of 1931 along the Yangtze River is acknowledged as the world’s deadliest flood on record due to a high rate of mortality. However, the physical mechanisms underlying flood event remain unknown. Based on the diagnosis of multiple lines of evidence from recently available historical observations, reanalysis datasets, and historical SST-driven AGCM simulations, we find that this flood was driven by the joint effects of tropical El Niño–related SST anomalies and extratropical BOC–SRP–EAP coupling wave activity. The major results are summarized below.
In the summer of 1931, there were two rain belts in eastern China, one located along the northern bank of the middle and lower reaches of the Yangtze River, and the other extending from the upper reaches of the Yangtze River to the Pearl River Basin. The rain that fell in July 1931 contributed more than half of the summertime precipitation along the YRV. However, the rainfall totals observed along the YRV in July 1931 were not the largest in history and rank second between 1901 and 2010. The July rainfall over the YRV in 1931 persisted over a long period of time, with the maximum consecutive 28-day precipitation (Rx28day) being 160% above the climate mean conditions.
The dominant atmospheric anomalies during the summer of 1931 consist of a southward shift of the upper-level subtropical westerly jet, a southwestward extension of the WPSH in the middle troposphere, and an anomalous anticyclone over the WNP in the lower troposphere. These anomalous circulation patterns enhance moisture transport toward the YRV and provide favorable moisture conditions for heavy rainfall. In the tropics, the SST anomalies exhibit a pattern typical for the summer of an El Niño decaying year, with basinwide warming in the tropical Indian Ocean. The intensity of the Indian Ocean basin mode ranked as the fifth-highest over the period of 1900–2010, demonstrating the contribution of El Niño–related SST forcing, particularly the Indian Ocean basin mode, to heavy rainfall. In the extratropics, the coupled wave trains over the Eurasian continent and western North Pacific, namely the coupling of BOC, SRP, and EAP, is conducive to a steady WPSH in July, which sustains persistent heavy rainfall by transporting sufficient moisture to the YRV.
Additionally, there are hints that the 1931 Yangtze River flood is similar to a preconditioned compound event, whose impacts are possibly aggravated by the preceding springtime heavy precipitation. In May 1931, the persistent rainfall was also extreme, with the Rx28day index being 90% above normal. The heavy springtime rainfall partly exacerbated the 1931 summer flooding by increasing soil moisture, making it easier for summer rainfall to generate a flood.
Acknowledgments.
This work is supported by the National Natural Science Foundation of China under Grant 41988101 and the K. C. Wong Education Foundation.
Data availability statement.
NUS Republican China Weather Database data are available at https://zenodo.org/record/3697956#.YxhOinZBzZc. Rain gauge observations from 756 stations were provided by the CMA at http://data.cma.cn/en. GPCC V2020 monthly precipitation data are available at https://opendata.dwd.de/climate_environment/GPCC/html/download_gate.html, and CRU V4.05 precipitation data can be downloaded from their website at https://crudata.uea.ac.uk/cru/data/hrg/. UDel V5.01 precipitation and 20CRv3 data were provided by the NOAA/OAR/ESRL PSL, Boulder, Colorado, USA, at https://psl.noaa.gov/data/gridded/data.UDel_AirT_Precip.html and https://psl.noaa.gov/data/gridded/data.20thC_ReanV3.html. ERA-20C and CERA-20C reanalysis data are available from the ECMWF website at https://www.ecmwf.int/en/forecasts/dataset/ecmwf-reanalysis-20th-century-using-surface-observations-only and https://www.ecmwf.int/en/forecasts/datasets/reanalysis-datasets/cera-20c. Monthly SST data HadISST V1.1 are available from the NCAR Climate Data Guide at https://climatedataguide.ucar.edu/climate-data/sst-data-hadisst-v11. The results of AMIP-hist simulations from the CMIP6 archive can be downloaded at https://esgf-node.llnl.gov/projects/cmip6/.
REFERENCES
Bamberg, S., T. Masson, K. Brewitt, and N. Nemetschek, 2017: Threat, coping and flood prevention—A meta-analysis. J. Environ. Psychol., 54, 116–126, https://doi.org/10.1016/j.jenvp.2017.08.001.
Bennett, B., M. Leonard, Y. Deng, and S. Westra, 2018: An empirical investigation into the effect of antecedent precipitation on flood volume. J. Hydrol., 567, 435–445, https://doi.org/10.1016/j.jhydrol.2018.10.025.
Berghuijs, W. R., R. A. Woods, C. J. Hutton, and M. Sivapalan, 2016: Dominant flood generating mechanisms across the United States. Geophys. Res. Lett., 43, 4382–4390, https://doi.org/10.1002/2016GL068070.
Bollasina, M. A., and G. Messori, 2018: On the link between the subseasonal evolution of the North Atlantic Oscillation and East Asian climate. Climate Dyn., 51, 3537–3557, https://doi.org/10.1007/s00382-018-4095-5.
Buck, J. L., 1932: The 1931 Flood in China: An Economic Survey by the Department of Agricultural Economics, College of Agriculture and Forestry, The University of Nanking, in Cooperation with the National Flood Relief Commission. The University of Nanking, 74 pp.
Chambers, D. P., B. D. Tapley, and R. H. Stewart, 1999: Anomalous warming in the Indian Ocean coincident with El Niño. J. Geophys. Res., 104, 3035–3047, https://doi.org/10.1029/1998JC900085.
Chang, C.-P., Y. Zhang, and T. Li, 2000a: Interannual and interdecadal variations of the East Asian summer monsoon and tropical Pacific SSTs. Part I: Roles of the subtropical ridge. J. Climate, 13, 4310–4325, https://doi.org/10.1175/1520-0442(2000)013<4310:IAIVOT>2.0.CO;2.
Chen, M., J.-Y. Yu, X. Wang, and W. Jiang, 2019: The changing impact mechanisms of a diverse El Niño on the western Pacific subtropical high. Geophys. Res. Lett., 46, 953–962, https://doi.org/10.1029/2018GL081131.
Chen, W., J.-Y. Lee, K.-J. Ha, K.-S. Yun, and R. Lu, 2016: Intensification of the western North Pacific anticyclone response to the short decaying El Niño event due to greenhouse warming. J. Climate, 29, 3607–3627, https://doi.org/10.1175/JCLI-D-15-0195.1.
Chen, Y., and P. Zhai, 2015: Synoptic-scale precursors of the East Asia/Pacific teleconnection pattern responsible for persistent extreme precipitation in the Yangtze River Valley. Quart. J. Roy. Meteor. Soc., 141, 1389–1403, https://doi.org/10.1002/qj.2448.
Chen, Y., J. Li, and A. Chen, 2021: Does high risk mean high loss: Evidence from flood disaster in southern China. Sci. Total Environ., 785, 147127, https://doi.org/10.1016/j.scitotenv.2021.147127.
Courtney, C., 2018: The Nature of Disaster in China: The 1931 Yangzi River Flood. Cambridge University Press, 312 pp.
Ding, Q., and B. Wang, 2005: Circumglobal teleconnection in the Northern Hemisphere summer. J. Climate, 18, 3483–3505, https://doi.org/10.1175/JCLI3473.1.
Ding, Y., and J. C. L. Chan, 2005: The East Asian summer monsoon: An overview. Meteor. Atmos. Phys., 89, 117–142, https://doi.org/10.1007/s00703-005-0125-z.
Ding, Y., P. Liang, Y. Liu, and Y. Zhang, 2020: Multiscale variability of Meiyu and its prediction: A new review. J. Geophys. Res. Atmos., 125, e2019JD031496, https://doi.org/10.1029/2019JD031496.
Ding, Y., Y. Liu, and Z.-Z. Hu, 2021: The record-breaking Mei-yu in 2020 and associated atmospheric circulation and tropical SST anomalies. Adv. Atmos. Sci., 38, 1980–1993, https://doi.org/10.1007/s00376-021-0361-2.
Du, Y., Y. Zhang, and Z. Xie, 2009: Impacts of the zonal position of the East Asian westerly jet core on precipitation distribution during Meiyu of China. J. Meteor. Res., 23, 506–516, http://jmr.cmsjournal.net/en/article/id/1206.
Du, Y., T. Li, Z. Xie, and Z. Zhu, 2016: Interannual variability of the Asian subtropical westerly jet in boreal summer and associated with circulation and SST anomalies. Climate Dyn., 46, 2673–2688, https://doi.org/10.1007/s00382-015-2723-x.
Enomoto, T., 2004: Interannual variability of the Bonin high associated with the propagation of Rossby waves along the Asian jet. J. Meteor. Soc. Japan, 82, 1019–1034, https://doi.org/10.2151/jmsj.2004.1019.
Enomoto, T., B. J. Hoskins, and Y. Matsuda, 2003: The formation mechanism of the Bonin high in August. Quart. J. Roy. Meteor. Soc., 129, 157–178, https://doi.org/10.1256/qj.01.211.
Gates, W. L., and Coauthors, 1999: An overview of the results of the Atmospheric Model Intercomparison Project (AMIP I). Bull. Amer. Meteor. Soc., 80, 29–56, https://doi.org/10.1175/1520-0477(1999)080<0029:AOOTRO>2.0.CO;2.
Gill, A. E., 1980: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc., 106, 447–462, https://doi.org/10.1002/qj.49710644905.
Gong, D.-Y., and C.-H. Ho, 2002: Shift in the summer rainfall over the Yangtze River valley in the late 1970s. Geophys. Res. Lett., 29, 1436, https://doi.org/10.1029/2001GL014523.
Grillakis, M. G., A. G. Koutroulis, J. Komma, I. K. Tsanis, W. Wagner, and G. Blöschl, 2016: Initial soil moisture effects on flash flood generation—A comparison between basins of contrasting hydro-climatic conditions. J. Hydrol., 541, 206–217, https://doi.org/10.1016/j.jhydrol.2016.03.007.
Gu, D., and S. G. H. Philander, 1995: Secular changes of annual and interannual variability in the tropics during the past century. J. Climate, 8, 864–876, https://doi.org/10.1175/1520-0442(1995)008<0864:SCOAAI>2.0.CO;2.
Guan, W., H. Hu, X. Ren, and X. Yang, 2019: Subseasonal zonal variability of the western Pacific subtropical high in summer: Climate impacts and underlying mechanisms. Climate Dyn., 53, 3325–3344, https://doi.org/10.1007/s00382-019-04705-4.
Guha-Sapir, D., D. Hargitt, and P. Hoyois, 2004: Thirty Years of Natural Disasters 1974–2003: The Number. UCL Presses Universitaires de Louvain, 188 pp.
Harris, I., T. J. Osborn, P. Jones, and D. Lister, 2020: Version 4 of the CRU TS monthly high-resolution gridded multivariate climate dataset. Sci. Data, 7, 109, https://doi.org/10.1038/s41597-020-0453-3.
He, C., A. Lin, D. Gu, C. Li, B. Zheng, B. Wu, and T. Zhou, 2018: Using eddy geopotential height to measure the western North Pacific subtropical high in a warming climate. Theor. Appl. Climatol., 131, 681–691, https://doi.org/10.1007/s00704-016-2001-9.
He, L., X. Hao, H. Li, and T. Han, 2021: How do extreme summer precipitation events over eastern China subregions change? Geophys. Res. Lett., 48, e2020GL091849, https://doi.org/10.1029/2020GL091849.
Hong, X., and R. Lu, 2016: The meridional displacement of the summer Asian jet, Silk Road pattern, and tropical SST anomalies. J. Climate, 29, 3753–3766, https://doi.org/10.1175/JCLI-D-15-0541.1.
Hsu, H.-H., and S.-M. Lin, 2007: Asymmetry of the tripole rainfall pattern during the East Asian summer. J. Climate, 20, 4443–4458, https://doi.org/10.1175/JCLI4246.1.
Hsu, P.-C., J.-Y. Lee, and K.-J. Ha, 2016: Influence of boreal summer intraseasonal oscillation on rainfall extremes in southern China. Int. J. Climatol., 36, 1403–1412, https://doi.org/10.1002/joc.4433.
Hu, K., S.-P. Xie, and G. Huang, 2017: Orographically anchored El Niño effect on summer rainfall in central China. J. Climate, 30, 10 037–10 045, https://doi.org/10.1175/JCLI-D-17-0312.1.
Hu, S., T. Zhou, B. Wu, and X. Chen, 2023: Seasonal prediction of the record-breaking northward shift of the western Pacific subtropical high in July 2021. Adv. Atmos. Sci., 40, 410–427, https://doi.org/10.1007/s00376-022-2151-x.
Huang, G., 2004: An index measuring the interannual variation of the East Asian summer monsoon—The EAP index. Adv. Atmos. Sci., 21, 41–52, https://doi.org/10.1007/BF02915679.
Huang, R., 1992: The East Asia/Pacific pattern teleconnection of summer circulation and climate anomaly in East Asia. J. Meteor. Res., 6, 25–37.
Huang, R., and Y. Wu, 1989: The influence of ENSO on the summer climate change in China and its mechanism. Adv. Atmos. Sci., 6, 21–32, https://doi.org/10.1007/BF02656915.
Huang, Y.-Y., and X.-F. Li, 2015: The interdecadal variation of the western Pacific subtropical high as measured by 500 hPa eddy geopotential height. Atmos. Oceanic Sci. Lett., 8, 371–375, https://doi.org/10.3878/AOSL20150038.
Iwao, K., and M. Takahashi, 2008: A precipitation seesaw mode between Northeast Asia and Siberia in summer caused by Rossby waves over the Eurasian continent. J. Climate, 21, 2401–2419, https://doi.org/10.1175/2007JCLI1949.1.
Klein, S. A., B. J. Soden, and N.-C. Lau, 1999: Remote sea surface temperature variations during ENSO: Evidence for a tropical atmospheric bridge. J. Climate, 12, 917–932, https://doi.org/10.1175/1520-0442(1999)012<0917:RSSTVD>2.0.CO;2.
Laloyaux, P., and Coauthors, 2018: CERA-20C: A coupled reanalysis of the twentieth century. J. Adv. Model. Earth Syst., 10, 1172–1195, https://doi.org/10.1029/2018MS001273.
Lee, S.-S., Y.-W. Seo, K.-J. Ha, and J.-G. Jhun, 2013: Impact of the western North Pacific subtropical high on the East Asian monsoon precipitation and the Indian Ocean precipitation in the boreal summertime. Asia-Pac. J. Atmos. Sci., 49, 171–182, https://doi.org/10.1007/s13143-013-0018-x.
Li, C.-Y., J.-T. Wang, S.-Z. Lin, and H.-R. Cho, 2004: The relationship between East Asian summer monsoon activity and northward jump of the upper westerly jet location. Chin. J. Atmos. Sci., 28, 641–658, https://doi.org/10.3878/j.issn.1006-9895.2004.05.01.
Li, J., J. Mao, and G. Wu, 2015: A case study of the impact of boreal summer intraseasonal oscillations on Yangtze rainfall. Climate Dyn., 44, 2683–2702, https://doi.org/10.1007/s00382-014-2425-9.
Li, J., B. Wang, and Y.-M. Yang, 2020: Diagnostic metrics for evaluating model simulations of the East Asian monsoon. J. Climate, 33, 1777–1801, https://doi.org/10.1175/JCLI-D-18-0808.1.
Li, T., B. Wang, B. Wu, T. Zhou, C.-P. Chang, and R. Zhang, 2017: Theories on formation of an anomalous anticyclone in western North Pacific during El Niño: A review. J. Meteor. Res., 31, 987–1006, https://doi.org/10.1007/s13351-017-7147-6.
Li, W., X. Cheng, Y. Liu, and M. Xia, 1994: Zhongguo jindai shi da zaihuang (The Ten Great Hazards of China’s Modern Period). Shanghai renmin chubanshe, 294 pp.
Li, X., and R. Lu, 2017: Extratropical factors affecting the variability in summer precipitation over the Yangtze River Basin, China. J. Climate, 30, 8357–8374, https://doi.org/10.1175/JCLI-D-16-0282.1.
Li, X., G. Gollan, R. J. Greatbatch, and R. Lu, 2018: Intraseasonal variation of the East Asian summer monsoon associated with the Madden–Julian Oscillation. Atmos. Sci. Lett., 19, e794, https://doi.org/10.1002/asl.794.
Liang, P., Z.-Z. Hu, Y. Ding, and Q. Qian, 2021: The extreme Mei-yu season in 2020: Role of the Madden–Julian Oscillation and the cooperative influence of the Pacific and Indian Oceans. Adv. Atmos. Sci., 38, 2040–2054, https://doi.org/10.1007/s00376-021-1078-y.
Liao, Q.-H., S.-Y. Tao, and H.-J. Wang, 2004: Interannual variation of summer subtropical westerly jet in East Asia and its impacts on the climate anomalies of East Asia summer monsoon. Chin. J. Geophys., 47, 12–21, https://doi.org/10.1002/cjg2.449.
Liu, B., Y. Yan, C. Zhu, S. Ma, and J. Li, 2020: Record-breaking Meiyu rainfall around the Yangtze River in 2020 regulated by the subseasonal phase transition of the North Atlantic Oscillation. Geophys. Res. Lett., 47, e2020GL090342, https://doi.org/10.1029/2020GL090342.
Liu, H., D.-L. Zhang, and B. Wang, 2008: Daily to submonthly weather and climate characteristics of the summer 1998 extreme rainfall over the Yangtze River Basin. J. Geophys. Res., 113, D22101, https://doi.org/10.1029/2008JD010072.
Liu, S., and A. Duan, 2017: Impacts of the leading modes of tropical Indian Ocean sea surface temperature anomaly on sub-seasonal evolution of the circulation and rainfall over East Asia during boreal spring and summer. J. Meteor. Res., 31, 171–186, https://doi.org/10.1007/s13351-016-6093-z.
Liu, Z., and M. Alexander, 2007: Atmospheric bridge, oceanic tunnel, and global climatic teleconnections. Rev. Geophys., 45, RG2005, https://doi.org/10.1029/2005RG000172.
Lu, R.-Y., 2004: Associations among the components of the East Asian summer monsoon system in the meridional direction. J. Meteor. Soc. Japan, 82, 155–165, https://doi.org/10.2151/jmsj.82.155.
Lu, R.-Y., J.-H. Oh, and B.-J. Kim, 2002: A teleconnection pattern in upper-level meridional wind over the North African and Eurasian continent in summer. Tellus, 54A, 44–55, https://doi.org/10.3402/tellusa.v54i1.12122.
Luo, C., 2006: The Flood Disasters in China (in Chinese). China Bookstore, 499 pp.
Mao, J., Z. Sun, and G. Wu, 2010: 20–50-day oscillation of summer Yangtze rainfall in response to intraseasonal variations in the subtropical high over the western North Pacific and South China Sea. Climate Dyn., 34, 747–761, https://doi.org/10.1007/s00382-009-0628-2.
National Flood Relief Commission, 1933: Report of the National Flood Relief Commission, 1931–1932 (in Chinese). Comacrib Press, 304 pp.
Nigam, S., and H.-S. Shen, 1993: Structure of oceanic and atmospheric low-frequency variability over the tropical Pacific and Indian Oceans. Part I: COADS observations. J. Climate, 6, 657–676, https://doi.org/10.1175/1520-0442(1993)006<0657:SOOAAL>2.0.CO;2.
Nitta, T., 1987: Convective activities in the tropical western Pacific and their impact on the Northern Hemisphere summer circulation. J. Meteor. Soc. Japan, 65, 373–390, https://doi.org/10.2151/jmsj1965.65.3_373.
Png, I. P. L., Y. Chen, J. Chu, Y. Feng, E. K. Lin, and W. Tseng, 2020: Temperature, precipitation and sunshine across China, 1912-51: A new daily instrumental dataset. Geosci. Data J., 7, 90–101, https://doi.org/10.1002/gdj3.91.
Poli, P., and Coauthors, 2016: ERA-20C: An atmospheric reanalysis of the twentieth century. J. Climate, 29, 4083–4097, https://doi.org/10.1175/JCLI-D-15-0556.1.