1. Introduction
During the cool and warm season transition from May to July, a rainy mei-yu (plum rain, baiu, or jangma) season occurs in East Asia (Webster et al. 1998; Wang 2006). The mei-yu season in Taiwan usually lasts for 1 month, from mid-May to mid-June (Wang et al. 1984; Chen and Chen 2003). Stationary mei-yu fronts, low-level jets (LLJs), and southwesterly monsoon flows are the main meteorological systems that play an important role in influencing precipitation in Taiwan during a mei-yu season (e.g., Chen and Liang 1992; Lin et al. 1992; Wang et al. 2014). Because of the southward movement of the front and the increase of the northwestward-pointing pressure gradient force around Taiwan, the LLJs embedded in the southwesterly monsoon flow (e.g., Chien and Chiu 2019; Chien et al. 2021) and the marine boundary layer jet (e.g., Tu et al. 2019; Chen et al. 2022) are often generated on the south side of the front (Lin et al. 1992; Li et al. 1997). The southwesterly flow and the marine boundary layer jet play an important role in the transport of warm humid air from the tropical ocean to the frontal zone, resulting in heavy precipitation events in Taiwan (e.g., Chen and Yu 1988; Li et al. 1997; Chen et al. 2005, 2008; Chien and Chiu 2019; Wang et al. 2023; Chien et al. 2021; Chen et al. 2022). When encountering the complex topography of Taiwan, the airflow associated with the LLJs is often modified such that the intensity and area of precipitation are largely determined by the characteristics of the jets (Xu et al. 2012; Sever and Lin 2017; Tai et al. 2020; Wang et al. 2022). The rainfall distribution usually exhibits a distinct spatial pattern, with more heavy precipitation occurring on the western windward slopes, especially in southwestern Taiwan (Henny et al. 2021; Chen and Chen 2003; Chen et al. 2007). In a set of experiments with idealized simulations of the southwesterly flow during the mei-yu season in Taiwan, Wang et al. (2022) showed that when the wind speed of the southwesterly flow reaches 12.5 m s−1 or higher, precipitation in Taiwan is dominated by the mechanical uplift of unstable air. Heavy precipitation occurs in the wind direction from the south (180°) to the west (270°), with the strongest precipitation occurring in wind directions of 240°–255°.
The Intergovernmental Panel on Climate Change Sixth Assessment Report (IPCC AR6) (IPCC 2023) states that global climate change has been an inevitable trend due to industrial development and natural climate variability. The rising global temperature will cause an increase in atmospheric moisture, which will further lead to an upsurge in precipitation (Kharin et al. 2013). This trend is not limited to the increase in total precipitation, but also to a growth in the frequency of extreme precipitation events (Liu et al. 2009; Kharin et al. 2013; Sillmann et al. 2013). Climate change has also been observed in the East Asian monsoon region. During the past half-century, not only has the distribution of precipitation changed, but also the frequency and intensity of extreme rain events have increased across China, along with a decrease in weak precipitation events (Ma et al. 2015, 2017; Li et al. 2017; Lu et al. 2020). Moreover, in the late twenty-first century, persistent heavy rainfall events are projected to occur more frequently in the middle and lower reaches of the Yangtze River due to the increase of greenhouse gases and decrease of aerosols (Zhou et al. 2021).
The intensity and frequency of extreme precipitation events in Taiwan have also increased, while the trend of consecutive dry days and consecutive wet days decreased from 1960 to 2017, especially noticeably in the southwestern part of Taiwan (Tung et al. 2022). Chen et al. (2011) demonstrated that sea surface temperature (Niño-3.4) anomalies are in phase with monsoon westerlies in northern Indochina, leading to the meridional shift of the North Pacific convergence zone and the interannual variability of its associated rainstorms. Henny et al. (2021) showed that extreme precipitation events in the mei-yu season, which significantly affect the water resources in Taiwan, have occurred more frequently in recent years (1988–2015) than in earlier years (1960–87). Kim and Kim (2020) studied the East Asian summer monsoon (EASM) from 1979 to 2018 and found that the warming of the EASM region caused the mei-yu front to move slightly southward to Taiwan in early summer. Huang et al. (2019b), in a future projection based on high-resolution simulations, found that the enhancement of the southwesterly monsoonal flow will lead to an increase in moisture convergence and a higher intensity and frequency of extreme precipitation on the west side of Taiwan in the late twenty-first century.
From the above reviews, it is clear that precipitation in Taiwan during a mei-yu season is greatly affected by the southwesterly flow. However, although there have been studies discussing the formation mechanism and the causes of interannual variability of the southwesterly flow (Chien and Chiu 2019; Chien et al. 2021), the long-term variability of the southwesterly flow has not been explored. Chien et al. (2021) used the low-level wind field around Taiwan to define the southwesterly flow events in two categories: northern (SWn) and southern (SWs). The former is the southwesterly flow events that occur primarily in northern Taiwan, and the latter, those that occur primarily in southern Taiwan. The SWs events occur primarily during mei-yu seasons and are highly correlated with the intraseasonal and interannual variability of precipitation. Therefore, this study, following Chien et al. (2021), aims to investigate the characteristics and long-term trends of SWs events and to analyze their correlation with precipitation in Taiwan. The key questions to be addressed are as follows:
-
What are the long-term trends and interannual variability of SWs around Taiwan during mei-yu seasons?
-
What are the factors responsible for the change in the characteristics of SWs?
-
Can the number of SWs be used to assess the interannual variability of precipitation in Taiwan?
Data sources and definitions are presented in section 2. Section 3 presents the long-term trends and the causes of SWs and precipitation. Section 4 discusses the feasibility of using the number of SWs and other monsoon indices to assess the interannual variability of precipitation in Taiwan. Last, conclusions are summarized in section 5.
2. Data and method
The ERA 5 reanalysis data (Hersbach 2016) provided by the European Centre for Medium-Range Weather Forecasts (ECMWF) were used for the analyses of climate statistics. A horizontal spatial resolution of 0.5° × 0.5° and a temporal resolution of 6-hourly intervals were selected for the analyzed period of 15 May–15 June from 1979 to 2022. In this study, the atmospheric state at every 6 h is labeled as an event. The number of events for a mei-yu season is thus 128, and the total amount of events evaluated for the 44 mei-yu seasons in this study is 5632. In the discussion that follows, these events of the entire analyzed period are referred to as the mei-yu climate (MYC).
The method for defining the observed rain intensity in Taiwan (RT) is the same as that in Chien et al. (2021). The average rainfall intensity [mm (6 h)−1] for each event was calculated using the hourly rainfall data from the 28 Central Weather Bureau (Taiwan) (CWB) surface weather stations (dots in Fig. 1a). The southern part of Taiwan, which is an essential agricultural region (Huang et al. 2020), is directly influenced by the southwesterly flow coming from the southern Taiwan Strait and the northern South China Sea (SCS). The annual precipitation in this region occurs mainly in the mei-yu (May–June) and typhoon (July–September) seasons (Tung et al. 2022). Since the number of typhoons varies from year to year, rain in the mei-yu season is very important to agriculture in southern Taiwan. Therefore, in addition to RT, precipitation intensity in southern Taiwan (Rs) was also examined in this study. Its intensity was defined in the same way as for RT, except that only seven stations in the south were used (red dots in Fig. 1a).
(a) The 14 boxes (2° × 2°) surrounding Taiwan are the areas used for average winds to identify SWs, with numbers showing the centers of the boxes. The 28 (blue and red) dots indicate the locations of the CWB weather stations used to calculate precipitation intensity in the entire Taiwan island (RT), with the 7 red dots being those used for southern Taiwan (Rs). The color shading denotes the terrain height (m). (b) The areas used for defining monsoon indices in section 2. Also shown are the 850-hPa wind (color shading is wind speed; arrows indicate wind vectors; m s−1) and geopotential height (contours, with interval of 10 gpm) at (c) 0600, (d) 1200, and (e) 1800 UTC 24 May 2015.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-23-0029.1
The definition of SWs in this study is based on the method developed by Chien et al. (2021) and is briefly described below. Basically, the 850-hPa winds from the ERA5 are averaged in fourteen 2° × 2° boxes surrounding Taiwan (Fig. 1a). A southwesterly flow event is identified if it meets the following two criteria: 1) at least one box is required to have a mean wind speed greater than 12 m s−1 with a southwesterly wind direction, and 2) at least one-half of the 14 boxes are required to have a southwesterly wind direction. If there were more boxes having a mean southwesterly wind speed exceeding 12 m s−1 (similar to the first criterion) near northern Taiwan (boxes 1–7) than southern Taiwan (boxes 8–14), the event was designated an SWn event; if not, it was designated as an SWs event. In this paper, the southwesterly wind direction is changed from 200°–260° to 200°–270°. The reason is that a west-southwesterly to westerly wind direction (260°–270°) is also favorable for precipitation in southern Taiwan according to a set of idealized simulations by Wang et al. (2022). Furthermore, in this study, we found that increasing the range of southwesterly wind direction to 270° provided more consistent results for analyzing continuous SWs events than the conventional definition. For example, during the period from 0600 to 1800 UTC 24 May 2015 (Figs. 1c–e), the three events are generally considered a continuous case of SWs. However, if the definition of Chien et al. (2021) is used, the three events are disconnected because the second event at 1200 UTC 24 May 2015 (Fig. 1d) would be excluded owing to its overall westerly wind direction (greater than 260°) to the southwest of Taiwan (boxes 12–14). For these reasons, we therefore applied the new definition in this study and found a total of 565 SWs events in the 44 mei-yu seasons, with an occurrence rate of about 10%.
The numbers of SWs cases and events in four groups divided by the case duration (h) during the 1979–2022 mei-yu seasons.
To examine the possibility of applying monsoon indices in assessing interannual rainfall in Taiwan, we selected three indices that have a high correlation with precipitation in East Asia from Wang et al. (2008). The He index, representing the north–south thermal contrast (He et al. 2001), is defined by an areal mean (0°–10°N, 100°–130°E; red-outlined box in Fig. 1b) of zonal wind difference between the 850- and 200-hPa levels (U850 − U200). The Wang–Fan (WF) index, corresponding to the shear vorticity of zonal winds (Wang and Fan 1999), is defined as the difference of the areal-mean U850 between the south (5°–15°N, 90°–130°E) and the north (22.5°–32.5°N, 110°–140°E) (green-outlined boxes in Fig. 1b). The Li–Zeng (LZ) index, representing the southwesterly monsoon (Li and Zeng 2002), is defined by the areal mean (10°–40°N, 110°–140°E; blue dash-outlined box in Fig. 1b) of the 850-hPa wind speed (WS850). In addition, since the western North Pacific subtropical high (PSH) also affects the EASM activity and the occurrence of southwesterly flows (Wang et al. 2013), the PSH index, H850, is also defined by the areal mean (15°–25°N, 115°–150°E; cyan dash-outlined box in Fig. 1b) of the 850-hPa geopotential height (Φ850). These four indices were calculated in this study using the mei-yu seasonal-mean meteorological fields. When the EASM is stronger, the He, WF, and LZ indices would be higher and the H850 index would be lower, and vice versa. Last, we also evaluate the feasibility of the SWs index, which was defined as the annual number of SWs, in assessing the interannual variability of precipitation in Taiwan.
3. Results
a. Climatology
The differences in environmental conditions between the composite means of the 565 SWs events and the MYC were first compared. As shown in the climatological mean of MYC (Fig. 2a), there are two major weather systems around Taiwan during mei-yu seasons; the low pressure associated with the East Asian monsoon is in southern China, and the high pressure of the PSH is in the western North Pacific. As a result of these two major systems, the geopotential height gradient near Taiwan has a direction pointing toward the southeast, and the wind direction is southwesterly (about 225°). Since the geopotential height gradient is not large in MYC, the southwesterly wind is weak near Taiwan. In the composite mean of SWs (Fig. 2c), however, the geopotential height gradient near Taiwan is large because the midlatitude troughing system moves southward to the north of Taiwan and the PSH ridge extends southwestward, resulting in strong southwesterly flows (with wind directions of about 240°) extending from the northern SCS, through Taiwan, and to the south of Japan. These composite results are similar to those of Chien and Chiu (2019), except that the main jet axis shifts a little more to the south because this study focuses only on the southwesterly flow events in southern Taiwan, the SWs. The strong southwesterly flows transport moisture from the northern SCS to the Taiwan area, resulting in increased moisture and precipitation in the vicinity of Taiwan. Furthermore, the 850-hPa moisture flux divergence is small around Taiwan in MYC (Fig. 2b), but it is strongly negative (convergence) over western Taiwan and strongly positive (divergence) over the ocean east of Taiwan in SWs (Fig. 2d). This pattern in SWs is due to the strong southwesterly flows because when they impinge upon the complex topography of Taiwan, the moisture flux convergence occurs on the windward side and the divergence on the lee side. On the other hand, the lapse rate between the 1000- and 700-hPa levels shows little difference between SWs (Fig. 2d) and MYC (Fig. 2b). These results suggest that the dynamic effect (e.g., flow convergence) is more important than the thermodynamic effect (e.g., vertical stability) in producing heavy rainfall in Taiwan during SWs.
The 44-yr climatological mean of (a) the 850-hPa geopotential height (contours, with interval of 8 gpm), wind (vectors; m s−1), and specific humidity (color shading; g kg−1) and (b) the lapse rate between 1000 and 700 hPa (contours, with interval of 0.3 K km−1) and the 850-hPa moisture flux divergence (color shading; 1 × 10−8 kg kg−1 s−1) in MYC. (c),(d) As in (a) and (b), but for the composite mean of SWs events. Red-outlined boxes A (19°–24°N, 116°–124°E) and B (22°–25.5°N, 120°–121.3°E) denote the area of domain average. The number in the upper-right corner of each plot indicates the number of events used in the average.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-23-0029.1
Figure 3a shows the areal mean of the 850-hPa horizontal wind speed in a region covering southern Taiwan, the southern Taiwan Strait, the northern SCS, and a small ocean area southeast to Taiwan (box A in Fig. 2a, hereinafter the southern vicinity of Taiwan) during mei-yu seasons from 1979 to 2022. In the MYC (blue dots), the average wind speed is about 4 m s−1, and the mean wind direction is southwesterly. However, both of them have a large spread. As for SWs (red dots), the average wind speed increases to about 11 m s−1, with a smaller spread, and the wind direction has more westerly components. It is therefore evident that during an SWs event, the 850-hPa wind speed increases and the wind direction tends to shift to more west-southwesterly. The areal-mean 850-hPa moisture flux divergence near western Taiwan (box B in Fig. 2b) further shows that SWs overall has much stronger moisture flux convergence than MYC (Fig. 3b). Such conditions are favorable for precipitation in Taiwan because the strong southwesterly flow can play an important role in moisture transport and moisture flux convergence in the surrounding area (Chien et al. 2021).
(a) Scatterplots of the areal-mean (box A of Fig. 2a) 850-hPa zonal (x axis) and meridional wind (y axis) components (m s−1) calculated from the 5632 events (blue dots; MYC) and the 565 SWs events (red dots) in 44 yr. Crosses indicate the range of the mean ± 1 standard deviation. Green circles from center to outside represent isotachs of 5, 10, and 15 m s−1. The number of outliers is denoted in the bottom-right corner. (b) Boxplots of the areal-mean (box B of Fig. 2b) 850-hPa moisture flux divergence (1 × 10−8 kg kg−1 s−1) for the MYC (blue) and the SWs (red) events. The box extends from the first quartile to the third quartile (interquartile range), with a line denoting the median value. The upper and lower error bars show the data value that is 1.5× the interquartile range above and below the third and first quartiles, respectively. Dots are outliers.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-23-0029.1
We further compared the precipitation intensity in Taiwan between the MYC and SWs using the exceedance probabilities (Fig. 4), which show that the RT and Rs curves of SWs (dashed lines) were both located to the right of the corresponding curves of MYC (solid lines), suggesting that the rain intensity of SWs events was larger and increased faster toward the extreme rain event than that of MYC. No matter whether for MYC or SWs, the rain intensity of Rs was mostly larger than that of RT. Furthermore, the mean rain intensities of RT and Rs were 2.8 and 3.7 mm (6 h)−1 for MYC, and 10.6 and 17.1 mm (6 h)−1 for SWs, respectively. These results suggest that the chance of heavy precipitation was much higher during SWs events, and more extreme rainfall had occurred in southern Taiwan than in other parts of the island. In the following discussions, a heavy rain event is defined to be when the rain intensity reaches 20 mm (6 h)−1 and above. Figure 4 shows that the probability of rain intensity exceeding 20 mm (6 h)−1 was 2.0% for RT and 5.4% for Rs in MYC, but it increased to 15.4% and 34.3% for SWs, respectively. In other words, the probability of heavy rain in Taiwan/southern Taiwan during an SWs event was about 7.5/6.3 times higher than that of MYC, respectively. In addition, the maximum precipitation intensity was 57.1 mm (6 h)−1 for RT and 105.1 mm (6 h)−1 for Rs, both of which occurred during an SWs event. It is thus evident that the southwesterly flow provided favorable precipitation conditions, particularly for southern Taiwan, which led to a significant increase in precipitation intensity and probability of heavy precipitation events.
The exceedance probability (y axis on a log scale; %) of precipitation intensity [x axis; mm (6 h)−1] of the MYC (solid lines) and SWs (dashed lines) events in 44 yr. Blue and red indicate RT and Rs, respectively. The black horizontal line denotes the median (50%).
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-23-0029.1
The characteristics of the southwesterly flow are somewhat similar to those of the atmospheric river (e.g., Dettinger et al. 2015; Ralph et al. 2016, 2019; Moore et al. 2021; Eiras-Barca et al. 2021) in terms of its capability of long-distance moisture transport and the role it can play in triggering continuous heavy precipitation. Duration of the continuous SWs events (an SWs case) and moisture transport can be two important factors affecting heavy precipitation. Therefore, we examined the relationship of rain intensity with SWs duration (Table 1), moisture flux, and the quantities associated with the moisture flux (Table 2) during the 565 SWs events. Figure 5a shows that for both RT and Rs, precipitation intensity in general increased with increasing SWs duration, judging by the means. If the medians were considered, the same conclusion can well be made for Rs; as for RT, precipitation intensity still increased with increasing duration except for the group of longest duration, in which the median was slightly smaller than that of 2–3 days. However, since the correlation coefficients of rain intensity with SWs duration were relatively low, 0.2 for RT and 0.27 for Rs (Table 3), SWs duration appears to be a minor factor in heavy precipitation. The 565 SWs events were further divided into four groups with equal amounts of events (Table 2) based on the areal average (box A in Fig. 2a) of the 850-hPa moisture flux, WS240, and specific humidity. The box plots show that the precipitation intensity for both RT and Rs overall increased with increasing moisture flux in terms of the medians (Fig. 5b), except for RT in the group of median quartile (Q2) to upper quartile (Q3). The correlation coefficients of rain intensity with moisture flux for the 565 SWs events were 0.43 for RT and 0.48 for Rs (Table 3). Both of these exceeded the 95% significance level, indicating that the amount of moisture flux was important in affecting heavy precipitation and that it was more correlated with rain intensity than the SWs duration during SWs events. Table 3 also shows that moisture flux was positively correlated with SWs duration, suggesting that SWs events with more continuous events before and after tended to have more intense moisture transport.
The 565 SWs events are divided into four groups with equal amounts of events (141), except that the first group has one extra event (142). The lower quartile (Q1), median (Q2), and upper quartile (Q3) that separate the four groups are presented for the areal average (box A in Fig. 2a) of the 850-hPa moisture flux, WS240, and specific humidity.
Boxplots of precipitation intensity (y axis; blue for RT, and red for Rs) vs (a) SWs duration, (b) moisture flux, (c) WS240, and (d) specific humidity during the 565 SWs events. The number of events in each group of (a) is shown in Table 1. Each quantity in (b)–(d) is averaged in box A of Fig. 2a at the 850-hPa level and is divided into four groups with equal amounts of events (141), except that the first group has one extra event (142). The lower quartile (Q1), median (Q2), and upper quartile (Q3) that separate the four groups are presented in Table 2. The box extends from the first quartile to the third quartile (interquartile range), with a line denoting the median value and a cross for the mean. The upper and lower error bars show the data values that are 1.5× the interquartile range above and below the third and first quartile, respectively. Dots are outliers.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-23-0029.1
The correlation coefficients between any two of the quantities including RT, Rs, SWs duration, the 850-hPa moisture flux, WS240, specific humidity, and moisture flux convergence during the 565 SWs events. Moisture flux, WS240, and specific humidity are averaged in box A of Fig. 2a, whereas moisture flux convergence is averaged in box B of Fig. 2b at the 850-hPa level. The asterisk indicates that the correlation passes the t test at a significance level of 95%.
The above analyses show that moisture flux in the southern vicinity of Taiwan is important in determining the rain intensity in Taiwan. However, it remains unknown whether the wind speed of the southwesterly flow or the moisture of the airflow plays a more important role in the moisture flux and rain intensity of RT and Rs. Figure 5c displays that the rain intensity of RT and Rs tended to increase with increasing WS240 during SWs. A minor exception occurred at the group of Q2–Q3 for RT in terms of the medians. The positive correlation is also evident in Table 3, which shows significant correlation coefficients: 0.4 between WS240 and RT, and 0.43 between WS240 and Rs. These results signify that for stronger SWs events, the rain intensity in Taiwan, particularly in southern Taiwan, tended to be heavier. The table also shows that WS240 was highly correlated to SWs duration (0.43) and very closely correlated to moisture flux (0.93). We also examined, although not shown in Table 3, the correlation between WS240 and the 850-hPa geopotential height gradient component pointing from 330° toward 150° (south-southeastward) for all SWs events. The significantly high correlation coefficient of 0.84 confirms that WS240 can mostly be determined by pressure systems around the vicinity of Taiwan. Figure 5d illustrates that rain intensity in general increased with increasing specific humidity during SWs for both RT and Rs. Table 3 also shows that the moisture of the airflow was positively correlated with RT (0.2) and Rs (0.25). However, the correlation coefficients were smaller than those between WS240 and rain intensity, suggesting that the wind speed of the southwesterly flow more dominantly influenced the rain intensity in Taiwan when compared with the moisture of the airflow. This result is also reflected in the smaller correlation coefficients between moisture and the quantities including SWs duration, moisture flux, and WS240. In particular, the correlation between moisture and WS240 was the only one in Table 3 without significance. These results suggest that the wind speed of the southwesterly flow plays a more critical role in moisture flux in comparison with moisture. Last, moisture flux alone cannot produce precipitation. It is the moisture flux convergence caused by orographic lifting that commences the action, which can easily be verified in Table 3 by the higher correlation coefficients of the areal-mean 850-hPa moisture flux convergence in box B either with RT (0.57) or with Rs (0.58) than those between moisture flux and the corresponding rain intensity. Moisture flux convergence also had a high correlation with moisture flux and WS240 but not with SWs duration or specific humidity.
b. Long-term trends of the SWs
This section examines the long-term trends of SWs by presenting the annual number of SWs events and several other associated quantities during the mei-yu seasons from 1979 to 2022, along with their corresponding 11-yr running means (Fig. 6). The long-term linear regression trend of the annual number of SWs (Fig. 6a; blue long-dashed line) in these 44 yr was increasing, with a slope of approximately 1.8 per decade. The smallest number of the 11-yr running mean (black curve) was 6.6 in 1991, while the peak was 18.8 in 2002. A relatively small number of 13.5 occurred in 2013, and the number increased to 14.8 toward the last year of 2017. This result suggests that an overall rising trend with approximately a decadal oscillation (about a 10 yr oscillation) was present in the number of SWs, as judged by the fact that the number was relatively small in the 1980s, rose in the 1990s, declined slowly after the early 2000s, and rose again in the 2010s.
Time series of annual (a) number of SWs (times), (b) event-mean wind direction (°), (c) event-mean wind speed (m s−1), (d) event-maximum wind speed (m s−1), (e) event-mean moisture flux (m s−1 g kg−1), (f) event-maximum moisture flux (m s−1 g kg−1), (g) number of SWs case lasting longer than 12 h (times), and (h) maximum duration of SWs cases (h) during mei-yu seasons in 1979–2022. Quantities in (b)–(f) are the areal average in box A of Fig. 2a. The gray bars show the annual values, with the 44-yr mean denoted on the rightmost of each diagram. The black curve represents the 11-yr running mean. The blue long-dashed line shows the best-fit linear trend, with its associated 95% confidence interval denoted in blue dashed curves. The slope of the linear trend (per decade) is shown in the upper-right corner of each plot. The 44-yr correlation coefficient between the quantity (gray bars) in (b)–(h) and the number of SWs (gray bars) in (a) is depicted in the upper-left corner of the corresponding panel, with an asterisk indicating that the correlation passes the 95% significance level. Since there were no SWs in 1980 and 1991, missing data were given to these two time points, except for (a) and (g).
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-23-0029.1
The following text discusses the areal-mean wind direction, wind speed, and moisture flux, which were computed in box A of Fig. 2a during the SWs. These quantities were then averaged using all the SWs events in a mei-yu season to obtain the corresponding event-mean quantities of that particular year. The maximum value of the quantity in the season was defined as the event maximum of that year. The annual number of long-lasting SWs cases was determined by the number of SWs cases that lasted longer than 12 h in the mei-yu season, and the annual maximum duration of the SWs case as the duration (h) of the longest-lasting SWs case of the season. Figure 6b shows that the annual event-mean wind directions of SWs only slightly fluctuated around 238.5 ± 3° in the 11-yr running mean. This little variation is perhaps not surprising because a wind direction of nearly southwesterly is a prerequisite for SWs. However, the other quantities related to SWs, including event-mean wind speed (Fig. 6c), event-maximum wind speed (Fig. 6d), event-mean moisture flux (Fig. 6e), event-maximum moisture flux (Fig. 6f), number of long-lasting SWs cases (Fig. 6g), and the longest SWs case duration (Fig. 6h), all exhibited a similar increasing trend as the number of SWs (Fig. 6a). Their 11-yr running means also displayed a similar decadal oscillating pattern as that of the number of SWs. The above similarity can be further verified in the highly positive correlation coefficients between these quantities and the number of SWs shown in the upper-left corner of each panel. Except for that of event-mean wind direction (Fig. 6b), all the correlation coefficients passed the 95% significance level. The results suggest that the years with more active SWs were more favorable for the occurrence of SWs events with stronger wind speeds, longer durations, and larger moisture fluxes. Due to global warming, there has been a long-term trend of increasing moisture in the atmosphere (Sun and Held 1996); this fact may be related to the recent rise of event-mean moisture flux, which shows a peak in 2017 in the 11-yr running mean (Fig. 6e).
We also examined the long-term trends of four other quantities related to precipitation: mei-yu seasonally accumulated precipitation, SWs event-mean precipitation intensity, SWs event-maximum precipitation intensity, and the percentage of SWs event-accumulated precipitation to the mei-yu seasonally accumulated precipitation during the 44 mei-yu seasons for both RT and Rs. Since the trends of these four quantities are quite similar between RT and Rs, only those of Rs are shown in Fig. 7. When compared with other quantities, the linear trend of mei-yu seasonally accumulated precipitation (Fig. 7a) had a relatively gentle slope during these 44 yr. However, its 11-yr running mean still had a decadal oscillation, with relatively larger values in the 2000s and smaller values in the late 1980s to early 1990s and in the 2010s. This trend was consistent with the number of SWs, as shown in the high correlation coefficient (0.83), suggesting that the decadal oscillation of the number of SWs can play a role in the amount of seasonal precipitation in Taiwan during mei-yu seasons. The SWs event-mean precipitation intensity (Fig. 7b) showed a relatively larger increasing trend in the 44-yr linear regression line, and its 11-yr running mean also exhibited a continuously increasing trend after the 1990s. Although the positive correlation with the number of SWs was not very high, it still passed the 95% significance level. Both the SWs event-maximum precipitation intensity (Fig. 7c) and the percentage of SWs event-accumulated precipitation to the mei-yu seasonally accumulated precipitation (Fig. 7d) had an increasing linear trend. The latter was more significant than the former, with a slope of 6.5% per decade in southern Taiwan. Consistent with the activity of SWs, these two quantities also exhibited a decadal oscillation in the 11-yr running means, with a relatively small value in the early 1990s, a relative peak in the 2000s, and then another relative minimum in the 2010s, which was why their correlations with the number of SWs were highly positive. These results indicate that the role of SWs in precipitation in southern Taiwan during a mei-yu season has become ever more important in recent years. As a result of the aforementioned increasing trends, stronger and longer-lasting precipitation events occurred during SWs in recent years, leading to the increasing potential threat of SWs to Taiwan.
As in Fig. 6, but for (a) mei-yu seasonally accumulated precipitation (mm), (b) SWs event-mean precipitation intensity [mm (6 h)−1], (c) SWs event-maximum precipitation intensity [mm (6 h)−1], and (d) the percentage of SWs event–accumulated precipitation to the mei-yu seasonally accumulated precipitation (%) during mei-yu seasons in 1979–2022. The precipitation evaluated is Rs.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-23-0029.1
c. Long-term trends of environmental conditions
Extreme weather events tend to occur not only when the averaged environmental fields (e.g., wind speed) increase, but also when the fields have strong intraseasonal oscillations (Li et al. 2015; Huang and Chang 2018; Huang et al. 2019a). The occurrence of SWs is primarily determined by the 850-hPa wind direction and wind speed. Since there was not much long-term variation in the wind direction of SWs and the mean direction was about 240° (Fig. 6b), we next analyzed the interannual variability of mean southwesterly wind speed (WS240) and its associated variance at 850 hPa. The 6-hourly areal mean of WS240 was calculated first in box A of Fig. 2a; the seasonal mean of WS240 and its variance were then calculated during the 44 mei-yu seasons, and last, an 11-yr running mean was taken for the mean and the variance. Since the first 5 yr (1979–83) and the last 5 yr (2018–22) lack sufficient data in the running mean, they are omitted in Fig. 8. The scatterplot shows that the mean wind speed of WS240 (x axis) overall had a long-term increasing trend. For example, the mean WS240 values were 3.23, 3.55, and 4.74 m s−1 in 1990, 2000, and 2010, respectively, which were closely related to the long-term increasing trend in the number of SWs. The mean WS240, however, did not exhibit a clear decadal oscillation like the number of SWs. For example, in the 2000s, when the number of SWs was at its highest (Fig. 6a), the mean WS240 was not the largest in 44 yr, suggesting that the interannual variability of wind speeds may not be the key factor determining the activity of SWs. The more related factor is actually the WS240 variance (the y axis), which showed a clear decadal oscillation. The wind speed variance was larger in the initial years, gradually decreased to a minimum in 1991, increased until 2001, and then decreased again with a small fluctuation toward 2017. This decadal variability of the WS240 variance can play an important role in the SWs occurrence, leading to an ∼10-yr oscillation of the number of SWs.
Scatterplot of the 11-yr running mean of seasonal- and areal-average (box A in Fig. 2a) WS240 (x axis; m s−1) and its seasonal variance (y axis; m2 s−2). The numbers indicate the last two digits of each year from 1984 to 2017, with black, red, green, and blue colors representing the periods of the 1980s, 1990s, 2000s, and 2010s, respectively. The gray dashed lines denote the 44-yr mean. The years 1979–83 and 2018–22 are omitted.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-23-0029.1
To investigate the causes of the above long-term WS240 trends, we examined the decadal average of the 850-hPa geopotential height and WS240 in four periods (Fig. 9). The mean WS240 in box A of Fig. 2a depends on the geopotential height gradient between the Asian summer monsoon circulation to the northwest and the PSH to the southeast of Taiwan. From the 1980s (Fig. 9a) to the 1990s (Fig. 9b), the PSH strengthened and pushed southwestward, resulting in an increase of geopotential height gradient near Taiwan. During the 2000s (Fig. 9c), although the PSH weakened near its center, the geopotential height around the Philippines area and around the Taiwan area was higher and lower, respectively, than in the 1980s (Fig. 9a). As a result, the geopotential height gradient in the southern vicinity of Taiwan was actually larger in the late 2000s than in the early years (Fig. 9f). Although in a short transitional period (e.g., 1998–2002), when the PSH weakened and the monsoonal trough had not yet extended to a southward latitude, the geopotential height gradient was relatively small, it increased soon after when the monsoonal trough well extended southward. The PSH strengthened again in the last decade (Fig. 9d), resulting in the increased geopotential height gradient. Due to this long-term increase trend of the south-southeastward-pointing geopotential height gradient near Taiwan (Fig. 9f), WS240 in the area around box A also increased (Figs. 9a–d). The 11-yr running means of the longitudinally averaged 850-hPa geopotential height and WS240 along a north–south line of box C in Fig. 9d show in more detail the decadal oscillation of the PSH and the increasing trend of WS240 in box A (Fig. 9e), including the southwestward extension of the PSH toward the Philippines that caused the pressure gradient increase in the 1990s and 2010s and the southward extension of the low pressure to the north of Taiwan when the PSH weakened slightly in the 2000s.
The decadal mean of the 850-hPa geopotential height (color shading; gpm) and WS240 (contours of 5 m s−1 and above, with interval of 1 m s−1) during the (a) 1979–89, (b) 1990–2000, (c) 2001–11, and (d) 2012–22 mei-yu seasons. (e) The Hovmöller diagram of the longitudinally averaged 850-hPa geopotential height (color shading; gpm) and WS240 (contours, with interval of 1 m s−1) along a north–south line shown in the purple-outlined box C of (d), with the width of the box showing the longitudinal range. Two black lines denote the north and south bounds of box A in Fig. 2a. (f) The south-southeastward-pointing component (a direction of 330°) of geopotential height gradient (perpendicular to WS240) averaged in box A of Fig. 2a. An 11-yr running mean is also performed on the data in (e) and (f) such that the years 1979–83 and 2018–22 are omitted.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-23-0029.1
Variances of the aforementioned 850-hPa geopotential height and WS240 are shown in Fig. 10. The horizontal domains are smaller than those in Fig. 9 because variances in regions far from Taiwan are not important for the current discussions. The geopotential height variance around Taiwan in all periods showed a northward-pointing gradient, with higher values in the north and low values in the south (Figs. 10a–d), indicating perhaps unsurprisingly that weather systems in higher latitudes that contributed to the most variances were more active than those in the subtropics. There was a decadal oscillation of the geopotential height variance near Taiwan, with larger variances in the periods of 1979–89 (Fig. 10a) and 2001–11 (Fig. 10c) and smaller in the other two periods of 1990–2000 (Fig. 10b) and 2012–22 (Fig. 10d). The larger variances in the 2000s were more evidently shown in Fig. 10e; this result was consistent with the decadal oscillation of WS240 variances in the area of box A (Fig. 10e; 19°–24°N), which was larger in the initial years, decreased in the early 1990s, became larger in the 2000s, and gradually decreased after reaching a peak. Before 1990 and between 2000 and 2010 (Fig. 10e), geopotential height variances to the north of Taiwan were relatively large, indicating more monsoonal or midlatitude troughs forming in this region. As these low pressure systems moved to the north of Taiwan, the pressure gradient around Taiwan increased, resulting in more SWs events. It is therefore concluded that strong intraseasonal oscillations can play an important role in the occurrence of southwesterly flows in the southern vicinity of Taiwan.
As in Figs. 9a–e, but for the geopotential height variance (color shading; gpm2) and the WS240 variance (contours, with interval of 6 m2 s−2). In (a–d), horizontal domains are smaller than those in Fig. 9, and only contours of 30 m2 s−2 and above are shown. In (e), the Hovmöller diagram is made along a north–south line shown in the purple-outlined box D of (d).
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-23-0029.1
To examine the causes of SWs occurrences in different decades, the composite means of environmental fields during SWs in the four decadal periods are shown in Fig. 11. When compared with the periods of weaker WS240 in 1979–2000 (Figs. 11a,b), the periods of stronger WS240 in 2001–22 (Figs. 11c,d) had relatively stronger wind speeds in the southern vicinity of Taiwan. The stronger WS240 can help in transporting moisture from the northern SCS to the Taiwan area and is one of the important reasons for the long-term increasing trend of most quantities presented in Fig. 6, particularly the event-mean or event-maximum moisture flux (Figs. 6e,f). During the periods in 1979–89 and 2000–11 when the WS240 variance was larger (Figs. 11a,c), stronger low pressure systems were located to the north of Taiwan during SWs and the PSH did not extend southwestward to the Philippines, indicating that the midlatitude low and the monsoonal trough played an important role in the occurrence of SWs during periods of higher variances. During mei-yu seasons, the activity of such troughs is mostly accompanied by the southward movement of mei-yu fronts, which lower the pressure to the north of Taiwan, leading to the formation of strong southwesterly flows on the south side of the fronts. The mei-yu fronts with their lingering characteristics can prolong the duration of the SWs cases, possibly explaining why long-lasting SWs cases were more likely to occur in the 1980s and particularly in the 2000s (Figs. 6g,h). During the periods (1990s and 2010s) when the WS240 variance was smaller (Figs. 11b,d), the PSH was stronger to the southeast of Taiwan, and the low pressure system was relatively weaker to the north of Taiwan, suggesting that the position and intensity of the PSH played a more important role in the occurrence of SWs events than the monsoonal or midlatitude troughs during these periods.
The composite means of the 850-hPa geopotential height (contours, with interval of 8 gpm), wind speed (color shading; m s−1), and streamlines for the SWs events during (a) 1979–89, (b) 1990–2000, (c) 2001–11, and (d) 2012–22. An inset on the upper left of each diagram highlights wind speeds and vectors in box A of Fig. 2a, and a number on the upper right indicates the number of SWs used in the composite. Red-outlined boxes N and S in (d) denote the regions of areal mean calculation.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-23-0029.1
The aforementioned results are further confirmed by Table 4, in which the relative importance of either the low system to the north or the high system to the south of Taiwan to the SWs formation is investigated. We first selected boxes N and S (see Fig. 11d for locations) based on the regions of maximum 850-hPa geopotential height differences between SWs and MYC. The areal means of 850-hPa geopotential height in boxes N and S were then computed for all of the 5632 events of MYC. Q1 of box N and Q3 of box S, which were 1462 and 1505 gpm, respectively, were chosen as the thresholds to determine the existence of either a low system in the north or a high system in the south. For each SWs event, the areal-mean 850-hPa geopotential height in box N/S was examined; if it was smaller than 1462/larger than 1505 gpm, we assumed that a low/high system existed in the north/south and led to the formation of the event, which was then determined as a low in the north (NL)/ high in the south (SH) event. Note that a particular SWs event may belong to both NL and SH events. Table 4 shows that the number of NL was always larger than that of SH for all periods. The ratio of NL to SH was 2.3 (414/183) in 44 yr, which was consistent with the larger variances in the north of Taiwan in Fig. 10. In addition, this ratio also exhibited a decadal oscillation. During the periods in 1979–89 and 2000–11 when the WS240 variance was larger, the ratios were 5.6 for the former and 2.9 for the latter. These numbers were larger than the mean ratio (2.3), suggesting that the low systems to the north of Taiwan were more dominant in the formation of SWs events. During the other two periods when the WS240 variance was smaller, both ratios reduced to about 1.6. This result suggests that although the monsoonal or midlatitude troughs still contributed to more occurrences of SWs, the PSH became more important in 1990–2000 and 2012–22 than during the periods of stronger WS240 variance.
The number of SWs events caused by the low in the north (NL) and the high in the south (SH) for the periods of 1979–89, 1990–2000, 2001–11, and 2012–22 and for all years (1979–2022). The ratio of NL to SH is also presented. Note that the combined number may exceed the total number of the SWs events because a particular SWs event may be caused by both NL and SH.
4. The SWs index
Monsoon indices are frequently used in regions influenced by monsoon to analyze favorable atmospheric conditions for precipitation and to assess the possible amount of precipitation during a rainy season (e.g., Wang and Fan 1999; He et al. 2001; Li and Zeng 2002). There have been dozens of such indices proposed for the East Asian monsoon regions; however, none of them is capable of estimating precipitation in Taiwan. To understand the possibility of applying the number of SWs as a monsoon index (the SWs index) to assess precipitation in Taiwan, we analyzed the correlation coefficients between mei-yu seasonally accumulated precipitation in Taiwan (RT and Rs) and the SWs index and then compared the results with those of several other East Asia monsoon indices (Table 5). The 44-yr correlation coefficient between the SWs index (Fig. 6a, gray bars) and RT for the 1979–2022 mei-yu seasons was 0.66, which was higher than 0.58 presented in Chien et al. (2021) in which the ERA-Interim reanalysis data were used for analyses. This finding suggests, on one hand, that the current generation of ECMWF reanalysis data (ERA5) with higher resolution can better reproduce the actual atmospheric conditions. On the other hand, it also indicates that the adjusted definition of the SWs index used in this study can better reflect the interannual variability of precipitation in Taiwan. As already presented in Fig. 7a, the correlation coefficient between the SWs index and Rs (Fig. 7a, gray bars) was even higher (0.83), meaning that seasonal precipitation in southern Taiwan can be more effectively guided by this index. In comparison with other indices, the SWs index has the highest correlation coefficients for RT and Rs, both of which passed the t test well, at a significance level of 95%. This result suggests that the SWs index, which best reflected the interannual variability of precipitation in Taiwan, can be used for assessing monsoonal precipitation associated with mei-yu in Taiwan, especially southern Taiwan. The other indices, except for the LZ index, which was the second best, failed the t test at a significance level of 95%. These traditional monsoon indices were barely capable of evaluating the long-term annual variability of precipitation in Taiwan.
The 44-yr correlation coefficients between the seasonal precipitation amount in Taiwan (RT; Rs) and the monsoon indices, including SWs, He, WF, LZ, and H850, during mei-yu seasons from 1979 to 2022. The asterisk indicates that the correlation passes the t test at a significance level of 95%.
In the previous section, it was shown that the main weather systems affecting precipitation in Taiwan may vary from one decade to another. To understand whether the SWs index is applicable not only to the entire 44-yr period, but also to periods of different lengths, the 21- and 11-yr sliding correlation coefficients between mei-yu seasonally accumulated precipitation in Taiwan (RT and Rs) and all monsoon indices were calculated (Fig. 12). The SWs index still had the highest 21-yr sliding correlation coefficients with either RT (Fig. 12a) or Rs (Fig. 12b) among the five indices during the whole period of investigation, and the coefficients all passed the t test at a significance level of 95%. The other indices, with the exception of the LZ index, which passed the significance test in most years, exhibited insignificant correlations with rainfall. In the 11-yr sliding correlations with RT (Fig. 12c), the SWs index was insignificant before 1988, during which period the WF and H850 indices were better, suggesting that precipitation in the entire Taiwan Island was more influenced by large-scale circulation during these less active years of SWs. However, after 1989, the SWs index again had the highest and most significant correlation coefficients with RT among all indices. Similar to the results in Fig. 12b, the correlation coefficients of the SWs index with Rs, which all passed the 95% significance level, were significantly higher than those of other indices (Fig. 12d).
The 21-yr sliding correlation coefficients between monsoon indices and seasonally accumulated precipitation of (a) RT and (b) Rs during mei-yu seasons from 1979 to 2022. The monsoon indices include SWs, He, WF, LZ, and H850. (c),(d) As in (a) and (b), but for 11-yr sliding correlation coefficients. The solid gray lines indicate the values of ±0.44 for (a) and (b) and ±0.61 for (c) and (d), as determined by the t test at a significance level of 95%.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-23-0029.1
The above analyses demonstrate that among all the indices, the SWs index is the most capable of assessing the interannual variability of precipitation in Taiwan during all different periods, regardless of whether it is applied to RT or Rs. It can therefore be used as an indicator of seasonal precipitation in Taiwan during mei-yu seasons. Because Taiwan is located at the boundary between the South Asian monsoon and the East Asian monsoon regions, the weather systems affecting this area are complex. Furthermore, Taiwan has steep topography that can complicate the causes of precipitation in a rainy season. As a result, although the traditional monsoon indices derived from seasonal mean fields can reflect the precipitation in many East Asian monsoon regions, they are unable to successfully assess the precipitation in Taiwan. On the other hand, because the SWs index is defined by the strong southwesterly wind events, it includes signals of both the seasonal mean wind speed and the intraseasonal wind speed oscillation that play an important role in moisture flux. Therefore, the index can successfully assess the interannual variability of precipitation in Taiwan, especially in southern Taiwan that is mostly located on the windward side of the southwesterly flow. Consequently, it is important to perform future studies related to SWs, such as investigating the seasonal SWs forecast and the influence of climate change on the SWs.
5. Conclusions
This paper investigates the characteristics and long-term trends of the SWs events around Taiwan and their relationship with precipitation in Taiwan during the mei-yu (15 May–15 June) seasons from 1979 to 2022, using ERA5 reanalysis data and rainfall data from the 28 CWB weather stations in Taiwan. The possibility of using the number of SWs events as a monsoon index in evaluating the interannual variability of rainfall in Taiwan is also discussed herein. The results show that when the SWs event occurred, on average, low-level winds strengthened with wind directions turning to west-southwesterly in the vicinity of southern Taiwan. The strong southwesterly flow transported moisture toward the Taiwan area, providing favorable environmental conditions for precipitation and increasing the probability of heavy rainfall events in Taiwan. During SWs events, mean precipitation intensity in Taiwan (RT) was about 3.8 times as high as that of the mei-yu climate, and 4.6 times as high in southern Taiwan (Rs). The probability of heavy precipitation events in RT was about 7.7 times as high as in the mei-yu climate, while in Rs it was 6.4 times as high. Rs had both a higher precipitation intensity and a higher probability of heavy precipitation events than RT because southern Taiwan was more affected by SWs. Furthermore, the strong moisture flux in the southern vicinity of Taiwan plays a more important role than the duration of SWs in leading to heavy rain.
The trend analyses of 44-yr data show that all the presented quantities associated with SWs, except event-mean wind direction, exhibit a similar long-term increasing trend, with a decadal oscillation. They include the number of SWs events, event-mean wind speed, event-maximum wind speed, event-mean moisture flux, event-maximum moisture flux, number of SWs cases lasting longer than 12 h, maximum duration of SWs cases, event-maximum precipitation, and the percentage of event-total precipitation to the total precipitation during a mei-yu season. Taking the number of SWs events as an example, the decadal oscillation consists of numbers that are relatively small before 1990, increase in the 1990s, reach a peak in the 2000s, slowly decrease until 2010, and increase again in the 2010s. In particular, both the SWs event-maximum precipitation intensity and the percentage of SWs event-accumulated precipitation to the mei-yu seasonally accumulated precipitation showed an increasing linear trend, indicating that the role of SWs in precipitation in southern Taiwan has become more and more important during a mei-yu season. In addition, due to the increasing oscillations in atmospheric conditions caused by climate change, it is likely that longer-lasting SWs and stronger precipitation may occur more often in the future, increasing the potential risk of SWs to Taiwan.
The long-term trend of SWs is closely related to the wind speed of the west-southwesterly flow (WS240) and its variance in the southern vicinity of Taiwan. The southwestward extension of the PSH toward the Philippines caused the pressure gradient increase near Taiwan in the 1990s and 2010s. Although the PSH weakened slightly in the 2000s, the low pressure to the north of Taiwan extended southward, maintaining a relatively large pressure gradient near Taiwan. This long-term increasing trend of geopotential height gradient lead to an increasing trend of WS240 in the southern vicinity of Taiwan. On the other hand, the decadal oscillation in the activity of SWs was a result of geopotential height variance around Taiwan. During the periods of stronger intraseasonal oscillations in 1979–89 and 2001–11, the midlatitude low and the monsoon trough to the north of Taiwan played an important role in the occurrence of SWs. As these low pressure systems moved to the north of Taiwan, the pressure gradient around Taiwan increased, resulting in more SWs events. It is therefore suggested that strong intraseasonal oscillations can provide a favorable condition for the occurrence of southwesterly flows in southern Taiwan. On the other hand, during weaker oscillations in 1990–2000 and 2012–22, the intensity and location of the PSH to the southeast of Taiwan played a relatively more important role in affecting the occurrence of SWs when compared with the periods of stronger oscillations.
It is exceptionally difficult to assess the interannual variability of precipitation in Taiwan, as the weather systems that mainly affect precipitation may vary from one period to another. In this study, the annual number of SWs events is considered as a monsoon index (the SWs index), which resembles well both the seasonal mean wind speed and the intraseasonal wind speed oscillation. As compared with other indices that only use the mei-yu seasonal-mean fields, the SWs index, which directly reflects the impact of strong southwesterly flows on precipitation, can therefore best assess the interannual variability of precipitation in Taiwan, especially southern Taiwan. Its correlation coefficients with precipitation are significantly higher than those of other indices for all different lengths of periods considered, such as 44, 21, and 11 yr. Because the SWs index can be easily obtained in both intraseasonal and interannual observations and forecasts, it is a useful index for evaluating precipitation in either the entire Taiwan Island or southern Taiwan. Therefore, it would be worthwhile to further investigate the seasonal prediction of SWs.
Acknowledgments.
The data used in this study are obtained from the European Centre for Medium-Range Weather Forecasts (ECMWF) and the Central Weather Bureau (Taiwan) (CWB). This research was supported by the Ministry of Science and Technology of Taiwan (Grants: MOST 111-2111-M-003-004, MOST 111-2625-M-003-002, and NSTC 112-2111-M-003-003).
Data availability statement.
The ERA5 dataset (https://www.ecmwf.int/en/forecasts/dataset/ecmwf-reanalysis-v5) and the dataset of surface weather stations (https://dbar.pccu.edu.tw/member/ParameterSearch.aspx) are available online.
REFERENCES
Chen, C.-S., and Y.-L. Chen, 2003: The rainfall characteristics of Taiwan. Mon. Wea. Rev., 131, 1323–1341, https://doi.org/10.1175/1520-0493(2003)131<1323:TRCOT>2.0.CO;2.
Chen, C.-S., W.-C. Chen, Y.-L. Chen, P.-L. Lin, and H.-C. Lai, 2005: Investigation of orographic effects on two heavy rainfall events over southwestern Taiwan during the mei-yu season. Atmos. Res., 73, 101–130, https://doi.org/10.1016/j.atmosres.2004.07.005.
Chen, C.-S., Y.-L. Chen, C.-L. Liu, P.-L. Lin, and W.-C. Chen, 2007: Statistics of heavy rainfall occurrences in Taiwan. Wea. Forecasting, 22, 981–1002, https://doi.org/10.1175/WAF1033.1.
Chen, G. T.-J., and C.-C. Yu, 1988: Study of low-level jet and extremely heavy rainfall over northern Taiwan in the mei-yu season. Mon. Wea. Rev., 116, 884–891, https://doi.org/10.1175/1520-0493(1988)116<0884:SOLLJA>2.0.CO;2.
Chen, G. T.-J., and C.-Y. Liang, 1992: A midlevel vortex observed in the Taiwan Area Mesoscale Experiment (TAMEX). J. Meteor. Soc. Japan, 70, 25–41, https://doi.org/10.2151/jmsj1965.70.1_25.
Chen, G. T.-J., C.-C. Wang, and S.-W. Chang, 2008: A diagnostic case study of mei-yu frontogenesis and development of wavelike frontal disturbances in the subtropical environment. Mon. Wea. Rev., 136, 41–61, https://doi.org/10.1175/2007MWR1966.1.
Chen, T.-C., W.-R. Huang, and M.-C. Yen, 2011: Interannual variation of the late spring–early summer monsoon rainfall in the northern part of the South China Sea. J. Climate, 24, 4295–4313, https://doi.org/10.1175/2011JCLI3930.1.
Chen, Y.-L., C.-K. Wang, C.-C. Tu, F. Hsiao, and P.-L. Lin, 2022: Revisiting a mei-yu front associated with heavy rainfall over Taiwan during 6–7 June 2003. Atmosphere, 13, 644, https://doi.org/10.3390/atmos13050644.
Chien, F.-C., and Y.-C. Chiu, 2019: A composite study of southwesterly flows and rainfall in Taiwan. J. Meteor. Soc. Japan, 97, 1023–1040, https://doi.org/10.2151/jmsj.2019-057.
Chien, F.-C., Y.-C. Chiu, and C.-H. Tsou, 2021: A climatological study of southwesterly flows and heavy precipitation in Taiwan during mei-yu seasons from 1979 to 2018. J. Meteor. Soc. Japan, 99, 913–931, https://doi.org/10.2151/jmsj.2021-044.
Dettinger, M. D., F. M. Ralph, and D. A. Lavers, 2015: Setting the stage for a global science of atmospheric rivers. Eos, 96, https://doi.org/10.1029/2015EO038675.
Eiras-Barca, J., and Coauthors, 2021: European west coast atmospheric rivers: A scale to characterize strength and impacts. Wea. Climate Extremes, 31, 100305, https://doi.org/10.1016/j.wace.2021.100305.
He, M., W. Song, and L. Xu, 2001: Definition of the South China Sea monsoon index and associated prediction. Dates of Summer Monsoon Onset in the South China Sea and Monsoon Indices (in Chinese), J. H. He, Y. H. Ding, and H. Gao, Eds., China Meteorological Press, 109–110.
Henny, L., C. D. Thorncroft, H.-H. Hsu, and L. F. Bosart, 2021: Extreme rainfall in Taiwan: Seasonal statistics and trends. J. Climate, 34, 4711–4731, https://doi.org/10.1175/JCLI-D-20-0999.1.
Hersbach, H., 2016: The ERA5 atmospheric reanalysis. 2016 Fall Meeting, San Francisco, CA, Amer. Geophys. Union, Abstract NG33D-01.
Huang, K.-C., C.-J. Lee, S.-L. Chan, and C.-H. Tai, 2020: Vulnerability assessment and adaptation strategies for the impact of climate change on agricultural land in southern Taiwan. Sustainability, 12, 4637, https://doi.org/10.3390/su12114637.
Huang, W.-R., and Y.-H. Chang, 2018: Impact of boreal summer intra‐seasonal oscillations on warm season diurnal convection activity in Taiwan. Int. J. Climatol., 38, 2187–2200, https://doi.org/10.1002/joc.5326.
Huang, W.-R., P.-Y. Liu, J.-H. Chen, and L. Deng, 2019a: Impact of boreal summer intra-seasonal oscillations on the heavy rainfall events in Taiwan during the 2017 meiyu season. Atmosphere, 10, 205, https://doi.org/10.3390/atmos10040205.
Huang, W.-R., P.-H. Huang, Y.-H. Chang, C.-T. Cheng, H.-H. Hsu, C.-Y. Tu, and A. Kitoh, 2019b: Dynamical downscaling simulation and future projection of extreme precipitation activities in Taiwan during the mei-yu seasons. J. Meteor. Soc. Japan, 97, 481–499, https://doi.org/10.2151/jmsj.2019-028.
IPCC, 2023: Climate Change 2021: The Physical Science Basis. Cambridge University Press, 2391 pp., https://doi.org/10.1017/9781009157896.
Kharin, V. V., F. W. Zwiers, X. Zhang, and M. Wehner, 2013: Changes in temperature and precipitation extremes in the CMIP5 ensemble. Climatic Change, 119, 345–357, https://doi.org/10.1007/s10584-013-0705-8.
Kim, K.-Y., and B.-S. Kim, 2020: The effect of regional warming on the East Asian summer monsoon. Climate Dyn., 54, 3259–3277, https://doi.org/10.1007/s00382-020-05169-7.
Li, H., H. Chen, and H. Wang, 2017: Effects of anthropogenic activity emerging as intensified extreme precipitation over China. J. Geophys. Res. Atmos., 122, 6899–6914, https://doi.org/10.1002/2016JD026251.
Li, J., and Q. Zeng, 2002: A unified monsoon index. Geophys. Res. Lett., 29, 1274, https://doi.org/10.1029/2001GL013874.
Li, J., Y.-L. Chen, and W.-C. Lee, 1997: Analysis of a heavy rainfall event during TAMEX. Mon. Wea. Rev., 125, 1060–1082, https://doi.org/10.1175/1520-0493(1997)125<1060:AOAHRE>2.0.CO;2.
Li, J., J. Mao, and G. Wu, 2015: A case study of the impact of boreal summer intraseasonal oscillations on Yangtze rainfall. Climate Dyn., 44, 2683–2702, https://doi.org/10.1007/s00382-014-2425-9.
Lin, Y.-J., R. W. Pasken, and H.-W. Chang, 1992: The structure of a subtropical prefrontal convective rainband. Part I: Mesoscale kinematic structure determined from dual-Doppler measurements. Mon. Wea. Rev., 120, 1816–1836, https://doi.org/10.1175/1520-0493(1992)120<1816:TSOASP>2.0.CO;2.
Liu, S. C., C. Fu, C.-J. Shiu, J.-P. Chen, and F. Wu, 2009: Temperature dependence of global precipitation extremes. Geophys. Res. Lett., 36, L17702, https://doi.org/10.1029/2009GL040218.
Lu, C., F. C. Lott, Y. Sun, P. A. Stott, and N. Christidis, 2020: Detectable anthropogenic influence on changes in summer precipitation in China. J. Climate, 33, 5357–5369, https://doi.org/10.1175/JCLI-D-19-0285.1.
Ma, S., T. Zhou, A. Dai, and Z. Han, 2015: Observed changes in the distributions of daily precipitation frequency and amount over China from 1960 to 2013. J. Climate, 28, 6960–6978, https://doi.org/10.1175/JCLI-D-15-0011.1.
Ma, S., and Coauthors, 2017: Detectable anthropogenic shift toward heavy precipitation over eastern China. J. Climate, 30, 1381–1396, https://doi.org/10.1175/JCLI-D-16-0311.1.
Moore, B. J., A. B. White, and D. J. Gottas, 2021: Characteristics of long-duration heavy precipitation events along the West Coast of the United States. Mon. Wea. Rev., 149, 2255–2277, https://doi.org/10.1175/MWR-D-20-0336.1.
Ralph, F. M., and Coauthors, 2016: CalWater field studies designed to quantify the roles of atmospheric rivers and aerosols in modulating U.S. West Coast precipitation in a changing climate. Bull. Amer. Meteor. Soc., 97, 1209–1228, https://doi.org/10.1175/BAMS-D-14-00043.1.
Ralph, F. M., and Coauthors, 2019: A scale to characterize the strength and impacts of atmospheric rivers. Bull. Amer. Meteor. Soc., 100, 269–289, https://doi.org/10.1175/BAMS-D-18-0023.1.
Sever, G., and Y.-L. Lin, 2017: Dynamical and physical processes associated with orographic precipitation in a conditionally unstable uniform flow: Variation in basic wind speed. J. Atmos. Sci., 74, 449–466, https://doi.org/10.1175/JAS-D-16-0077.1.
Sillmann, J., V. V. Kharin, X. Zhang, F. W. Zwiers, and D. Bronaugh, 2013: Climate extremes indices in the CMIP5 multimodel ensemble: Part 1. Model evaluation in the present climate. J. Geophys. Res. Atmos., 118, 1716–1733, https://doi.org/10.1002/jgrd.50203.
Sun, D.-Z., and I. M. Held, 1996: A comparison of modeled and observed relationships between interannual variations of water vapor and temperature. J. Climate, 9, 665–675, https://doi.org/10.1175/1520-0442(1996)009<0665:ACOMAO>2.0.CO;2.
Tai, S.-L., Y.-C. Liou, S.-F. Chang, and J. Sun, 2020: The heavy rainfall mechanism revealed by a terrain-resolving 4DVar data assimilation system—A case study. Mon. Wea. Rev., 148, 2307–2330, https://doi.org/10.1175/MWR-D-19-0244.1.
Tu, C.-C., Y.-L. Chen, P.-L. Lin, and Y. Du, 2019: Characteristics of the marine boundary layer jet over the South China Sea during the early summer rainy season of Taiwan. Mon. Wea. Rev., 147, 457–475, https://doi.org/10.1175/MWR-D-18-0230.1.
Tung, Y.-S., C.-Y. Wang, S.-P. Weng, and C.-D. Yang, 2022: Extreme index trends of daily gridded rainfall dataset (1960–2017) in Taiwan. Terr. Atmos. Oceanic Sci., 33, 8, https://doi.org/10.1007/s44195-022-00009-z.
Wang, B., 2006: The Asian Monsoon. Springer, 845 pp.
Wang, B., and Z. Fan, 1999: Choice of South Asian summer monsoon indices. Bull. Amer. Meteor. Soc., 80, 629–638, https://doi.org/10.1175/1520-0477(1999)080<0629:COSASM>2.0.CO;2.
Wang, B., Z. Wu, J. Li, J. Liu, C.-P. Chang, Y. Ding, and G. Wu, 2008: How to measure the strength of the East Asian summer monsoon. J. Climate, 21, 4449–4463, https://doi.org/10.1175/2008JCLI2183.1.
Wang, B., B. Xiang, and J.-Y. Lee, 2013: Subtropical high predictability establishes a promising way for monsoon and tropical storm predictions. Proc. Natl. Acad. Sci. USA, 110, 2718–2722, https://doi.org/10.1073/pnas.1214626110.
Wang, C.-C., J. C.-S. Hsu, G. T.-J. Chen, and D.-I. Lee, 2014: A study of two propagating heavy-rainfall episodes near Taiwan during SoWMEX/TiMREX IOP-8 in June 2008. Part I: Synoptic evolution, episode propagation, and model control simulation. Mon. Wea. Rev., 142, 2619–2643, https://doi.org/10.1175/MWR-D-13-00331.1.
Wang, C.-C., P.-Y. Chuang, S.-T. Chen, D.-I. Lee, and K. Tsuboki, 2022: Idealized simulations of mei-yu rainfall in Taiwan under uniform southwesterly flow using a cloud-resolving model. Nat. Hazards Earth Syst. Sci., 22, 1795–1817, https://doi.org/10.5194/nhess-22-1795-2022.
Wang, C.-C., T.-Y. Yeh, M.-S. Li, K. Tsuboki, and C.-H. Liu, 2023: A modeling study of an extreme rainfall event along the northern coast of Taiwan on 2 June 2017. Atmos. Chem. Phys., 23, 501–521, https://doi.org/10.5194/acp-23-501-2023.
Wang, S., H. Cheng, and Y. Chao, 1984: Natural seasons of the weather in the Taiwan area (in Chinese with English abstract). Atmos. Sci., 11, 101–120.
Webster, P. J., V. O. Magaña, T. N. Palmer, J. Shukla, R. A. Tomas, M. Yanai, and T. Yasunari, 1998: Monsoons: Processes, predictability, and the prospects for prediction. J. Geophys. Res., 103, 14 451–14 510, https://doi.org/10.1029/97JC02719.
Xu, W., E. J. Zipser, Y.-L. Chen, C. Liu, Y.-C. Liou, W.-C. Lee, and B. J.-D. Jou, 2012: An orography-associated extreme rainfall event during TiMREX: Initiation, storm evolution, and maintenance. Mon. Wea. Rev., 140, 2555–2574, https://doi.org/10.1175/MWR-D-11-00208.1.
Zhou, T., L. Ren, and W. Zhang, 2021: Anthropogenic influence on extreme meiyu rainfall in 2020 and its future risk. Sci. China Earth Sci., 64, 1633–1644, https://doi.org/10.1007/s11430-020-9771-8.