Northeast China (NEC) is the largest agricultural production region in China, and its high-latitude location means that its climate is affected by both tropical and subtropical circulation systems, especially the East Asian monsoon and cold-air activity from the northern polar region. The precipitation in NEC features quite significant monthly differences, with maximum amounts occurring in boreal summer (Gong et al., 2006) and minimum values in winter. The precipitation in summer directly affects grain production and economic development in the region. In recent years, research has increasingly focused on the changes in precipitation features and the methods used for forecasting seasonal precipitation. Sun et al. (2000) pointed out that the spatial distribution of the summer precipitation anomaly shows consistent interannual variation over the whole region in most years, but in some years, it also exhibits opposite variation in northern and southern parts of the region, or eastern and western parts. Due to the concentrated rainy season, heavy rainfall events in July and August are more frequent than in other regions at the same latitudes. Similar to the amount of precipitation, the number of rainfall days in the region shows an obvious decreasing trend from the southeast to the northwest (Sun et al., 2010). Zou and Ding (2010) found that the first heavy rainfall date in NEC has an advancing trend during 1961–2005, and the heavy rainfall frequency since the 1990s is significantly greater than the frequency in the previous period of 1961–90, especially in its northwest and northeast regions. Studies have also indicated that the intensity of both heavy rainfall and severe dry events have increased in line with global warming (Yang and Zhao, 2011). This means that regional flooding and drought events have become increasingly frequent. Hence, in recent years, the forecasting technologies for summer precipitation in the region have received increasing attention.

The climate in NEC is influenced by many factors, but particularly by the sea surface temperature (SST) in previous seasons (Sun and Wang, 2006). The close relationship between SST and precipitation has been applied widely in seasonal forecasts — for example, the use of SSTs in North Pacific ( Sun and An, 2003) and the Indian Ocean (Feng et al., 2006). Jia and Wang (2003) revealed the role played by the North Atlantic Ocean in the previous spring, and this role is considered to be partly due to the impacts of the North Atlantic SST tripole on the East Asian summer monsoon (EASM) (Zuo et al., 2013). The SST near the southeast coast of China also has an important effect on precipitation. He et al. (2006) found that the SST anomalies near China in the previous winter can be used as a precursory signal of the intensity of the Northeast Cold Vortex (NECV) in summer. The NECV is thought to be one of the most important circulation systems affecting the climate in NEC. Gao and Gao (2014) analyzed the possible physical mechanism involved in the effect of Kuroshio SSTs on summer precipitation in NEC by simulating the NECV.

Precipitation in NEC is also strongly related to the local general circulations. The intensity of the southerly winds at 850 hPa in northeastern Asia can directly lead to abnormal precipitation in the region. Furthermore, both the position and intensity of the western Pacific subtropical high (WPSH) show obvious differences between flood and drought summers (Sun et al., 2002a, b, 2003). Jia and Wang (2006) pointed out that there are cyclonic (anticyclonic) wind anomalies over the southeast area of Mongolia at 850 hPa in flood (drought) summers, and southwesterly (northeasterly) flow over eastern China. Corresponding to this circulation pattern, there are anticyclonic (cyclonic) wind anomalies in the south of Japan. Thus, the moisture convergence in the lower-level atmosphere over the region is strengthened (weakened), resulting in more (less) precipitation in summer. Based on the Lamb–Jenkinson atmospheric circulation classification method, Jia et al. (2006) explored the eight primary circulation patterns for abnormally higher or lower precipitation in NEC. Statistical analyses have also indicated that severe droughts in NEC could be attributable to a weaker and more southward WPSH, a stronger Mongolian high, weaker moisture transport by westerly winds from higher latitudes, and southwest summer monsoon flow from the Bay of Bengal (Wei and Zhang, 2009). Sun et al. (2016) defined a new EASM index to quantify the effects of the EASM on NEC summer precipitation. The results indicated that the new EASM index is closely linked to anomalous rainfall in NEC and could be used to measure the physical processes that affect the regional dry or wet condition in the region.

Atmospheric circulation anomalies at high latitudes constitute another key factor affecting precipitation in NEC. Wu et al. (2008) found that a more westward Arctic polar vortex usually leads to an increase in summer precipitation in NEC. Liu et al. (2012) revealed that the 10–30-day oscillation of the atmospheric circulation associated with the NECV has a clear phase-lock relationship with summer precipitation. Shen et al. (2011) investigated the large-scale circulation anomalies related to monthly summer rainfall in NEC and found that the interannual variation in rainfall over NEC was mainly dominated by the NECV in early summer (May–June) and by the EASM in late summer (July–August). Considering the relationship between the NECV and the mid-summer rainy period and their corresponding atmospheric circulations, Gong et al. (2015) developed an objective identification method to define the annual start and end dates of the NECV and mid-summer rainy periods.

In summer 2013, northern China was much wetter than normal (Gong, 2014), especially in NEC. Owing to heavy rainfall, severe floods and landslide events occurred in the region. In late July, four consecutive heavy rainstorms attacked the region, resulting in the highest levels of precipitation since 1997. The average precipitation during that period was about 54% more than the normal. The floods resulted in huge economic losses and serious casualties. Synoptic analysis of this single case indicated that the NECV was one of the main systems of influence (Lin et al., 2013). However, few studies have analyzed the influence of the NECV on a long-term timescale. Therefore, it was necessary to analyze the atmospheric circulation characteristics in 2013 and in other abnormal precipitation summers in NEC to distinguish the roles played by the main systems influencing precipitation in the region on both seasonal and diurnal timescales (particularly the influence of the NECV and WPSH at different rainy stages), and to provide useful information for seasonal forecasts in the region.

The observed daily rainfall amounts at 2513 stations in China were provided by the China Meteorological Administration. For data quality, only those stations for which the data were complete during 1981–2013 were selected for further study. Considering the fact that the climate, especially summer precipitation, in eastern Inner Mongolia is quite similar to that in Heilongjiang, Jilin, and Liaoning provinces, in this study, the representative stations for NEC were chosen as those in the region (38°–55°N, 120°–135°E). There were a total of 194 stations in this region (Gao and Gao, 2015). The atmospheric variables were obtained from the NCEP/NCAR reanalysis (Kalnay et al., 1996; Kistler et al., 2001) and included 500-hPa geopotential height and 850-hPa horizontal winds on a 2.5° × 2.5° latitude–longitude resolution. According to the World Meteorological Organization (WMO) standard, the climate norm is the latest three-decade average of a climatological variable. Thus, in this study, all the anomalies were derived based on the climate norm of 1981–2010.

In summer 2013, the regional-average precipitation for the whole of China was 339.9 mm, which was about 4.5% more than the climatology (Gong et al., 2014). The pattern of summer rainfall showed a typical spatial distribution of wet in northern China and dry in southern China. There were two rainfall belts in China (Fig. 1). The major rainfall belt was located in northern China. In this region, the precipitation anomaly percentages (PAPs) at most stations exceeded 20%, especially in North China and NEC. In NEC, there were 37 stations at which the PAPs were greater than 50%. As mentioned above, severe floods occurred in NEC in summer, particularly in July. The average PAP for NEC was about 25.4%, which was the highest value since 1995. The minor rainfall belt was located in South China. However, the PAPs in South China were much lower than those in North China. At most stations in South China, the PAPs were less than 20%. Figure 2 shows the normalized time series of the regional-average summer precipitation in NEC from 1981 to 2013. The seasonal flooding in 2013 can be seen quite clearly from Fig. 2. Summer precipitation in NEC shows obvious interdecadal variations over the past 33 years. Before 1999, precipitation in NEC was greater than the climatological mean in most years, and in seven of the years, the standard deviation value was more than 0.5. Standard deviation values less than –0.5 before 1999 only occurred in 1982, 1989, 1992, and 1997. Since 1999, there are only three years (2010, 2012, and 2013) in which the precipitation standard deviations were above 0.5. Furthermore, the interannual variation in summer precipitation is also quite significant, for example, in 1998 and 1999. In this study, 1985, 1991, 1994, and 1998 were selected and considered as typical flood years in NEC according to the WMO anomaly criteria. The normalized precipitation value exceeded 1.0 in these four years.

Figure 1 Precipitation anomaly percentages in summer 2013 in China.

Figure options

Figure 2 Time series of normalized summer precipitation in NEC during 1981–2013.

Figure options

The daily variation of precipitation in NEC was also analyzed (Fig. 3). From the climatology (black solid curve in Fig. 3), it can be seen that the daily regional-average precipitation was about 2–7 mm in summer. The maximum intensity of the annual cycle occurred mainly from mid July to early August, which is the climatological period of the rainy season in North China and NEC (Chen and Zhao, 2000). Compared with the climatology, the rainy season in 2013 began earlier, with the first period occurring from late June to early July and the second period from mid July to early August. In these two periods, the precipitation amounts were much greater than the climatological values. In summer 2013, the greatest daily rainfall occurred on 2, 16, 19, 24, and 28 July, and 16 August. The total precipitation for these six days contributed more than 20% of the total summer precipitation. The temporal variation of heavy rainfall days was similar to Fig. 3 (figure omitted). For example, on 2 July, there were 45 stations at which the daily rainfall amounts were greater than 50 mm.

Figure 3 Daily precipitation amount (mm) in NEC in summer 2013 (histogram) and the climatology (black solid curve).

Figure options

As mentioned above, the NECV and WPSH in the middle troposphere and the southerly winds associated with them in the lower troposphere are the main systems influencing summer rainfall in NEC. In addition, the northward-moving tropical cyclone plays an important role (Jin et al., 2006). During the northward-marching process, some tropical cyclones can develop into extratropical cyclones and cause heavy rainfall in NEC (Lin and Jiang, 1992). For example, in 2011, Super Typhoon Muifa (No. 1109) brought heavy rainstorms to southern NEC (Xu et al., 2011). However, in summer 2013, the main influence of tropical cyclones on China was limited to the south and southeast coasts, and none is considered to have brought heavy rainfall to NEC (Lin et al., 2013; Yang and He, 2013; Zhang and He, 2013). Therefore, the influence of tropical cyclones on the seasonal flooding in 2013 is not considered in the following analyses.

In summer 2013, the WPSH was located much farther west than normal (Fig. 4). The 5880-gpm contour at 500 hPa indicates that the western boundary of the WPSH was near 125°E, about 7° farther west than its climatological position (132°E). In contrast to its climatology, in summer 2013, the WPSH was not in an ellipse-type shape. Both the northern boundary and latitudinal axis were farther north, owing to the uplifted part on the western side of the WPSH. This circulation pattern favored the enhancement of southerly winds on the western side and the transport of more moisture to North China and NEC (thick solid arrows in Fig. 4). Another moisture transport pathway to NEC can be traced to the cross-equatorial flow (CEF) over the Bay of Bengal (BOBCEF). In 2013, the BOBCEF was significantly farther west than normal, and its central channel shifted from its usual position of 80°–90°E to 60°–70°E, i.e., in the middle of the Somalia CEF channel and the BOBCEF channel. Owing to the Coriolis force, the westward BOBCEF first strengthened the westerly winds over the Indian Ocean, and then enhanced the moisture transport from the BOB to NEC. Therefore, both the westerly winds and moisture transport were stronger owing to the northward WPSH and the westward BOBCEF. In the mid–high latitudes, a cyclonic wind anomaly center clearly existed over the northern part of NEC, with a central geopotential height less than –20 gpm (marked by “C” in Fig. 4). In the eastern part of the NECV, there was an anticyclonic wind anomaly center (marked by “A” in Fig. 4). The geopotential height anomaly in the anticyclonic center exceeded +30 gpm. This helped the NECV to remain stable over NEC, causing more frequent cold-air activity from the northern polar region. Thus, the moisture convergence at lower levels in the atmosphere was much stronger than normal in NEC, resulting in seasonal flooding in the region.

Figure 4 850-hPa vapor flux (arrows; g s–1 cm–1 hPa–1) and 500-hPa geopotential height (contours; gpm) and its anomaly (only positive values are shaded and shown; gpm). The thick solid and dashed contours indicate the 5880-gpm contours in summer 2013 and the climatology, respectively. The thick dashed and solid arrows represent the vapor transport pathway from the BOB and from the western Pacific, respectively. “A” and “C” indicate the anticyclonic and cyclonic wind anomaly centers, respectively.

Figure options

Correlation analyses showed consistent conclusions for the case study of summer 2013 analyzed above. Figure 5 shows the spatial distribution of the correlation coefficients (CCs) between precipitation in NEC and both the 500-hPa geopotential height and the 850-hPa horizontal wind in boreal summer. For the 500-hPa geopotential height, a significant negative CC center exists over the northern area of NEC, i.e., the climatological position of the NECV (figure omitted). The minimum value in the center is less than –0.45, which passed the t-test at the 95% confidence level. The correlation pattern means that a stronger (weaker) NECV is related to more (less) rainfall in NEC. Similar to the correlation pattern between precipitation and the 500-hPa geopotential height, there is a cyclonic CC center over the same region for the 850-hPa wind fields, further confirming that the NECV is a relatively deep circulation system. In the subtropical region, a remarkable positive CC center can be seen near 30°N over the western Pacific warm pool, which is farther north than the climatological position of the WPSH by 5° latitudes (shown in Fig. 4). This means that a more northward WPSH can cause greater rainfall over NEC in summer, which is consistent with the results revealed in Fig. 4 and with predictions from the statistical seasonal forecast model operating in China (Chen and Zhao, 2000; Zhao, 1999). However, compared with the relationship between rainfall and the NECV, the relationship between rainfall and the WPSH is obviously weaker. The CC between summer precipitation in NEC and the regional-average 500-hPa geopotential height over 20°–30°N, 120°–150°E is only 0.20, while the value is 0.42 for the 500-hPa vorticity over 45°–55°N, 110°–130°E. The latter value passed the t-test at the 95% confidence level.

The partial correlation method was also used here to separate the individual effects of the NECV and WPSH on precipitation. The results showed that the partial correlation coefficient (PCC) between the NECV and precipitation was 0.51 excluding the WPSH effect, which passed the t-test at the 99% confidence level. However, the PCC was only 0.15 between the WPSH and precipitation without the effect of the NECV. Therefore, from the statistical results, it can be considered that both the NECV and the WPSH have an effect on abnormal precipitation in NEC, but the role of the NECV is more direct. It can also be seen from Fig. 5 that summer precipitation in NEC has a negative correlation with the equatorial meridional wind speed in the range of 50°–80°E, and a positive correlation with the BOBCEF, especially in its eastern part. This conclusion is consistent with the results obtained from the case analysis of summer 2013, which also suggests that, for summer precipitation in NEC, the high-latitude circulation systems over the Asian continent, especially the NECV, play a more direct role.

Figure 5 Correlation coefficients between summer precipitation in NEC and 850-hPa horizontal wind (arrows) and 500-hPa geopotential height (contours). Values greater than the 95% confidence level (t-test) are shaded.

Figure options

In summer 2013, the two heaviest rainfall events occurred on 2 and 16 July (Fig. 3). Therefore, the atmospheric circulation characteristics on these two days are shown here. On 2 July, the heavy rainfall was mainly concentrated in the southern and eastern areas of NEC (figure omitted). On this day, the WPSH extended westward to cover eastern China, with its western boundary reaching 111°E and its northern boundary locating over the northern area of the Yangtze River basin. However, due to the relative zonal shape of the WPSH in its northern part, the southwesterly flow of warm and moist air was brought directly to southern Japan instead of NEC. At high latitudes, NEC was controlled by a deep NECV, and the water vapor was transported from the Japan Sea to southern NEC by the southwesterly and southeasterly air flows on its eastern side (Fig. 6a).

On 16 July, the distribution of rainfall was similar to that on 2 July, but the heavy rainfall area was more extensive. The WPSH was in its eastward-retreating stage, which was different from that on 2 July, but its northern boundary reached farther northward to 35°N (Fig. 6b). This circulation pattern was favorable for the transport of more moisture to southern NEC. On this day, NEC was also affected by the southwest air flow from the eastern side of the NECV. Analyses of other heavy rainfall days also supported the conclusion that the NECV exerts the most important influence on heavy rainfall in the region (figure omitted). The systems influencing the heavy storm processes during 1–2 and 14–16 July 2013 from synoptic analyses were also checked (Zhang et al., 2014). For the first event (Fig. 6a), the system at higher levels was a cold vortex. For the second event (Fig. 6b), the system at higher levels was a trough dividing the NEC in its north. Therefore, from both the statistical results and the synoptic analyses, it can be concluded that the NECV played a more direct role than other circulations in the flooding event of 2013.

Figure 6 850-hPa wind (arrows; m s–1) and 500-hPa geopotential height (contours; gpm) on the two heaviest rainfall days in summer 2013 in NEC: (a) 2 July and (b) 16 July. The 5880-gpm contours are thickened in the figure.

Figure options

Figure 7 As in Fig. 4, but for (a) 1985, (b) 1991, (c) 1994, and (d) 1998. In each panel, “C” indicates the center of the NECV. The solid and dashed contours indicate the 5880-gpm contour for summer 2013 and the climatology, respectively.

Figure options

To obtain the common circulation features and to better reveal the impact of the NECV, the 850-hPa wind anomalies and 500-hPa geopotential height in 1985, 1991, 1994, and 1998 are shown in Fig. 7. In 1985, the geopotential height anomaly over the whole of East Asia was negative (Fig. 7a). In particular, the subtropical region in the summer was controlled by an unusually deep cyclone. In the center of the cyclone, the anomaly was less than –15 gpm. In summer 2013, the western boundary of the WPSH was close to 143°E, farther east than its climatology by about 10° longitudes. Over the western area of NEC, there was also an anomalous cyclone with a central anomaly less than –25 gpm. From the moisture transport path, it can be seen that the moisture in NEC was mainly caused by easterly winds from the south side of the anticyclonic circulation over the Japan Sea. Therefore, the circulation analyses in Fig. 7a indicate that, in summer 1985, the WPSH was weaker and the NECV was stronger than the climatological mean. In summer 1991 (Fig. 7b), the distribution of the geopotential height anomaly in East Asia was basically opposite to that in 1985, with most areas controlled by a positive anomaly. The WPSH was obviously stronger than normal, and its western boundary reached as far as Taiwan Island. However, in contrast to 2013 (Fig. 4), the southwesterly winds on the west side of the subtropical high appeared mainly south of 35°N; thus, the vapor on the southwest side of the WPSH was brought to the Yangtze River basin. However, from there, the vapor flow changed its direction to be westerly and the moisture was thus transported to the western Pacific instead of NEC. In North China, especially NEC, there was a cyclonic wind anomaly, indicating that the easterly winds on the northern side of the NECV were the major moisture transport pathway to the region.

In 1994 (Fig. 7c), the WPSH was weaker and more eastward than normal, and on its west side, there was a northwesterly anomalous flow. NEC was controlled by a very strong positive geopotential height anomaly center. Compared with the other years studied in this paper, the NECV was weaker in this year. Among the five flooding seasons selected in this study, the 1994 season was unique in terms of a weaker NECV. However, it can be determined from Fig. 7c that a weaker cyclonic circulation anomaly existed on the northwest side of the anticyclone center, which suggests a potential indirect influence of the NECV on flooding in this year. Further analyses of the circulations in June, July, and August 1994 indicated that both the monsoon circulation and the monsoon intensity had quite remarkable intraseasonal differences (figures omitted). In June 1994, the position and intensity of the WPSH were quite close to the climatology. The wind anomaly along 30°N was easterly instead of southeasterly. Therefore, the wet moisture could not be transported to NEC from the west boundary of the WPSH. In June, the EASM index (Zhang et al., 2003) was negative. However, in July and August, the WPSH was much farther west than its climatology. In these two months, the wet moisture could be directly transported to NEC from the western boundary of the WPSH. In these two months, the EASM index was positive. The circulation in 1998 (Fig. 7d) was quite similar to that in 1991. East Asia was almost completely controlled by a positive geopotential height anomaly, except in the southern area of NEC. Many studies have revealed that, in this summer, the WPSH was unusually strong and much farther west because of the influence of the El Niño event at that time. In contrast to 1991, the WPSH in 1998 was extremely strong and its 5880-gpm contour extended west of 110°E. Thus, the southwesterly moisture on its west side was brought to the Yangtze River basin and converged with the cold air from the west side of the NECV, and finally formed the extremely rare flooding in the basin. In contrast, some of the southwesterly water vapor was also transported to the Japan Sea after passing the Yangtze River, and met with the northerly winds from the east of the NECV. In the lower-level atmosphere over NEC, there was a moisture convergence center. Thus, abnormal rainfall occurred in the region.

Therefore, based on the analyses of summer 2013 and the four wettest summers before 2000, it can be concluded that the NECV plays a more direct role in seasonal flooding in NEC than the WPSH.

In summer 2013, severe flooding occurred in NEC. The regional-average precipitation was the second highest since 1981. The floods caused huge economic losses and serious casualties in the region. In this study, the summer precipitation was analyzed by using both case studies and statistical methods to explore the primary features and possible systems of influence. The results indicated that summer precipitation in NEC has obvious interdecadal variations. Before 1999, NEC was wetter than the climatological mean in most years; whereas, after 1999, NEC became relatively drier. Significant interannual change also existed in the study period. In 2013, the rainy season in NEC began earlier, with the first stage occurring from late June to early July and the second stage from mid July to early August. The heaviest rainfall occurred on 2, 16, 19, 24, and 28 July, and 16 August, and the total precipitation for these six days contributed more than 20% to the total amount of summer precipitation.

The summer 2013 case study indicates that both the WPSH and the NECV influence seasonal flooding in NEC. However, the role of the NECV is more direct. In 2013, the southwesterly flow of moisture to NEC by the BOBCEF was also vital, but its contribution barely appeared in the statistical results. The correlation coefficient between the regional-average precipitation and the WPSH was only 0.20, while the value increased to 0.42 for the NECV. Synoptic analyses of the two heaviest rainfall days also supported the conclusion that the NECV exerts the most influence on rainfall in the region. To obtain the common circulation features, four other flood years (1985, 1991, 1994, and 1998) were also studied. In three of the years, the NECV was much stronger than the climatological mean; in 1994, the vortex was slightly weaker.

Seasonal forecasting in NEC remains a big challenge in China. It should be noted that the conclusions of this paper are mainly based on the five wettest case studies and statistical results; severe drought events in NEC are not discussed here, although the major circulations are almost opposite to those for flooding events. Furthermore, the detailed mechanisms by which the NECV influences intraseasonal processes still require further analysis, especially for different timescales (from short to long terms). In addition, extreme heavy rainfall events have a substantial impact on, and are a major contributor to, the seasonal flooding in the region, but few studies have investigated this relationship.