Few studies of extreme precipitation have been conducted in Northeast China, particularly at multi-timescales. We aim to enhance the understanding of changes and variability in extreme precipitation over the past 54 years (1961–2014) in this region. We have investigated the potential relationship among extreme precipitation, climate and agricultural drought focusing on several timescales in this region. Thirteen extreme precipitation indices at seasonal, annual, and growing-period scales were estimated on the basis of daily precipitation data from 70 meteorological stations. The results indicate that all extreme precipitation indices that reflect the features of extreme wet events showed increasing trends in spring, and the trends of these indices were almost negative for the other timescales. Spatially, the frequency, duration and intensity of extreme wet events decreased gradually from south to north. The range of influence and the duration of extreme dry events increased continuously in Northeast China. In Northeast China, extreme precipitation was more easily influenced by the polar climate than the monsoon. Furthermore, correlation between the extreme precipitation indices and comprehensive crop failure ratios of agricultural drought disasters (C index) confirmed that agricultural drought was heavily influenced by precipitation anomalies in this area.

INTRODUCTION

Changes in precipitation are considered to be among the most important elements of climate change, and multiple aspects of these changes have been confirmed to vary spatially over regional and global scales (Donat et al. 2016). Studies have also indicated that extreme precipitation has increased in frequency in more regions than it has decreased worldwide (Alexander 2016). Furthermore, the damage of anomalous precipitation events has affected a wide range of domains, such as society, economics, agricultural and industrial production, and natural ecosystems (Fu et al. 2013; Donat et al. 2014; Stott et al. 2016). The risks for severe disasters associated with extreme precipitation events, such as droughts, floods, debris flows and landslides, have risen (Fu et al. 2013; Fischer & Knutti 2015). Therefore, the potential impacts of these events, which may trigger ripple effects across the global economy, cannot be ignored (Goodess 2013; Vavrus et al. 2015). However, most existing studies of extreme precipitation have concentrated on the characteristics of mean precipitation values because of limitations to both the quality and the accessibility of data (Croitoru et al. 2013). Overall changes in extreme precipitation must also be detected to provide a more comprehensive understanding of the mechanisms of natural climate variability to guide agricultural planning and mitigate losses incurred by extreme precipitation events.

Many studies of extreme precipitation at global, national and regional scales have been conducted since the 2000s (Rosenzweig et al. 2002; Zhang et al. 2011; Croitoru et al. 2013; Donat et al. 2013; 2016; Ramos et al. 2015), because of its profound impacts on the development of society and economics. At the global scale, the features of seasonal extreme precipitation present clear differences, i.e. wet areas have tended to become wetter, and dry areas have tended to become drier (Alexander et al. 2006; Russo & Sterl 2012). On a national scale, the majority of studies focused on the changing characteristics of extreme precipitation events have been conducted in developed countries, especially the USA, Canada and European nations (Zhang et al. 2001; Van den Besselaar et al. 2013; Cannon 2015). Although such studies are gradually spreading to developing counties, the advancement of these studies cannot yet meet the economic, developmental and agricultural industry needs of these regions (Roy & Balling 2004; Zhang et al. 2012a). Therefore, these studies must be expanded to developing countries, particularly in nations that depend on agriculture. Regional precipitation patterns have presented many new features due to continuous global warming, especially in terms of differences in extreme precipitation events (Stott et al. 2016). In particular, the impacts of extreme precipitation on crop yields in these regions are more significant and must not be ignored (Bekele et al. 2016; Lesk et al. 2016): stronger focus is required, especially for rainfed croplands.

Two distinguishing features should be noted of existing investigations on extreme precipitation. First, it is clear that study of spatial variability of precipitation at different timescales (e.g. annual, seasonal and monthly) is essential whether at global, national or regional scales, and further studies of extreme precipitation in rainfed agricultural regions should not be overlooked (Zhang et al. 2011; Bouchard & Qi 2016). Secondly, most previous studies have employed the indices of extreme precipitation proposed by the joint World Meteorological Organization (WMO) Commission for Climatology (CCl)/Climate Variability and Predictability (CLIVAR) Expert Team (ET) on Climate Change Detection, Monitoring and Indices (ETCCDMI) (Zhang et al. 2005; Alexander et al. 2006). These indices enhance the monitoring of climate anomalies and mapping the spatial variability of extreme climatic events with broad spatial coverage (Zhang et al. 2005, 2011; Alexander et al. 2006; Goodess 2013; Vavrus et al. 2015). Although these indices have been widely used to investigate the annual trends of extreme precipitation, the characteristics of extreme precipitation at other timescales are not fully understood. Therefore, it is important to expand application ranges of the extreme precipitation indices to multiple timescales and investigate the multi-timescale features of extreme precipitation for rainfed agricultural regions.

In recent years, there has been an increase in research into extreme precipitation in China, and various fields of researches have been involved in these studies (Chen & Zhai 2015; Liu & Xu 2015; Yuan et al. 2015; Sun et al. 2016). However, most of these studies focused on the annual feature of extreme precipitation. The historical features of extreme precipitation in other timescales have rarely been investigated. Northeast China, which is one of the main grain-producing regions in China, is dominated by rainfed agriculture. The security of grain production in this region will be easily affected by abnormal precipitation events, especially in spring and summer. However, few studies have focused on the features of extreme precipitation in Northeast China at seasonal and crop growth period scales. One example is the study by Han et al. (2015), which investigated changes in summer precipitation and the atmospheric circulation background circulation over the Northeast China. The most relevant studies on these issues have focused on changes in extreme precipitation at annual scales, whereas the potential relationships between precipitation and other factors were not considered (Du et al. 2013; Wang et al. 2013). It is, therefore, necessary to fill these gaps which refer to multiple timescale features of extreme precipitation and the potential relationship between extreme precipitation events and typically atmospheric oscillations in Northeast China.

In order to provide a more comprehensive understanding of extreme precipitation in Northeast China, this study was conducted with the following objectives: (1) to modify the extreme precipitation indices, developed by ETCCDMI to explore extreme precipitation at the annual scale, and expand their applicable scope and timescale for application at multi-timescales to reveal the spatiotemporal characteristics of extreme precipitation events in Northeast China; (2) to originally reveal potential teleconnections among extreme precipitation indices, climate indices and crop failure; and (3) to choose suitable extreme precipitation indices for application in Northeast China.

MATERIALS AND METHODS

Study area

Northeast China is one of the major grain-producing areas of China; most farms of this region produce single season crops, e.g. wheat, maize or rice. This region includes Heilongjiang Province, Jilin Province, Liaoning Province and the eastern part of Inner Mongolia. There are just three cites (Chifeng, Tongliao and Hulunbeir) and Hinggan League of Inner Mongolia, which belong to Northeast China. In consideration of the integrity of administrative region, Heilongjiang, Jilin and Liaoning Provinces have been adopted as case studies (38°43′–53°24′N, 115°20′–135°E) (Figure 1). The total planting area in these three provinces is about 2,538.01 million hectares, which occupies 20.24% of the research area. This area also comprises 19.52% of the total planting area of China and produces 19.00% of China's total food production. From north to south, the climate zones of this region are cold temperate, mid-temperate and warm temperate. The annual average precipitation ranges from 400 to 1,000 mm.
Figure 1

Study area and locations of meteorological stations in the Northeastern China.

Figure 1

Study area and locations of meteorological stations in the Northeastern China.

Data

Daily precipitation data were obtained from the China Meteorological Data Sharing Service System (http://cdc.cma.gov.cn/home.do). To calculate extreme precipitation indices, data from 70 representative meteorological stations in the study area for 54 years (1961–2014) were used (Figure 1); these stations were chosen based on their locations and the length and completeness of their datasets.

Ten climate indices were selected to investigate relationships between extreme precipitation indices and atmospheric circulation. The basic principle of choosing climatic indices is the potential impacts of atmospheric circulation in Northeast China, which has close connection with the geographic location of study area. Based on their definitions, these indices can be divided into two categories: annular mode indices and monsoon indices. Annular mode indices include the Arctic Oscillation Index (AOI), North Atlantic Oscillation (NAO), Southern Oscillation Index (SOI), Indian Ocean Dipole (IOD) (http://www.jamstec.go.jp/frcgc/research/d1/iod/iod/dipole_mode_index.html), Pacific Decadal Oscillation (PDO) and El Niño Southern Oscillation (ENSO). The features of ENSO can be represented by the Niño 3.4 indices, which can be downloaded from the Climate Prediction Center of National Oceanic and Atmospheric Administration (NOAA) website (http://www.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ensoyears.shtml). Four monsoon indices were included in this study. The Indian Summer Monsoon Index (ISMI) and the Western North Pacific Monsoon Index (WNPMI) were obtained from the Monsoon Monitoring Page (http://apdrc.soest.hawaii.edu/projects/monsoon/seasonal-monidx.html). The East Asian Summer Monsoon Index (EASMI) and the South Asian Summer Monsoon Index (SASMI) can be downloaded from http://ljp.gcess.cn.

Methodology

Extreme precipitation indices

Thirteen extreme precipitation indices were used in this study and were classified into three timescales categories: seasonal, crop-growth and annual scales. The seasonal scale includes spring (March–April–May), summer (June–July–August), autumn (September–October–November) and winter (December–January–February). The crop-growing season ranges from April to September because single crops were cultivated in the majority of the research area during this period. A further 13 extreme precipitation indices were used that are classified into three types (amount indices, duration indices and intensity indices) to characterize the amount, duration and intensity of precipitation. The definitions of these indices are shown in Table 1.

Table 1

List of extreme precipitation indices

No. Index Indicator name Definition Unit Category 
R95p Very wet days Annual/seasonal/ the crop growth season /monthly total PRCP when RR > 95th percentile mm Amount index 
R99p Extremely wet days Annual/seasonal/ the crop growth season /monthly total PRCP when RR > 99th percentile mm 
PRCPTOT Total wet-day precipitation Annual/seasonal/ the crop growth season /monthly total PRCP in wet days (RR ≥ 1 mm) mm 
R10 Number of heavy precipitation days Annual/seasonal/ the crop growth season /monthly count of days when PRCP ≥ 10 mm Days Duration index 
R20 Number of very heavy precipitation days Annual/seasonal/ the crop growth season /monthly count of days when PRCP ≥ 20 mm Days 
R25 Number of extreme heavy precipitation days Annual/seasonal/ the crop growth season /monthly count of days when PRCP ≥ 25 mm Days 
CDD Consecutive dry days Maximum number of consecutive days with RR < 1 mm Days 
CWD Consecutive wet days Maximum number of consecutive days with RR ≥ 1 mm Days 
Dry day Dry days index Number of days with PRCP <1 mm Days 
10 Wet day Wet days index Number of days with PRCP ≥ 1 mm Days 
11 SDII Simple daily intensity index Annual/seasonal/ the crop growth season /monthly total precipitation divided by the number of wet days (defined as PRCP ≥ 1.0 mm) in the year mm/day Intensity index 
12 R1day Maximum 1-day precipitation The maximum 1-day value in year/season/ the crop growth season /month mm 
13 R5days Maximum 5-day precipitation Maximum consecutive 5-day precipitation in year/season/ the crop growth season /month mm 
No. Index Indicator name Definition Unit Category 
R95p Very wet days Annual/seasonal/ the crop growth season /monthly total PRCP when RR > 95th percentile mm Amount index 
R99p Extremely wet days Annual/seasonal/ the crop growth season /monthly total PRCP when RR > 99th percentile mm 
PRCPTOT Total wet-day precipitation Annual/seasonal/ the crop growth season /monthly total PRCP in wet days (RR ≥ 1 mm) mm 
R10 Number of heavy precipitation days Annual/seasonal/ the crop growth season /monthly count of days when PRCP ≥ 10 mm Days Duration index 
R20 Number of very heavy precipitation days Annual/seasonal/ the crop growth season /monthly count of days when PRCP ≥ 20 mm Days 
R25 Number of extreme heavy precipitation days Annual/seasonal/ the crop growth season /monthly count of days when PRCP ≥ 25 mm Days 
CDD Consecutive dry days Maximum number of consecutive days with RR < 1 mm Days 
CWD Consecutive wet days Maximum number of consecutive days with RR ≥ 1 mm Days 
Dry day Dry days index Number of days with PRCP <1 mm Days 
10 Wet day Wet days index Number of days with PRCP ≥ 1 mm Days 
11 SDII Simple daily intensity index Annual/seasonal/ the crop growth season /monthly total precipitation divided by the number of wet days (defined as PRCP ≥ 1.0 mm) in the year mm/day Intensity index 
12 R1day Maximum 1-day precipitation The maximum 1-day value in year/season/ the crop growth season /month mm 
13 R5days Maximum 5-day precipitation Maximum consecutive 5-day precipitation in year/season/ the crop growth season /month mm 

PRCP is the daily precipitation amount; is the daily precipitation amount on day i in period j.

Modified Mann–Kendall trend test

The statistical distribution of extreme precipitation index is not fixed. In order to improve the reliability of trend test, the nonparametric and automatic features of extreme precipitation indices have been fully considered. The Mann–Kendall (MK) trend test is a nonparametric trend test and has been widely applied for hydrometeorological data (Zhang et al. 2014; Sun et al. 2015). However, previous research has indicated that precipitation exhibits self-correlation, and this component may impact the results of this trend test for rainfall (Hamed & Rao 1998; Hamed 2008; Hamed 2009). The modified Mann–Kendall (MMK) trend test was developed to avoid the effects of self-correlation by varying the calculation method of variance (Hamed & Rao 1998). For a time series , the statistic S can be obtained with Equation (1): 
formula
1
where 
formula
2
represents the data values at times i and j, and n is the number of observational data. The empirical formula for the variance of S that may be impacted by autocorrelation of data is: 
formula
3
where indicates the correction that originates from autocorrelation in the data. 
formula
4
is the autocorrelation formula of the rank of the observational data. The significance of the observational data was ascertained by comparing the results for the standardized test Z with the standardized normal variate at the 0.05 significance level. The range of the trend (β) is the median of the set of slopes (Equation (5)): 
formula
5
The nonparametric change point test method should be applied to detect the change points of each extreme climate index. Pettitt's test (Pettitt 1979) is a nonparametric test that has been widely applied to detect change points in climatic records (Reiter et al. 2012; Duhan & Pandey 2013) and was selected to explore abrupt changes in the extreme climate indices. This method is also able to confirm the sustainability of a change in trend. It is also necessary to analyze the spatial distinctions of the extreme precipitation index, which can provide an overview of a region at high risk for extreme climatic events. In order to show the spatial characteristics of the typical indices, the Inverse Distance Weighting method (IDW) was chosen to present the spatial distributions of extreme precipitation indices based on stationary data. Pearson correlation, Kendall rank correlation, Spearman correlation and the point-biserial correlation have been widely utilized to test the correlation relationship (Gholami et al. 2015). In view of the uncertainty distribution of the extreme precipitation index, Spearman's rank correlation was employed to reveal potential relationships among the extreme precipitation indices, climate indices and comprehensive crop failure ratios of agricultural drought disasters (C index).

Comprehensive crop failure ratios of agricultural drought disasters (C index)

To reveal the potential relationships between extreme precipitation events and agricultural drought, an index that can directly present losses of agricultural drought should be adopted. The comprehensive crop failure ratios of agricultural drought disasters (the C index), which was proposed by the Office of State Flood Control and Drought Relief Headquarters in 2006, can directly represent the influence of drought on crop production in a region. So the C index was adopted in this study. Details of this index were presented in the Drought Assessment Criteria (http://www.dyswater.gov.cn/News/News.asp?id=1477). The C index can be obtained with Equation (6): 
formula
6
where is the ratio between the drought-affected area and the sowing area, is the ratio of the drought-inundated area to the sowing area and is the ratio of the total crop failure area to the sowing area. If C is greater than 0.4, the region has suffered an extreme drought disaster. C value between 0.3 and 0.4 indicates a heavy drought disaster. Moderate and weak feeble droughts correspond to C values of 0.2–0.3 and 0.1–0.2, respectively. If the value of C is below 0.1, there is no drought in the evaluated area. All data used to calculate C index values were obtained from the Ministry of Agriculture of the People's Republic of China (http://202.127.42.157/moazzys/nongqing.aspx).

RESULTS AND DISCUSSION

Temporal variation of extreme precipitation

Amount indices

At the seasonal level, very wet days (R95p), extremely wet days (R99p), and total precipitation (PRCPTOT) presented similar change trends, although the fluctuation ranges varied (Table 2). In spring, significant rises in R95p and R99p were detected at rates of 1.40 and 0.87 mm/decade, respectively. Meanwhile, the magnitude of trend for PRCPTOT was 1.65 mm/decade, but the change was not significant. In the remaining seasons, the change trends of the three amount indices were negative, and significant decreasing trends were detected for autumn and winter. For the growing season, these amount indices also showed declining trends, but only the downward trend of PRCPTOT was verified as significant at −7.09 mm/decade. Annually, R95p, R99p and PRCPTOT showed significant decreasing trends at −8.78 mm/decade, −4.43 mm/decade and −12.17 mm/decade, respectively. The trend results for PRCPTOT at the annual level were similar to those reported in previous studies (Zhang et al. 2012b; Fu et al. 2013; Han et al. 2015). The values of the amount indices from 1961–2014 are given in Figure 2. For the different timescales, the amounts of precipitation ranked from highest to lowest were the annual, growing season, summer, autumn, spring and winter amounts.
Table 2

Trend per decade and the significant mutation year of the amount indices in the Northeastern China during 1961–2014

Test Timescale R95p R99p PRCPTOT 
MMK trend test Spring 1.404* 0.874*** 1.648 
Summer −3.919 −1.764 −4.688 
Autumn −4.084*** −1.356 −5.887*** 
Winter −2.872*** −1.178** −3.269*** 
Growth season −5.428 −3.063 −7.091* 
Annual −8.785* −4.418* −12.168* 
Significant mutation year (Pettitt) Spring   2007 
Summer    
Autumn    
Winter 2000 1969  
Growth season    
Annual    
Test Timescale R95p R99p PRCPTOT 
MMK trend test Spring 1.404* 0.874*** 1.648 
Summer −3.919 −1.764 −4.688 
Autumn −4.084*** −1.356 −5.887*** 
Winter −2.872*** −1.178** −3.269*** 
Growth season −5.428 −3.063 −7.091* 
Annual −8.785* −4.418* −12.168* 
Significant mutation year (Pettitt) Spring   2007 
Summer    
Autumn    
Winter 2000 1969  
Growth season    
Annual    

*Significance level 0.1; **significance level 0.05; ***significance level 0.01.

Figure 2

Time series and smoothing averages of R95p (Very wet days), R99p (Extremely wet days) and PRCPTOT (Total wet-day precipitation) in multi-timescale during 1961–2014.

Figure 2

Time series and smoothing averages of R95p (Very wet days), R99p (Extremely wet days) and PRCPTOT (Total wet-day precipitation) in multi-timescale during 1961–2014.

In addition to analysis of the trends of the amount indices, the contributions of these trends to annual PRCPTOT were also evaluated. The proportion of spring R95p to annual PRCPTOT shows non-significant increasing trends during 1961–2014. The rates of R95p for the other timescales relative to the annual PRCPTOT decline with non-significant trends. The ratio of R99p to annual PRCPTOT showed upward trends for most timescales, except for autumn and winter. The contributions of seasonal/growing season PRCPTOT to annual PRCPTOT were also explored to further characterize extreme wet events. The results indicated that the spring PRCPTOT accounted for 15.1% of annual PRCPTOT (within the range of 11.7–20.8%), and this ratio increased at a rate of 0.323% per decade at the 0.1 significance level. However, the proportions of summer, autumn, winter and growing season PRCPTOT to annual PRCPTOT showed downward, non-significant trends, with average ratios of 47.1%, 23.6%, 14.1% and 68.2%, respectively.

Based on the abovementioned results, the temporal features of the amount indices at these six timescales reveal several aspects of the characteristics of extreme wet events. First, the values of extreme wet events (represented by R95p and R99p) increased significantly in spring during 1961–2014, which illustrates that Northeast China has suffered more severe extreme wet events in spring over 1961–2014. However, such trends were negative at all other timescales. Secondly, the impacts of extreme wet events, as indicated by the contributions of the amount indices to annual PRCPTOT, have slightly increased at all timescales, except for autumn and winter. These findings also demonstrate that Northeast China had endured more severe extreme wet events in spring, summer, the growing season and annually over the course of the evaluated period of time.

Duration indices

Seven duration indices were applied that can be divided into two categories based on their definitions, which demonstrate the capability of climate indices to reflect the features of extreme wet or dry events. The numbers of heavy precipitation days (R10), very heavy precipitation days (R20) and extremely heavy precipitation days (R25), as well as consecutive wet days (CWD) and Wet day, can represent the durations of extreme wet events. Another type of duration index includes consecutive dry days (CDD) and Dry day, which can demonstrate the features of extreme dry events.

Table 3 provides the trend test results on the basis of MMK test for all the duration indices. Seasonally, R10, R20, R25, CWD and Wet day have similar increasing and decreasing trends over time, but the slopes of these trends were distinct for each season. The changing trends of these indices demonstrate that the duration of extreme wet events tended to increase over 1961–2014 in spring, whereas the trend was negative for other seasons. Particularly, the duration of extreme wet events appeared to decline severely in autumn and winter. Furthermore, marked drops in the area means of R10, R20, R25, CWD and Wet day were detected both in the growing season and annually, and the trends for these indices at the annual scale were significant at the 0.1 level. The fluctuations of these duration indices are shown in Figures 35.
Table 3

Trend per decade and the significant mutation year of the duration indices in the Northeastern China during 1961–2014

Test Timescale R10 R20 R25 CDD CWD Dry day Wet day 
MMK trend test Spring 0.060* 0.024 0.016 0.090 0.015 −0.086 0.088 
Summer −0.084 −0.042 −0.031 0.147 −0.049 0.355* −0.375* 
Autumn −0.190*** −0.093*** −0.048** 0.299 −0.088*** 0.476*** −0.476*** 
Winter −0.118*** −0.068*** −0.043*** −0.149 −0.028** 0.137 −0.235*** 
Growth season −0.162 −0.083 −0.062 −0.117 −0.051** 0.577*** −0.599** 
Annual −0.345* −0.181* −0.128* 0.363 −0.071** 0.815*** −0.814** 
Significant mutation year (Pettitt) Spring 2007   1976  2004 2004 
Summer       1997 1997 
Autumn 1998 1996    1997 1999 1999 
Winter  2005 2000     
Growth season    1985  1996 1996 
Annual    2013    
Test Timescale R10 R20 R25 CDD CWD Dry day Wet day 
MMK trend test Spring 0.060* 0.024 0.016 0.090 0.015 −0.086 0.088 
Summer −0.084 −0.042 −0.031 0.147 −0.049 0.355* −0.375* 
Autumn −0.190*** −0.093*** −0.048** 0.299 −0.088*** 0.476*** −0.476*** 
Winter −0.118*** −0.068*** −0.043*** −0.149 −0.028** 0.137 −0.235*** 
Growth season −0.162 −0.083 −0.062 −0.117 −0.051** 0.577*** −0.599** 
Annual −0.345* −0.181* −0.128* 0.363 −0.071** 0.815*** −0.814** 
Significant mutation year (Pettitt) Spring 2007   1976  2004 2004 
Summer       1997 1997 
Autumn 1998 1996    1997 1999 1999 
Winter  2005 2000     
Growth season    1985  1996 1996 
Annual    2013    

*Significance level 0.1; **significance level 0.05; ***significance level 0.01.

Figure 3

Time series and smoothing averages of R10 (Number of heavy precipitation days), R20 (Number of very heavy precipitation days) and R25 (Number of extreme heavy precipitation days) in multi-timescale during 1961–2014.

Figure 3

Time series and smoothing averages of R10 (Number of heavy precipitation days), R20 (Number of very heavy precipitation days) and R25 (Number of extreme heavy precipitation days) in multi-timescale during 1961–2014.

Figure 4

Time series and smoothing averages of CWD (Consecutive wet days) and CDD (Consecutive dry days) in multi-timescale during 1961–2014.

Figure 4

Time series and smoothing averages of CWD (Consecutive wet days) and CDD (Consecutive dry days) in multi-timescale during 1961–2014.

Figure 5

Inter-annual variation of Wet day and Dry day in multi–timescale during 1961–2014.

Figure 5

Inter-annual variation of Wet day and Dry day in multi–timescale during 1961–2014.

The regional average of CDD showed declining trends for winter and the growing season. For other timescales, there was a non-significant increasing tendency. The change trends of CDD and CWD were opposite at each timescale except for spring, and the Dry day trends for all timescales were opposite to the corresponding Wet day trends (Figure 5).

The values of the duration indices (R10, R20, R25, CWD and Wet day), which reflects extreme wet events, was far less than a third of CDD in all timescales. The mean proportions of regional average Dry day to 365/366 for all six timescales (spring, summer, autumn, winter, growing season and annual) were 21.7%, 18.8%, 20.5%, 21.8%, 39.5% and 82.8%. The ratios of annual Dry day to a year showed a decreasing trend at the 0.05 significance level in spring, and an increasing trend at the 0.05 significance level was detected at the annual scale over 1961–2014. This finding indicates that Northeast China faced more extreme dry events than extreme wet events, and that the occurrence frequency of these events showed increasing trends at all timescales, except for the downward trends detected in spring and winter. This also manifests that the risks of agricultural drought were getting higher in this area.

Intensity indices

To reveal the trends of change in intensity indices, the MMK test was employed to detect trends in the simple daily intensity index (SDII), maximum 1-day precipitation (R1day) and maximum 5-day precipitation (R5days) at multiple timescales (seasonal, growing season and annual); the results of the trend tests are displayed in Table 4. The regional average SDII decreased in autumn and annually (Figure 6); the downward trend of autumn SDII in particular is significant at the 0.1 significance level (−0.12 mm/decade). For other timescales, there were no significant upward trends in the SDII. The trend of the SDII is widely regarded as dominated by precipitation and the Wet day of the corresponding timeframe. The trends in PRCPTOT and the Wet day present analogous tendencies at the same timescale with different significance levels. Therefore, the increase in the spring SDII was caused by the more notable increase in precipitation than the Wet day. The decrease in the autumn and annual SDII was caused by the severe decrease in PRCPTOT, whereas the increases of the SDII in summer, winter and the growing season were caused by greater decline of the Wet day. These findings further illustrate that Northeast China underwent more intense extreme precipitation as well as longer extreme dry events over time from 1961 to 2014.
Table 4

Trend per decade and the significant mutation year of the intensity indices in the Northeastern China during 1961–2014

Test Timescale SDII (mm/day) R1day (mm) R5days (mm) 
MMK trend test Spring 0.070 0.399 0.526 
Summer 0.055 −0.598 −1.855** 
Autumn −0.121* −0.788*** −1.809*** 
Winter 0.011 −0.373* −0.903*** 
Growth season 0.022 −0.828* −2.040** 
Annual −0.039 −0.9785* −2.3292** 
Significant mutation year (Pettitt) Spring    
Summer 1994   
Autumn    
Winter    
Growth season    
Annual   1967 
Test Timescale SDII (mm/day) R1day (mm) R5days (mm) 
MMK trend test Spring 0.070 0.399 0.526 
Summer 0.055 −0.598 −1.855** 
Autumn −0.121* −0.788*** −1.809*** 
Winter 0.011 −0.373* −0.903*** 
Growth season 0.022 −0.828* −2.040** 
Annual −0.039 −0.9785* −2.3292** 
Significant mutation year (Pettitt) Spring    
Summer 1994   
Autumn    
Winter    
Growth season    
Annual   1967 

*Significance level 0.1; **significance level 0.05; ***significance level 0.01.

Figure 6

Inter-annual variation of SDII (Simple daily intensity index) in multi-timescale during 1961–2014.

Figure 6

Inter-annual variation of SDII (Simple daily intensity index) in multi-timescale during 1961–2014.

The regional average of R1day and R5days indices show similar downward trends for all timescales except for spring (Figure 7 and Table 4). The ratios of R1day and R5days to PRCPTOT in the corresponding timescales are displayed in Table 5. These results indicate that the R1day and R5days are important components of PRCPTOT, as these extreme wet events usually cause flooding. In spring, the ratios of these two indices to the corresponding PRCPTOT values showed slight decreasing trends from 1961 to 2014. The ratios of R1day at the other timescales indicated increasing trends; these trends were found to be significant (0.05 level) for summer and the growing season in particular. The contributions of the R5days index to PRCPTOT in summer, autumn, winter and the growing season showed slow increases, and the annual R5days values showed a negative trend without statistical significance.
Table 5

Rates of R1day and R5days to PRCPTOT at corresponding timescale in the Northeastern China

Timescale R1day (%)
 
R5days (%)
 
Mean Maximum Minimum Mean Maximum Minimum 
Spring 24.9% 31.2% 19.5% 37.0% 46.2% 28.8% 
Summer 17.3% 21.9% 13.8% 27.3% 32.8% 23.0% 
Autumn 22.6% 26.9% 18.1% 33.6% 40.6% 27.7% 
Winter 21.3% 39.2% 16.6% 32.5% 56.8% 19.7% 
Growth season 13.4% 16.8% 10.8% 21.1% 25.3% 16.4% 
Annual 11.0% 14.0% 8.9% 17.4% 20.6% 14.5% 
Timescale R1day (%)
 
R5days (%)
 
Mean Maximum Minimum Mean Maximum Minimum 
Spring 24.9% 31.2% 19.5% 37.0% 46.2% 28.8% 
Summer 17.3% 21.9% 13.8% 27.3% 32.8% 23.0% 
Autumn 22.6% 26.9% 18.1% 33.6% 40.6% 27.7% 
Winter 21.3% 39.2% 16.6% 32.5% 56.8% 19.7% 
Growth season 13.4% 16.8% 10.8% 21.1% 25.3% 16.4% 
Annual 11.0% 14.0% 8.9% 17.4% 20.6% 14.5% 
Figure 7

Inter-annual variation of R1day (Maximum 1-day precipitation) and R5days (Maximum 5-day precipitation) in multi–timescale during 1961–2014.

Figure 7

Inter-annual variation of R1day (Maximum 1-day precipitation) and R5days (Maximum 5-day precipitation) in multi–timescale during 1961–2014.

The results of the trend test for the intensity indices at multiple timescales indicate that the frequency of extreme precipitation increased only in spring, whereas the contribution of extreme precipitation to spring precipitation showed a slight decrease. Furthermore, the proportions of R1day and R5days to total precipitation showed significant upward trends in summer and the growing season, which means that particularly in these seasons, Northeast China experienced more severe extreme precipitation over time.

The temporary features of extreme indices presented that higher frequency of extreme wet events were detected in spring. However, the extreme dry events were more prevalent at the other timescales. Generally, Northeast China was suffering drier condition that could lead to higher frequency of severe drought events.

Abrupt change analysis

The mutation test results for the extreme amount indices indicate that significant abruption in winter R95p (2000), winter R99p (1969) and spring PRCPTOT (2007) (Table 2). The trends of R95p and R99p were found to show more notable decreasing trends after the abruption years. There was also a sharper increasing trend of spring PRCPTOT after 2007. The significant change points of the duration indices are shown in Table 3. Similar to spring PRCPTOT, there was also a more noticeable rise in spring R10 after 2007. For autumn, lower increases in R10 and R20 were detected after the significant change points. The significant change years for R20 and R25 were 2005 and 2000, respectively, and the trends for these indices showed steep declines following these mutation years. For CDD, significant change points were detected for spring (1976), the growing season (1985) and the annual scale (2013). It is notable that this significant change point for annual CDD was detected for 2013, and that a sharper increase in CDD was discovered for 2014, while Northeast China had been experiencing its worst drought in 60 years (Wang & He 2015). The significant change points were the same for the Dry day and Wet day for the same timescale because the properties of these indices are opposite. Significant abruption of the Dry day has not been detected for the winter or annual timescales; it can therefore be inferred that the Dry day presented more a continuous increasing trend for these scales. Overall, the amount indices (R95p, R99p and PRCPTOT) and intensity indices (SDII, R1day and R5days) showed more consistent increasing/decreasing tendencies than the duration indices, for which mutation features were more significant over 1961–2014 in autumn and winter.

The results of change point tests can manifest the consistent trend of extreme precipitation indices. It is clear that Northeast China faced continuously increasing extreme wet events in spring. On the whole, amount indices and intensity indices displayed unanimous trends during 1961–2014.

Spatial features of extreme precipitation

Northeast China is one of the most important grain-producing areas in China and is dominated by single-season crop farming. The growing season includes key periods for these crops, and anomalous precipitation during this season can be devastating to agricultural production. Therefore, the vital importance of precipitation in the growing season is clear compared to the other timescales based on its potential impacts on agriculture. Precipitation during the growing season (April–September) may reflect a combination of the trends in precipitation of the spring, summer and autumn. The temporal features of precipitation described above illustrate that extreme precipitation in winter has been found to fluctuate less than in other seasons; therefore, more attention should be given to other seasons in future studies. In order to highlight the influence of extreme precipitation in the growing season on agricultural production, the spatial analysis of all the extreme precipitation indices evaluated were focused on the growing season.

Amount indices

The values of R95p, R99p and PRCPTOT have similar spatial distributions in which values gradually decrease from south to north (Figure 8). Decreasing trends of R95 and PRCPTOT were detected at more stations than R99p; increasing trends were detected at 42.9% and 44.3% of stations, respectively. For R95p, the proportion of stations with increasing trends (42.9%) is markedly higher than the number with downward trends (31.4%), which means that the study area suffered more severe extreme wet events during 1961–2014. For the major meteorological stations, both increasing and decreasing trends were detected, and the proportions of trends significant at a level below 0.1 was around 10%, except for the proportion of increasing trends in PRCPTOT. It is clear that the significant trends in Northeast China have not yet been detected by most stations. However, a remarkable phenomenon is that the wetter regions (Liaoning Province and Jilin Province) were found to have lower increases in magnitude than the drier regions, whereas the trend for the center of Hei Longjiang Province was found to have a higher slope and significance level. However, the decrease in regional PRCPTOT can be explained by the decreased PRCPTOT in Liaoning and Jilin Provinces.
Figure 8

Spatial distribution of the mean annual amount indices and the trend of each meteorological station in the growth season in the Northeastern China.

Figure 8

Spatial distribution of the mean annual amount indices and the trend of each meteorological station in the growth season in the Northeastern China.

Duration indices

The spatial distribution of mean values of the duration indices in the growing season are presented in Figure 9. The mean annual values of R10, R20 and R25 in the growing season decreased gradually from the south to north, and the most notable decreasing trends were detected in Liaoning Province. Moreover, the trend rangeability of the three indices was clearer in Hei Longjiang where the average values of R10, R20 and R25 were lower than those in the other two provinces. In addition, it was easily determined that the duration of extreme wet conditions in Northeast China, represented by the duration indices (R10, R20, R25 and CWD), are far shorter than the duration of dry conditions, which are shown by the CDD and Dry day. Based on comparison of the trend slopes for the duration indices, larger increases in the intensity of dry periods were concentrated in the southern region, whereas larger increases in the wet durations was detected in the north, especially on the Songnen Plain, one of primary grain-producing areas of Northeast China.
Figure 9

Spatial distribution of the mean annual duration indices and the trend of each meteorological station in the growth season in the Northeastern China.

Figure 9

Spatial distribution of the mean annual duration indices and the trend of each meteorological station in the growth season in the Northeastern China.

Intensity indices

The spatial distributions of the intensity indices are given in Figure 10. The mean values of the intensity indices in the growing season show similar spatial distribution patterns with values gradually decreasing from south to north. The mean values of the SDII in the growing season ranged from 3.65 mm/day to 17.42 mm/day, and increased over 1961–2014 at 70% of the stations in Northeast China. Meteorological stations with significant increasing trends were mainly those in Liaoning Province and the western part of Hei Longjiang Province; the proportion of significant increasing trends (17.10%) was notably higher than that of significant decreasing trends (2.9%). At the regional scale, the increase of the SDII resulted from a decrease in dry days and increase in PRCPTOT in the growing season, especially in Liaoning Province where the local contributions to regional PRCPTOT were high compared to the other two provinces. The mean values of R1day (9.61 to 116.55 mm) and R5days (12.52 to 199.4 mm) in the growing season decreased from south to north with an opposite trend to the slope. It was clear that the impact of the intensity of extreme precipitation events is highest in Liaoning Province and lowest in Hei Longjiang Province, while the change trends of intensity presented inverse tendencies in these areas.
Figure 10

Spatial distribution of the mean annual intensity indices and the trend of each meteorological station in the growth season in the Northeastern China.

Figure 10

Spatial distribution of the mean annual intensity indices and the trend of each meteorological station in the growth season in the Northeastern China.

Overall, there are three obviously spatial distribution features of extreme precipitation that were shown by the indices. First, the multiple mean annual amount indices and intensity indices gradually decreased from south to north in Northeast China. Second, the wetter region was becoming drier and the drier region was suffering a higher frequency of extreme wet events. Third, the higher frequency of extreme wet events was mainly concentrated in Songnen Plain.

Correlation analysis

The potential correspondence relationships of 13 extreme precipitation indices were explored with Spearman's rank correlation for all six timescales (spring, summer, autumn, winter, growing season and annual). The amount indices (R95p, R99p and PRCPTOT), which have similar correlation features, showed significant positive correlations with the SDII, R1day, R5days, R10, R20 and R25, except that the winter positive correlations were below 0.22 and non-significant. The eight duration indices are divided into two categories: those that represent extreme wet events (R10, R20, R25, CWD and Wet day) and those that represent extreme dry events (CDD and Dry day). R10, R20 and R25 have similar features at all the timescales; the differences between them are in their fluctuating ranges. The correlations between the intensity indices (SDII, R1day and R5days) are always significantly positive except in winter. The intensity indices presented significantly positive correlations with R10, R20, R25, R95p, R99p and PRCPTOT, except in winter. This finding can account for the contributions of seasonal precipitation to annual precipitation. Generally, the correlations between the intensity indices and other indices were positive with the indices associated with precipitation days (R10, R20, R25, CWD and Wet day) and the amount of precipitation (R95p, R99p PRCPTOT). The correlations among all extreme precipitation indices were mainly determined based on the definitions of each index. Note that the CDD and CWD did not correlate significantly with the other indices because their values present more randomness than other indices between different years. After comparing the correlations between these indices for different timescales, it was further confirmed that the correlation features of these indices express the relationships between precipitation and other factors and that these relationships would not change easily with the timescales of these indices.

Correlations between extreme precipitation indices and climate indices

R95p, R99p, PRCPTOT, R10, Wet day, SDII and R5days were used to evaluate relationships with climate indices based on the inter-correlations among the 13 extreme indices (Table 6). The 10 climate indices used in this study can be divided to annular mode indices (AOI, NAO, IOD, PDO, SOI and ENSO3.4) and monsoon indices (ISMI, WNPMI, EASMI and SASMI). The correlations between the extreme precipitation indices and annular mode indices were notably closer than the correlations between the extreme precipitation indices and the monsoon indices. Significant correlation was detected only between the AOI, NAO and SOI and specific extreme precipitation indices for certain timescales. Generally, all seven extreme precipitation indices correlated positively with the AOI in spring and winter, and these correlations were negative at the remaining timescales. However, significant correlations with the AOI were detected only with the annual Wet winter day (−2.91) and the winter SDII (0.279). The NAO correlated positively with the amount indices (R95p and PRCPTOT) and the Wet day duration index in winter. At the other timescales, significant correlations between the NAO and extreme precipitation indices were not detected; the correlation coefficients ranged from −0.134 to 0.267. The SOI showed significant positive correlation with only with one duration index (R10) in spring.

Table 6

Correlation coefficients between the extreme precipitation indices and climatic indices

Climatic index Timescale R95p R99p PRCPTOT R10 Wet day SDII R5days 
AOI Spring 0.094 0.03 0.108 0.122 0.073 0.07 0.017 
Summer −0.183 −0.109 −0.198 −0.253 −0.24 −0.071 −0.162 
Autumn −0.093 −0.033 −0.148 −0.165 −0.227 −0.019 −0.072 
Winter 0.069 0.032 0.095 0.109 0.119 0.279* −0.084 
Growth season −0.126 −0.147 −0.152 −0.167 −0.196 −0.09 −0.127 
Annual −0.174 −0.16 −0.231 −0.245 − 0.291* −0.125 −0.122 
NAO Spring 0.027 0.079 0.008 −0.022 −0.035 −0.042 −0.024 
Summer −0.134 −0.124 −0.13 −0.128 −0.077 −0.129 −0.084 
Autumn 0.064 0.129 0.024 0.008 −0.078 0.098 0.084 
Winter 0.288* 0.267 0.320* 0.263 0.272* 0.169 0.033 
Growth season −0.057 −0.104 −0.067 −0.072 −0.052 −0.086 −0.023 
Annual −0.056 −0.054 −0.092 −0.115 −0.099 −0.012 0.041 
IOD Spring 0.137 0.206 0.13 0.128 0.068 0.251 0.188 
Summer −0.056 0.068 −0.092 −0.173 −0.22 0.185 0.038 
Autumn −0.024 −0.013 −0.05 −0.013 −0.034 0.019 −0.029 
Winter −0.185 −0.178 −0.196 −0.122 −0.144 0.04 −0.157 
Growth season −0.02 0.039 −0.094 −0.097 −0.178 0.142 0.021 
Annual −0.043 −0.008 −0.072 −0.092 −0.162 0.002 −0.022 
PDO Spring −0.034 0.038 −0.019 −0.08 0.046 −0.099 −0.024 
Summer −0.021 −0.065 −0.036 0.089 −0.057 0.062 −0.138 
Autumn −0.046 −0.144 −0.034 −0.038 0.031 −0.036 0.042 
Winter 0.092 0.075 0.09 0.055 −0.052 −0.11 −0.021 
Growth season −0.03 −0.055 −0.039 −0.002 −0.073 0.069 −0.088 
Annual −0.024 −0.035 −0.017 0.005 −0.012 0.023 −0.028 
SOI Spring 0.189 −0.037 0.223 0.272* 0.174 0.113 0.081 
Summer 0.011 0.073 0.027 −0.051 0.035 −0.006 0.05 
Autumn −0.043 −0.057 −0.054 −0.011 −0.127 −0.084 −0.215 
Winter −0.041 0.019 −0.09 −0.077 −0.044 0.19 −0.033 
Growth season 0.035 0.039 0.068 0.046 0.118 −0.011 
Annual 0.007 0.022 −0.014 −0.001 −0.065 0.015 −0.048 
ENSO3.4 Spring −0.128 0.075 −0.201 −0.198 −0.235 −0.029 −0.021 
Summer −0.05 −0.101 −0.055 0.011 −0.029 −0.024 −0.083 
Autumn 0.042 0.047 0.068 0.029 0.137 0.08 0.21 
Winter 0.019 −0.006 0.084 0.04 0.102 −0.107 −0.011 
Growth season −0.062 −0.067 −0.088 −0.06 −0.12 0.008 −0.041 
Annual −0.025 −0.037 0.001 −0.003 0.07 −0.036 0.008 
ISMI Spring −0.004 −0.088 −0.03 −0.017 −0.064 0.013 −0.103 
Summer 0.064 0.013 0.045 0.028 −0.063 0.098 0.05 
Autumn 0.176 0.152 0.086 0.095 −0.092 0.266 0.158 
Winter 0.078 0.013 0.021 0.082 −0.047 0.08 0.043 
Growth season 0.081 0.027 0.056 0.029 −0.039 0.05 0.011 
Annual 0.091 0.027 0.042 0.027 −0.131 0.145 0.002 
WNPMI Spring −0.173 −0.064 −0.177 −0.209 −0.147 −0.134 −0.105 
Summer −0.033 −0.004 −0.036 −0.071 −0.001 −0.054 −0.024 
Autumn −0.057 0.011 −0.035 −0.038 0.005 −0.064 −0.013 
Winter 0.082 0.104 0.185 0.082 0.268 0.107 0.001 
Growth season −0.077 −0.045 −0.09 −0.103 −0.056 −0.083 −0.043 
Annual −0.079 −0.057 −0.08 −0.11 0.009 −0.123 −0.026 
EASMI Spring −0.186 −0.114 −0.157 −0.196 −0.141 −0.162 −0.127 
Summer 0.024 0.117 −0.001 −0.103 −0.049 0.007 0.127 
Autumn 0.043 0.114 0.057 0.057 0.046 −0.042 −0.012 
Winter 0.117 0.067 0.152 0.195 0.099 0.018 0.072 
Growth season −0.019 0.083 −0.036 −0.08 −0.049 −0.036 0.098 
Annual 0.01 0.069 0.013 −0.038 0.014 0.014 0.135 
SASMI Spring −0.068 −0.18 −0.065 −0.029 −0.106 −0.03 −0.153 
Summer 0.104 0.07 0.064 −0.002 −0.025 0.068 0.014 
Autumn 0.041 0.035 −0.01 0.036 −0.082 0.103 0.003 
Winter 0.042 −0.03 0.023 0.128 0.007 0.101 −0.065 
Growth season 0.049 0.053 0.018 0.009 −0.039 −0.014 −0.068 
Annual 0.052 −0.009 −0.002 −0.003 −0.135 0.046 −0.082 
Climatic index Timescale R95p R99p PRCPTOT R10 Wet day SDII R5days 
AOI Spring 0.094 0.03 0.108 0.122 0.073 0.07 0.017 
Summer −0.183 −0.109 −0.198 −0.253 −0.24 −0.071 −0.162 
Autumn −0.093 −0.033 −0.148 −0.165 −0.227 −0.019 −0.072 
Winter 0.069 0.032 0.095 0.109 0.119 0.279* −0.084 
Growth season −0.126 −0.147 −0.152 −0.167 −0.196 −0.09 −0.127 
Annual −0.174 −0.16 −0.231 −0.245 − 0.291* −0.125 −0.122 
NAO Spring 0.027 0.079 0.008 −0.022 −0.035 −0.042 −0.024 
Summer −0.134 −0.124 −0.13 −0.128 −0.077 −0.129 −0.084 
Autumn 0.064 0.129 0.024 0.008 −0.078 0.098 0.084 
Winter 0.288* 0.267 0.320* 0.263 0.272* 0.169 0.033 
Growth season −0.057 −0.104 −0.067 −0.072 −0.052 −0.086 −0.023 
Annual −0.056 −0.054 −0.092 −0.115 −0.099 −0.012 0.041 
IOD Spring 0.137 0.206 0.13 0.128 0.068 0.251 0.188 
Summer −0.056 0.068 −0.092 −0.173 −0.22 0.185 0.038 
Autumn −0.024 −0.013 −0.05 −0.013 −0.034 0.019 −0.029 
Winter −0.185 −0.178 −0.196 −0.122 −0.144 0.04 −0.157 
Growth season −0.02 0.039 −0.094 −0.097 −0.178 0.142 0.021 
Annual −0.043 −0.008 −0.072 −0.092 −0.162 0.002 −0.022 
PDO Spring −0.034 0.038 −0.019 −0.08 0.046 −0.099 −0.024 
Summer −0.021 −0.065 −0.036 0.089 −0.057 0.062 −0.138 
Autumn −0.046 −0.144 −0.034 −0.038 0.031 −0.036 0.042 
Winter 0.092 0.075 0.09 0.055 −0.052 −0.11 −0.021 
Growth season −0.03 −0.055 −0.039 −0.002 −0.073 0.069 −0.088 
Annual −0.024 −0.035 −0.017 0.005 −0.012 0.023 −0.028 
SOI Spring 0.189 −0.037 0.223 0.272* 0.174 0.113 0.081 
Summer 0.011 0.073 0.027 −0.051 0.035 −0.006 0.05 
Autumn −0.043 −0.057 −0.054 −0.011 −0.127 −0.084 −0.215 
Winter −0.041 0.019 −0.09 −0.077 −0.044 0.19 −0.033 
Growth season 0.035 0.039 0.068 0.046 0.118 −0.011 
Annual 0.007 0.022 −0.014 −0.001 −0.065 0.015 −0.048 
ENSO3.4 Spring −0.128 0.075 −0.201 −0.198 −0.235 −0.029 −0.021 
Summer −0.05 −0.101 −0.055 0.011 −0.029 −0.024 −0.083 
Autumn 0.042 0.047 0.068 0.029 0.137 0.08 0.21 
Winter 0.019 −0.006 0.084 0.04 0.102 −0.107 −0.011 
Growth season −0.062 −0.067 −0.088 −0.06 −0.12 0.008 −0.041 
Annual −0.025 −0.037 0.001 −0.003 0.07 −0.036 0.008 
ISMI Spring −0.004 −0.088 −0.03 −0.017 −0.064 0.013 −0.103 
Summer 0.064 0.013 0.045 0.028 −0.063 0.098 0.05 
Autumn 0.176 0.152 0.086 0.095 −0.092 0.266 0.158 
Winter 0.078 0.013 0.021 0.082 −0.047 0.08 0.043 
Growth season 0.081 0.027 0.056 0.029 −0.039 0.05 0.011 
Annual 0.091 0.027 0.042 0.027 −0.131 0.145 0.002 
WNPMI Spring −0.173 −0.064 −0.177 −0.209 −0.147 −0.134 −0.105 
Summer −0.033 −0.004 −0.036 −0.071 −0.001 −0.054 −0.024 
Autumn −0.057 0.011 −0.035 −0.038 0.005 −0.064 −0.013 
Winter 0.082 0.104 0.185 0.082 0.268 0.107 0.001 
Growth season −0.077 −0.045 −0.09 −0.103 −0.056 −0.083 −0.043 
Annual −0.079 −0.057 −0.08 −0.11 0.009 −0.123 −0.026 
EASMI Spring −0.186 −0.114 −0.157 −0.196 −0.141 −0.162 −0.127 
Summer 0.024 0.117 −0.001 −0.103 −0.049 0.007 0.127 
Autumn 0.043 0.114 0.057 0.057 0.046 −0.042 −0.012 
Winter 0.117 0.067 0.152 0.195 0.099 0.018 0.072 
Growth season −0.019 0.083 −0.036 −0.08 −0.049 −0.036 0.098 
Annual 0.01 0.069 0.013 −0.038 0.014 0.014 0.135 
SASMI Spring −0.068 −0.18 −0.065 −0.029 −0.106 −0.03 −0.153 
Summer 0.104 0.07 0.064 −0.002 −0.025 0.068 0.014 
Autumn 0.041 0.035 −0.01 0.036 −0.082 0.103 0.003 
Winter 0.042 −0.03 0.023 0.128 0.007 0.101 −0.065 
Growth season 0.049 0.053 0.018 0.009 −0.039 −0.014 −0.068 
Annual 0.052 −0.009 −0.002 −0.003 −0.135 0.046 −0.082 

*Significance level 0.05; **significance level 0.01.

The potential influences of monsoon indices must be considered despite the lack of identified significant correlations between these indices and the extreme precipitation indices. It is clear that the ISMI and WNPMI had more consistent impacts on the extreme precipitation indices than did the other monsoon indices. There were nearly identical negative correlations between the ISMI and Wet day for all six timescales. Meanwhile, there is a remarkable result that the correlation between the ISMI and SDII was positive no matter in which timescale. The correlations between the ISMI and the other extreme precipitation indices were similar for the corresponding timescales, which were detected negative correlations between them in spring and the positive correlations in the other timescales. The WNPMI generally showed negative correlation with all extreme precipitation indices except in winter. Based on comparison of the correlations between the climate indices and the extreme precipitation indices for the six timescales, it is clear that the amounts and durations of extreme wet events were more consistently impacted by the AOI, NAO, IOD, PDO, ISMI, WNPMI and SASMI because these climate indices show similar correlations with the amount and duration indices for each corresponding timescales. Furthermore, the SDII shows a positive relationship with the IOD and ISMI at all timescales, which indicates that the intensity of extreme wet events was consistently impacted by the IOD and ISMI. However, the mechanism by which these climactic indices and extreme precipitation events are linked should be investigated in future research.

Correlations between extreme precipitation indices and the C index

The values of the C index increased slightly over 1961–2014, which indicates that the effects of agricultural droughts were more severe in Northeast China, although this increase in the C index did not pass the significance test. Significant correlations were not detected between the extreme precipitation indices and the C index in winter, but the relationships between each extreme precipitation index and the C index were nearly consistently positive or negative for the six timescales (Table 7). The amount indices (R95p, R99p and PPRCPTOT) had significant negative correlations except in winter. The duration indices (R10, R20, R25, CWD and Wet day), which express the durations of extreme wet events, generally displayed significant negative relationship with the C index; the highest correlation coefficients were detected in the growing season. The intensity indices (SDII, R1day and R5days) showed significant negative correlations in spring, growing season and annually.

Table 7

Correlation coefficients between the extreme precipitation indices and C index

Index Spring Summer Autumn Winter Growth season Annual 
R95p −0.415** −0.507** −0.328* −0.149 −0.574** −0.566** 
R99p −0.353** −0.394** −0.398** −0.165 −0.500** −0.521** 
PRCPTOT −0.390** −0.559** −0.361** −0.153 −0.627** −0.563** 
R10 −0.320* −0.594** −0.317* −0.081 −0.626** −0.559** 
R20 −0.423** −0.504** −0.315* −0.194 −0.573** −0.578** 
R25 −0.332* −0.464** −0.333* −0.217 −0.570** −0.602** 
CDD 0.024 0.228 0.150 −0.102 −0.070 −0.123 
CWD −0.124 −0.354** −0.314* −0.158 −0.359** −0.336* 
Dry day 0.196 0.615** 0.250 0.092 0.652** 0.535** 
Wet day −0.195 −0.618** −0.250 −0.050 −0.654** −0.513** 
SDII −0.316* −0.189 −0.273* −0.012 −0.365** −0.440** 
R1day −0.360** −0.249 −0.287* −0.134 −0.278* −0.294* 
R5days −0.385** −0.312* −0.243 −0.092 −0.304* −0.308* 
Index Spring Summer Autumn Winter Growth season Annual 
R95p −0.415** −0.507** −0.328* −0.149 −0.574** −0.566** 
R99p −0.353** −0.394** −0.398** −0.165 −0.500** −0.521** 
PRCPTOT −0.390** −0.559** −0.361** −0.153 −0.627** −0.563** 
R10 −0.320* −0.594** −0.317* −0.081 −0.626** −0.559** 
R20 −0.423** −0.504** −0.315* −0.194 −0.573** −0.578** 
R25 −0.332* −0.464** −0.333* −0.217 −0.570** −0.602** 
CDD 0.024 0.228 0.150 −0.102 −0.070 −0.123 
CWD −0.124 −0.354** −0.314* −0.158 −0.359** −0.336* 
Dry day 0.196 0.615** 0.250 0.092 0.652** 0.535** 
Wet day −0.195 −0.618** −0.250 −0.050 −0.654** −0.513** 
SDII −0.316* −0.189 −0.273* −0.012 −0.365** −0.440** 
R1day −0.360** −0.249 −0.287* −0.134 −0.278* −0.294* 
R5days −0.385** −0.312* −0.243 −0.092 −0.304* −0.308* 

*Significance level 0.05; **significance level 0.01.

Originally, the C index, which is used to reflect the magnitude of agricultural drought disasters at a regional scale, was co-determined by crop yield, sowing areas, drought-attacked areas and drought-inundated areas. It does not have a direct relationship with any extreme precipitation index. However, the results of correlation analysis revealed potential indirect relationships between extreme precipitation indices and the C index. This analysis further demonstrated the capability of the C index, which can not only represent the severity of agricultural drought disaster directly but also is largely impacted by extreme precipitation events, especially in the growing season. The significant response of the C index to extreme precipitation indices depended largely on the characteristics of the agricultural region of Northeast China, where mass rainfed farming and agricultural drought disasters are heavily influenced by precipitation anomalies.

CONCLUSIONS

Thirteen extreme precipitation indices, classified into amount indices, duration indices and intensity indices based on the characteristics of the extreme precipitation events, were used in this study, and revised to adapt to a variety of timescales. The temporal features of these indices were evaluated based on the results of trend and mutation tests for seasonal, growing-season and annual scales. The spatial distribution features of all extreme precipitation indices in the growing season were explored in the study area. To expand the understanding of extreme precipitation, potential correlations between atmospheric circulation and extreme precipitation indices were investigated. The potential relationships between extreme precipitation events and agricultural drought disasters were also analyzed at multiple timescales. The main conclusions were as follows:

  • 1.

    The main temporal features of extreme precipitation detected were that extreme dry events increased in Northeast China between 1961 and 2014. The magnitudes of extreme day events were higher than that of extreme wet events, which also presented a downward trend at all timescales, except for an increase in spring. The effects of extreme wet events in the study area were more severe in summer and the growing season, even though the duration of extreme wet events showed a significant decline. In addition, the results of the change point test further confirmed the continuous decrease in the frequency of extreme wet events in the study area.

  • 2.

    Spatially, the extreme precipitation indices that reflect extreme wet events (R95p, R99p, PRCPTOT, R10, R20, R25, Wet day, SDII, R1day and R5days) have similar distributions, and the average values of each of these indices over 1961–2014 presented gradually decreasing trends from south to north in the growing season. The most significant decreasing trends in the amount indices were detected in the Heilongjiang Province in the growing season. The wetter regions (Liaoning Province and Jilin Province) experienced lower frequencies of extreme wet events, and drier regions experienced higher frequencies of extreme wet events. Meanwhile, the Songliao Plain experienced notably increased frequency and intensity of extreme wet events and significant decrease in dry periods during 1961–2014, based on the spatial distribution of extreme precipitation in the growing season.

  • 3.

    The results of the correlation analysis among the 13 extreme precipitation indices at multiple timescales further confirmed that correlations between these indices were determined by the definitions of each index, and that the correlations between indices were impacted and limited by various timescales. Furthermore, the extreme precipitation indices had closer relationships with the annular mode indices (AOI, NAO, IOD, PDO, SOI and ENSO3.4) than that with the monsoon indices (ISMI, WNPMI, EASMI and SASMI). This finding may be attributed to the high-latitude position of Northeast China, which may be more easily impacted by the polar climate than by the monsoon. In addition, the relationships between extreme precipitation and the C index illustrate that agricultural droughts were more sensitive to extreme precipitation events in the growing season in Northeast China. However, only potential correlations between extreme precipitation, climatic conditions and agricultural disasters based on the teleconnection relationships were analyzed in this study. Further study should focus on identifying the corresponding mechanisms.

  • 4.

    Based on the spatiotemporal distribution and correlation features of the 13 extreme precipitation indices, R95p, R99p, CDD, CWD, Dry day, Wet day, R1day and R5days are strongly recommended as representative extreme precipitation indices for Northeast China.

The spatiotemporal features of extreme precipitation in Northeast China obtained in this study provide valuable information for flood control and drought relief. Further research should pay more attention on quantitative relation between the extreme precipitation events and the annular mode indices with more precise data.

ACKNOWLEDGEMENTS

This work was financially supported by the Special Fund for Research on Public Interests, Ministry of Water Resources (No. 201401036). We also acknowledge the comments and suggestions by the editor and anonymous reviewers.

REFERENCES

REFERENCES
Alexander
L. V.
Zhang
X.
Peterson
T. C.
Caesar
J.
Gleason
B.
Klein Tank
A. M. G.
Haylock
M.
Collins
D.
Trewin
B.
Rahimzadeh
F.
Tagipour
A.
Rupa Kumar
K.
Revadekar
J.
Griffiths
G.
Vincent
L.
Stephenson
D. B.
Burn
J.
Aguilar
E.
Brunet
M.
Taylor
M.
New
M.
Zhai
P.
Rusticucci
M.
Vazquez-Aguirre
J. L.
2006
Global observed changes in daily climate extremes of temperature and precipitation
.
J. Geophys. Res.-Atmos.
111
,
D05109
.
Bekele
D.
Alamirew
T.
Kebede
A.
Zeleke
G.
Melese
A. M.
2016
Analysis of rainfall trend and variability for agricultural water management in Awash River Basin, Ethiopia
.
J. Water Clim. Change
.
DOI: 10.2166/wcc.2016.044
.
Bouchard
E.
Qi
Z.
2016
Long-term trends of climate change and its impact on crop growing season on Montreal Island
.
J. Water Clim. Change
.
DOI: 10.2166/wcc.2016.139
.
Croitoru
A.-E.
Chiotoroiu
B.-C.
Todorova
V. I.
Torică
V.
2013
Changes in precipitation extremes on the Black Sea Western Coast
.
Global Planet. Change
102
,
10
19
.
Donat
M. G.
Alexander
L. V.
Yang
H.
Durre
I.
Vose
R.
Dunn
R. J. H.
Willett
K. M.
Aguilar
E.
Brunet
M.
Caesar
J.
Hewitson
B.
Jack
C.
Tank
A.
Kruger
A. C.
Marengo
J.
Peterson
T. C.
Renom
M.
Rojas
C. O.
Rusticucci
M.
Salinger
J.
Elrayah
A. S.
Sekele
S. S.
Srivastava
A. K.
Trewin
B.
Villarroel
C.
Vincent
L. A.
Zhai
P.
Zhang
X.
Kitching
S.
2013
Updated analyses of temperature and precipitation extreme indices since the beginning of the twentieth century: the HadEX2 dataset
.
J. Geophys. Res.-Atmos.
118
,
2098
2118
.
Donat
M. G.
Peterson
T. C.
Brunet
M.
King
A. D.
Almazroui
M.
Kolli
R. K.
Boucherf
D.
Al-Mulla
A. Y.
Nour
A. Y.
Aly
A. A.
Nada
T. A. A.
Semawi
M. M.
Dashti
H. A.
Salhab
T. G.
Fadli
K. I.
Muftah
M. K.
Eida
S. D.
Badi
W.
Driouech
F.
Rhaz
K.
Abubaker
M. J. Y.
Ghulam
A. S.
Erayah
A. S.
Ben Mansour
M.
Alabdouli
W. O.
Dhanhani
J. S.
Al Shekaili
M. N.
2014
Changes in extreme temperature and precipitation in the Arab region: long-term trends and variability related to ENSO and NAO
.
Int. J. Climatol.
34
,
581
592
.
Donat
M. G.
Lowry
A. L.
Alexander
L. V.
Ogorman
P. A.
Maher
N.
2016
More extreme precipitation in the world's dry and wet regions
.
Nature Clim Change
6
,
508
513
.
Fu
G.
Yu
J.
Yu
X.
Ouyang
R.
Zhang
Y.
Wang
P.
Liu
W.
Min
L.
2013
Temporal variation of extreme rainfall events in China, 1961–2009
.
J. Hydrol.
487
,
48
59
.
Gholami
V.
Chau
K. W.
Fadaee
F.
Torkaman
J.
Ghaffari
A.
2015
Modeling of groundwater level fluctuations using dendrochronology in alluvial aquifers
.
J. Hydrol.
529
,
1060
1069
.
Ramos
A. M.
Trigo
R. M.
Liberato
M. L. R.
Tome
R.
2015
Daily precipitation extreme events in the iberian peninsula and its association with atmospheric rivers
.
J. Hydrometeorol.
16
,
579
597
.
Rosenzweig
C.
Tubiello
F. N.
Goldberg
R.
Mills
E.
Bloomfield
J.
2002
Increased crop damage in the US from excess precipitation under climate change
.
Global Environ. Change
12
,
197
202
.
Roy
S. S.
Balling
R. C.
Jr.
2004
Trends in extreme daily precipitation indices in India
.
Int. J. Climatol.
24
,
457
466
.
Stott
P. A.
Christidis
N.
Otto
F. E. L.
Sun
Y.
Vanderlinden
J.-P.
van Oldenborgh
G. J.
Vautard
R.
von Storch
H.
Walton
P.
Yiou
P.
Zwiers
F. W.
2016
Attribution of extreme weather and climate-related events
.
WIREs Clim. Change
7
,
23
41
.
Van den Besselaar
E. J. M.
Klein Tank
A. M. G.
Buishand
T. A.
2013
Trends in European precipitation extremes over 1951–2010
.
Int. J. Climatol.
33
,
2682
2689
.
Vavrus
S. J.
Notaro
M.
Lorenz
D. J.
2015
Interpreting climate model projections of extreme weather events
.
Weather Clim. Extremes
10
(
Part B
),
10
28
.
Wang
H.
He
S.
2015
The North China/Northeastern Asia severe summer drought in 2014
.
J. Climat.
28
,
667
668
.
Wang
B. L.
Zhang
M. J.
Wei
J. L.
Wang
S. J.
Li
X. F.
Li
S. S.
Zhao
A. F.
Li
X. S.
Fan
J. P.
2013
Changes in extreme precipitation over Northeast China, 1960–2011
.
Quatern. Int.
298
,
177
186
.
Yuan
Z.
Yang
Z.
Yan
D.
Yin
J.
2015
Historical changes and future projection of extreme precipitation in China
.
Theor. Appl. Climatol.
127
,
1
15
.
Zhang
X.
Aguilar
E.
Sensoy
S.
Melkonyan
H.
Tagiyeva
U.
Ahmed
N.
Kutaladze
N.
Rahimzadeh
F.
Taghipour
A.
Hantosh
T. H.
Albert
P.
Semawi
M.
Karam Ali
M.
Said Al-Shabibi
M. H.
Al-Oulan
Z.
Zatari
T.
Al Dean Khelet
I.
Hamoud
S.
Sagir
R.
Demircan
M.
Eken
M.
Adiguzel
M.
Alexander
L.
Peterson
T. C.
Wallis
T.
2005
Trends in Middle East climate extreme indices from 1950 to 2003
.
J. Geophys. Res.-Atmos.
110
,
D22104
.
Zhang
X. B.
Alexander
L.
Hegerl
G. C.
Jones
P.
Tank
A. K.
Peterson
T. C.
Trewin
B.
Zwiers
F. W.
2011
Indices for monitoring changes in extremes based on daily temperature and precipitation data
.
WIREs Clim. Change
2
,
851
870
.
Zhang
Q.
Singh
V. P.
Li
J.
Jiang
F.
Bai
Y.
2012a
Spatio-temporal variations of precipitation extremes in Xinjiang, China
.
J. Hydrol.
434–435
,
7
18
.