Statistical analysis of the characteristics of typhoons approaching Japan from 2006 to 2019

Abstract

This study investigated the characteristics of 60 typhoons approaching Japan over the past 14 years (2006–2019) by conducting statistical analysis of their temporal evolution, active hours, intensity, frequency, size, duration, and translation speed. By dividing the time period into the earlier (before 2012) and the most recent (after 2013) years, the analysis indicated that the annual frequency of typhoons is higher in the most recent years than in the earlier years. The typhoons in recent years took relatively less time to reach Japan and remained active for shorter time over the land are of Japan than those in the earlier years. The intensity of the typhoons in the recent years showed stronger winds and considerably lower pressures at the landfall time than that in the earlier years. Typhoons in recent years carry more frequent and intense rainfall compared to those in the earlier years in this study. The analysis inferred that the higher sea surface temperature, weaker vertical wind shears, and a larger amount of moisture around the centers of the recent typhoons were responsible for making them stronger.

Keywords:

• Typhoon
• typhoon intensity
• typhoon hazard
• environmental conditions
• climate change

1. Introduction

Tropical cyclones (TCs), also known as hurricanes, typhoons, or cyclones depending on the region, are among the most devastating natural disasters in the world. These TCs generally form over warm ocean surface near the equator and cause significant socio-economic damages when they make landfall in populated areas (Aon Benfield 2016). TCs have caused significant damages, and losses of lives and properties in many countries across the world. In Bangladesh, frequent TCs caused notable damage and displacement, specifically with Cyclone Sidr in 2007 and Cyclone Amphan in 2020 (Paul 2010; Hassan et al. 2020; Ahsan and Özbek 2022). The Philippines is also highly vulnerable to TCs, for instance, Super Typhoon Haiyan in 2013 caused significant damages and losses of lives and properties (Takayabu et al. 2015; Takagi and Esteban 2016). In the United States, storms such as Hurricane Katrina in 2005 and Hurricane Sandy in 2012 caused significant damages and alerted the need for improved preparedness and response measures (Padgett et al. 2008; Halverson and Rabenhorst 2013; Xian et al. 2015; Gall and Cutter 2019). The Caribbean islands are also highly vulnerable to TCs, for instance, Hurricane Maria in 2017 caused substantial damages, and losses of lives and properties in Puerto Rico and other islands (Cox et al. 2019; Kim and Bui 2019). These cases illustrate the significant impacts that TCs can have on different countries and regions, and highlight the need for further research and preparedness measures to mitigate the risks of future storms.

In Japan, typhoons are responsible for causing significant damage each year, which includes the threat to human life, destruction of infrastructure, and damage to agriculture (Aon Benfield 2016; Takemi et al. 2016a; Guan et al. 2018). In recent decades, more typhoons have made landfall and some have increased in intensity, resulting in prominent typhoon-related losses of lives, crops, and infrastructure (Mei et al. 2015; Hornyak 2020; Yamaguchi and Maeda 2020). As a result, several studies have focused on understanding typhoon characteristics from disaster prevention and mitigation perspectives, including the changes in typhoon frequency, formation, development, structure, and intensity in recent decades or future changes due to climate change (Ishikawa et al. 2013; Mei and Xie 2016; Takemi et al. 2016b, 2016c; Zhan et al. 2017; Gao et al. 2018; Kamahori and Arakawa 2018; Kanada et al. 2019; Nayak and Takemi 2019a, 2019b, 2020a; Hatsuzuka et al. 2020; Nayak and Takemi 2020b). These studies highlight the need for continued research and improved disaster prevention and mitigation measures to mitigate the risks posed by typhoons in Japan.

According to Yamaguchi and Maeda (2020), the number of typhoons approaching the southern coast of Japan has increased from 1980 to 2019 due to favorable environmental conditions for typhoon development. As a result, the number of typhoons forming over the western North Pacific region has also increased, leading to more landfalls in Japan. Mei and Xie (2016) found that typhoons striking east and southeast Asia have intensified by 12%–15% between 1977 and 2014. Several studies have investigated the effects of climate change on individual typhoons, including Typhoon Vera (1959), Mireille (1991), Songda (2004), Talas (2011), Haiyan (2013), Lionrock (2016), Jebi (2018), and Hagibis (2019), to understand their severity if similar typhoons occur in the future with a warming climate (Takayabu et al. 2015; Ito et al. 2016; Takemi et al. 2016b, 2016c; 2019; 2019; Nayak and Takemi 2019a, 2019b; Takemi and Unuma 2020). Hatsuzuka et al. (2020) indicated that typhoon-induced extreme precipitation is likely to increase throughout Japan under a 4 K warmer climate. These studies indicated that the characteristics of typhoons approaching Japan have changed in recent years and are projected to change in their intensity and impacts under future climate changes. However, the detailed characteristics of typhoons such as the temporal evolution, duration, geographical locations over Japan are not well documented in observational or modeling studies. Similarly, the durations and frequencies of typhoon-associated rainfall over Japan have not been thoroughly explored. Understanding these features of approaching typhoons, especially in recent years, is crucial for disaster prevention and mitigation of typhoon-related hazards in Japan.

This study aimed to investigate several characteristics of typhoons, i.e. TCs that occur in the western North Pacific Ocean, specifically those that approached or passed over Japan from 2006. The reason for selecting 2006 as the starting year for the analysis period is due to the availability of high-quality precipitation data for the Japanese islands from radar-raingauge products, which have been available since 2006 (explained more about this data in Section 2). The characteristics examined include: what was the duration of typhoons when they were active over Japan land; how did the strength and location of typhoons change as they approached or passed over Japan; what was the frequency of typhoons during the study period; what was the duration and size of typhoons as they approached or passed over Japan; how fast did typhoons move across the ocean and towards Japan; and finally, are there any changes in the typhoon characteristics from an earlier to a later period? This study also examined the environmental conditions that might influence typhoon development and intensification. These factors include sea surface temperature (SST), which affects the amount of heat energy available for typhoon intensification; moisture, which impacts the formation and structure of typhoons; and vertical wind shear, which disrupts or enhances typhoon formation and growth.

2. Data and methods

We examined typhoons that approached and/or landed Japan between 2006 and 2019 from the Regional Specialized Meteorological Center (RSMC) Tokyo best-track datasets, the China Meteorological Administration (CMA) tropical cyclone best-track datasets (Ying et al. 2014; Lu et al. 2021), and the Joint Typhoon Warning Center (JTWC) Western North Pacific Ocean best-track datasets. We chose to focus on a specific subset of typhoons that could have the greatest impact on Japanese landmasses. We selected a total of 60 typhoons that met specific criteria. These typhoons either made landfall or passed within 100 kilometers of a Japanese landmass and did not weaken before making landfall, with a maximum wind speed of 15 m s⁻¹ or higher. Typhoons that decayed or weakened before making landfall were not included in the study as their impact was assumed to be negligible. Table 1 shows the detailed information of the 60 typhoons.

Table 1. List of typhoons with maximum intensity from RSMC best-track datasets.

Sl. No.	Year	International Number ID	Name	Maximum wind speed (knots)	Minimum central pressure (hPa)
1	2006	0607	Maria	70	975
2		0610	Wukong	50	980
3		0613	Shanshan	110	919
4	2007	0704	Man-Yi	95	930
5		0705	Usagi	90	945
6		0709	Fitow	70	965
7	2008	0813	Sinlaku	100	935
8	2009	0909	Etau	40	992
9		0911	Krovanh	60	975
10		0918	Melor	110	910
11	2010	1004	Dianmu	50	985
12		1007	Kompasu	80	960
13		1009	Malou	50	992
14		1014	Chaba	95	930
15	2011	1102	Songda	105	920
16		1106	Ma-On	95	935
17		1112	Talas	50	970
18		1115	Roke	85	940
19	2012	1204	Guchol	100	930
20		1216	Ganba	110	900
21		1217	Jelawat	110	905
22	2013	1318	Man-Yi	65	960
23		1324	Danas	90	935
24		1326	Wipha	90	930
25	2014	1408	Neoguri	100	930
26		1411	Halong	105	920
27		1418	Phanfone	95	935
28		1419	Vongfong	115	900
29	2015	1511	Nangka	100	925
30		1515	Goni	100	930
31		1518	Etau	50	985
32	2016	1606	Conson	45	985
33		1607	Chanthu	55	980
34		1609	Mindulle	65	975
35		1610	Lionrock	90	940
36		1611	Kompasu	35	994
37		1612	Namtheun	70	955
38		1616	Malakas	95	930
39		1618	Chaba	115	905
40	2017	1703	Nanmadol	55	985
41		1705	Noru	95	935
42		1718	Talim	95	935
43		1721	Lan	100	915
44		1722	Saola	60	975
45	2018	1807	Prapiroon	65	960
46		1812	Jongdari	75	960
47		1813	Shanshan	70	970
48		1815	Leepi	50	994
49		1819	Soulik	85	950
50		1820	Cimaron	85	950
51		1821	Jebi	105	915
52		1824	Trami	105	915
53		1825	Kong-Rey	115	900
54	2019	1903	Sepat	40	994
55		1906	Nari	35	998
56		1908	Francisco	70	970
57		1910	Krosa	75	965
58		1915	Faxai	85	955
59		1917	Tapah	65	970
60		1919	Hagibis	105	915

In order to analyze the paths taken by typhoons, we used data on the latitude and longitude positions of the storm at regular intervals between 1 and 6 h. The exact interval used likely depended on the availability of the data, as it may not be possible to collect data on a storm’s position every hour in all cases. We calculated the center of the storm based on the available latitude and longitude data at 6-h intervals. To analyze the intensity of a typhoon, we used the maximum wind speed (the highest speed at which the wind is blowing within the typhoon) and minimum central pressure (the lowest atmospheric pressure at the center of the typhoon) at 6-h intervals. The latitudinal and longitudinal locations of each typhoon center were also investigated in 6-h intervals to analyze its track. The annual frequencies of typhoons were calculated by counting the number of typhoons per year that met the defined criteria for inclusion in this study. To calculate the time for each typhoon to reach land, we estimated the duration from the initiation point of the typhoon, which is the location where the storm begins to form until it reaches a distance of 200 km from Japan. Once the typhoon center is within 200 km of Japan, its active period over Japan is computed from the time taken by each typhoon’s center to travel between 200 km from land and after passing Japan. This is because typhoons can have different intensities and effects depending on their distance from land and their trajectory over Japan. In this study, the rainfall amounts associated with each typhoon were computed based on the longest radius of 15 m s⁻¹ or higher winds from the typhoon center. This radius is often referred to as R_L15. However, the R_L15 information was taken from the RSMC best-track datasets only for this particular analysis as it was not available in the other two datasets used in the study.

To determine the rainfall amount associated with each typhoon, we utilized observed rainfall product generated by combining radar and raingauge observations, called radar/raingauge analyzed precipitation data (Nagata 2011). This rainfall product provides gridded, accurate precipitation fields throughout the Japanese islands. The horizontal resolution has upgraded in the past; the current data have spatial resolution of 1 km, which is available since 2006. In order to take advantage of this high-resolution precipitation data, the analysis period is set to be after 2006. The time period of the data is 1 h. This high-resolution precipitation data is a valuable source of information about the intensity and spatial distribution of rainfall associated with typhoons. With the aid of high-resolution precipitation data, it is possible to obtain more accurate information about the heaviest rainfall and their frequency. This information can help local authorities and emergency responders to prepare for and respond to the storm more effectively.

In order to assess the severity of recent typhoons, we divided our data into two distinct periods of seven years each, covering the periods of 2006 to 2012 (hereafter referred to as "earlier years") and 2013 to 2019 (hereafter referred to as ‘recent years’). We then analyzed the temporal evolution of various typhoon characteristics, including maximum wind speed, minimum central pressure, translation speed, size, and rainfall amount, for each typhoon in both time periods, as well as the mean features for each period. These two periods were chosen because the years between 2013 and 2018 are among the six warmest on record between 1880 and 2018 (Arguez et al. 2020). Additionally, the period 2014–2016 was characterized by the occurrence of El Niño-Southern Oscillation (ENSO). ENSO can also impact the intensity of storms. For example, during El Niño events, there is a trend of a decrease in the intensity of TCs in the Atlantic basin (Landsea et al. 2010), while the opposite is true for the western Pacific basin (Fudeyasu et al. 2006). Research studies have provided evidence for the impact of ENSO on TCs (Girishkumar et al. 2015; Zhao and Wang 2019: Albert et al. 2022; Song et al. 2022). Hence, it is of interest to examine the behavior of typhoons between 2013 and 2019 and compare it with another 7-year period (2006–2012). To investigate changes that may have occurred in recent years, we analyzed the mean properties of various typhoon variables at 6-h intervals, and compared the results between the two time periods.

The rainfall durations and associated rainfall amounts were explored by categorizing the independent rainfall durations caused by each typhoon into different duration bins. Here, an independent rainfall duration refers to a continuous period of rainfall over Japanese land while a typhoon passed. The rainfall duration samples collected from each typhoon were independent. Hence, we considered the accumulated rainfall in each independent duration and explored their frequencies. The environmental conditions during the time evolution of each typhoon and the mean features in the two periods were examined by analyzing the 3-h surface brightness temperature as SST, 6-h vertical wind shear between 850 and 300 hPa, and 3-h precipitable water from the 55-year Japanese Reanalysis (JRA-55) data (Kobayashi et al. 2015). These features were investigated in 6-h intervals in the R_L15 regions from the typhoon center. The spatial differences in these features were investigated by averaging SST, wind, and moisture field for each period. The mean values at 6-h intervals were compared between two time periods in order to investigate any changes that may have occurred in recent years. The mean values of certain variables (e.g. wind speed, precipitation) were calculated at 6-h intervals for each of the two time periods under consideration. The mean value here was the average of all the values collected over a specific time interval. By comparing the mean values at 6-h intervals, we aimed to detect any significant changes in the variables of interest over time for understanding the characteristics and impacts of the recent typhoons. The Student’s t-test is used to assess the significance of differences between the two time periods. The null hypothesis (H₀) assumes that there is no change, with a significance level of 95%. This means that if the calculated p-value from the t-test is less than 0.05, we can reject the null hypothesis and conclude that there is a significant difference between the two periods.

3. Results

First, we analyzed the tracks of the 60 typhoons, their annual frequencies, and time durations. Next, we evaluated the intensities and locations of these 60 typhoons. We also investigated the temporal evolution of the intensity, translation speed, location, and size. Finally, we discussed the intensity and rainfall duration of typhoons in recent years and the associated environmental conditions.

3.1. Typhoon tracks, frequencies, and time durations

Typhoons approaching Japan between 2006 and 2019 were mostly formed in the region of 5–15 N and initially moved northwestward before recurving and moving northeastward towards the southern coast of Japan (Figure 1). The term ‘recurving of TCs’ refers to the process where a storm initially moving westward/northwestward turns towards the north or northeast and begins moving in that direction, which is common in the western North Pacific basin. During this period, most typhoons made landfall on the Pacific Ocean side of Japan, with more landfall occurring on that side than the Sea of Japan side. All three best-track datasets, namely RSMC, CMA, and JTWC, showed similar tracks until the typhoons reached the Hokkaido region. The JTWC best-track datasets did not typically show tracks over Hokkaido, while the RSMC and CMA best-track datasets showed tracks over Hokkaido, but the CMA best-track datasets showed tracks mostly up to 155E after landfall.

Figure 1. Track of 60 typhoons from (a) RSMS, (b) CMA, and (c) JTWC best-track datasets.

The annual frequencies of the typhoons between 2006 and 2019 indicated that the occurrence of the typhoons substantially increased in recent years, especially after 2015 (Figure 2). Prior to 2015, there were approximately four typhoons, but the number has almost doubled in recent years. The 5-year moving averages computed from the previous 5-points and center ±2-points along with the linear trend of the frequencies showed a steady increase in typhoons approaching Japan in recent years. Previous studies on the global tropical TCs (Kossin et al. 2020) also highlighted significant increases in the frequency of TCs making landfall. Takagi and Esteban (2016) reported that the frequency of TCs making landfall over the Philippines increased 0.02 times per year between 1945 and 2013 and most TCs made landfall in the 10–12 N latitudinal region.

Figure 2. (a) Annual frequency of typhoons; (b) Time taken by typhoons to reach the Japanese landmass from their initiation; and (c) Duration of typhoons while active over Japan. The smooth lines (dotted/solid) in (b-c) represent the annual mean durations.

Most typhoons during 2006–2019 took 5–8 days, on average, to reach Japan and remained active for 1–3 days over Japan (Figure 2b). A few typhoons took longer to reach Japan (>8 days) and remained active over land for a longer duration (>3 days) (Figure 2c). The reason could be associated with the translation speed, which may be slower in long-lived typhoons. All three best-track datasets yielded similar results. However, JTWC showed a slight variation, which may be due to its insufficient information over the Hokkaido region.

3.2. Intensities and locations

We analyzed the intensity of the 60 typhoons in 6-h intervals from their initiation (T = 0) until their decay in terms of the relationship between maximum wind speed and minimum central pressure (Figure 3a). Typhoons approaching Japan between 2006 and 2019 possessed a maximum wind speed of 20–60 m s⁻¹ in the RSMC best-track datasets, while they showed maximum wind speeds of 15–70 m s⁻¹ and 10–75 m s⁻¹ in the CMA and JTWC best-track datasets, respectively. The typhoons approaching Japan possessed a minimum pressure between 900 and 1005 hPa in all three best-track datasets. The mean intensities of all the typhoons analyzed in 6-h intervals showed a maximum wind speed up to 40 m s⁻¹ and a minimum pressure above 955 hPa.

Figure 3. (a) Typhoon intensity (wind speed vs. central pressure) and (b) hourly rainfall amount within R_L15 (longest radius of wind ≥ 15 m/s from the typhoon center) along the track of each typhoon over Japan land. (c) The latitudinal and (d) longitudinal location of each typhoon at each 6-hours interval. The rainfall intensity in (b) correspond to results calculated with RSMC best-track locations. The bold lines represent the mean values of various typhoon characteristics at each time step.

Most typhoons brought hourly rainfall up to 6 mm in the R_L15 regions over Japan. A few typhoons carried heavy rainfall up to 18 mm h⁻¹ (Figure 3b). The mean rainfall intensities associated with these typhoons were mostly 3 mm h⁻¹ averaged in the R_L15 regions over land. Figures 3(c,d) illustrate the latitudinal and longitudinal locations of the typhoons in 6-h intervals. Typhoons approaching Japan between 2006 and 2019 originated between 5 and 25 N latitude and 125 and 170E longitude. Most passed through 20–40 N and 125–145E within 7–8 days. All three best-track datasets indicated similar locations with a ∼1–2° variation and had the same mean latitudinal locations up to 7 days. This indicated that the typhoons which approached and/or landed the Japanese islands do not simultaneously approach and/or made landfall on coastal areas in the East Asian continent.

3.3. Temporal evolutions

The analysis of wind speed and central pressure indicated that typhoons during 2006–2019 showed peak intensities around the 5^th day after their initiation. The maximum wind speeds were 60, 70, and 75 m s⁻¹ in the RSMC, CMA, and JTWC best-track datasets, respectively (Figure 4a–c). The typhoons showed a minimum central pressure of 900 hPa in all three datasets (Figure 4d–f). The translation speed of each typhoon was calculated in 6-h intervals from the time and distance traveled in ±6 h (Figure 5a–c). Typhoons approaching Japan between 2006 and 2019 typically traveled at a speed of 2–10 m s⁻¹. However, a few traveled at much faster speeds (>20–25 m s⁻¹). The translation speed of typhoons was nearly constant from their initiation until their decay, but a few typhoons had an increased speed on the 5^th day. All the three best-track datasets gave similar translation speeds.

Figure 4. Time evolution of wind speed (upper panel) and minimum central pressure (lower panel) from (a, d) RMSC, (b, e) CMA and (c, f) JTWC best-track datasets during 2006–2019. The open circle represents the magnitude at each time step for the periods 2006–2012 (in blue) and 2013–2019 (in red). The solid blue (red) lines correspond to the mean magnitude at each time step for the period 2006–2012 (2013–2019).

Figure 5. Time evolution of translation speed (upper row), latitudinal location (middle row) and longitudinal location (lower row) from (a, d, g) RMSC, (b, e, h) CMA and (c, f, i) JTWC best-track datasets during 2006–2019. The open circle represents the magnitude at each time step for the periods 2006–2012 (in blue) and 2013–2019 (in red). The solid blue (red) lines correspond to the mean magnitude at each time step for the period 2006–2012 (2013–2019).

In all three best-track datasets, the typhoons between 2006 and 2019 approaching Japan originated in 5–25 N latitude. However, typhoons traveled at higher latitudes up to 60 N in RSMC and CMA datasets. The same trend was noted up to 42 N in the JTWC best-track datasets (Figure 5d–f). In all best-track datasets, the typhoons originated in 125–170E. After making landfall, they went beyond 170E in RSMC datasets, but they mostly went up to 155E in the other two datasets (Figure 5g–i). The size of the typhoons with R_L15 was observed frequently 200–600 km during 2006–2019, although a few were up to 1,000 km (Figure 6a).

Figure 6. Time evolution of (a) typhoon size and (b) rainfall amount with R_L15 along the typhoon track at RMSC best-track location during 2006–2019. The open circle represents the magnitude at each time step for the periods 2006–2012 (in blue) and 2013–2019 (in red). The solid blue (red) lines correspond to the mean magnitude at each time step for the period 2006–2012 (2013–2019).

3.4. Characteristics of typhoons in recent years

To understand the characteristics of typhoons over Japan in the recent years, we compared the intensity, translation speed, location, size, duration, and associated rainfall of typhoons in the earlier years (2006–2012) with those in the recent years (2013–2019). In Figure 4, it is seen that the intensity (maximum wind speed and minimum central pressure) of the typhoons in the two periods were mostly the same from the date of their initiation until the 8th day; after the 8th day, recent typhoons showed stronger winds and lower minimum pressures. All the three best-track datasets provided qualitatively similar results, but the JTWC best-track datasets quantitatively showed higher wind speeds. The student’s t-test was used to evaluate the 95% statistical significance of the mean difference in the intensities of the typhoons for the two periods. The mean intensity in terms of wind and central pressure of recent typhoons was significantly higher (lower in case of central pressure) than that of earlier typhoons (p < 0.05). Mei et al. (2015) also highlighted an increasing trend in the intensity of TCs over the western North Pacific during 1950–2010. Other studies also noted increased TC intensities in recent decades (Mei and Xie 2016; Yoshida et al. 2017; Song et al. 2018; Yamaguchi and Maeda 2020).

The translation speeds were not significantly different between the two periods among the three best-track datasets. Here the student’s t-test did not yield a statistically significant result at the 95% level of confidence. There are some variations in the differences of the translation speed on different days in the RSMC, CMA, and JTWC datasets (Figure 5a–c). Recent studies (e.g. Kim et al. 2020) have highlighted that the translation speed of TCs has increased, whereas other studies (e.g. Chan 2019; Kossin et al. 2020) reported that the TC translation speed decreased. Our overall analysis did not find any significant changes in the translation speed of typhoons approaching Japan between 2006 and 2019 until they made landfall. Thus, it is difficult to provide a definite conclusion on this point.

The latitudinal locations of the typhoons did not show noticeable changes from their initiation until decay between the two periods in the RSMC and CMA best-track datasets. By contrast, that in the JTWC best-track datasets showed 2–5° lower latitudes after the 10th day in recent years (Figure 5d–f). The longitudinal locations of the typhoons were at ∼5° lower longitudes from their initiation until the 5–7th day in recent years, but the changes during the 8–12th day were negligible (Figure 5g–i). None of the datasets showed clear changes in the longitudinal locations for typhoons active over Japan. Additionally, the typhoon size in the RSMC data did not show substantial changes between the two periods (Figure 6a). The rainfall intensity within the R_L15 region over Japan in the two periods indicated almost the same magnitudes (Figure 6b). However, typhoons in earlier years showed rainfall until the 12th day, while rainfall continued even after the 12th day for typhoons in recent years.

Figure 7a depicts the rainfall duration and the total rainfall along the RSMC tracks of the 60 typhoons over the Japanese landmass. The rainfall duration refers to continuous hours of rainfall along the track within the R_L15 region over Japan. Typhoons approaching Japan between 2006 and 2019 brought rainfalls on various time scales with a total amount up to 300 mm. Our analysis showed that, although the frequency of typhoons in recent years was higher (Figure 7a), they brought a slightly lower amount of total rainfall, up to 270 mm. However, typhoons in the recent years caused more frequent rainfall events lasting up to 3 days (Figure 7b), and more frequent total rainfall amounts up to 250 mm than those in the earlier years (Figure 7c).

Figure 7. (a) Duration versus total rainfall amount (accumulated in each duration), and frequency of (b) duration and (c) total rain during 2006–2012 and 2013–2019.

These analyses indicate that typhoons are more intense in the recent years than in the earlier years in terms of maximum wind speed and minimum central pressure. However, in terms of overall rainfall intensity, there is no noticeable change. Notably, typhoons approaching Japan in the recent years have been associated with more frequent rainfall events that last up to 3 days, with a total rainfall amount of up to 250 mm.

4. Discussion

Overall results indicated that that the typhoons in recent years are associated with stronger winds and longer lifespan compared to those in earlier years. To comprehend the mechanism of this severity of typhoons in recent years, we compared the environmental conditions, including SST (Figure 8), vertical wind shear (Figure 9), and moisture (Figure 10) for typhoons approaching Japan between the two periods. The best-track data from RSMC were used here. The time evolution of the mean SST did not show significant differences between the two periods until the 8^th day after initiation. Then SSTs within the typhoon regions were higher in recent years than those in earlier years (Figure 8a). The spatial distribution of the mean SST between 2013 and 2019 showed a higher SST in typhoon active areas over the Pacific Ocean and the Sea of Japan compared to that between 2006 and 2012 (Figure 8b), indicating that warm SST intensified typhoons and prevented decay in recent years (Sun et al. 2017; Mohanty et al. 2019). Takagi and Esteban (2016) also documented that warm SST played an important role in intensified Typhoons Haiyan and Zeb.

Figure 8. (a) Time evolution of surface brightness temperature with R_L15 along each typhoon during 2006–2012 and 2013–2019, and (b) spatial distribution of their mean change. Dotted regions indicate 95% significance level.

Figure 9. (a) Time evolution of vertical wind shear between 850 hPa and 300 hPa with R_L15 along each typhoon during 2006–2012 and 2013–2019, and (b) spatial distribution of their mean change. Dotted regions indicate 95% significance level.

Figure 10. (a) Time evolution of precipitable water with R_L15 along of each typhoon during 2006–2012 and 2013-2019, and (b) spatial distribution of their mean change. Dotted regions indicate 95% significance level.

The time evolution of the vertical wind shear between 850 and 300 hPa indicated that typhoons in recent years were associated with relatively lower vertical wind shears on the 3rd–6th, 8th, and 11th day onwards compared to those in earlier years, indicating more favorable conditions for the intensification and maintenance of typhoons (Figure 9a). Wang et al. (2015) also highlighted that the rapid intensified TCs over the western North Pacific possess relatively weaker vertical wind shear. Fudeyasu et al. (2022) found that the Typhoon Faxai (2019) reached the tropical storm intensity when the vertical shear decreased. We find that the spatial mean difference between the two periods indicated much lower vertical wind shears over the region where most typhoons formed (150–160E and 10–20 N) and over the regions before and after they approached land (121–140E and 22–30 N) (Figure 9b). When the wind shear is low, it not only creates a more stable environment for tropical storms to develop, but also beneficial for various sectors such as aviation as they can reduce turbulence, leading to smoother flights and safer conditions for pilots and passengers. Low wind shear conditions can lead to better oceanic conditions, such as calmer seas and reduced upwelling, which can improve fishing conditions. Low wind shear can create a more stable atmosphere, reducing the chance of severe thunderstorms, hail, and tornadoes. Low wind shear can improve air quality by reducing the mixing of pollutants in the atmosphere. When winds are weak, pollutants can accumulate in the same area, leading to poor air quality.

The time evolution of mean moisture showed a similar pattern as that of SST, but the spatial mean difference of moisture between the two periods over the regions of typhoon formation and southern Japan showed a pattern similar to that of vertical wind shear. The amount of moisture was almost the same from initiation to the 8th day. After the 8th day, it was higher in recent years (Figure 10a). Higher moisture availability was also noticed over the region where a much lower vertical wind shear was observed in recent typhoons (Figure 10b).

Takemi et al. (2019) investigated the impacts of future climate change on typhoon intensities and the resulting hazards by conducting pseudo-global warming experiments on Typhoon Talas (2011). They found that a significant increase in precipitable water vapor contributed to the intensification of heavy rainfalls generated by a typhoon. Recent extreme typhoons and rainfall events (Takemi and Unuma 2019, 2020) had a much larger moisture content than the climatological amount from past observations (Unuma and Takemi 2016). Thus, the environmental moisture content seen for typhoons in recent years and the conditions in other recent extreme events appear to signal that a warmed climate state anticipated to occur in the future has already begun.

The El Niño/Southern Oscillation (ENSO) may be another cause for strengthening the typhoons during 2013–2019. Because the years 2014–2016 were El Niño years and may contribute to stronger typhoons during this period. Multitudes of studies have examined the relationship between TC genesis and ENSO and have provided the evidences supporting the ENSO effect on strengthening the TC (Girishkumar et al. 2015; Zhang et al. 2018; Zhao and Wang 2019: Albert et al. 2022; Song et al. 2022). In the El Niño years, the water temperature over Pacific Ocean remains warmer than other years which creates low vertical shear. This provides a favorable environment for typhoon formations and their maintenance for longer lifespan (Patricola et al. 2018; Wang and Li 2022). In a study, Fudeyasu et al. (2006) reported that the typhoons over Japan during El Niño years tend to have stronger intensities and longer lifespan. The impact of ENSO on typhoons during the period 2006–2019 may have played a similar role towards strengthening the typhoons in recent years. However, it’s worth noting that not all studies have found a significant impact of ENSO on TC intensity. For example, a study by Chan (2005) found no significant correlation between El Niño events and the intensity of TCs in the western North Pacific. Overall, the impact of ENSO on TC intensity is complex and varies depending on a range of factors. More research is needed to fully understand the mechanisms underlying this relationship.

5. Conclusions and recommendations

In this study, 60 typhoons that approached Japan in the past 14 years (2006–2019) were analyzed to explore their temporal evolution, active hours, intensity, latitude, longitude, frequency, size, duration, and translation speed. Typhoons carried winds up to 60 m s⁻¹ around the 5th day after typhoon initiation and had central pressures as low as 900 hPa. Most typhoons during this period traveled at a speed of 2–10 m s⁻¹ with the size of 200–600 km within the R_L15 region. However, the annual frequency of the recent typhoons increased substantially compared to earlier years. In fact, it almost doubled after 2015.

Most typhoons between 2006 and 2019 took an average of 5–8 days to reach Japan and an additional 1–3 days to travel through Japan. Some typhoons took longer to reach Japan (>8 days) and remained over land for longer (>3 days). Typhoons in recent years showed stronger winds and considerably lower pressures at landfall compared to those in earlier years. Recent typhoons also brought frequent continuous heavy rainfalls up to 250 mm total rainfall with a duration of up to 3 days.

This implies that typhoons in the near future may be a serious concern and a threat to human lives and properties. ENSO effect may be a possible cause for this due to its influential interplay in the Pacific Basin. Many previous studies highlighted about the influence of ENSO towards strengthening the TCs. To understand the mechanism for stronger typhoons with heavy rainfall in recent years, we analyzed the SST, vertical wind shear between 850 and 300 hPa, and the precipitable water between the two periods. Environmental conditions in recent years such as higher SST, relatively lesser vertical wind shears, and increased availability of moisture around the typhoon centers have become more favorable to the development of typhoons, making them stronger.

These findings provide valuable information to assess the potential impacts of approaching typhoons in Japan and devise effective adaptation strategies. The study’s results can also be used to evaluate model simulations that assess typhoon activities under warming conditions. However, due to the short length of the analyzed data (only 14 years), drawing a definitive conclusion about climate change would be speculative. Therefore, further investigation using historical and future data is recommended, focusing on anomalies such as unusual strength, direction, and track that may be responsible for potential damages. By studying these factors, we can better understand and prepare for the impacts of future typhoons.

Authors’ contributions

SN proposed the topic, designed the study, analyzed the data and drafted the manuscript. TT helped in the interpretation and the construction of the manuscript. SM helped in analyzing data. All authors read and approved the final manuscript.

Acknowledgments

The authors would like to thank the comments by anonymous reviewers for improving the original manuscript. This study was supported by the Integrated Research Program for Advancing Climate Models (Grant Number JPMXD0717935498) funded by the Ministry of Education, Culture, Sports, Science and Technology of Japan and the Environment Research and Technology Development Fund (Grant Number JPMEERF20192005) of the Environmental Restoration and Conservation Agency. The Japan Meteorological Agency (JMA) is acknowledged for providing the Radar/Rain gauge: Analyzed Precipitation product. The RSMC Tokyo (https://www.jma.go.jp/jma/indexe.html), the CMA Tropical Cyclone Data Center (https://www.typhoon.org.cn), and the JTWC (https://www.metoc.navy.mil/jtwc/jtwc.html?best-tracks) are recognized for providing the best-track datasets.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The dataset supporting the conclusions of this article includes the Japan Meteorological Agency (JMA) based Radar/Rain gauge – Analyzed Precipitation product, and the RSMC Tokyo (https://www.jma.go.jp/jma/indexe.html), the CMA Tropical Cyclone Data Center (https://www.typhoon.org.cn), and the JTWC (https://www.metoc.navy.mil/jtwc/jtwc.html?best-tracks) best-track datasets.

Additional information

Funding

This study was supported by the TOUGOU program Grant Number JPMXD0717935498 and the advanced studies of climate change projection (SENTAN) Grant Number JPMXD0722678534 funded by the Ministry of Education, Culture, Sports, Science, and Technology, Government of Japan.