A Contrast of the Monsoon–Tropical Cyclone Relationship between the Western and Eastern North Pacific

Abstract

The monsoon and tropical cyclone (TC) are principal components of global climate variability. The relationship between the monsoon intensity and the TC genesis frequency (TCGF) in different major monsoon regions has not been fully studied. Here, we compared the relationship of monsoon intensity and TCGF during the extended boreal summer between the western and eastern North Pacific, results of which revealed different monsoon–TC relationships (with opposite-sign correlations) in these two regions. A significant positive correlation could be found between the western North Pacific summer monsoon (WNPSM) index and the TCGF over the western North Pacific (WNP). In contrast, a significant negative correlation was identified between the North American summer monsoon (NASM) index and the TCGF over the eastern North Pacific (ENP). The observed different monsoon–TC relationships could be explained by the monsoon-associated changes in the environmental factors over the regions where TCs were formed and the influences from sea surface temperature (SST) anomalies across tropical ocean basins. By comparing the environmental factors in the TC genesis potential index (GPI), the mid-level relative humidity (vertical wind shear) was the factor to make the largest contribution to the monsoon-associated TC genesis changes over the WNP (ENP). In strong (weak) WNPSM years, the high (low) atmospheric mid-level relative humidity could promote (inhibit) the TCGF over the WNP, resulting in a significant positive monsoon–TC correlation. In contrast, in strong (weak) NASM years, the strong (weak) vertical wind shear could inhibit (promote) the TCGF over the ENP, thus leading to a significant negative monsoon–TC correlation. In addition, the WNPSM and the TCGF over the WNP could be modulated by the similar tropical Pacific–Atlantic SST anomalies jointly, thus leading to a significant positive correlation between the WNPSM and the WNP TCGF. In contrast, the signs of tropical Pacific–Atlantic SST anomalies influencing the NASM were almost opposite to those affecting the TCGF over the ENP, thus resulting in a significant negative correlation between the NASM and the ENP TCGF. The results obtained herein highlight the differences of the monsoon–TC relationship between the WNP and the ENP, which may provide useful information for the prediction of monsoon intensity and TC formation number over these two regions.

1. Introduction

The monsoon and tropical cyclone (TC) are principal components of global climate variability [1,2]. Both the western North Pacific (WNP) and the eastern North Pacific (ENP) can breed active monsoons and frequent TC genesis activities. The WNP is home to the world’s largest number of TCs [3] and also fosters the western North Pacific summer monsoon (WNPSM), which is one important subcomponent of the Asian summer monsoon [4,5]. Similarly, the ENP is also an important TC formation area with the second largest number in the world [6]; at the same time, there is the North American Summer Monsoon (NASM) in the adjacent region [7,8]. It is of great scientific and practical significance to understand the variation process and physical mechanism of the monsoons and TCs, especially considering the impact of natural disasters caused by the monsoon and TC activity changes on agriculture, economy, and society.

The interannual variability of the TC genesis frequency (TCGF) over the WNP may have a close relationship with the WNPSM [9,10,11]. It has long been recognized that a high proportion (about 70%) of TCs over the WNP are formed within the monsoon trough, which can provide favorable environmental conditions for breeding the TC formation [12,13,14,15]. Choi et al. [10] found a high positive correlation coefficient of 0.62 at a 99% confidence level between the WNPSM index and the TCGF over the WNP during the boreal summer (June–July–August; JJA) spanning a period from 1977 to 2013. They further conducted composite analyses of large-scale environmental fields between strong and weak WNPSM years to find out the reasons for this high positive correlation, results of which indicated that anomalous cyclonic circulations in the south of 30° N in the strong WNPSM years could fuel more TCs to be formed. Moreover, the local air-sea interaction and positive feedback could be included in the related processes. Zhao et al. [11] documented a strengthening of the interannual relationship between the WNPSM and the JJA TCGF over the WNP during recent decades, which further confirmed the significant positive correlation between the WNPSM index and the WNP TCGF, especially after the late 1990s. Basconcillo et al. [16] reported that the highest recorded WNPSM intensity and the highest TCGF over the WNP since 1984 could be observed simultaneously in JJA 2018, results of which also supported that there was expectedly higher WNP TCGF during JJA when the WNPSM was more intense.

However, the interannual relationship between the monsoon intensity and the TCGF in other major monsoon regions has not been fully studied, considering that previous studies on the monsoon–TC relationship mainly focused on the WNP monsoon region rather than other monsoon regions. The relationship between the NASM and the TCGF over the ENP has not been systematically investigated. Boos and Pascale [8] recently provided a new perspective on the processes that drive the formation of the NASM, suggesting that the formation of the NASM is a unique case in the world and may be different from other typical monsoons. The environmental field conditions corresponding to the formation of TCs over the ENP also have some unique attributes and characteristics [17]. Considering the uniqueness of the NASM and the ENP TC [8,17,18], the relationship between the NASM and the ENP TCGF may be different from that between the WNPSM and the WNP TCGF, which deserves more detailed investigations. In this study, we perform statistical analyses to compare the interannual relationship of the monsoon intensity and the TCGF during the extended boreal summer between the WNP and ENP. The rest of this paper is organized as follows. The data and methods used in this study are introduced in Section 2. The differences in the monsoon–TC relationship between the WNP and ENP are compared in Section 3. Section 4 illustrates possible explanations for different monsoon–TC relationships between the WNP and ENP. Conclusions and discussion are given in Section 5.

2. Data and Methods

In this study, we use the TC data spanning a period from 1979 to 2020 from the International Best Track Archive for Climate Stewardship (IBTrACS) version v03r10 [19], to depict and quantify the TC genesis activities over the WNP and ENP. Our analyses here are focused on TCs formed over the WNP and ENP during the extended boreal summer (June–September). Only TCs with at least the intensity of tropical storms are counted in this analysis, considering that there may be some uncertainties in the identification of weak systems that do not reach the intensity of tropical storms. The position where the TC first reaches the tropical storm intensity with the maximum sustained wind speed exceeding 17.2 m s⁻¹ is defined as the genesis location of the TC. The results reported in this study are not sensitive to the definition of the TC genesis location, and very similar results can be obtained by defining the TC genesis location as the position of the first record of TCs in the IBTrACS.

The monthly-mean atmospheric fields (including relative humidity, horizontal and vertical wind velocity, sea level pressure, etc) are derived from the National Center for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis product [20], which has a 2.5° × 2.5° horizontal resolution and extends from 1000 to 10 hPa with 17 vertical pressure levels. The monthly sea surface temperature (SST) used here is obtained from the National Oceanic and Atmospheric Administration (NOAA) Extended Reconstructed SST dataset version 5 (ERSSTv5) [21], with a spatial resolution of 2.0° × 2.0°. The NCEP–NCAR reanalysis data and the ERSSTv5 data are mainly used to examine the atmospheric and oceanic environmental factors related to the TC genesis.

To document and quantify the time evolution of the WNPSM and the NASM, the WNPSM index and the NASM index are utilized in this analysis. According to Yim et al. [22], the NASM index is defined as the differences in 850-hPa zonal winds during the extended boreal summer between a southern region (5°–15° N, 130°–100° W) and a northern region (20°–30° N, 110°–80° W). Similarly, the WNPSM index is defined as the differences in 850-hPa zonal winds during the extended boreal summer between a southern region (5°–15° N, 100°–130° E) and a northern region (20°–35° N, 110°–140° E). When the defined monsoon index is positive (negative), it means that there is a cyclonic (anticyclonic) circulation anomaly over the subtropical WNP and ENP.

The signal of interannual variability is obtained by applying a 9-yr high-pass filter to the anomalies, which are acquired by removing the long-term trend and mean seasonal cycle from the original time series. The statistical analysis methods used in this study include the Pearson correlation, composite, and regression analyses. The two-tailed Student’s t test is applied to determine the statistical significance levels.

The modified genesis potential index (GPI) [23] is used in this analysis to diagnose the contributions of different environmental factors to the monsoon-associated TC genesis changes over the WNP and ENP. According to Murakami et al. [23], the modified GPI is defined as

$$\mathrm{GPI}=|{10}^5{\mathsf{\zeta }}^{\frac{3}{2}}|{\left(\frac{\mathrm{H}}{50}\right)}^3{\left(\frac{{\mathrm{V}}_{\mathrm{pot}}}{70}\right)}^3{\left(1+0.1{\mathrm{V}}_{\mathrm{shear}}\right)}^{-2}\left(\frac{-\mathsf{\omega }+0.1}{0.1}\right)$$

(1)

where $\mathsf{\zeta }$ is the absolute vorticity at 850-hPa, H is the relative humidity at 600-hPa, V_pot is the potential intensity, V_shear stands for the magnitude of the vertical wind shear, and $\mathsf{\omega }$ represents the atmospheric vertical velocity at 500-hPa. The vertical wind shear, calculated as the magnitude of the vector difference between horizontal winds at 850-hPa and 200-hPa, is defined as

$${\mathrm{V}}_{\mathrm{shear}}=\sqrt{{\left({\mathrm{u}}_{850}-{\mathrm{u}}_{200}\right)}^2+{\left({\mathsf{\nu }}_{850}-{\mathrm{v}}_{200}\right)}^2}$$

(2)

where ${\mathrm{u}}_{850}$ and ${\mathrm{u}}_{200}$ are the horizontal zonal winds at 850-hPa and 200-hPa, respectively; ${\mathrm{v}}_{850}$ and ${\mathrm{v}}_{200}$ are the horizontal meridional winds at 850-hPa and 200-hPa, respectively. According to Bister and Emanuel [24], the potential intensity is defined as

$${\mathrm{V}}_{\mathrm{pot}}^2=\frac{{\mathrm{C}}_{\mathrm{k}}{\mathrm{T}}_{\mathrm{s}}}{{\mathrm{C}}_{\mathrm{D}}{\mathrm{T}}_0}\left({\mathrm{CAPE}}^{*}-{\mathrm{CAPE}}^{\mathrm{b}}\right)$$

(3)

where ${\mathrm{C}}_{\mathrm{k}}$ is the exchange coefficient for enthalpy and ${\mathrm{C}}_{\mathrm{D}}$ is the drag coefficient,${\mathrm{T}}_{\mathrm{s}}$ is SST and ${\mathrm{T}}_0$ is the mean outflow temperature. The convective available potential energy (CAPE) is the vertical integral of parcel buoyancy as a function of parcel temperature, pressure, and specific humidity. ${\mathrm{CAPE}}^{*}$ is the CAPE for an air parcel at the radius of maximum winds and ${\mathrm{CAPE}}^{\mathrm{b}}$ denotes the CAPE for ambient boundary layer air with its pressure being reduced to its value at the radius of maximum winds.

Following analysis methods used in previous studies [25,26], the GPI index is further split into five items ($\mathrm{A}=|{10}^5{\zeta }^{\frac{3}{2}}|$, $\mathrm{B}={\left(\frac{\mathrm{H}}{50}\right)}^3$, $\mathrm{C}={\left(\frac{{\mathrm{V}}_{\mathrm{pot}}}{70}\right)}^3$, $\mathrm{D}={\left(1+0.1{\mathrm{V}}_{\mathrm{shear}}\right)}^{-2}$, $\mathrm{E}=\left(\frac{-\mathsf{\omega }+0.1}{0.1}\right)$), to facilitate the diagnosis of the relative contributions from different environmental factors to TC genesis changes. Further, each term is expressed as the sum of the average and the perturbation. From this, we can get the following expression

$${\mathrm{GPI}}^{\prime }=\overline{\mathrm{ABCD}}{\mathrm{E}}^{\prime }+\overline{\mathrm{ABCE}}{\mathrm{D}}^{\prime }+\overline{\mathrm{ABDE}}{\mathrm{C}}^{\prime }+\overline{\mathrm{ACDE}}{\mathrm{B}}^{\prime }+\overline{\mathrm{BCDE}}{\mathrm{A}}^{\prime }+\mathrm{F}$$

(4)

where, the overbar denotes a time average and the prime denotes a perturbation. The first five terms on the right-hand side of Equation (4) represent the contribution of each factor in the modified GPI, and the last item (F) represents the non-linear term.

The relative humidity is defined as $\frac{\mathrm{q}}{{\mathrm{q}}_{\mathrm{s}}}$, where q and q_s denote the specific humidity and the statured specific humidity, respectively. In response to local warming, the q and q_s tend to increase simultaneously. Thus, the relative humidity is not determined by the SST [27]. Accordingly, the moisture budget is calculated to diagnose the physical processes contributing to the mid-level atmospheric relative humidity. The water vapor budget equation can be expressed as

$$\frac{\partial {\mathrm{q}}^{\prime }}{\partial \mathrm{t}}=-{\left(\overrightarrow{\mathrm{V}}·\nabla \mathrm{q}\right)}^{\prime }-{\left(\mathsf{\omega }\frac{\partial \mathrm{q}}{\partial \mathrm{p}}\right)}^{\prime }-{\left(\frac{{\mathrm{Q}}_2}{\mathrm{L}}\right)}^{\prime }$$

(5)

where q is the specific humidity, $\overrightarrow{\mathrm{V}}$ is the horizontal velocity, $\mathsf{\omega }$ is the p-vertical velocity, ${\mathrm{Q}}_2$ is apparent moisture sink, L is the latent heat constant. $-{\left(\overrightarrow{\mathrm{V}}·\nabla \mathrm{q}\right)}^{\prime }$ represents anomalous horizontal moisture advection. $-{\left(\mathsf{\omega }\frac{\partial \mathrm{q}}{\partial \mathrm{p}}\right)}^{\prime }$ indicates anomalous vertical moisture advection. $-{\left(\frac{{\mathrm{Q}}_2}{\mathrm{L}}\right)}^{\prime }$ denotes anomalous moisture source or sink. In this study, we find that the contribution of $-{\left(\mathsf{\omega }\frac{\partial \mathrm{q}}{\partial \mathrm{p}}\right)}^{\prime }$ to the change of 600-hPa relative humidity is particularly important, and this term can be further divided into three terms:

$$-{\left(\mathsf{\omega }\frac{\partial \mathrm{q}}{\partial \mathrm{p}}\right)}^{\prime }=-\overline{\mathsf{\omega }}\frac{\partial {\mathrm{q}}^{\prime }}{\partial \mathrm{p}}- {\mathsf{\omega }}^{\prime }\frac{\partial \overline{\mathrm{q}}}{\partial \mathrm{p}}- {\mathsf{\omega }}^{\prime }\frac{\partial {\mathrm{q}}^{\prime }}{\partial \mathrm{p}}$$

(6)

where, $-\overline{\omega }\frac{\partial {\mathrm{q}}^{\prime }}{\partial \mathrm{p}}$ indicates the vertical anomalous vapor transport caused by the mean vertical velocity, $- {\mathsf{\omega }}^{\prime }\frac{\partial \overline{\mathrm{q}}}{\partial \mathrm{p}}$ represents the vertical mean vapor transport caused by the anomalous vertical velocity, and ${\mathsf{\omega }}^{\prime }\frac{\partial {\mathrm{q}}^{\prime }}{\partial \mathrm{p}}$ denotes the residual term.

3. Contrasting the Monsoon–TC Relationship between the WNP and ENP

In the following, we first calculated the monsoon indexes (the WNPSM index and the NASM index) and the TCGF anomalies over the WNP and ENP during the extended boreal summer. Figure 1a depicted the time series of the NASM index and the TCGF over the ENP spanning the period of 1979–2020, which exhibited prominent interannual variations. We noticed that fluctuations of the two-time series in Figure 1a tended to have an out-of-phase relation, implying that the interannual variation of the NASM intensity and the TCGF over the ENP might be closely tied with each other. The direct calculation of the correlation coefficient (r = $-0.45$) could confirm a significant negative correlation between the NASM index and the TCGF over the ENP (Figure 1c). The time series of the WNPSM index and the TCGF over the WNP were presented in Figure 1b, in which we noticed that fluctuations of the two-time series tended to be in phase. There was a significant positive correlation of 0.51 between the intensity of the WNPSM and the TCGF in the WNP (Figure 1d), which were consistent with the results in previous studies [10,11]. Therefore, different monsoon–TC relationships (with opposite-sign correlations) could be identified in the WNP and ENP, with a significant positive (negative) correlation between the WNPSM and the WNP TCGF (between the NASM and the ENP TCGF). Over the WNP, the significant positive correlation indicated that a stronger (weaker) WNPSM tended to be accompanied by more (less) TCs. In contrast, over the ENP, the negative correlation indicated that the TCGF tended to be inhibited (promoted) in the year with a stronger (weaker) NASM. The monsoon–TC interannual synchronization and co-variability could be found over both the WNP and the ENP. However, the WNPSM and the WNP TCGF (over the WNP) were in-phase co-variability; in contrast, the NASM and the ENP TCGF (over the ENP) were out-of-phase co-variability.

Next, the differences in TC genesis activities between strong and weak monsoon years were further compared. The years with a monsoon index larger (lower) than 0.55 standard deviation are defined as strong (weak) monsoon years. According to these thresholds, the sample sizes of NASM and WNPSM in weak, normal, and strong monsoon years are approximately the same. Totally, 13 strong WNPSM years, 11 weak WNPSM years, 12 strong NASM years, and 13 weak NASM years were chosen in this analysis (Figure 2). The selected strong WNPSM years included 1982, 1985, 1986, 1990, 1994, 2001, 2002, 2004, 2009, 2011, 2012, 2018 and 2019; while the selected weak WNPSM years were 1983, 1988, 1993, 1995, 1996, 1998, 2007, 2008, 2010, 2017 and 2020. Meanwhile, the chosen strong NASM years included 1979, 1981, 1988, 1995, 1998, 1999, 2005, 2008, 2010, 2011, 2013 and 2017; while the selected weak NASM years were 1982, 1986, 1987, 1990, 1991, 1992, 1994, 2001, 2006, 2009, 2015, 2019 and 2020.

The genesis numbers of TCs were significantly different in the selected strong and weak monsoon years. Over the WNP, the mean TCGF during the strong WNPSM years was 17.15, while the mean value during the weak WNPSM years was reduced to 13.85. These observed changes in the TCGF in strong and weak WNPSM years were consistent with the positive correlation between the WNPSM and the WNP TCGF as shown in Figure 1d. Over the ENP, the mean TCGF during the strong NASM years was 10.92, while the mean value during the weak NASM years was boosted to 15.54. These observed changes in the TCGF in strong and weak NASM years were also consistent with the negative correlation between the NASM and the ENP TCGF as shown in Figure 1c. The above differences in the mean TCGF between strong and weak monsoon years were both significant at the 95% confidence level.

The changes of the TCGF in each sub-region in strong and weak monsoon years were further examined. According to Liu et al. [28], we further divided the WNP into five sub-domains (Figure 3), including the I sub-region (0°–40° N, 100° E–120° E), the II sub-region (20° N–40° N, 120° E–140° E), the III sub-region (0°–20° N, 120° E–140° E), the IV sub-region (20° N–40° N, 140° E–180°) and the V sub-region (0°–20° N, 140° E–180° E). Similarly, four sub-regions in the ENP were also selected to facilitate our analysis (Figure 3), including the VI region (15° N–40° N, 180–115° W), the VII region (0°–15° N, 180°–115° W), the VIII region (15° N–40° N, 115° W–80° W), and the IX region (0°–15° N, 115° W–80° W). The mean TCGF in all these sub-regions were listed and compared for weak and strong monsoon years in

Table 1. Comparisons of the mean TCGF in each sub-regions between the strong monsoon years and weak monsoon years.

Sub-Region	Strong Monsoon Years	Weak Monsoon Years	The Difference
I	3.54	3.64	−0.10
II	2.15	2.91	−0.76
III	4.92	3.27	1.65
IV	1.85	1.91	−0.06
V	4.69	2.09	2.60 *
VI	0.50	1.69	−1.19
VII	1.50	4.38	−2.88 *
VIII	3.83	2.92	0.91
IX	4.75	6.54	−1.78

Note: the asterisk (*) represent that the difference between the strong and weak monsoon years is statistically significant at the 99% confidence level.

, to evaluate the relative contribution of changes in each sub-region to the total TC number changes in the entire region. The results revealed that the spatial distribution of the TC number change was uneven, with obvious spatial heterogeneity.

According to Table 1, the V sub-region (the southeast quadrant of the WNP) showed the most obvious changes between weak and strong monsoon years, indicating that this sub-region in the southeast quadrant may be the sub-region making the largest contribution to the total TC number changes of the entire WNP region. The average TC number formed in the V sub-region was 4.69 during the strong WNPSM years, whereas the value decreased to 2.09 during the weak WNPSM years. This difference of 2.60 TCs between strong and weak WNPSM years was significant at the 99% confidence level. Over the ENP, the VII sub-region (the southwest quadrant of the ENP) was identified to be the sub-region with the most pronounced changes in the TCGF between strong and weak NASM years (Table 1), suggesting that this sub-region in the southwest quadrant may make the greatest contribution to the overall TC number changes of the entire ENP region. The mean TC number formed in the VII sub-region increased from 1.50 in the strong NASM years to 4.38 in the weak NASM years, which indicated an increase of more than 190% in the mean TCGF. Thus, in the following analysis, we mainly focused on these two important sub-regions (the V sub-region and VII sub-region) that make the greatest contributions to the overall TC number changes of the entire region.

This spatial heterogeneity in the spatial distribution of the TC number change might lead to the shifts of mean TC genesis locations in strong and weak monsoon years (Figure 4). According to Figure 4, the differences in the mean TC genesis locations between strong and weak monsoon years were notable. Over the WNP, the mean TC genesis location was (17.91° N, 135.25° E) in the strong WNPSM years and (19.13° N, 132.18° E) in the weak WNPSM years. Therefore, in the strong WNPSM year, the mean TC genesis location over the WNP tended to shift to the southeast, which was consistent with more TCs formation in the V sub-region (the southeast quadrant of the WNP) during the strong WNPSM years (Figure 3 and Table 1). This difference in the TC genesis location between the positive monsoon years and negative monsoon years also manifested itself over the ENP. The mean TC genesis location over the ENP was (13.86° N, 116.83° W) in the strong NASM years and (14.76° N, 109.62° W) in the weak NASM years. Thus, the mean TC genesis location over the ENP tended to move to the southwest in the weak NASM years, which was consistent with more TCs formation in the VII sub-region (the southwest quadrant of the ENP) during the weak NASM years (Figure 3 and Table 1). Overall, the mean genesis locations of TCs tended to show an obvious spatial shift over both the WNP and the ENP, along with the variations in the monsoon intensity. The strengthened WNPSM might result in a southeastward shift of the mean TC genesis location over the WNP, while the weakened NASM might lead to a southwestward shift of the TC genesis location over the ENP.

4. Possible Explanations for Different Monsoon–TC Relationships over the WNP and ENP

The above analyses in Section 3 show that the TCGF over the WNP displayed a robust positive relationship with the intensity of the WNPSM. In contrast, the TCGF over the ENP showed a significant negative correlation with the intensity of the NASM. A question to be addressed is why there are different monsoon–TC relationships over the WNP and ENP (a significant positive correlation between the WNPSM and the WNP TCGF, but a significant negative correlation between the NASM and the ENP TCGF). This scenario could be understood from two possible perspectives, including the monsoon-associated changes in the environmental factors over the regions where TCs were formed and the influences from SST anomalies across tropical ocean basins.

4.1. Monsoon-Associated Changes in the Environmental Factors Defined in the GPI

In the following section, monsoon-associated changes in the environmental factors defined in the modified GPI would be further examined, which could provide more insights into the reasons why different monsoon–TC relationships were observed over the WNP and the ENP. Different changes in environmental factors defined in the modified GPI associated with variations of the WNPSM and the NASM might lead to different monsoon–TC relationships.

4.1.1. The Modified GPI

Composite differences in the calculated modified GPI (according to Equation (1)) between the strong and weak monsoon years were depicted in Figure 5. Lower values of the modified GPI were observed in the VII sub-region over the ENP in the strong NASM years than those in the weak NASM years (Figure 5a), which was consistent with the fact that fewer TCs were formed over the ENP during the strong NASM years. In contrast, higher values of the modified GPI were identified in the V sub-region over the WNP in the strong WNPSM years than those in the weak WNPSM years (Figure 5b), which was consistent with the increased TCGF over the WNP during the strong WNPSM years.

In addition, significant correlations could be found between the area-mean modified GPI and the TCGF (the monsoon index) (

Table 2. Correlations between the local monsoon index, the TCGF over the sub-regions and the area-mean modified GPI over the sub-regions.

Sub-Region	Local Monsoon Index and GPI	TCGF and GPI
V	0.69 *	0.49 *
VII	−0.58 *	0.72 *

Note: the asterisk (*) represents that the correlation is statistically significant at the 95% confidence level.

). The area-mean GPI over the VII sub-region of the ENP was positively correlated with the TCGF over the VII sub-region of the ENP (r = 0.72) and negatively correlated with the NASM index (r = −0.58). In contrast, the area-mean GPI over the V sub-region of the WNP was both positively correlated with the TCGF (r = 0.49) and the WNPSM index (r = 0.69). These results suggested that the modified GPI could capture and reproduce the main features of monsoon-associated TC genesis variations over both the ENP and the WNP. The calculated modified GPI could account for the changes of TCGF over the WNP and ENP in the strong and weak monsoon years.

How do different environmental factors contribute to the monsoon-associated TC genesis changes over the WNP and ENP? Which factor is the most important compared with other factors in the modified GPI over the WNP and ENP? The relative contributions of different environmental factors in the modified GPI were further diagnosed (Figure 6). By comparing the environmental factors in the GPI, the vertical wind shear (dynamic factor) was the factor to make the largest contribution to the monsoon-associated changes in the environmental factors affecting the TC activity over the ENP. In contrast, the mid-level relative humidity (thermodynamic factor) was the factor to make the largest contribution over the WNP as implied by the modified GPI. These results suggest the different roles of large-scale environmental factors in the monsoon-associated TC genesis changes over the WNP and the ENP.

4.1.2. Vertical Wind Shear

The vertical wind shear is known to be an important dynamic environmental factor affecting the TC genesis [29,30], with a weak vertical wind shear being necessary for TC genesis and a strong vertical wind shear tending to be a major impediment to TC formation. Low vertical wind shear may be conducive to the accumulation of moisture and enthalpy in a vertical air column, thus helping the system to organize itself into a vertically stacked tropospheric system by losing little moisture and heat energy. Instead, the strong vertical wind shear may hinder the TC genesis by breaking the warm core structure of the TC due to the ventilation effect. According to Figure 6, the vertical wind shear was the factor making the largest contribution to the monsoon-associated changes in the environmental factors affecting the TC activity over the ENP. Therefore, the changes in the vertical wind shear associated with the monsoon will be examined in more detail in the following.

The differences in vertical wind shear between the strong and weak NASM years showed that larger vertical wind shear was found over the VII sub-region of the ENP (Figure 7a), which could explain the decrease of the TCGF over this region during the strong NASM monsoon years. In the strong NASM years, westerly wind anomalies at 200-hPa (Figure 8a) and easterly wind anomalies at 850-hPa (Figure 9a) dominated in the VII sub-region over the ENP. The observed increase of the vertical wind shear in this area was mainly due to the changes in the upper-level winds, considering that the anomalous winds in the upper level were relatively stronger than those in the lower level. The variations of the vertical wind shear over the ENP TC formation region were considered to be largely modulated by the tropical Pacific–Atlantic SST anomalies [31,32], by triggering the zonal wind change across Central America. Significant signals of SST anomalies in the tropical Pacific and Atlantic could also be observed in the results shown in Figure 9a.

For the WNP, the differences in vertical wind shear between the strong and weak WNPSM years showed an east-west dipole pattern over the V sub-region (Figure 7b). The enhanced vertical wind shear resided over the west flank of the V region, whereas the east flank was dominated by the weakened vertical wind shear. This uneven spatial distribution in the vertical wind shear changes might lead to inconsistencies in the effects of the vertical wind shear across the whole region. Therefore, the vertical wind shear made a relatively small contribution to the monsoon-associated changes in the environmental factors affecting the TC activity over the WNP (Figure 6b), which was obviously different from the contribution of vertical wind shear over the ENP (Figure 6a).

4.1.3. Mid-Level Atmospheric Relative Humidity

The mid-level atmospheric relative humidity is one key thermodynamic environmental factor affecting the TC formation [29], as high values of relative humidity could provide sufficient moisture to fuel the TC genesis. Sufficient water moisture is necessary to support a persistent large amount of latent heat release [33], thus providing an energy source for the TC formation. According to the above analyses, the mid-level relative humidity was found to be the environmental factor making the largest contribution to the monsoon modulation of the WNP TC genesis (Figure 6b), as implied by the modified GPI. In addition, the 600-hPa relative humidity was also found to be one important environmental factor for the TC genesis over the ENP, making the second largest contribution to the monsoon-associated TC genesis changes (Figure 6a). Accordingly, the changes in the mid-level relative humidity associated with the monsoons will be further analyzed in the following.

Figure 10b presented the difference of 600-hPa relative humidity between the strong and weak WNPSM years. Enhanced 600-hPa relative humidity was observed over the V sub-region of the WNP, which was consistent with more active TC genesis activities over the V sub-region of the WNP during strong WNPSM years. This observed strengthened relative humidity over the V sub-region was conducive to the genesis of the local TCs, by providing sufficient moisture to fuel the TC formation. Consistent with the increase of mid-level humidity, the intensity of precipitation over the WNPSM monsoon area was also strengthened during strong WNPSM years (Figure 8b). To diagnose the physical processes contributing to the mid-level atmospheric relative humidity change, the moisture budget was further calculated according to Equation (5) and comparative analyses were conducted for the moisture budget between the strong and weak WNPSM monsoon years. By comparing the contributions of different terms in the water vapor budget equation (Figure 11), the vertical moisture advection term ($-{\left(\mathsf{\omega }\frac{\partial \mathrm{q}}{\partial \mathrm{p}}\right)}^{\prime }$) dominantly contributed to the monsoon-associated variations of the mid-level relative humidity over the V sub-region of the WNP. By further comparing the three terms for the vertical moisture advection (Equation (6)), the term associated with the anomalous vertical velocity and the mean specific humidity ($- {\mathsf{\omega }}^{\prime }\frac{\partial \overline{\mathrm{q}}}{\partial \mathrm{p}}$) made the largest contribution.

It should be noted that dynamic factors (the 500-hpa vertical velocity and the 850-hPa relative vorticity) could also make important contributions to the TC genesis over the WNP (Figure 6b). As suggested by Murakami and Wang [34], the 600-hPa relative humidity tended to be highly correlated with the 500-hPa vertical velocity. The changes in 600-hPa relative humidity were partly attributable to changes in dynamic factors (e.g., the 500-hPa vertical velocity). Thus, the role of the 600-hPa relative humidity in the TC formation might actually involve some contributions from dynamic factors (e.g., the 500-hPa vertical velocity).

For the ENP, the differences in the 600-hPa relative humidity between the strong and weak NASM years were depicted in Figure 10a. Negative relative humidity anomalous were evident over the VII sub-region, which was consistent with suppressed TC genesis activities there during the strong NASM years. However, positive anomalous relative humidity resided over the NASM monsoon area (Figure 10a). The composite difference of the precipitation displayed a very similar spatial pattern (Figure 8a). The moisture budget analyses revealed that the vertical advection associated with anomalous vertical velocity dominantly contributed to the variations of the mid-level relative humidity (Figure 11a). During the strong NASM years, the vertical mean vapor transport caused by the anomalous vertical velocity over the VII sub-region of the ENP tended to reduce the 600-hPa relative humidity, thus suppressing the formation of local TCs.

In summary, the strong WNPSM (NASM) could increase (decrease) the mid-level relative humidity over the V sub-region (VII sub-region) by enhancing (weakening) the vertical moisture transports caused by the anomalous vertical velocities, thus being able to support (suppress) the formation of TCs over the WNP (ENP).

4.1.4. Other Related Environmental Factors

In addition to the vertical wind shear and mid-level relative humidity, there are three other environmental factors in the modified GPI (Equation (1)): the low-level atmospheric relative vorticity, the mid-level atmospheric vertical velocity, and TC potential intensity. These other related environmental factors will be examined in the following.

The low-level relative vorticity is also one dynamic environmental factor affecting the TC genesis [29], with a large positive low-level cyclonic vorticity being considered as one necessary condition for TC formation. The low-level relative vorticity is considered to be important for TC genesis over the WNP [35,36,37]. According to the differences in the 850-hPa relative vorticity between the strong and weak WNPSM years (Figure 7b), positive lower-level vorticity anomalies were evident over the V sub-region of the WNP, which was associated with the cyclonic circulation anomalies over the WNP during the strong WNPSM years (Figure 9b). The tropical Pacific and Atlantic SST anomalies might play an important role in initiating these low-level positive relative vorticity and cyclonic circulation anomalies over the WNP [38,39]. These positive low-level relative vorticity anomalies were conducive to the TC formation over the WNP. Therefore, positive contributions could also be found from the low-level relative vorticity to the monsoon modulation of TC genesis over the WNP (Figure 6b). In contrast, over the ENP, the contribution from the low-level relative vorticity was small to the monsoon-associated TC genesis changes (Figure 6a).

According to the moisture budget analyses above, atmospheric vertical velocity is important for vertical moisture transport. The composite differences in the 500-hPa vertical velocity between the strong and weak monsoon years were shown in Figure 10. Significant positive 500-hPa vertical velocity anomalies (anomalous descending motion) were evident over the VII sub-region of the ENP during the strong NASM years (Figure 10a), which could hinder the formation of TCs. These observed enhanced positive vertical velocity anomalies (descending motion) over the VII sub-region were concurrent with negative vertical velocity anomalous (ascending motion) over both the tropical WNP and north Atlantic (Figure 10a). The anomalous descending motion over the VII sub-region of the ENP was consistent with the observed suppressed precipitation (Figure 8a), low-level divergence (Figure 9a) and upper-level convergence (Figure 8a) of winds. In contrast, significant negative 500-hPa vertical velocity anomalies (anomalous ascending motion) were evident over the V sub-region of the WNP during the strong WNPSM years (Figure 10b), which could promote the formation of TCs. The anomalous ascending motion over the V sub-region of the WNP was consistent with observed enhanced precipitation there (Figure 8b).

The potential intensity is one thermodynamic parameter used in the modified GPI to quantify the effects of SSTs and moist static stability on TC genesis. The composite differences of the potential intensity (not shown) were very similar to the distribution pattern of the composite differences for SST (Figure 9). For the ENP, cold SST anomalies, along with the reduced values of potential intensity were found over the VII sub-region of the ENP (Figure 9a), which tended to hinder the TC formation there. For the WNP, an east-west dipole pattern in the composite differences of SST was identified over the V sub-region (Figure 9b), the uneven spatial distribution of which might lead to the inconsistencies of the effects from the SST and potential intensity across the whole region. In general, compared with other environmental factors, the potential intensity made a relatively small contribution to the monsoon-associated TC genesis changes over both the WNP and the ENP (Figure 6).

4.2. Influences from SST Anomalies across Tropical Ocean Basins

The influences from tropical SST anomalies across tropical ocean basins may also contribute to the significant monsoon–TC relationships over both the WNP and ENP. The related SST configurations across the tropical ocean basins may be the primary driving force to modulate the monsoon and the TCGF together and thus trigger the monsoon–TC synchronization and co-variability. It is well known that local warm SSTs are important for the TC genesis by modulating TC formation locally and directly [29]. The remote SST anomalies can also modulate the variation of TC genesis to a large extent, by affecting the large-scale environmental factors in regions where TC are formed via atmospheric teleconnection [39,40,41,42,43,44]. For example, the vertical wind shear was found to be the factor making the largest contribution to the monsoon-associated changes in the environmental factors affecting the TC activity over the ENP. But what drives the variations of the vertical wind shear? The tropical Pacific and tropical Atlantic SST anomalies were considered to be the key modulator to control the variability of the vertical wind shear over this region [31,32]. Therefore, it is necessary and important to reveal the influences of remote SST anomalies across tropical ocean basins on TC and monsoon activities.

Very similar spatial distribution characteristics of SST anomalies could be identified in the regressed SST anomalies during the extended boreal summer with respect to the WNPSM index (Figure 12a) and the TCGF over the V sub-region over the WNP (Figure 12b), including notable warm SST anomalies in the tropical central Pacific and cold SST anomalies in the tropical Atlantic. The warm SST anomalies in the tropical central Pacific, accompanied by the cold SST anomalies in the western Pacific at the same time, resemble SST anomalies associated with El Niño. The warm SST anomalies in the tropical central Pacific could promote cyclonic circulation anomalies over the WNP through a Gill-type atmospheric response [45], thus being able to foster a stronger WNPSM [38] and more active TC genesis activities over the WNP [44]. The tropical Atlantic SST anomalies could excite the atmospheric Kelvin wave [40] and Rossby wave [39] to modulate the atmospheric circulation anomalies over the WNP. Meanwhile, the tropical Atlantic could also induce trans-basin zonal vertical circulation anomalies [46,47]. Therefore, the tropical Atlantic SST anomalies could also modulate the variations of the WNPSM [47,48] and the TC genesis over the WNP [39,40]. The cold tropical Atlantic SST anomalies were helpful to a stronger WNPSM and more TCs formed over the WNP. These results imply that the WNPSM and the WNP TCGF could be modulated by the similar tropical Pacific–Atlantic SST anomalies jointly (Figure 12), thus leading to a significant positive correlation between the WNPSM and the WNP TCGF.

In contrast, similar spatial distribution characteristics of SST anomalies (but with opposite signs) could be observed in the regressed simultaneous SST anomalies with respect to the NASM index (Figure 13a) and the TCGF over the VII sub-region of the ENP (Figure 13b), including prominent SST anomalies in the tropical central Pacific and the tropical Atlantic. Noticeable cold (warm) SST anomalies in the tropical central Pacific (in the tropical Atlantic and western Pacific) were evident in the regressed SST anomalies during the extended boreal summer with respect to the NASM index (Figure 13a). Conversely, obvious warm (cold) SST anomalies in the tropical central Pacific (in the tropical Atlantic and western Pacific) were present in the regressed simultaneous SST anomalies with respect to the TCGF over the VII sub-region of the ENP (Figure 13b). The regressed SST anomalies with respect to the TCGF over the VII region (Figure 13b) were nearly the mirror image of those with respect to the NASM (Figure 13a). Previous studies [31,32] have demonstrated the combined effects of the tropical Atlantic and Pacific SST anomalies on modulating the ENP TC genesis, by inducing the zonal wind change across Central America and then modulating the vertical wind shear over the region where TCs are formed. The warm tropical Atlantic and cold tropical central-east Pacific SST anomalies could induce wind anomalies at 850-hPa and 200-hPa and then increase the vertical wind shear over the region where the ENP TCs are formed, thus being able to inhibit the TCGF over the ENP. At the same time, the warm tropical Atlantic and cold tropical central-east Pacific SST anomalies could trigger cyclonic circulation anomalies in the lower troposphere over the NASM region, thus helping to promote a stronger NASM. The opposite situation would occur in the case of cold tropical Atlantic and warm cold tropical central-east Pacific SST anomalies, with the weakened NASM and enhanced TC genesis activities over the ENP. The signs of tropical Pacific–Atlantic SST anomalies influencing the NASM were almost opposite to those affecting the ENP TCGF (Figure 13), thus resulting in a significant negative correlation between the NASM and the ENP TCGF.

The results of the composite analyses for the differences in SSTs and 850-hPa winds between the strong and weak monsoon years (Figure 9) were very similar to those obtained by the regression analyses (Figure 12a and Figure 13a). Therefore, the results of both regression analyses (Figure 12 and Figure 13) and composite analyses (Figure 9) could support the close associations between tropical Pacific and Atlantic SST anomalies and variations of the WNPSM and the NASM. Based on these above analyses, both the statistical relationships and physical mechanisms could jointly support the influences from remote SST anomalies across tropical ocean basins.

5. Conclusions and Discussion

In this study, we compared the interannual relationship of the monsoon intensity and the TCGF during the extended boreal summer between the WNP and the ENP, both regions of which could breed active monsoons and frequent TC genesis activities. Although the close relationship between the WNPSM and the TCGF over the WNP has been noticed in previous studies, the relationship between the NASM and the TCGF over the ENP is rarely involved and considered in previous studies. Considering that previous studies on the monsoon-TC relationship mainly focused on the WNP monsoon region rather than other monsoon regions, the interannual relationship between the monsoon intensity and the TCGF in other major monsoon regions has not been fully studied. The results of this study are dedicated to filling the knowledge gap in this area.

The results obtained herein revealed different monsoon–TC relationships (with opposite-sign correlations) over the WNP and the ENP (Figure 14). A significant positive correlation could be found between the WNPSM index and the TCGF over the WNP. In contrast, a significant negative correlation was identified between the NASM index and the TCGF over the ENP. The observed different monsoon–TC relationships could be explained from two possible perspectives, including the monsoon-associated changes in the environmental factors over the regions where TCs were formed and the influences from SST anomalies across tropical ocean basins (Figure 14).

From the perspective of the monsoon-associated changes in the environmental factors over the regions where TCs were formed, different changes in environmental factors defined in the modified GPI associated with variations of the WNPSM and the NASM could explain the observed different monsoon–TC relationships. By comparing the environmental factors in the GPI, the vertical wind shear (dynamic factor) was the factor to make the largest contribution to the monsoon-associated changes in the environmental factors affecting the TC activity over the ENP. In contrast, the mid-level relative humidity (thermodynamic factor) was the factor to make the largest contribution to the monsoon-associated TC genesis changes over the WNP. The moisture budget analyses revealed that the vertical advection associated with the anomalous vertical velocity dominantly contributed to the variations of the mid-level relative humidity. In strong (weak) WNPSM years, the high (low) atmospheric mid-level relative humidity could promote (inhibit) the TCGF over the WNP, resulting in a significant positive monsoon–TC correlation. In contrast, in strong (weak) NASM years, the strong (weak) vertical wind shear could inhibit (promote) the TCGF over the ENP, thus leading to a significant negative monsoon–TC correlation.

From the perspective of the influences from SST anomalies across tropical ocean basins, the influences from tropical SST anomalies could also contribute to the different monsoon–TC relationships over the WNP and the ENP. The tropical Pacific and Atlantic SST anomalies could modulate the monsoon intensity and the TCGF together and trigger the monsoon–TC synchronization and co-variability, thus leading to significant monsoon–TC relationships over both the WNP and the ENP. The WNPSM and the TCGF over the WNP could be modulated by the similar tropical Pacific–Atlantic SST anomalies jointly, thus leading to a significant positive correlation between the WNPSM and the WNP TCGF. In contrast, the signs of tropical Pacific–Atlantic SST anomalies influencing the NASM were almost opposite to those affecting the TCGF over the ENP, thus resulting in a significant negative correlation between the NASM and the ENP TCGF.

In addition, the genesis locations of TCs tended to show an obvious spatial shift over both the WNP and the ENP, along with the variations in the monsoon intensity. In the strong WNPSM year, the increase in the TCGF mainly occurred in the southeast quadrant of the WNP; in contrast, in the strong NASM, the decrease in TCGF over the ENP was mainly due to the change of TCGF in the southwest quadrant of the ENP. Therefore, in the strong WNPSM year, more TCs could be formed over the WNP and the mean TC genesis location tended to shift to the southeast. In contrast, in the weak NASM, more TCs could be generated over the ENP and the mean TC genesis location tended to move to the southwest.

The results show that the regulation of tropical Pacific–Atlantic SST anomalies is an important driving force for the formation of a significant monsoon–TC relationship over the WNP and ENP. It is not only the SST anomalies in the tropical Pacific that plays a role, but also the SST anomalies in the tropical Atlantic can also have an important impact on the monsoon and TC in the Pacific through the cross-basin processes. In addition, the cross-basin influences from the tropical Atlantic on the Pacific have tended to be enhanced substantially since the 1990s [48,49,50]. Thus, the cross-basin influences from the tropical Atlantic may need to be given more attention when assessing the variations of the Pacific monsoons and TC activities.

It is also interesting to note that the spatial distribution patterns of SST anomalies in the tropical Pacific and Atlantic, which influence the WNPSM and the NASM, were very similar (Figure 12, Figure 13 and Figure 14). These similar SST anomalies might support possible associations between monsoons and TCs over the WNP and the ENP. We calculated the correlation coefficient between the time series of the WNPSM index and the NASM index (r = −0.54) and between the time series of the TCGF over the V sub-region and the TCGF over the VII sub-region (r = 0.59), implying a significant negative (positive) correlation between the WNPSM and the NASM (between the WNP TCGF and the ENP TCGF). The possible relationship between the WNPSM and the NASM was also supported in the previous study [51]. During strong WNPSM years, a Rossby wave train that emanated from the WNP could cross the North Pacific and extend to North America [4,52,53,54]. Meanwhile, the heating source associated with the WNPSM could disturb the mid-latitude jet stream and further excite circulation anomalies over North America [4]. In addition, Lee et al. [51] suggested that the more frequent occurrences of the central Pacific (CP) type of El Niño-Southern Oscillation (ENSO) may enhance the relationship between the WNPSM and the NASM during recent decades. These factors mentioned above may promote the linkage between the WNPSM and the NASM. Also, previous studies also suggested possible associations between TC activity variations in different ocean basins [31,55,56]. According to the results in this study, significant monsoon–TC relationships were confirmed for both the ENP and the WNP. Therefore, the relationship between monsoons and TCs in different regions (such as the WNP and the ENP) may be recognized and understood as a whole in a unified climate system.

Our findings herein highlight the differences in the monsoon–TC relationship between the WNP and the ENP, which may provide useful information for the prediction of monsoon intensity and TC formation number over these two regions. Since there is a close monsoon–TC relationship in both the WNP and the ENP, the intensity of monsoon and the TCGF may be considered as a closely related whole system in climate prediction.

Recently, Murakam and Wang [34] reported that the projected future TCGF over the WNP (ENP) may be decreased (increased) under global warming, mainly due to the influences of atmospheric vertical motion. In addition, the NASM and the WNPSM may also show different responses to global warming. He et al. [18] revealed that the Asian monsoon precipitation (including the WNPSM) may tend to be increased under the global warming scenario, while the precipitation related to the NASM may tend to be decreased. However, the response of the monsoon–TC correlation to global warming is still unknown. Possible future changes in the monsoon–TC relationship under different warming scenarios are worthy of future investigations.