This section further examines the characteristics of TC RAI in terms of translational speed, intensity, minimum sea level pressure, TC radius, and lifetime with respect to similar landfalling TCs in the Philippines to identify potential unique features of TC RAI.
3.2.1. Translational Speed, Intensity, and Direction of Movement
We analyzed TC events that made landfall in the Philippines from 1979–2020. We identified the TCs that made landfall in Luzon, Visayas, and Mindanao to represent the northern, central, and southern areas of the Philippines, respectively (Figure S2). Table 1 shows the count of all TCs at different categories upon landfall in these three areas.
Table 1
Tropical Cyclone Category of Philippine TCs (1979–2020) upon landfall using WMO-IBTrACS.
Category
|
Philippines
|
Luzon
|
Visayas
|
Mindanao
|
TD
|
123
|
49
|
46
|
28
|
TS
|
88
|
53
|
28
|
7
|
C1
|
33
|
27
|
6
|
0
|
C2
|
31
|
23
|
6
|
2
|
C3
|
8
|
5
|
3
|
0
|
As mentioned previously, since the IBTrACS dataset only records MSW > 30 kts, TCs that do not have any MSW entry in the dataset were assumed to be in the TD category and were excluded, resulting in 160 TCs that were used in the analysis. Next, the MSLP, translational speed, and MSW were compared as shown in Figs. 2a and 2b to ascertain any trends. Based on the results, a strong negative correlation between MSLP and MSW was found (r=-0.9561; p < 0.0001), while no linear relationship between the translational speed and MSW could be established (r=-0.02; p = 0.8011). This observation agrees with the findings from Choi et al. (2016), in which MSW and MSLP were found to have an inverse relationship, while no particular relationship between translational speed and MSW (Takagi and Esteban, 2012) could be confirmed. Upon closer analysis of the TCs, TC RAI had the second highest MSW, tied with BETTY (1987) and HAIMA (2016), and the 3rd lowest MSLP from the landfalling TCs, after HAIYAN (2013) and BETTY (1987) (see color codes in Fig. 2). One key finding is that among the TCs in the STY category upon landfall, TC RAI had a speed of 16.7 kts, which is the second fastest landfalling TC next to HAIYAN (2013).
Figures 2c and 2d further illustrate the landfalling TCs in terms of days until landfall and TC lifetime. In this study, days until landfall is defined as the duration of TC activity from genesis to landfall in the Philippines, while lifetime is defined as the duration of TC activity from genesis to dissipation. Based on Fig. 2c, TC RAI had 4.75 days from genesis until its landfall in the Philippines, which is comparable to HAIYAN (2013), BETTY (1987), and HAIMA (2016). Also, based on Fig. 2c, TCs make landfall in the Philippines about 4.3 ± 3.2 days from genesis, and that TC RAI is within the 56th percentile among the TCs considered. These findings are comparable with locations such as Japan, where Nayak and Takemi (2023) observed that TCs from 2006–2019 take about 5 to 8 days to make landfall in Japan. In terms of TC lifetime from genesis to dissipation, TC RAI had about 10 days, which is comparable to HAIYAN (2013) and BETTY (1987). TCs that make landfall in the Philippines are active in about 8.6 ± 3.5 days, with TC RAI within the 71st percentile. On average, TCs make landfall in less than 5 days with a lifetime of less than 8 days during the 1979–2020 period (Figure S3). So far, there are no clear trends, yet there are peaks in the time to reach land and TC lifetime in 1997 and 2018.
While TC radius is usually not included in other climatological studies, the estimated radius of gale-force wind (in km), R30, of all landfalling TCs in the Philippines are examined in this study. R30 is often used to determine TC impact and assess disaster risk (Kim et al., 2022). Figures 2e and 2f show the long and short TC radius of the 30 kt wind, respectively, during landfall wherein TC RAI has a 240 km long and 150 km short radius for its 30 kt winds. In terms of other TCs, the average landfalling TC radii are about 188 ± 78 km for long and 157 ± 68 km for short radii. The long (short) TC radius for TC RAI is within the 83rd (64th) percentile. Based on the aforementioned findings, TC RAI’s lifetime, days until landfall, and its radius fall within the usual characteristics of a typical landfalling TC in the Philippines. When comparing TC size, it is worth noting that HAIYAN had a 250 (180) km long (short) 30 kt radius, while the largest TC in terms of radius was ZEB (1999) having both 450 km long and short 30 kt radii.
Further examining the intensity and temporal distribution of the landfalling TCs, Fig. 3 shows the frequency distributions of the 160 TCs per year from 1979–2020 and intensity distributions according to MSW at 5 kt intervals. Although there were 17 and 14 TC landfalls in 1993 and 1995, respectively, there are no significant trends in the peaks of annual TC count from 1979–2020 (r = 0.017; p = 0.92) (Fig. 3a). This agrees with Takagi and Esteban (2016), who found no significant trend in most areas of the Philippine archipelago except along 10°–12°N. Moreover, for Fig. 3b, using Kruskal-Wallis (Hollnader and Wolfe, 1973) and the Anderson-Darling test (Thode, 2002), the data is not normally distributed. Albeit, based on Fig. 3b, it will be noted that TC RAI is near the rightmost tail of the distribution, which is around the 96th percentile.
To identify the unique features of TC RAI, we examine December TCs having similar track, area of landfall, and direction of movement, using the IBTrACS data set and found seven TCs: NORRIS (1986), MARGE (1986), NELL (1993), AXEL (1994), KAJIKI (2001), WUKONG (2012), and JANGMI (2015). Figure 4 shows the track of these TCs with their corresponding genesis and dissipation points. The tracks of these TCs differ in their genesis points, yet resemble a similar track west of 130°E. It can also be seen that TC RAI formed closer to the Philippines and at a lower latitude at 5.3°N.
In terms of intensity, TC RAI's MSW is comparable to the composite TCs at LF-24 and LF-18 as shown in Fig. 5a. TC RAI further intensified, peaking to 105 kts (194.4 kph) at LF, and decreased its intensity before plateauing at 80 kts at LF + 18 onwards, maintaining its TY category. These observations coincide with TC RAI’s decrease in MSLP starting at 950 hPa at LF-12, and reaching the minimum of 915 hPa at LF. In contrast, the intensity of the composite TCs consistently decreases from LF-24 onwards. In terms of MSLP, TC RAI had a consistently lower MSLP with respect to other TCs across LF-24 to LF + 24 (Fig. 5b). This indicates TC RAI’s unique feature, which shows how it continued to intensify until landfall. The decrease in intensity is expected as a result of the TC encountering topography, and increased friction and loss of latent heat (Wu and Choy, 2015). Unique to the central Philippines is the absence of a mountain range (Fig. 1c) that would heavily affect the intensity and rainfall distribution of the TC. Brand and Belloch (1973) have noted that landfalling TCs in central Philippines are more prone to intensification due to the archipelagic nature of this region. The island-and-sea configuration has lesser effect on TCs due to the smaller surface areas of the land (Brand and Belloch, 1973). Petilla et al. (2023) noted that about 6 (4) TCs passing through central Philippines had increased (maintained) in intensity upon approach and departure. Hence, one probable cause to the sustenance of TC RAI above the TY category may be attributed to the island-and-sea mix in this region (Brand and Belloch, 1973). Additionally, the area of rich moisture along the path of the TC may have allowed it to produce heavier precipitation and more energy, as will be discussed later.
The speed of TC RAI is comparable to that of the composite TCs until LF at which point TC RAI's speed continued to increase until peaking at LF + 6 at around 13.1 kts (Fig. 5c). This behavior has been observed by Petilla et al. (2023) for TCs crossing Mindanao Island and central Visayas. The speed remained consistently greater than 10 kts until LF + 24. At LF + 24, the speed of TC RAI is comparable to the mean of the 7 TCs. More specifically, both the speed of the composite TCs and TC RAI increased starting at LF and peaking on LF + 6. This is in agreement with the observations from Chang (1982) and Bender et al. (1987), where both hypothesized an increase of speed when approaching land, which may be caused by the interaction between the TC and topography.
In terms of direction of movement, TC RAI's north-westward direction is comparable to the composite TCs until LF to LF + 18, where the TC moved westward upon landfall. The significant changes in direction of the TC track may be explained by the minimal amount of mountain ranges along the path of the TC. Figure 1c shows that the island-and-sea mix along the path of the TC is a key feature. The absence of mountain ranges may have influenced the TC’s direction as noted by Maw and Min (2017) who experimented the impact of topography on ROANU (2016), which made landfall in Myanmar. They noted that changes in the altitude of the Rakhine Mountain have an influence on the track of the TC such that the absence (or presence) of the mountain shifted the track away from (into) Myanmar. Additionally, the straight and westward track upon approach also agrees with the hypothesis proposed by Corporal-Lodangco et al. (2016) that TCs formed during a La Nina season often follow a straight-moving track (more details discussed on section 3.2.2). Additionally, Wu and Choy (2015) have suggested that topography is more effective in deflecting TCs with weaker intensity. As such, the westward track combined with fast speed allowed TC RAI to traverse central Visayas and make 8 landfalls within 24 hours as observed by NDRRMC (2022).
The TC 30 kt radius (long) was also examined to understand the evolution and scope of the TC event. As shown in Fig. 5e, TC RAI had a consistent radius of 240 km from LF-24 to LF + 24, and yet the composite TCs show a consistent decrease in size during the same period. Specifically, the radius of TC RAI was still comparable to that of the composite TCs before LF + 6, at which point the composite TC size decreased. The large TC radius contributed to the wider scope of damage caused by TC RAI.
To further quantify the TC’s intensification/dissipation and energy, Table 2 shows the ACE and the LFDR of the composite TCs, including HAIYAN (2013). Based this, TC RAI had the highest ACE when compared with the composite TCs. TCs such as NELL (1993), WUKONG (2012), and JANGMI (2014) had a negative LFDR, indicating that these TCs did intensify after landfall, while NORRIS (1986), MARGE (1986), and AXEL (1994) had a positive LFDR. A unique finding from the composite TCs is that KAJIKI (2001) had zero LFDR, indicating that this particular TC maintained its intensity before and after landfall. However, what sets TC RAI apart from other TCs is that although it had a higher intensity, its LFDR was only about 0.02. This is in contradiction to the findings from Kaplan and Demaria (1995) wherein they observed that TC's rate of wind speed decay is proportional to wind speed. Yet, this TC was able to maintain its intensity albeit its STY category as exhibited to its near zero LFDR, which indicates that its dissipation rate was very low relative to other composite TCs.
Table 2. Monthly Ocean Heat Content (OHC) Anomaly, Accumulated Cyclone Energy (ACE), and Landfalling Dissipation Rate (LFDR) of the 7 Composite TCs and TC RAI.
Name
|
Yeara
|
Monthly OHC Anomalyb (x±s.d.) kJ cm -2
|
Accumulated Cyclone Energy (ACE)
|
Landfalling Dissipation Rate (LFDR)
|
NORRIS
|
1986
|
-101±12
|
11.08
|
0.46
|
MARGE
|
1986
|
-101±12
|
12.78
|
0.54
|
NELL
|
1993
|
-98±19
|
3.89
|
-0.15
|
AXEL
|
1994
|
-113±6
|
9.89
|
0.42
|
KAJIKI
|
2001
|
79±8
|
1.35
|
0
|
WUKONG
|
2012
|
201±14
|
1.71
|
-0.08
|
JANGMI
|
2014
|
-15±19
|
1.02
|
-0.77
|
HAIYANc
|
2013
|
233±19
|
20.69
|
0.36
|
RAI
|
2021
|
123±14
|
18.94
|
0.02
|
aBlue (red) years indicate La Nina (El Nino) years
bOHC anomaly computed by solving the mean inside the white box (7.5-10.5°N, and 127.5-135.5°E) of Figure 1
c Daily OHC anomaly computed to be 115-135 kJ cm-2 (Lin et al., 2021)
|
Based on these findings, there are five features unique to TC RAI when compared to landfalling TCs in the Islands of Siargao-Dinagat Islands. Upon landfall, TC RAI is the TC case having: a) MSW of 105 kts and an MSLP of 915 hPa and is the strongest TC to make landfall in Siargao-Dinagat Island; b) a high ACE and a near-zero LFDR indicating a low dissipation rate of the TC during landfall, c) the second fastest translational speed upon landfall (next to NORRIS,1986); d) having a westward track (~ 270°) from LF to LF + 24; and e) having the second largest TC 30 kt radius (next to MARGE, 1986) at 240 km.
3.2.2. Analysis of Environmental Factors
This section covers the environmental factors that contributed to the development, intensification, and maintenance of TC RAI. The ERA5 dataset was used for the analysis of the moisture and vertical wind shear while the GHRSST dataset for SST. This section focuses on the time steps between LF-24 to LF due to the consistent intensification of TC RAI.
SST plays a vital role in the development and intensification of TCs (Kuroda et al., 1998). Although other factors such as weak vertical wind shear, increase in moisture flux, high relative humidity, etc. contribute to TC intensification, TCs moving over waters warmer than 28.5°C (Kuroda et al., 1998) for one to two days often acquire higher intensities eventually. Figure 1a shows that TC RAI passed over waters warmer than 29°C for at least two days before landfall. Another impact in terms of SSTa is that TC RAI developed and intensified during a La Nina season at which temperatures in the ENSO 3.4 region during OND, NDJ, DJF were about − 1.1, -0.9, and − 0.8°C, respectively (Diamond and Schreck, 2022; Null and CCM, 2024). At the time of TC RAI, the SSTa over the Philippine Sea was about + 0.5 to 1.5°C. According to Corporal-Lodangco et al. (2016), during La Nina conditions, TCs in the fourth quarter (OND) have approximately straight-moving tracks. Also, the genesis of these TCs is concentrated to lower latitudes and are formed closer to the Philippines, about west of 160°E (Corporal-Lodangco et al., 2016). This hypothesis is demonstrated in the tracks of TC RAI and the composite TCs shown in Fig. 1b and Fig. 4, respectively.
Table 2 shows the monthly ocean heat content (OHC), which was averaged over the white box in Fig. 1, of the composite TCs, TC RAI, and HAIYAN (2013). The white box covers the area where TC RAI intensified from TS to STY TC within 36 hours Table 2 shows a positive OHC anomaly during TC RAI in this area of about 123 ± 14 kJ cm− 2. It is also worth noting that HAIYAN (2013) had the highest monthly OHC anomaly at 233 ± 19 kJ cm− 2, while AXEL (1994) had the lowest OHC anomaly at -113 ± 6kJ cm− 2. Compared to other STYs, the daily OHC anomaly for HAIYAN (2013) and HAGIBIS (2019) were estimated to be 115–135 kJ cm− 2 and 140–160 kJ cm− 2, respectively (Lin et al., 2021). TC RAI had SST and OHC conditions that favored its intensification to STY in the WNP (Lin et al., 2021).
Vertical wind shear is a major contributor to TC generation, weakening, and maintenance of vertical structure (Wong and Chan, 2004). From LF-24 until LF-6, the TC moved through areas of high vertical wind shear (10–20 m s− 1) within the 1° radius of the TC as shown in Fig. 6. Zehr (1992) and Flatau et al. (1994) noted that TCs can usually withstand a vertical wind shear of 9 m s− 1–12.5 m s− 1. Above this threshold, TC development is already impeded. This suggests that TC RAI was able to intensity despite moving in an area of high vertical wind shear.
Figure 7 shows the vertically integrated water vapor (VIWV) from 1000 to 700 hPa (Zomeren and Delden, 2007) during TC RAI’s approach in the Philippine Sea. It can be seen that the TC passed through an area of rich moisture content with about 60 to 80 kgm− 2 of water vapor from LF-24 to LF-6. High environmental moisture, specifically in the rear quadrants of the TC, provides a favorable condition for TC intensification (Wu et al., 2015) as shown in Fig. 7d.
High specific humidity is observed along the eastern Philippine coast as the TC approaches at LF-24 to LF + 12 (Fig. 8). This enabled ample intensification time for the TC through moisture transport, as described earlier. Moreover, moisture from the northeast of the Philippines was also drawn towards the central Philippines as depicted at LF-24 and LF-12. This allowed the TC to consistently gain moisture as it approached the Philippines, which resulted in greater precipitation (Chien and Kuo, 2011) and higher energy. In terms of moisture flux, a positive and strong convergence can be seen along the TC center (Fig. 9). A consistent moisture supply from the central Philippine region may also explain why the TC maintained its intensity during and after its initial landfall.