Abstract

This study investigates changes in the destructiveness of landfalling tropical cyclones (TCs) over China during 1975–2014. Using four different TC datasets, it is found that TCs making landfall over east China (TCEC) have tended to be more destructive in recent decades, with a significant increase in the power dissipation index (PDI) after landfall. Both time series analysis and diagnostic analysis reveal that such an increase in the PDI of TCEC is associated with concomitant enhancement in landfall frequency as well as landfall intensity over east China. In contrast, changes in the PDI of TCs making landfall over south China (TCSC) are less apparent. Examination of different TC-related parameters shows no obvious changes in terms of landfall frequency, duration, and maximum intensity of TCSC. Diagnostic analysis further suggests that the reduction in TC occurrence over south China offsets considerably the positive effects of the intensity and the nonlinear term.

Further examination of the environmental parameters reveals significant changes in the large-scale steering flow in recent decades, which is characterized by a prominent cyclonic circulation centered over southeast China. The southeasterly flows on the eastern flank of the cyclonic circulation tend to favor subsequent landfall of TCs over east China, resulting in an increase in landfall frequency, which contributes in part to the enhanced PDI of TCs over this region. Meanwhile, the slowing down of the mean translation speed of TCEC and the weakening of vertical wind shear coupled with warmer SSTs in the WNP tend to favor the intensification of TCEC, leading to an increase in intensity and hence the PDI of TCs over east China.

1. Introduction

China, one of the most densely populated countries in the world, is adversely affected by tropical cyclones (TCs). About seven TCs make landfall in China every year, inflicting huge losses of life and property (Zhang et al. 2009; Lu and Zhao 2013). During the period 1983–2006, an average of 472 people were killed each year by the landfalling TCs, with the greatest casualties occurring in coastal cities in Zhejiang, Fujian, and Guangdong Provinces (Zhang et al. 2009). Super Typhoon Fred, landfalling over Zhejiang in August 1994; Super Typhoon Herb, striking Fujian in August 1996; and Tropical Storm Bilis, invading Fujian in July 2006, are some of the deadliest TCs that have affected China in recent decades (Zhang et al. 2009). More recently, Super Typhoon Rammasun also wreaked havoc in Hainan Island, western Guangdong, Guangxi, and Yunnan in July 2014, resulting in at least 30 deaths and over 26.5 billion renminbi (RMB) direct economic loss (HKO 2015). Throughout the world, TC-induced socioeconomic losses have also shown a remarkable increase over the last few decades as a result of population growth, urbanization, and climate change (Pielke et al. 2008; Zhang et al. 2009; Xiao and Xiao 2010). Understanding changes in the characteristics of landfalling TCs thus becomes essential for climate change adaptation and disaster mitigation.

The relationship between climate change and TC activity has long been a topic of active research and debate. Based on the power dissipation index (PDI), several studies have reported a significant increase in the destructive power of TCs in the western North Pacific (WNP) and North Atlantic since the mid-1970s (Emanuel 2005; Webster et al. 2005), which they argued were induced by anthropogenic warming. By using the anthropogenic climate change index (ACCI), Holland and Bruyère (2014) investigated the potential global warming contribution to global tropical cyclone activity from 1975 to 2010 and, after accounting for analysis and observing system changes, found substantial relationships between ACCI and the observed increase in the proportion of very intense TCs (Saffir–Simpson category 4 and 5) in all ocean basins. On the other hand, other research groups have indicated that there are still large uncertainties in detecting the human influence on TC activity in the WNP basin due to the considerable interdecadal natural variability and the issues of homogeneity and consistency of TC records kept by different warning centers (Landsea 2005; Chan 2006, 2008). For future projections of TC activities in WNP, climate models mostly suggest decreases in TC frequency but increases in TC intensity and related precipitation rates in the twenty-first century (e.g., Ying et al. 2012; Wu et al. 2014; Tsuboki et al. 2015; Knutson et al. 2015; Walsh et al. 2016). Most of the previous studies have focused primarily on basinwide TCs; of these, relatively few have tried to investigate changes in the properties of TCs after landfall. An increase in TC activity after landfall is particularly hazardous since most TC-related casualties occur during this time. A recent study by Chan and Xu (2009) analyzed the annual frequency of landfalling TCs over East Asia during 1945–2004 but found no obvious trends. Park et al. (2011), on the other hand, showed that the PDI, as well as the rainfall of landfalling TCs, has increased significantly in recent decades over Japan and the Korean peninsula. Park et al. (2014) further noted a shift in the location of maximum intensity of TCs on the East Asian coastline during 1977–2011 and the potentially growing threat of TCs over East Asia. Tu et al. (2009) also reported that an abrupt shift in the TC activities in the vicinity of Taiwan occurred in 2000, mainly due to a northward shift of the typhoon track over the WNP–East Asian region. Owing to the vast socioeconomic influences of landfalling TCs and their uncertain variation, a comprehensive study is essential to investigate changes in the destructiveness of landfalling TCs over China, a topic that has not been examined in detail in previous studies.

The rest of this paper is organized as follows: section 2 introduces the data and methodology used in this study, and the climatology and distributions of landfalling TCs over China are presented in section 3. Section 4 examines the changes in destructiveness of these landfalling TCs, and section 5 investigates the possible factors contributing to such changes. Finally, section 6 discusses and summarizes the results.

2. Data and methodology

a. Data

The TC dataset acquired from the Joint Typhoon Warning Center (JTWC; https://metoc.ndbc.noaa.gov/web/guest/jtwc/best_tracks/western-pacific) at 6-h intervals is primarily used for investigation in this study unless otherwise stated. Other best-track datasets from the Hong Kong Observatory (HKO), the Regional Specialized Meteorological Center of the Japan Meteorological Agency (JMA), and the China Meteorological Administration–Shanghai Typhoon Institute (CMA) were also used to validate our results. Since the maximum sustained wind speed is recorded differently for different TC datasets (the JTWC and CMA datasets are based on 1-min and 2-min sustained wind speed, while both the HKO and JMA use 10-min sustained wind speed), an adjustment factor of 1.14 (1.01) was applied to the HKO and JMA (CMA) datasets for 10- to 1-min (2- to 1-min) conversion (Knapp et al. 2010; Park et al. 2011; Barcikowska et al. 2012). It should be noted that using the 10-min sustained wind speed is the World Meteorological Organization (WMO) recommended practice for depicting tropical cyclone intensity. The use of 1-min mean wind as the basis in computing the PDI in this study is only for the sake of aligning with the original definition of PDI by Emanuel (2005). It does not have any implications with regard to the wind averaging period practice as recommended by the WMO. We do not expect that the use of the 1-min wind versus 10-min mean wind in this study will have significant impact on the observed trend and the salient findings, as the conversions between 1-min and 10-min mean winds involve a linear factor (Harper et al. 2009). To ensure data reliability, the analysis period of the present study is restricted primarily to 1975–2014, which roughly corresponds to the period when routine satellite data are available (Park et al. 2011). Monthly atmospheric data for the same period were archived from the National Centers for Environmental Prediction (NCEP)–National Center for Atmospheric Research (NCAR) reanalysis (Kalnay et al. 1996), while the monthly National Oceanic and Atmospheric Administration 2° × 2° Extended Reconstructed Sea Surface Temperature (ERSST) version 4 data were obtained from the website (http://www.esrl.noaa.gov/psd/) of NOAA. A landfalling TC in this study refers to any TC that crosses the coastline of China at least once during its lifetime. Although we consider primarily all the TCs making landfall over China in the present study, it is worth noting that the results below will not be affected after excluding those weak tropical depressions with maximum sustained wind speed less than 34 kt.

b. Statistical analysis of the potential destructiveness of TCs

One common measure of the potential destructiveness of TCs is the PDI, which is defined as the sum of the cubes of the maximum sustained wind speed over the entire lifetime of a TC (Emanuel 2005). Since the main focus of this study is the changes in TC destructiveness after landfall, the PDI is derived specifically based on the TCs’ maximum sustained wind speed after landfall. The annual PDI is then calculated by summing the individual PDIs over a particular year. Calculated this way, the annual PDI thus takes into account the frequency, intensity, and duration of landfalling TCs and can be used to represent the activity and destructive potential of these TCs.

To further confirm the changes in the annual PDI and to quantitatively assess the relative contributions of different incorporated parameters (TC frequency, TC intensity, and the nonlinearity of the previous two factors) to the overall PDI changes, an alternative diagnostic analysis is also carried out. In this method, the PDI is recalculated for each 5° × 5° grid cell such that its value can be expressed in terms of the occurrence frequency (F) and the maximum sustained wind speed (υ) of the TCs in that particular grid. The climatological PDI (denoted by overbars) for each 5° × 5° grid cell A can thus be written as

 
formula

where F is the TCs’ occurrence frequency, and υ is the TCs’ maximum sustained wind speed in grid cell A. The PDI anomaly (denoted by prime symbols), with respect to its climatology, can then be evaluated as

 
formula

Equation (2) consists of three terms, which illustrate the different contributions of these factors to the overall changes in PDI. The first term reveals the contribution from anomalous TC occurrence frequency to the overall PDI changes under the condition that the maximum sustained wind speed is unchanged. The second term represents the contribution from anomalous maximum sustained wind speed, while the third term is the nonlinear term associated with changes in both the frequency and maximum sustained wind. Through such a decomposition, changes in PDI can be quantitatively decomposed and assessed in terms of the three factors, namely the frequency effect (first term), the intensity effect (second term), and the nonlinear effect (third term).

3. Distribution of landfalling TCs over China

In this section, we will first take a look at the climatic characteristics of landfalling TCs in China. Figure 1 shows the distribution of landfalling TCs in China during 1975–2014. Following previous studies (Kim et al. 2008; Li and Zhou 2013), the large territory of China is subdivided into south China (SC) and east China (EC) by an artificial boundary line along 25°N. A total of 211 TCs made landfall in China during 1975–2014. Of these, 150 TCs (71%) made landfall over SC, while the remaining 61 TCs (29%) struck the EC coast. Substantial differences can also be found in the prevailing tracks of these two groups of TCs. As shown in Figs. 1b and 1c, TCs making landfall over SC (TCSC) take mainly a northwestward track, whereas TCs landfalling over EC (TCEC) are associated primarily with a recurving track. Figure 2 further shows the monthly variation in landfalling TCs in China. On average, the landfalling frequency of TCs is the highest during June–October, accounting for 97% of the annual total. Therefore, in the following, the PDI associated with TCEC and TCSC during the peak season (June–October) will be investigated to infer any possible changes in recent decades.

Fig. 1.

(a) The locations (indicated by the triangles) and the mean frequency (indicated by the contours) of TCs making landfall in China during 1975–2014, and the associated tracks of (b) TCEC and (c) TCSC. The bold curves in (b) and (c) denote the mean regression trajectories of TCEC and TCSC.

Fig. 1.

(a) The locations (indicated by the triangles) and the mean frequency (indicated by the contours) of TCs making landfall in China during 1975–2014, and the associated tracks of (b) TCEC and (c) TCSC. The bold curves in (b) and (c) denote the mean regression trajectories of TCEC and TCSC.

Fig. 2.

Monthly-averaged frequency of (a) TCEC, (b) TCSC, and (c) their sum during 1975–2014.

Fig. 2.

Monthly-averaged frequency of (a) TCEC, (b) TCSC, and (c) their sum during 1975–2014.

4. Changes in destructiveness of landfalling TCs in China

a. East China

Figure 3a first shows the variation in the PDI associated with TCEC during 1975–2014. Apart from clear interannual variation, a marked positive trend can be observed in the PDI time series, with an increase of 1.91 × 105 kt3 decade−1, which is significant at 95% confidence based on the Student’s t test. In other words, in recent decades, TCs have tended to be more active and destructive after making landfall over EC. A closer look at the TC-related parameters suggests that both the landfall frequency and the mean maximum TC intensity show noticeable increases during 1975–2014, whereas the mean duration after landfall reveals no obvious change (Figs. 3b–d and Table 1). This suggests that the increase in the PDI of TCEC after landfall might be closely related to the concomitant enhancement in landfall frequency as well as the mean maximum intensity in recent decades.

Fig. 3.

Variations of (a) PDI after landfall (105 kt3), (b) landfall frequency (number), (c) mean duration after landfall (h), and (d) mean maximum intensity after landfall (kt) of TCEC during 1975–2014. The corresponding 40-yr linear trends are indicated by the dashed lines. (e) Associated tracks and landfall locations.

Fig. 3.

Variations of (a) PDI after landfall (105 kt3), (b) landfall frequency (number), (c) mean duration after landfall (h), and (d) mean maximum intensity after landfall (kt) of TCEC during 1975–2014. The corresponding 40-yr linear trends are indicated by the dashed lines. (e) Associated tracks and landfall locations.

Table 1.

The 40-yr linear trends of the PDI after landfall, landfall frequency, duration after landfall, and mean maximum intensity after landfall of TCEC during 1975–2014. Trends that are statistically significant at 95% and 90% confidence are denoted by two asterisks (**) and one asterisk (*), respectively.

The 40-yr linear trends of the PDI after landfall, landfall frequency, duration after landfall, and mean maximum intensity after landfall of TCEC during 1975–2014. Trends that are statistically significant at 95% and 90% confidence are denoted by two asterisks (**) and one asterisk (*), respectively.
The 40-yr linear trends of the PDI after landfall, landfall frequency, duration after landfall, and mean maximum intensity after landfall of TCEC during 1975–2014. Trends that are statistically significant at 95% and 90% confidence are denoted by two asterisks (**) and one asterisk (*), respectively.

To further confirm the changes in various TC-related parameters, the results here have also been tested and verified using different TC datasets, as depicted in Fig. 4 and Table 1. Similar to the JTWC, the HKO, JMA, and CMA datasets depict a significant increasing trend in the PDI of TCEC, with a value of 2.66 × 105 kt3 decade−1, 1.97 × 105 kt3 decade−1, and 1.44 × 105 kt3 decade−1, respectively. The consistency among different TC datasets suggests that the observed increase in the PDI of TCEC is robust and detectable. As for the landfall frequency, a prominent increase can be observed in both the HKO and JMA datasets, which is consistent with the results of JTWC. Meanwhile, all four agencies consistently depict a marked increase in mean maximum intensity after landfall, while revealing insignificant trends in the mean duration of TCEC. Comparisons between different TC datasets similarly reveal a positive trend in the potential destructiveness of TCEC, which aligns with an associated increase in landfall frequency and landfall intensity.

Fig. 4.

Variations and linear trends of (a) PDI after landfall (105 kt3), (b) landfall frequency (number), (c) mean duration after landfall (h), and (d) mean maximum intensity after landfall (kt) of TCEC based on the datasets of JTWC, HKO, and JMA during 1975–2014.

Fig. 4.

Variations and linear trends of (a) PDI after landfall (105 kt3), (b) landfall frequency (number), (c) mean duration after landfall (h), and (d) mean maximum intensity after landfall (kt) of TCEC based on the datasets of JTWC, HKO, and JMA during 1975–2014.

To quantitatively assess the factors responsible for the overall increase in the PDI over EC, an alternative diagnostic analysis is also carried out by decomposing the PDI in each 5° × 5° grid into frequency, intensity, and nonlinear terms based on Eq. (2). Figure 5a shows the linear trend of the PDI calculated in each 5° × 5° grid for TCEC during 1975–2014. A pronounced positive trend in the PDI can be found in areas extending from the ocean to the coastal region of EC, suggesting an increase in destructiveness of TCs over EC. Such an increase in the PDI of TCEC can be attributed primarily to the positive frequency effect over EC, accounting for 43% of the local positive trend in the PDI (Figs. 5b,e). The positive intensity effect ranks second (41%) and also contributes considerably to the overall positive trend in the PDI over EC (Figs. 5c,e). The nonlinear term, on the other hand, is relatively small and plays only a marginal role compared with the other two terms (Figs. 5d,e). The results here further indicate that TCEC has tended to be more destructive in recent decades, which is a consequence of increases in frequency as well as intensity over EC.

Fig. 5.

Linear trends of (a) PDI (105 kt3 decade−1) for TCEC and the contribution of the (b) frequency effect, (c) intensity effect, and (d) nonlinear effect to the PDI trend during 1975–2014. Regions with trends that are significant at 90% confidence are shaded by dots. (e) The associated percentage contributions of each of the term to the overall trend in PDI over EC (as indicated by the rectangles).

Fig. 5.

Linear trends of (a) PDI (105 kt3 decade−1) for TCEC and the contribution of the (b) frequency effect, (c) intensity effect, and (d) nonlinear effect to the PDI trend during 1975–2014. Regions with trends that are significant at 90% confidence are shaded by dots. (e) The associated percentage contributions of each of the term to the overall trend in PDI over EC (as indicated by the rectangles).

b. South China

Compared with EC, changes in the PDI associated with TCSC appear to be less pronounced. The time series of the PDI reveals an insignificant trend of 1.65 × 105 kt3 decade−1 during 1975–2014 (Fig. 6 and Table 2). Examination of different TC-related parameters also shows no obvious change in terms of landfall frequency, duration, or maximum intensity of TCSC (Fig. 6 and Table 2). It should be noted that such results are still valid even if different TC datasets are used (Fig. 7 and Table 2). Consistent with the results using JTWC datasets, changes in the PDI and other TC-related parameters are not evident for the HKO, JMA, or CMA datasets. By decomposing the PDI in each 5° × 5° grid into frequency, intensity, and nonlinear terms, it is similarly noted that changes in the PDI over SC are not significant. The reduction in TC occurrence over SC contributes negatively to the overall PDI increase in SC and offsets considerably the positive effects of the intensity and the nonlinear term (Fig. 8). The results here suggest that changes in the destructiveness of TCSC are less apparent than those of TCEC.

Fig. 6.

Variations of (a) PDI after landfall (105 kt3), (b) landfall frequency (number), (c) mean duration after landfall (h), and (d) mean maximum intensity after landfall (kt) of TCSC during 1975–2014. The corresponding 40-yr linear trends are indicated by the dashed lines. (e) Associated tracks and landfall locations.

Fig. 6.

Variations of (a) PDI after landfall (105 kt3), (b) landfall frequency (number), (c) mean duration after landfall (h), and (d) mean maximum intensity after landfall (kt) of TCSC during 1975–2014. The corresponding 40-yr linear trends are indicated by the dashed lines. (e) Associated tracks and landfall locations.

Table 2.

As in Table 1, except for TCSC.

As in Table 1, except for TCSC.
As in Table 1, except for TCSC.
Fig. 7.

Variations and linear trends of (a) PDI after landfall (105 kt3), (b) landfall frequency (number), (c) mean duration after landfall (h), and (d) mean maximum intensity after landfall (kt) of TCSC based on the datasets of JTWC, HKO, and JMA during 1975–2014.

Fig. 7.

Variations and linear trends of (a) PDI after landfall (105 kt3), (b) landfall frequency (number), (c) mean duration after landfall (h), and (d) mean maximum intensity after landfall (kt) of TCSC based on the datasets of JTWC, HKO, and JMA during 1975–2014.

Fig. 8.

Linear trends of (a) PDI (105 kt3 decade−1) for TCSC and the contribution of the (b) frequency effect, (c) intensity effect, and (d) nonlinear effect to the PDI trend during 1975–2014. Regions with trends that are significant at 90% confidence are shaded by dots. (e) The associated percentage contributions of each of the term to the overall trend in PDI over SC (as indicated by the rectangles).

Fig. 8.

Linear trends of (a) PDI (105 kt3 decade−1) for TCSC and the contribution of the (b) frequency effect, (c) intensity effect, and (d) nonlinear effect to the PDI trend during 1975–2014. Regions with trends that are significant at 90% confidence are shaded by dots. (e) The associated percentage contributions of each of the term to the overall trend in PDI over SC (as indicated by the rectangles).

5. Possible factors contributing to the enhanced PDI of TCs over east China

The previous section has identified an evident increasing trend in the PDI of TCEC, which is associated with enhanced landfall frequency as well as landfall intensity in recent decades. In this section, we identify and discuss several factors that may be closely related to the PDI changes of TCEC in recent decades.

a. Changes in environmental steering flow

As suggested by previous studies (Gray 1979; Chan 2005), variation in TC tracks and landfall positions is governed predominantly by changes in the environmental steering flows. Figure 9a shows the 40-yr linear trend of the steering flow (i.e., wind averaged over 850 to 300 hPa) during 1975–2014. The linear trend is characterized by a prominent cyclonic circulation centered over southeast China. The southeasterly flows at the eastern flank of the cyclonic circulation tend to favor subsequent landfall of TCs over EC, while the westerly flows at the southern flank tend to suppress the number of TCs landfalling over SC. Such a change in the steering flow is one possible factor leading to the recent increase in landfall frequency over EC, contributing in part to the enhanced PDI of TCs over this region. The prevailing track shift due to cyclonic circulation anomaly centered over southeast China was also reported in the studies by Wu et al. (2005), Lee et al. (2012), and Zhao and Wu (2014). Based on numerical model results of global warming experiments, Wu and Wang (2004) suggested that the warming trend of sea surface temperature would cause the shift of prevailing TC tracks. Moreover, by using a singular value decomposition (SVD) analysis and IPCC AR4 historical forcing runs, Wang et al. (2011) suggested that the observed shift of TC tracks was linked to the leading SVD mode of global sea surface temperature warming and the associated changes in large-scale steering flows. However, since there are considerable interannual and interdecadal variations in the TC tracks in the WNP (Liu and Chan 2008; Choi et al. 2010), further observations and research will still be required to understand the influence and contribution of natural variability and anthropogenic warming on the TC track changes in the WNP.

Fig. 9.

Linear trends of (a) steering flow (m s−1 yr−1), (b) vertical wind shear (m s−1 yr−1), and (c) SST (°C decade−1) during June–October 1975–2014. Regions with trends that are significant at 90% confidence are shaded by dots.

Fig. 9.

Linear trends of (a) steering flow (m s−1 yr−1), (b) vertical wind shear (m s−1 yr−1), and (c) SST (°C decade−1) during June–October 1975–2014. Regions with trends that are significant at 90% confidence are shaded by dots.

b. Weakened vertical wind shear over EC and basinwide SST warming in the WNP

On the other hand, the increasing intensity of TCEC might be related to a remarkable reduction in vertical wind shear over EC (Fig. 9b). The weakened wind shear helps maintain TC structure, sustains TC intensity, and inhibits extratropical transition of TCs (Gray 1968; Baik and Paek 2001; Chan 2008). This was also identified by Park et al. (2011), who revealed that the weakened wind shear over this region as a result of the weakening of the East Asian jet has also contributed to stronger landfalling TCs over the Korean peninsula and Japan in recent decades. Apart from the favorable dynamic factor of weakened vertical wind shear, it is also noted that the SST in the WNP has revealed a significant warming trend in the recent decades (Fig. 9c) accompanied by weak cooling in the central Pacific. Such warming in SST is particularly evident over the northern part of the WNP and is consistent with that found in the previous studies (Park et al. 2013, 2014). The basinwide SST warming in the WNP also provides favorable thermodynamic background for the intensification of TCs in the WNP. As pointed out previously by Park et al. (2014), such increasing zonal SST gradient over the tropical Pacific is also closely linked to the strengthened Walker circulation, which results in strengthened cyclonic flows and weakened vertical wind shear along the East Asian coastline during the recent decades.

c. Reduction in mean translation speed

Apart from the weakened wind shear and the warmer SST, it is also worth noting that there is a significant decreasing trend in the mean translation speed of TCEC over the open ocean before landfall, while changes in the mean translation speed of TCSC are less apparent (Fig. 10). With favorable background of increasing SST (Fig. 9c) and deepening of the 26°C isotherm across the WNP in recent decades (Park et al. 2013), the slowing down of the mean translation speed allows TCEC to stay longer over the ocean with warmer SST and deepened warm mixed layer and favors the intensification of TCEC, which might also help explain the increase in TC landfall intensity over EC. Yet the exact cause of this slowing down of the mean translation speed of TCEC is unknown at present and deserves further study.

Fig. 10.

Linear trends of mean translation speed (km h−1 yr−1) of (a) TCEC and (b) TCSC during 1975–2014. Regions with trends that are significant at 90% confidence are shaded by dots.

Fig. 10.

Linear trends of mean translation speed (km h−1 yr−1) of (a) TCEC and (b) TCSC during 1975–2014. Regions with trends that are significant at 90% confidence are shaded by dots.

6. Discussion and summary

This study investigates changes in the destructiveness of landfalling TCs over China during 1975–2014. Using four different TC datasets, it is found that TCEC has tended to be more destructive in recent decades, with a significant increase in PDI after making landfall over EC. Both time series analysis and diagnostic analysis suggest that this increase in the PDI of TCEC can be attributed to the concomitant enhancement in landfall frequency as well as landfall intensity over EC. In contrast, changes in the PDI of TCSC are less apparent. Examination of different TC-related parameters shows no obvious changes in terms of landfall frequency, duration, or maximum intensity of TCSC. Diagnostic analysis further suggests that the reduction in TC occurrence over SC offsets considerably the positive effects of the intensity and the nonlinear term.

Examination of large-scale environmental parameters reveals significant changes in the environmental steering flow in recent decades, which is characterized by a prominent cyclonic circulation centered over southeast China. The southeasterly flows on the eastern flank of the cyclonic circulation tend to favor subsequent landfall of TCs over EC, resulting in an increase in landfall frequency, which contributes in part to the enhanced PDI of TCs over this region. Meanwhile, the slowing down of the mean translation speed of TCEC and the weakening of vertical wind shear coupled with warmer SST in the WNP tend to favor the intensification of TCEC, leading to an increase in intensity and hence the PDI of TCs over EC. Wu et al. (2014) found the prevailing TC tracks have shifted westward significantly in recent decades, which leads to growing TC influence over east China, while Park et al. (2013) and Park et al. (2014) have similarly identified a strengthening in TC intensity in southern Japan and northeast Asia, which they attributed to the changes in intensification rate and genesis frequency over these regions. Through a new approach of diagnostic analysis of the PDI, the results of the present study further substantiate and extend the results of these previous studies by quantitatively assess the relative contributions of different factors (TC frequency, TC intensity, and the nonlinearity of these two factors) to the overall PDI changes of landfalling TCs over EC and SC, which helps further enhance our understanding on the changes in TC destructiveness by spotting out the key factors contributing to the overall PDI changes in recent decades.

Overall, this study has highlighted a potential increase in the destructiveness of TCs making landfall over EC, which is coincident with the corresponding changes in large-scale environmental factors in recent decades. It should be noted that although PDI is a widely adopted parameter for assessing the potential destructiveness of a storm, it may not fully reflect all hazardous impacts of TCs, including torrential rain and storm surge induced by landfalling TCs. Given the significant socioeconomic impacts of landfalling TCs, follow-up studies will still be necessary to keep in view the changes in these TCs and to further explore the underlying factors and mechanisms by means of numerical experiments, in particular on the connection between global warming and the shift in the prevailing track of TCs in the WNP. Moreover, against the background of global warming and sea level rise, the risk of extreme weather and storm surge induced by landfalling TCs to coastal cities should be further investigated to assist in developing relevant disaster mitigation and adaptation measures.

Acknowledgments

This research is supported by National Natural Science Foundation of China (NSFC 41675062).

REFERENCES

REFERENCES
Baik
,
J.-J.
, and
J.-S.
Paek
,
2001
:
Relationship between vertical wind shear and typhoon intensity change, and development of three‐predictor intensity prediction model
.
J. Meteor. Soc. Japan
,
79
,
695
700
, doi:.
Barcikowska
,
M.
,
F.
Feser
, and
H.
von Storch
,
2012
:
Usability of best track data in climate statistics in the western North Pacific
.
Mon. Wea. Rev.
,
140
,
2818
2830
, doi:.
Chan
,
J. C. L.
,
2005
:
The physics of tropical cyclone motion
.
Annu. Rev. Fluid Mech.
,
37
,
99
128
, doi:.
Chan
,
J. C. L.
,
2006
:
Comment on “Changes in tropical cyclone number, duration, and intensity in a warming environment.”
Science
,
311
,
1713
, doi:.
Chan
,
J. C. L.
,
2008
:
Decadal variations of intense typhoon occurrence in the western North Pacific
.
Proc. Roy. Soc.
,
464
,
249
272
, .
Chan
,
J. C. L.
, and
M.
Xu
,
2009
:
Inter-annual and inter-decadal variations of landfalling tropical cyclones in East Asia. Part I: Time series analysis
.
Int. J. Climatol.
,
29
,
1285
1293
, doi:.
Choi
,
K.-S.
,
B. J.
Kim
,
D. W.
Kim
, and
H.-R.
Byun
,
2010
:
Interdecadal variation of tropical cyclone making landfall over the Korean Peninsula
.
Int. J. Climatol.
,
30
,
1472
1483
, doi:.
Emanuel
,
K. A.
,
2005
:
Increasing destructiveness of tropical cyclones over the past 30 years
.
Nature
,
436
,
686
688
, doi:.
Gray
,
W. M.
,
1968
:
Global view of the origin of tropical disturbances and storms
.
Mon. Wea. Rev.
,
96
,
669
700
, doi:.
Gray
,
W. M.
,
1979
: Hurricanes: Their formation, structure and likely role in the tropical circulation. Meteorology over the Tropical Oceans, D. B. Shaw, Ed., Royal Meteorological Society, 155–218.
Harper
,
B. A.
,
J. D.
Kepert
, and
J. D.
Ginger
,
2009
: Guidelines for converting between various wind averaging periods in tropical cyclone conditions. World Meteorological Organization Rep. TCM-VI/Doc. 2.3, 62 pp.
HKO
,
2015
: Tropical cyclones in 2014. Hong Kong Observatory Rep., 120 pp.
Holland
,
G.
, and
C. L.
Bruyère
,
2014
:
Recent intense hurricane response to global climate change
.
Climate Dyn.
,
42
,
617
627
, doi:.
Kalnay
,
E.
, and Coauthors
,
1996
:
The NCEP/NCAR 40-Year Reanalysis Project
.
Bull. Amer. Meteor. Soc.
,
77
,
437
471
, doi:.
Kim
,
J.
,
C.
Ho
,
H.
Kim
,
C.
Sui
, and
S. K.
Park
,
2008
:
Systematic variation of summertime tropical cyclone activity in the western North Pacific in relation to the Madden–Julian oscillation
.
J. Climate
,
21
,
1171
1191
, doi:.
Knapp
,
K. R.
,
M. C.
Kruk
,
D. H.
Levinson
,
H. J.
Diamond
, and
C. J.
Neumann
,
2010
:
The International Best Track Archive for Climate Stewardship (IBTrACS): Unifying tropical cyclone data
.
Bull. Amer. Meteor. Soc.
,
91
,
363
376
, doi:.
Knutson
,
T. R.
,
J. J.
Sirutis
,
M.
Zhao
,
R. E.
Tuleya
,
M.
Bender
,
G. A.
Vecchi
,
G.
Villarini
, and
D.
Chavas
,
2015
:
Global projections of intense tropical cyclone activity for the late twenty-first century from dynamical downscaling of CMIP5/RCP4.5 scenarios
.
J. Climate
,
28
,
7203
7224
, doi:.
Landsea
,
C. W.
,
2005
:
Hurricanes and global warming
.
Nature
,
438
,
E11
E12
, doi:.
Lee
,
T.-C.
,
Y.-Y.
Leung
,
M.-H.
Kok
, and
H.-S.
Chan
,
2012
:
The long term variations of tropical cyclone activity in the South China Sea and the vicinity of Hong Kong
.
Trop. Cyclone Res. Rev.
,
1
,
277
292
, doi:.
Li
,
R. C. Y.
, and
W.
Zhou
,
2013
:
Modulation of western North Pacific tropical cyclone activities by the ISO. Part II: Tracks and landfalls
.
J. Climate
,
26
,
2919
2930
, doi:.
Liu
,
K. S.
, and
J. C. L.
Chan
,
2008
:
Interdecadal variability of western North Pacific tropical cyclone tracks
.
J. Climate
,
21
,
4464
4476
, doi:.
Lu
,
X.
, and
B.
Zhao
,
2013
:
Analysis of the climatic characteristics of landing tropical cyclones in East China
.
J. Trop. Meteor.
,
19
,
145
153
.
Park
,
D.-S. R.
,
C.-H.
Ho
,
J.-H.
Kim
, and
H.-S.
Kim
,
2011
:
Strong landfall typhoons in Korea and Japan in a recent decade
.
J. Geophys. Res.
,
116
,
D07105
, doi:.
Park
,
D.-S. R.
,
C.-H.
Ho
,
J.-H.
Kim
, and
H.-S.
Kim
,
2013
:
Spatially inhomogeneous trends of tropical cyclone intensity over the western North Pacific for 1977–2010
.
J. Climate
,
26
,
5088
5101
, doi:.
Park
,
D.-S. R.
,
C.-H.
Ho
, and
J.-H.
Kim
,
2014
:
Growing threat of intense tropical cyclones to East Asia over the period 1977–2010
.
Environ. Res. Lett.
,
9
, 014008, doi:.
Pielke
,
R. A.
, Jr.
,
J.
Gratz
,
C. W.
Landsea
,
D.
Collins
,
M. A.
Saunders
, and
R.
Musulin
,
2008
:
Normalized hurricane damage in the United States: 1900–2005
.
Nat. Hazards Rev.
,
9
,
29
42
, doi:.
Tsuboki
,
K.
,
M. K.
Yoshioka
,
T.
Shinoda
,
M.
Kato
,
S.
Kanada
, and
A.
Kitoh
,
2015
:
Future increase of supertyphoon intensity associated with climate change
.
Geophys. Res. Lett.
,
42
,
646
652
, doi:.
Tu
,
J.-Y.
,
C.
Chou
, and
P.-S.
Chu
,
2009
:
The abrupt shift of typhoon activity in the vicinity of Taiwan and its association with western North Pacific–East Asian climate change
.
J. Climate
,
22
,
3617
3628
, doi:.
Walsh
,
K. J. E.
, and Coauthors
,
2016
:
Tropical cyclones and climate change
.
Wiley Interdiscip. Rev.: Climate Change
,
7
,
65
89
, doi:.
Wang
,
R.
,
L.
Wu
, and
C.
Wang
,
2011
:
Typhoon track changes associated with global warming
.
J. Climate
,
24
,
3748
3752
, doi:.
Webster
,
P. J.
,
G. J.
Holland
,
J. A.
Curry
, and
H. R.
Chang
,
2005
:
Changes in tropical cyclone number, duration, and intensity in a warming environment
.
Science
,
309
,
1844
1846
, doi:.
Wu
,
L.
, and
B.
Wang
,
2004
:
Assessing impacts of global warming on tropical cyclone tracks
.
J. Climate
,
17
,
1686
1698
, doi:.
Wu
,
L.
,
B.
Wang
, and
S.
Geng
,
2005
:
Growing typhoon influence on East Asia
.
Geophys. Res. Lett.
,
32
,
L18703
, doi:.
Wu
,
L.
, and Coauthors
,
2014
:
Simulations of the present and late-twenty-first-century western North Pacific tropical cyclone activity using a regional model
.
J. Climate
,
27
,
3405
3424
, doi:.
Xiao
,
F.
, and
Z.
Xiao
,
2010
:
Characteristics of tropical cyclones in China and their impacts analysis
.
Nat. Hazards
,
54
,
827
837
, doi:.
Ying
,
M.
,
T. R.
Knutson
,
H.
Kamahori
, and
T. C.
Lee
,
2012
:
Impacts of climate change on tropical cyclones in the western North Pacific basin. Part II: Late 21st century projections
.
Trop. Cyclone Res. Rev.
,
1
,
231
241
, doi:.
Zhang
,
Q.
,
L.
Wu
, and
Q.
Liu
,
2009
:
Tropical cyclone damages in China 1983–2006
.
Bull. Amer. Meteor. Soc.
,
90
,
489
495
, doi:.
Zhao
,
H.
, and
L.
Wu
,
2014
:
Inter-decadal shift of the prevailing tropical cyclone tracks over the western North Pacific and its mechanism study
.
Meteor. Atmos. Phys.
,
125
,
89
101
, doi:.

Footnotes

© 2017 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).