IIE, the interface between the Indian Summer Monsoon (ISM) and the East Asian Summer Monsoon (EASM), is defined using the equivalent potential temperature and summer long-term mean reanalysis data provided by NOAA/OAR/ESRL PSD. The June–July–August reanalysis data for the period 1951–2008 and empirical orthogonal function analysis are further applied to obtain the IIE index at the near-surface isobaric level. The index has a prominent interannual variation that is strongly correlated with the seesaw variation between the ISM and EASM. When a strong EASM and weak ISM occur, this interface index is higher than the normal, with the interface between the two summer monsoons shifting farther eastward than normal. When a weak EASM and strong ISM appear, the index is lower than normal, with the interface moving farther westward than normal. The western North Pacific subtropical high, a major factor in the EASM system, plays an important role in the year-to-year variation of the IIE. Compared with approaches taken in previous studies, this index objectively and quantitatively describes the IIE variation and better represents the two teleconnection patterns associated with the Asian summer monsoon, thus enhancing interpretations of the interaction between the ISM and EASM and its effects on regional droughts and floods in East Asia.
 The Asian summer monsoon system encompassing the Indian summer monsoon (ISM) and the East Asian summer monsoon (EASM) is the most complex and powerful monsoon system in the world. Within the Asian summer monsoon system, the principal components of the Indian monsoon circulation system include a monsoon trough over northern India, cross-equatorial airflows, the Somali low-level jet, the Mascarene high, and the South Asia high [Krishnamurti and Bhalme, 1976; Wu et al., 2012]. Meanwhile, the EASM circulation system consists of a monsoon trough in the South China Sea and western equatorial Pacific, cross-equatorial airflows to the east of 100°E, a cold anticyclone in Australia, a western Pacific subtropical high, an upper-level northeasterly flow, tropical easterlies, convection along the monsoon trough, Meiyu/Baiu frontal zones, and midlatitude disturbances [Tao and Chen, 1987]. The southwesterly flow and southeasterly flow of the Asian summer monsoon system respectively transporting water vapor from the North Indian Ocean and the western Pacific to East Asian directly affect the drought and flood patterns in East Asia [Ding and Sun, 2001; Duan and Wu, 2005; Li, 1999; Liu et al., 2007, 2012; Ninomiya, 1999; Ninomiya and Kobayashi, 1999; Zhou and Yu, 2005]. Additionally, the oxygen isotope ratio of precipitation detected from speleothems fluctuates with the different sources of water vapor between the Indian Ocean and Pacific Ocean [e.g., Cheng et al., 2009; Wang et al., 2001]. Therefore, studies on the variation of the interface between the ISM and East Asia summer monsoon (namely, the IIE) are crucial in understanding the interactions between the two monsoon subsystems, and their effects on local climate and weather around the IIE.
 Since the early 1980s, scientists have explored the interaction between the ISM and EASM, and the associated interface shift. Jin and Chen suggested that in terms of both air pressure and low-level geopotential height, the EASM system significantly affects the ISM system, whereas the ISM system does not affect the EASM system significantly. The IIE normally appears between 95° and 100°E.Zhu et al.  pointed out that the interplay between the ISM and EASM is complicated in forms of wave propagation, vapor transport and energy exchange. Lau and Li , Tao and Chen  and Ding  pointed out that the EASM is affected not only by the ISM flow but also by the western Pacific subtropical high and midlatitude circulations. More recently, Wang and LinHo  argued that a broad corridor over the Indochina Peninsula separates the ISM, EASM, and Northwest Pacific summer monsoon. Wang et al.  studied the effects of internal and external forcing of the monsoon system by contrasting the annual cycle and interannual variability of Indian and East Asian monsoons. They defined the Indian monsoon over the sector 40°–105°E and the East Asian monsoon over the sector 105°–160°E, implying that the IIE is located at 105°E. Hong and Liu  found that abnormal rainfall in the Huaihe River valley and the regions north of it in China is closely linked with the location of subtropical anticyclone over Western Pacific and the latent heating anomaly over the Arabian Sea and Indian peninsula.
 Although the previous studies have proposed the existence of the IIE and shed light on the potential interaction between the summer monsoon subsystems and the possible IIE position, there are differences in their conclusions to some extent [e.g., Jin and Chen, 1982; Wang et al., 2003]. Furthermore, there has been no quantitative definition of the IIE with physical meaning. To better reveal the complicated nature of the IIE, it is necessary to develop an objective and quantitative index with physical meaning that can describe the interaction between the two monsoon systems. This would also help us understand the effects of the IIE variation on the local climate and weather variability around the IIE.
 In this paper, through systemic analysis of the long-term averaged spatial structure of equivalent potential temperature (EPT) over the potential area of the IIE, we define a physically meaningful index for the IIE, and then explore its relationship with large-scale monsoon processes. Finally, its relationships with ISM and EASM indices are analyzed.
 We use atmospheric circulation reanalysis data and their long-term mean provided by NOAA/OAR/ESRL PSD for the summers (June–July–August; JJA) of 1951–2008 [Compo et al., 2006, 2011; Whitaker et al., 2004]. The resolution of the reanalysis data is 2° in latitude and longitude and there are 16 pressure levels from 1000 to 200 hPa. The global terrestrial precipitation data provided by CRU for the period 1951–2006 have resolution of 0.5° × 0.5°. The China Climate Center provides the summer precipitation data collected by 496 Chinese stations during the period 1958–2008. The Climate Center of Yunnan Province supplies summer precipitation data collected by 124 stations during the period 1961–2008. The studied geographic area is 60°–135°E and 0°–50°N, in which the core area is 90°–110°E and 18°–28°N (Figure 1).
 Since the EPT θEis a composite variable comprising temperature and humidity, and it is conservative during dry and moist adiabatic and pseudo-adiabatic processes [Bolton, 1980; Holton, 2004, pp. 290–404], it is frequently used to study the march of monsoons [e.g., Ding, 2004; Zhang et al., 2004] and can be applied to distinguish the sources of summer monsoon air masses. EPT θE is calculated as
where TK, p and r are the absolute temperature, pressure and mixing ratio at the initial level, and TL is the absolute temperature at the lifting condensation level [Bolton, 1980], the units of θE, TK and TL are Kelvin (hereafter referred as the K), the units of p and r are hPa and g/kg respectively. Analysis methods include composite analysis, correlation analysis, regression analysis and empirical orthogonal function (EOF) analysis.
3. Interface Between the ISM and EASM
 Previous studies have indicated that there should be an interface between the two monsoon subsystems. Since the IIE variability may synthetically reflect the effect of the ISM and EASM on the climate and weather in East Asia, especially the climate and weather around the interface, it is necessary to quantitatively define the IIE. Using equation (1)and the long-term averaged relative humidity and temperature data, we obtain a climatological distribution of the summer EPT in the core region (Figure 2). Figure 2illustrates that the EPT at the same isobaric level tends have a low-high-low pattern with an increase in longitude. The EPT at the same longitude tends to have a high-low-high pattern as air pressure decreases. Consequently, the long-term averaged summer EPT has a typical saddle-shaped pattern, in which two high-value centers with values above 351 K are located at the lower and mid-high levels of the troposphere, respectively. Two low-value centers with values below 345 K are distributed near the mid-troposphere at 90°E and 110°E, respectively. The typical saddle-shaped distribution of EPT indicates the existence of the monsoon air masses from the Bay of Bengal and those from the South China Sea. Affected by the local terrain, the air masses from the Bay of Bengal glide up from the southwest to the northeast, while those from the South China Sea glide up from the southeast to the northwest along the EPT curve. The maximum lifting height of air masses occurs near 100°E.
 As shown in Figure 3, the pattern of the distribution is consistent with that of EPT in Figure 2, presenting an asymmetric distribution against the central axis near 100°E where , with λ denoting longitude. To the west of 100°E, is positive, and the corresponding maximum above 8 K/° is located at the near-surface level between 92°E and 96°E. In contrast, is negative to the east of 100°E, and the corresponding minimum below −8 K/° is located at the near-surface level between 104°E and 108°E. From the agreement betweenFigure 2 and Figure 3, we define the IIE to be the surface where , mathematically representing the EPT extremum of the saddle-shaped distribution. The corresponding physical meaning is that the summer monsoon air masses from the Bay of Bengal and those from the South China Sea merge at the EPT extremum where the sources of the two monsoon air masses cannot be distinguished from each other.
 Using the above definition for the IIE, the long-term averaged IIE position is calculated and mapped along 18°N, 900 hPa; 20°N, 850 hPa; 22°N, 800 hPa; 24°N, 750 hPa; 26°N, 650 hPa; and 28°N, 600 hPa (hereafter referred to as the near-surface isobaric level).Figure 4shows that the long-term averaged IIE has a wavy pattern, along 102.98°E, 18°N, 900 hPa; 101.68°E, 20°N, 850 hPa; 100.88°E, 22°N, 800 hPa; 100.96°E, 24°N, 750 hPa; 101.08°E, 26°N, 650 hPa; and 99.32°E, 28°N, 600 hPa. The IIE position at the near-surface isobaric level mainly agrees with the previous understanding that the IIE is situated near 100°E [e.g.,Jin and Chen, 1982].
4. Definition of the IIE Index
 When calculating the summer EPT and values from 1951 to 2008, the IIE position in each year can be determined at the near-surface isobaric level in the range 18°–28°N. The IIE variability is analyzed using the empirical orthogonal function (EOF). The EOF results show that the first three modes respectively account for 64.1%, 27.7% and 5.0% of the IIE variability; hence, the leading EOF mode is a major IIE variation pattern.
Figure 5ashows that the distribution of the leading mode of the IIE has the same sign, suggesting that the IIE variability is in-phase at different latitudes. The corresponding time series (Figure 5b) is dominated mainly by interannual variation, and has no apparent trends. When the time coefficient changes by ±1 unit, the IIE shifts eastward or westward by 0.4°, with the largest variation of about 0.6° sitting between 22°N and 26°N. Because the second and third EOF modes account for much less of the variance than the first mode, we define the IIE index as the time coefficient corresponding to the first mode.
5. Relationship Among the IIE, ISM and EASM
 We now investigate the relationships of the IIE variation with the ISM and EASM. The time series of the IIE index from 1951 to 2008 yields five positive-anomaly years and six negative-anomaly years, with an anomaly being defined as a value outside a standard deviation of ±0.9 of the index (Figure 5b). The five positive-anomaly years are 1964, 1969, 1993, 1998, and 2007, while the six negative-anomaly years are 1957, 1960, 1962, 1963, 1965, and 2003.
Figure 6ashows that in the positive-anomaly years, the whole IIE shifts eastward on average by 0.7° (dashed line) from the long-term average (black solid line), agreeing with the variation characteristics inFigure 5a; the extremely positive IIE in 1964 is shown by the red solid line. The positive anomalies are most prominent between 22°N and 24°N where the IIE shifts eastward by more than 1.0° relative to the long-term average. Even though the positive anomaly is smallest at 28°N, the IIE still shifts eastward by 0.3°. For the positive-anomaly composite of surface winds (Figure 6c), the significant anomalous surface easterlies are primarily confined to the tropical region between 60°E and 135°E. Meanwhile, the region of the South China Sea to the east of 105°E is dominated by anomalous anticyclonic circulation, and there is a convergence zone to the north of the anomalous anticyclone over the lower and middle reaches of the Yangtze River Valley. Another anomalous northwest-southeast convergence zone appears near Yunnan in China with anomalous westerlies in the west and anomalous easterlies in the east. The anomalous pattern of surface wind indicates that the southwest summer monsoon over the Bay of Bengal is weaker than normal and the position of the western North Pacific subtropical high (WNPSH) is farther to the south and east than normal. At 500 hPa, the regions over 75°–135°E and 5°–30°N are dominated by significant positive anomalies (Figure 6e), suggesting that the WNPSH is more intense than normal and its position is farther south than normal. The relatively low geopotential height near Yunnan is also distributed from the northwest to the southeast (Figure 6e), conforming to the surface wind convergence anomalies (Figure 6c). At 200 hPa, significant convergence anomalies appear over the west of the Indochina Peninsula and the western Pacific Ocean (Figure 6g), contributing to anomalous anticyclones in the lower and middle levels of the troposphere across the two regions. Additionally, the region of 90°–100°E and 20°–30°N is dominated by convergence anomalies, which are also favorable to the anomalous anticyclone in the lower and middle levels of the troposphere in the same region (Figure 6c). Corresponding to the anomalous atmospheric circulation patterns, the summer rainfall (Figure 6i) reduces over part of India and most of the Indochina Peninsula, while summer rainfall is significantly enhanced over Bangladesh, southeastern Tibet, and the lower and middle reaches of the Yangtze River. The results obtained from station precipitation data, where 68 stations pass the significance test at a level of 0.1, agree with the distribution of significant rainfall anomalies across mainland China derived from the CRU precipitation data analysis (Figure 6i).
Goswami et al.  defined MHISA by the meridional wind (V) shear between 850 and 200 hPa (V850-V200) averaged over a larger region (10°–30°N, 70°–110°E). While Wang and Fan  defined WFIEA by U850 in (5°–15°N, 90°–130°E) minus U850 in (22.5°–32.5°N, 110°–140°E). MHISA and reversed WFIEA represent respectively the ISM and EASM strength in terms of both rainfall and monsoon circulation [Wang et al., 2008]. Here we calculate MHISA and reversed WFIEAas −0.03 and 1.24 under the positive-anomaly condition. The positive-anomaly composites of atmospheric circulation and rainfall resemble the case of a moderately weak ISM and stronger East Asian subtropical summer monsoon (Meiyu/Baiu) front over the lower and middle reaches of the Yangtze River Valley [e.g.,Huang et al., 2004, 2007; Wang et al., 2001; Wang and LinHo, 2002]. It is worth noting that the anomalous atmospheric circulation pattern is conducive to the formation of an anomalous anticyclone over 80°–98°E and 15°–30°N and an anomalous cyclone across the lower and middle reaches of the Yangtze River Valley, which results in the IIE being situated amid the anomalous anticyclone and the anomalous cyclone shifting eastward.
 In contrast to the case for the positive-anomaly composite,Figure 6bshows that the IIE shifts westward by 0.8° (dashed line) from the long-term average (black solid line), again agreeing with the variation characteristics inFigure 5a; the extremely negative IIE in 1957 is even farther westward (as shown by the blue solid line). The negative anomalies are most prominent between 22°N and 26°N, where the IIE shifts westward by at least 1.0° relative to the long-term average. Even though the negative anomaly is smallest at 18°N, the IIE still moves westward by 0.2° relative to the long-term average position.Figure 6dshows that significant anomalous westerlies extend from about 60°E and 20°N to 95°E and 20°N, significant anomalous easterlies of surface wind appear immediately east of 95°E and 20°N, and there is an anomalous northwest-southeast convergence zone along the west coast of the Indochina Peninsula. The Yangtze River Valley is dominated by an anomalous anticyclone. The anomalous pattern of the surface wind indicates that the southwest summer monsoon is stronger than normal over the Bay of Bengal and the WNPSH position is farther west than normal. At 500 hPa (Figure 6f), the negative geopotential height anomalies extend from northern India to the Indochina Peninsula, there are positive anomalies over the northeast of the Indochina Peninsula, and the positive anomalies pass the significance test at a level of 0.1 over the upper reach of the Yellow River Valley. At 200 hPa (Figure 6h), significant convergence anomalies of divergent wind prevail over the Indian subcontinent, southern China, and the middle reach of the Yellow River Valley, and they are conducive to the formation of anomalous anticyclones at the lower and middle levels of the troposphere in the three regions. Divergence anomalies mainly appear over the Indochina Peninsula and the region between the lower reach of the Yellow River Valley and the lower reach of Yangtze River Valley (Figure 6h). Corresponding to the anomalous atmospheric circulation pattern, the summer rainfall (Figure 6j) reduces in most parts of India, the northeastern Indochina Peninsula, and most parts of China, while the summer rainfall increases in central Indochina Peninsula, and between the lower reach of the Yellow River and the lower reach of the Yangtze River. The summer rainfall anomalies across mainland China obtained from station precipitation data, where 67 stations pass the significance test at a level of 0.1, agree with those obtained from CRU rainfall data (Figure 6j). We also calculated MHISA and reversed WFIEAas 0.23 and −0.31 for the negative-anomaly years. Therefore, the negative-anomaly composites of circulation and rainfall correspond to the moderately strong ISM and weaker EASM over southern China. It should be mentioned that the anomalous atmospheric circulation patterns are favorable to an anomalous cyclonic circulation appearing over 90°–100°E and 15°–28°N and an anomalous anticyclonic circulation over 102°–135°E and 15°–30°N, which in turn contribute to the IIE being located amid the anomalous cyclone and the anomalous anticyclone shifting westward.
 The composite results presented above were obtained with the IIE anomaly. In turn, the IIE variations for both ISM and EASM anomalies are analyzed. The ±0.5 standard deviation of MHISA and reversed WFIEA are set as criteria. We find that there are four anomalous states between the ISM and EASM, namely the positive or negative anomaly of the ISM with the positive or negative anomaly of the EASM. The values of anomalous IIE corresponding to the four conditions are given in Table 1.
Table 1. The Value of the IIE Index Under Different Configuration of the ISM and the EASMa
PISM and NISM are positive and negative anomaly of ISM respectively, while PEASM and NEASM are positive and negative anomaly of EASM respectively.
 In Table 1, the absolute values of the IIE index are equal to or larger than 0.5 under the first two conditions, but smaller than 0.5 under the last two conditions, indicating that the position anomaly of the IIE is strongly related to the out-of-phase variation between the ISM and EASM.
 To reveal the effects of ISM and EASM upon IIE, a bilinear regression model, in which MHISA and reversed WFIEA are independent variables and the IIE index is a dependent variable, is built as equation (2) with normalized data.
where the multiple correlation coefficient of equation (2) is 0.46 (α = 0.005); the regression coefficient of MHISAhas a student-t test value of 2.30 and that of reversed WFIEAhas a student-t test value of 3.29, and both correlations pass the significance test at a level of 0.05.
Equation (2) suggests that both the ISM and EASM have a significant effect on the interannual variability of the IIE, but the effect of the EASM is stronger than that of the ISM. The signs of the regression coefficients in equation (2)also indicate that the interannual variability of the IIE is mainly determined by the out-of-phase variation between the ISM and EASM. When reversed WFIEA and MHISA are lower (higher) than normal, suggesting that the EASM is stronger (weaker) than normal over the middle and low reaches of the Yangtze River Valley and the ISM is weaker (stronger) than normal, then the IIE index is higher (lower) and the IIE shifts eastward (westward).
 The abovementioned composite analysis results show there exists zonal movement of the IIE with the eastward and westward shifts of the WNPSH. This might implies that the WNPSH movement in the east-west direction may be an important physical process affecting the spatial distribution of the IIE. To further verify the relationship between the IIE and WNPSH and reveal the physical processes of IIE variability, we calculate the coefficients of correlation between the IIE index, the zonal WNPSH index defined by the JJA-mean anomalies of geopotential heights at 850 hPa averaged over 10°–30°N and 110°–150°E, and the meridional WNPSH index defined by the JJA-mean anomalies of geopotential heights at 850 hPa averaged over 30°–40°N and 120°–150°E [Lu, 2002]. The correlation coefficient between the IIE index and zonal WNPSH index reaches −0.29, passing the significance test at the 97.3% confidence level. However, the correlation coefficient between the IIE index and meridional WNPSH index is only 0.01. These results indicate that WNPSH movement in the west-east direction significantly affects the IIE position to a larger degree. When the zonal WNPSH index is higher (lower), meaning that the WNPSH position is more westward (eastward), the IIE index is lower (higher), meaning that the IIE position is more westward (eastward). The WPNSH movement in the north-south direction does not seem to affect the IIE position.
6. Conclusion and Discussion
 This study took advantage of the EPT's ability to detect different sources of wet air masses, and defines an IIE at the near-surface isobaric level according to the spatial distribution of the long-term averaged EPT. The long-term averaged position of the IIE at the near-surface isobaric level passes along 102.98°E, 18°N; 101.68°E, 20°N; 100.88°E, 22°E; 100.96°E, 24°N; 101.08°E, 26°N; and 99.32°E, 28°E with a wavy pattern. We further defined an objective and quantitative IIE index that can describe the year-to-year zonal movement of the IIE using 58-year NOAA reanalysis data and the EOF method.
 The results obtained from composite, correlation and regression studies agree with each other, indicating that the IIE index well describes the interannual variation in the interface between the ISM and EASM, and better identifies the driving force behind the zonal movement of the IIE. The eastward shift of the IIE is found to be associated with a stronger-than-normal EASM, a convergence zone accompanied with a heavier-than-normal summer rainfall over the lower and middle reaches of the Yangtze River Valley, and a weaker ISM. By the same token, the westward shift of the IIE is associated with a weaker-than-normal EASM, less-than-normal summer rainfall over the lower and middle reaches of the Yangtze River Valley, heavier-than-normal summer rainfall between the lower and middle reaches of the Yangtze River and the lower and middle reaches of the Yellow River, and a stronger ISM. When both the ISM and EASM are weak or strong, the IIE is located at its normal position.
 The circulation anomalies and rainfall anomalies well represent two teleconnection patterns associated with the Asian summer monsoon, agreeing with the results of some other studies [e.g., Guo and Wang, 1988; Hu and Nitta, 1996; Huang and Li, 1987; Kripalani and Kulkarni, 1997, 2001; Lau, 1992; Lau et al., 2000; Nitta, 1987]. This suggests that the IIE variation reflects to a large extent the interaction between the ISM and EASM. The analysis results for the monsoon index further indicate that, although the seesaw relation between the ISM and EASM mainly contributes to the IIE position anomaly, the relationship between the IIE and the EASM is much closer than that between the IIE and ISM; i.e., there is a stronger effect of the EASM on the IIE.
 In the East Asian Monsoon system, the zonal movement of the WNPSH may well be one of the key physical processes affecting the year-to-year variation in the IIE. When the WNPSH prevails over the South China Sea in the summer, the large-scale quasi-stationary rainbands to its north pull the IIE eastward (Figure 6c). By contrast, when the WNPSH dominates the lower and middle reaches of the Yangtze River Valley, the WNPSH pushes the IIE farther westward (Figure 6d). These results agree with previous studies that have already highlighted that the WNPSH is the most dominant component in EASM subsystem, and its position, shape and strength determinate the large-scale quasi-stationary frontal zones in East Asia [e.g.,Ding, 1994; Guo and Liu, 2008; Huang and Sun, 1992; Tao and Chen, 1987].
 The newly defined thermodynamic index for the IIE refines our understanding of the interaction between the ISM and EASM and its effects on the weather and climate over East Asia during summer. It may also be helpful for explaining the paleoclimate evidence provided by oxygen isotope records of speleothems in caves located around the IIE, such as the famous Dongge Cave in China (25.28°N, 108.08°E) and Longquan Cave (25.52°N, 107.83°E) [Yuan et al., 2004; Zhang et al., 2009], and thus broadens our perspectives in studying the drought and flood patterns associated with the Asian summer monsoon.
 This study was supported by the National Natural Science Foundation of China under contract U0933603 and the Science Foundation of Yunnan Province under contract 2009CC002.