1. Introduction
Tropical cloud clusters (TCCs) develop into tropical cyclones (TCs) in environments with high sea surface temperature, high low-level vorticity, and high midlevel moisture (Gray 1968, 1998), which is referred to as the TC formation process. In the western North Pacific (WNP) TCs can form in the monsoon trough (Wu et al. 2012b; Molinari and Vollaro 2013; Zong and Wu 2015), monsoon confluence (Holland 1995; Kuo et al. 2001), monsoon gyre (Lander 1994; Wu et al. 2013), easterly wave (Shapiro 1977; Dunkerton et al. 2009), and monsoon and easterly interaction environments (Chen et al. 2008). TCs that form in different environmental patterns have different size, intensity, and landfall characteristics (Fudeyasu and Yoshida 2018; Teng et al. 2021). Although characteristics of TC formation under each environment type have been studied, the relative importance of each environmental factor between monsoon and easterly environments is not well understood.
Because environmental conditions substantially control TC formation, they are used to forecast and evaluate how favorable environments are for TC formation. For example, Gray (1968, 1984a, b) used environmental parameters including the characteristics of El Niño–Southern Oscillation (ENSO) and the quasi-biennial oscillation to forecast seasonal TC formation. Following this concept, Emanuel and Nolan (2004) defined a TC genesis potential index (GPI) using a statistical method to determine the overall environmental conditions favorable for TC formation. The GPI consists of four terms: low-level vorticity, midlevel relative humidity, vertical wind shear, and maximum potential intensity (Bister and Emanuel 2002). Camargo et al. (2007) used GPI to investigate the critical factors for TC formation under the phases of ENSO. They attributed changes in the locations and numbers of TC formation in the WNP to changes in the large-scale patterns of relative humidity and vorticity associated with ENSO. Bruyère et al. (2012) applied the GPI to assess the interannual variability of TC formation in climate models, and found TC variability in the North Atlantic to be more correlated with the maximum potential intensity and vertical wind shear than the other terms. Both Fu et al. (2012) and Kerns and Chen (2013) suggested that the TC formation probability is more sensitive to the environmental dynamic conditions (e.g., low-level vorticity) for the WNP, while the thermodynamic conditions (e.g., water vapor content) appear to be more important for the North Atlantic. This is because TCs forming in the WNP are mostly associated with the monsoon system but associated with easterly waves in the North Atlantic (Teng et al. 2019). However, given that TCs can also form under easterly conditions in the WNP, it is important to understand the contribution of each environmental parameter to TC formation separately for easterly- and monsoon-type TCs. To date, only the sensitivity of TC formation to the low-level vorticity between the easterly- and monsoon-type TCs has been explored (Hsieh et al. 2017). To the authors’ knowledge, no previous study has analyzed the importance of moisture for TC formation within the monsoon and easterly environments in the WNP.
Higher environmental moisture favors TCs with a larger size, larger rainfall radius, and higher intensification rate (Hill and Lackmann 2009; Wu et al. 2012a; Lin et al. 2015). The TC formation forecast is also highly sensitive to the moisture pattern and accuracy in the initial condition (Sippel and Zhang 2008; Zhang and Sippel 2009; Doyle et al. 2012). The moisture underestimation has been shown to substantially reduce the predictability of TC formation in global models (Kuo et al. 2018, 2020) and regional models (Chen et al. 2020). Data assimilation (DA) of local and satellite observations have been used to improve the initial moisture fields of TC simulations and increase the forecast accuracy of TC track, intensity, and formation (Greeshma et al. 2015; Choi et al. 2017; Minamide and Zhang 2018; Lim et al. 2019). Because TCs form and develop over oceans, satellite data are the main data source for the DA. Among different satellite data, the global positioning system (GPS) radio occultation (RO) provides high vertical-resolution moisture, temperature, and pressure data with less contamination by clouds and precipitation (Anthes et al. 2000, 2008; Kuo et al. 2000; Kursinski et al. 2000). The assimilation of GPS RO data mainly affects the thermodynamic terms in the initial condition and improves the simulation performance of TC track, rainfall, and intensity (Huang et al. 2005, 2010; Chen et al. 2009; Kueh et al. 2009; Chen et al. 2015). Liu et al. (2012) found that the assimilation of GPS RO data enhanced the moisture and cyclonic circulation around the pre-TC disturbance of Ernesto (2006) in the lower troposphere, and improved TC formation prediction. Chen et al. (2020) further quantified the potential contributions of GPS RO assimilation to the TC formation forecast for ten TCs. They found that the assimilation of GPS RO data improved the midlevel moisture in the initial condition, and increased the detection rate of TC formation in 2-day lead-time simulations from 30% to 70%. Building on this work, this study assesses the importance of midlevel moisture for the simulation performance of TC formation in different large-scale environments based on the results of GPS RO DA.
High moisture at midlevels around the pre-TC disturbance center supplies an unstable condition for deep convection development and maintenance at the initial stage of TC formation. The associated diabatic heating released at mid- and low-levels supports strengthening of convergence, midlevel vortex genesis, disturbance intensification, and TC formation (Emanuel 1986; Gray 1998; Hendricks et al. 2004; Houze 2010). However, the relative importance of midlevel moisture for TC formation may vary between monsoon and easterly environments. This is the hypothesis explored in this study. Compared to vorticity, the moisture impact on TC formation is rarely compared systematically based on multiple cases. Although midlevel moisture is a critical term in the GPI formulation, the sensitivity of TC formation in monsoon and easterly environments to synoptic-scale moisture patterns and its mechanism are not clear. Therefore, the objectives of this study are to (i) characterize the climatological moisture patterns of TC formation in monsoon and easterly environments, (ii) construct a database of TC formation simulation with different moisture patterns based on the assimilation of GPS RO data, and (iii) assess the importance of synoptic-scale moisture for the simulation performance (15-km scale) of TC formation under different environments in a regional model. These objectives will contribute to the understanding of the physical processes captured in a 15-km model associated with moisture and vorticity evolution and how the assimilation of GPS RO data improves the TC formation forecast skill.
The data used in this study, definitions of monsoon and easterly environments, and regional model settings are described in section 2. The climatological comparisons of environmental parameters between the monsoon- and easterly-type TCs are shown in section 3. The statistical characteristics of TC formation simulation under different environments are also discussed in this section. The importance of midlevel moisture for TC formation is addressed through additional sensitivity experiments in section 4. A discussion and conclusions are presented in section 5.
2. Data and methodology
a. Data
In this study, climatological characteristics of TC formation in monsoon and easterly environments are compared during July–October, 1981–2009, and the simulation characteristics of TC formation are analyzed during July–October, 2006–10. The 6-hourly TC data are obtained from the Joint Typhoon Warning Center (JTWC) best track dataset (Guard et al. 1992; Chu et al. 2002). The earliest time at which a tropical disturbance with maximum sustained surface wind speed equal to or exceeding 25 kt (1 kt ≈ 0.51 m s−1) is defined as the time of TC formation. TCCs are the pre-TC disturbances and have a specific convective structure and low-level cyclonic circulation (Hennon et al. 2011; Kerns and Chen 2013; Teng et al. 2014). TCCs that can develop into TCs are categorized as developing TCCs, and the others are nondeveloping TCCs. TCC track data used in this study are obtained from Teng et al. (2014). For the climatological analyses and environmental classifications (shown in section 3), global gridded atmospheric parameters are derived from the Climate Forecast System Reanalysis V1 (CFSR; Saha et al. 2010) of National Centers for Environmental Prediction (NCEP), which is a global, 0.5° × 0.5°, 37-level, and 6-hourly dataset from 1979 to 2011.
To be consistent with operational forecasts, the initial conditions for simulations and sensitivity experiments use analysis data (shown in section 4), rather than reanalysis data. They are obtained from the NCEP Final (FNL) Operational Global Analysis Data (NCEP 2000), which is a global, 1° × 1°, 26-level, and 6-hourly dataset from 1999 to the present. NCEP FNL is also regarded as the reference to compare with different simulations in section 4. The observations used in the Weather Research and Forecasting (WRF) Model Data Assimilation System are obtained from NCEP (2015), including the conventional, GPS RO, and satellite radiance data. However, the satellite radiance is not assimilated in this study. The GPS RO data include the Constellation Observing System for Meteorology, Ionosphere and Climate (COSMIC), the Challenging Minisatellite Payload (CHAMP), Gravity Recovery and Climate Experiment (GRACE), Meteorological Operational Satellites-A (MetOp-A), Satellite for Scientific Applications-C (SAC-C), and Terra Synthetic Aperature Radar at X-Band (TerraSAR-X) satellite. The horizontal distribution of GPS RO data is approximately 2000–2700 soundings globally per day, and the vertical resolution is approximately 200 m for each sounding covering from 40 km to the surface. The simulated TCs are selected from the COSMIC peak period (2006–09), corresponding to high GPS RO data volume.
b. Environmental patterns and TC selection
This study uses the environmental classification procedure developed by Teng et al. (2020) to define the TC formations associated with easterly and monsoon environments. They used the k-means cluster analysis (MacQueen 1967), and separated TCs that form in different large-scale circulation patterns based on 850-hPa wind fields. Teng et al. (2020) classified the formation processes of 531 TCs occurring in July–October, 1981–2009 into 40 cluster types. Five main environmental cluster types (highest percentages) for TC formation in the WNP were defined: Easterly flow west and southwest of the subtropical high, monsoon trough, monsoon confluence, and north of monsoon trough. In these five cluster types, each TCC that developed into a TC (i.e., the entire formation process) was required to develop in one specific environment without crossing into a different environment. In total, 82 TCs that developed in easterly flow west and southwest of subtropical high were categorized into easterly-type TCs (ETCs), while 131 TCs were categorized into monsoon-type TCs (MTCs).
In this study, the environmental characteristics for the 82 ETCs and 131 MTCs are defined as the climatological characteristics for ETCs and MTCs, respectively. Out of these 213 TCs, 33 TCs occur during 2006–09 (the peak period of GPS RO data), and are selected for further simulation. Because the ratio of ETC number to the total number (33%) is lower than the climatological ratio (38%), two more ETCs in 2010 are considered in this study, increasing the ETC ratio to 37%. A total of 13 ETCs and 22 MTCs are analyzed and simulated in this study (Table 1), and their simulation performance are defined as model characteristics for ETCs and MTCs. Figure 1 shows the circulation and relative humidity composites for the 35 TCs during their formation process under Earth-relative and comoving frames. The circulation around the disturbance center removed the system phase speed (averaged over the TC formation process) is defined as the comoving streamlines. The monsoon environments provide TCCs and TCs with stronger large-scale cyclonic circulation and more widely dispersed moisture than the easterly environments. This agrees with the characteristics of large-scale monsoon trough found by Wang et al. (2001). The pattern of comoving streamlines for ETCs is similar to the pouch structure of easterly wave TCs in the North Atlantic (Dunkerton et al. 2009), which may prevent dry air intrusion and protect disturbance development.
List of TCs selected in this study and simulation performance of TC formation. The circles (O) indicate the TC formation is detected in the simulation, while the crosses (X) are not. The last column shows the type of simulation performance (defined in section 3b).
Composites of 850-hPa streamlines and 700-hPa relative humidity (colored shading; %) for (a),(b),(e),(f) 13 ETCs and (c),(d),(g),(h) 22 MTCs at the time of (top) TCC formation and (bottom) TC formation. The streamlines under the Earth-relative frame are shown in (a), (c), (e), and (g), whereas (b), (d), (f), and (h) show the streamlines under the comoving frame. The phase speed of ETCs removed is 5.2 m s−1, while 2.6 m s−1 is removed for MTCs. The origin in each panel represents the location of TCC or TC formation.
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
c. Model setting and numerical experiments
The Advanced Research version of the WRF Model (version 4.0.3; Skamarock et al. 2019) and the WRF three-dimensional variational (3DVAR) DA method (Barker et al. 2004, 2012) are used in this study. For all simulations, a single domain with 15-km horizontal grid spacing, 610 × 406 grid points, and 41 vertical levels is used (Fig. 2a). The initial and lateral boundary conditions are derived from the NCEP FNL. Model physics includes the WRF double-moment 6-class microphysics scheme (Lim and Hong 2010), the Rapid Radiative Transfer Model for general circulation models (Iacono et al. 2008), the unified Noah land surface model (Chen and Dudhia 2001), the Monin–Obukhov surface-layer scheme (Monin and Obukhov 1954; Beljaars and Holtslag 1991), the Yonsei University boundary layer scheme (Hong et al. 2006), and the multiscale Kain–Fritsch cumulus scheme (Zheng et al. 2016). Although simulations with 12–15-km grid spacings cannot capture some mesoscale and convective-scale processes (Kukkonen et al. 2012), the synoptic and large-scale characteristics of TC formation should be resolved (Li and Pu 2014; Hsieh et al. 2017; Chen et al. 2020).
(a) Model domain setting and distribution (red dots) of GPS RO soundings used in 3-day cycling DA for Typhoon Kajiki. (b) As in (a), but zooms in on Typhoon Kajiki. The top number shows the total number of GPS RO soundings in each panel. Blue circles indicate the 6-hourly TCC locations during the DA process, and blue dots indicate the 6-hourly TCC locations during the free simulation. Open TC symbols indicate the 6-hourly locations of TCs with 34-kt (or higher) wind speed based on the JTWC best track, while closed TC symbols for TCs with 65-kt (or higher) wind speed.
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
Figure 3 shows a schematic diagram for the experiment design in this study. The observed TC formation time is regarded as the time reference (t = 0 h). A 3-day cycling DA procedure is performed from 5 days before (t = −120 h) to 2 days before (t = −48 h) the observed TC formation time, in which four cycling assimilation cycles are conducted per day with a 6-h window. A free simulation is conducted after the DA procedure (t = −48 h), and this time is defined as the initial time of free simulations.
Schematic diagram for the experiment procedure.
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
Generally, the operational global and regional analyses have a dry bias at 48–120-h lead times. This dry bias may suppress the development of disturbances in forecast models or delay the TC formation process in model simulations (Chen et al. 2020; Kuo et al. 2020). The forecast that fails to capture TC occurrence or captures a TC with large spatial and temporal error decreases the accuracy of TC formation forecasts. Similar dry biases are found in this study (shown in section 4a). To assess the relative importance of moisture for TC formation in different environments, this study compares the TC formation simulations of which the dry bias is reduced by GPS RO DA to the simulations associated with the original dry bias. For each TC, two experiment simulations are performed: one with the additional assimilation of conventional observation data (i.e., Global Telecommunication System, labeled as GTS) and the other with the additional assimilation of GPS RO data in addition to conventional observations. Although FNL may consider conventional observations and GPS RO data, the additional 3-day cycling DA performed in this study can further improve moisture and other parameters at the initial time of free simulations (t = −48 h; Fig. 3) than those without additional DA. Previous studies (Chen et al. 2009; Huang et al. 2016) have shown that the model performance of TCs and heavy rainfall prediction is improved with the use of a nonlocal RO operator (i.e., the excess phase; Sokolovskiy et al. 2005; Schreiner et al. 2010) than a local RO operator. The nonlocal RO operator better adjusts the temperature and moisture fields throughout the troposphere in the simulation (Chen et al. 2020). Therefore, the excess phase operator is used for the simulation with the assimilation of GPS RO data (labeled as EPH; Table 2). In this study, the additional 3-day cycling DA of GPS RO and its nonlinear effects improve the initial conditions of free simulations and impact model performance. The importance of the main parameter (moisture) improved by GPS RO DA for TC formation simulation is explored based on the difference in model performance between GTS and EPH.
Description for each label.
Due to variation in the daily number of soundings, the number of assimilated GPS RO soundings in the 3-day cycling DA process differs for each TC. The assimilation window of each TC includes approximately 400–700 GPS RO soundings in the domain. For instance, a total of 698 GPS RO soundings are assimilated in the simulation of typhoon Kajiki (2007; Fig. 2). A more detailed description of the assimilation of GPS RO data is given by Chen et al. (2009, 2020).
Sensitivity experiments are performed for selected TCs to examine the contribution of an individual parameter improved by EPH to TC formation simulation (section 4b). Because the relative importance of midlevel moisture for TC formation under different environments may differ (shown later in sections 3 and 4), the moisture is the parameter tested in the sensitivity experiment. After the 3-day cycling DA (i.e., the initial time of free simulation), the moisture fields within a 20° radius of the pre-TC disturbance center at all levels in GTS are replaced by those in EPH. This experiment is defined as GTS-E. In turn, the EPH with replaced GTS moisture fields is defined as EPH-G (Table 2). In each sensitivity experiment, only the moisture parameters (water vapor mixing ratio, cloud water mixing ratio, rainwater mixing ratio, snow mixing ratio, ice mixing ratio, and graupel mixing ratio) are exchanged, rather than other dynamic or thermodynamic parameters. Similar procedures have been used in previous studies to explore the moisture contribution to TC structure (Hill and Lackmann 2009; Chen et al. 2012).
d. Analysis of diabatic potential vorticity term
3. Environmental characteristics and simulation skill of TC formation
Environmental characteristics of TC formation, such as environmental vorticity, angular momentum flux, and thermodynamic conditions, affect the probability, simulation skill, and efficiency of TC formation (Ooyama 1982; Hsieh et al. 2017; Teng et al. 2019, 2020). In section 3a, the environmental characteristics of ETC and MTC formations are compared first based on reanalysis data (NCEP CFSR). Then, statistical relationships between TC formation environments and simulation skill are analyzed in section 3b, and the potential sensitivities of ETC and MTC formations to midlevel moisture and vorticity are discussed.
a. Environmental characteristics in climatology
The relative importance of environmental parameters to TC formation differ between monsoon and easterly circulations. Four parameters constituting the GPI are compared between the 82 ETCs and 131 MTCs at the time of TCC formation (i.e., the initial time of TC formation process). Figure 4
Climatological difference in azimuthal-averaged (a) relative humidity (colored shading; %) and (b) relative vorticity (colored shading; 10−6 s−1) between MTCs and ETCs at the time of TCC formation. Black dots indicate the differences that pass the t test at a 95% confidence level. The origin in each panel represents the location of TCC formation.
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
Climatological difference in composite-averaged (a) maximum potential intensity (colored shading; m s−1) and (b) 850–200-hPa vertical wind shear (colored shading; m s−1) between MTCs and ETCs at the time of TCC formation. Black dots indicate the differences that pass the t test at a 95% confidence level. The origin in each panel represents the location of TCC formation.
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
A TCC that develops into a TC must have some essential environmental conditions, e.g., sufficient magnitudes of low-level vorticity, midlevel humidity, and sea surface temperature (Gray 1968, 1998; Emanuel and Nolan 2004; Fu et al. 2012; Kerns and Chen 2013). For developing TCCs, the background environments already offer higher moisture for MTC formation and higher vorticity and lower wind shear for ETC formation. Once a developing TCC is associated with stronger vorticity and weaker wind shear in monsoon environments or associated with higher relative humidity in easterly environments, the probability and efficiency of TC formation may increase substantially. In summary, the moisture term is the key factor for TC formation in easterly environments, while the vorticity term is critical in monsoon environments.
b. Simulation skill of ETC and MTC formation
To compare model simulation skill for ETC and MTC formation, the objective TC criteria and TC detection rate are defined. Before TC formation, the maximum 850-hPa vorticity center of the disturbance is recorded as the track of the pre-TC disturbance in simulations. The definition of TC formation in WRF simulations is based on Hsieh et al. (2017), Peng et al. (2017), and Chen et al. (2020). A TC must have a closed circulation center at 850 hPa, 700-hPa vorticity greater than 10−4 s−1, 700-hPa temperature higher than 274 K, 10-m wind speed higher than 25 kt, and persist more than 12 h before landfall. ETCs must form in easterly environments and MTCs must form in monsoon environments (defined by Teng et al. 2020). If the simulated TC forms in a different environment than the observation, it is regarded as a different TC. The earliest time a disturbance meets these criteria is defined as the time of TC formation, and the 850-hPa circulation center is defined as the location of TC formation. Cases where TC formation is detected within the period 24 h before to 24 h after the observed TC formation time and within a 1200-km radius of the observed TC formation location are regarded as successful TC formation simulations (Fig. 3), and the TC is defined as a detected TC. Otherwise, the TC is defined as a nondetected TC. The 1200-km radius is the value of two standard deviations of JTWC 120-h forecast track error (Peng et al. 2017), which is used to search TC formation in 120-h simulations if the simulated and observed background synoptic environments have different motion directions (Hsieh et al. 2017). The 48-h targeted window and other criteria are taken from Chen et al. (2020). The TC detection rate is defined as the number of simulations in which the TC formation is detected successfully divided by the total number of simulations. A higher TC detection rate indicates the model has a higher simulation skill for TC formation. Although the TC detection rate may slightly change if the targeted window, TC criteria, and model physical scheme change, the qualitative characteristics of TC detection rate under different formation environment types are not affected (Hsieh et al. 2017). For nondetected TCs, the pre-TC disturbance track is also recorded to analyze the environmental differences from the detected TCs. The time at which the disturbance reaches the maximum intensity is set as the time reference for the formation process composites of nondetected TCs.
Table 1 shows the statistics of simulation performance for the 35 TC formations in GTS and EPH. Thirteen TC formations are detected in both GTS and EPH (defined as type-1 TCs), 11 TC formations cannot be detected either in GTS or EPH (type-2 TCs), and 11 TC formations are only detected in EPH (type-3 TCs). Ninety-two percent of type-1 TCs belong to MTCs, while 92% of ETCs belong to type-2 and -3 TCs. Table 3 shows the detection rate for TCs that form in each environment in GTS and EPH. TCs that form in easterly flow southwest of the subtropical high have the largest difference in TC detection rate between GTS and EPH. In turn, TCs that form in the monsoon trough have the highest detection rate in GTS and their detection rate is not sensitive to the assimilation of GPS RO data. All TCs that form in the monsoon trough are type-1 TCs, except typhoon Goni (2009). Because the simulated pre-TC disturbance of Goni touches the topography of the Philippines earlier than the observation, its structure is destroyed and the TC formation cannot be detected in either GTS or EPH.
Percentage (%) of TC number in each type and detection rate (%) of TC formation for GTS and EPH. Asterisks (*) indicate the difference in annual detection rate between GTS and EPH passes the t test at a 95% confidence level.
The TC detection rate in EPH is significantly higher than that in GTS. This indicates that the assimilation of GPS RO data improves some key environmental patterns, leading to an increase in the TC detection rate from 37% to 69%. However, the contribution of GPS RO data mainly affects ETCs, rather than MTCs. With the assimilation of GPS RO, the TC detection rate is increased by 54% in ETC formations, but only 18% for MTC formations. Although MTCs also have a higher detection rate in EPH than in GTS, the difference is not statistically significant (Table 3).
In summary, the WRF Model has a higher simulation skill of TC formation for MTCs than ETCs. This finding agrees with Hsieh et al. (2017). Because TC formations in monsoon and easterly environments have different background environmental conditions (described in section 3a), the difference in simulation skill of TC formation among types 1–3 is likely affected by the difference in their initial environmental conditions. Specifically, based on statistics using reanalysis data (shown in the appendix), most type-1 TCs develop in monsoon environments and have more favorable environmental conditions for TC formation than other types. This is particularly true for the monsoon trough. Conversely, most ETCs belong to types 2 and 3. Type-3 TCs have higher moisture but weaker vorticity at midlevels than type-2 TCs during their formation process. The favorable environmental condition for type-3 TC formation mainly depends on midlevel moisture. Therefore, type-3 TCs can be assumed as moisture-sensitive TCs, while type-2 TCs are vorticity-sensitive TCs. The contribution of midlevel moisture to TC formation process for each type TC is further examined using numerical simulations in the next section.
4. Importance of midlevel moisture for TC formation
This section compares the moisture patterns among GTS, EPH, and NCEP FNL to show the difference in moisture evolution throughout the TC formation process between the observation and numerical simulation for each TC type. In addition, this section examines the moisture importance for TC formation using the difference in moisture patterns between GTS and EPH for moisture-sensitive TCs (type 3) in sensitivity experiments. To simplify and highlight the difference, this section focuses on figures showing azimuthal averages and time–height sections. Horizontal patterns and evolutions at specific levels are shown in supplement 1 in the online supplemental material.
a. Importance of midlevel moisture
Figures 6 and 7 show the difference in azimuthal average between the initial fields of free simulations of GTS and EPH (the time of DA completion; t = −48 h in Fig. 3) and NCEP FNL and the difference between EPH and GTS. The FNL composites consider the average based on the TCC center. The simulation composites consider the average based on the pre-TC disturbance center. As Fig. 6 shows, the relative humidity around the disturbance center at the initial time of free simulations is significantly underestimated relative to FNL for all TCs, regardless of GTS or EPH. The relative humidity within 2° radius above 850 hPa of the disturbance center for EPH is 2%–5% higher than GTS (Fig. 6c). This difference is statistically significant and improves 10%–23% of the dry bias in simulations relative to FNL. Its impact on TC formation and detection rate is further examined in the next section. However, the difference in vorticity around the disturbance center between the simulations and FNL is not statistically significant (Fig. 7). Also, the GTS and EPH vorticities do not differ. This indicates the GPS RO DA mainly affects the moisture fields around the disturbance center, rather than the vorticity, although the simulations with the assimilation of GPS RO data are still drier than FNL. Although assimilation of GPS RO data may increase or decrease moisture for different cases, EPH significantly increases the moisture magnitudes around the disturbance center relative to GTS for the 35 TCs in this study (Fig. 8). For moisture-sensitive TCs (type 3), TC formation can be detected in the EPH with the moisture increases above 850 hPa around the disturbance center (Fig. 8c). For the other TCs (types 1 and 2), EPH mainly increases the moisture at mid- (700–450 hPa; Fig. 8a) and high levels (500–250 hPa; Fig. 8b) but its skill in TC detection is the same with GTS.
Difference in azimuthal-averaged relative humidity (colored shading; %) at the initial time of free simulations between (a) GTS and FNL, (b) EPH and FNL, and (c) EPH and GTS for 35 TCs. Black dots indicate the differences that pass the t test at a 90% confidence level. The upper color bar is for (a) and (b), and the lower one is for (c). The origin in each panel represents the disturbance location.
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
As in Fig. 6, but for relative vorticity (colored shading; 10−6 s−1).
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
Difference in azimuthal-averaged relative humidity (colored shading; %) at the initial time of free simulations between EPH and GTS for (a) type 1, (b) type 2, and (c) type 3. Black dots indicate the differences that pass the t test at a 90% confidence level. The origin in each panel represents the disturbance location.
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
Further, to compare the moisture patterns during free simulation, Figs. 9 and 10 show the time–height section of the difference in relative humidity following the disturbance during the TC formation process. The time–height section considers the simulated parameters averaged within a 5° radius of the pre-TC disturbance center. For detected TCs, the time–height section is composited based on the time relative to TC formation (t = 0 h) in each simulation. However, for nondetected TCs, reference time is relative to the time that the disturbance reaches maximum intensity. As Fig. 9 shows, the underestimated moisture around the disturbance center in both GTS and EPH simulations persists throughout the TC formation process. This moisture underestimation agrees with the findings of Chen et al. (2020) and Kuo et al. (2020). Although the moisture around the disturbance center is underestimated relative to FNL, TC formations are still detected in GTS and EPH for type 1 and in EPH for type 3. Compared to GTS, the relative humidity around the disturbance center at midlevels in EPH is increased by 2%–5% at the initial time of free simulations (Fig. 6c). During the TC formation process, the maximum difference (6%–9%) occurs between 700 and 500 hPa over the period 30–6 h before TC formation (Fig. 9c). This is a likely explanation for higher TC detection rate in EPH than GTS.
Difference in time–height section of relative humidity (colored shading; %) averaged within a 5° radius of the disturbance center before TC formation (t = 0 h) between (a) GTS and FNL, (b) EPH and FNL, and (c) EPH and GTS for 35 TCs. Black dots indicate the differences that pass the t test at a 90% confidence level. The upper color bar is for (a) and (b), and the lower one is for (c).
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
Difference in time–height section of relative humidity (colored shading; %) averaged within a 5° radius of the disturbance center before TC formation (t = 0 h) between EPH and GTS for (a) type 1, (b) type 2, and (c) type 3. Black dots indicate the differences that pass the t test at a 90% confidence level.
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
As Fig. 10a shows, the moisture improvement during the formation process in EPH is not significant for type-1 TCs. Type-1 TCs develop in environments with higher moisture, and their simulated moisture performance during TC formation process is not sensitive to GPS RO data. Although EPH of type-2 TCs has higher moisture around the disturbance center at mid- (850–650 hPa) and high levels (450–300 hPa) over 48–36 h before the disturbance reaches the maximum intensity than GTS, this higher moisture characteristic does not persist during the entire formation process (Fig. 10b). TC formation cannot be detected in EPH either. For type-3 TCs, the significant difference in moisture improvement between EPH and GTS is observed at 48 h before TC formation and is enhanced at midlevels (850–500 hPa) during the formation process (Fig. 10c). Because the dry bias within the disturbance center at midlevels in simulations is substantially reduced by EPH, the model may better capture the vertical development of strong convergence from low-levels, deep convection maintenance, and self-sustaining process of disturbance (feedback between vertical motion and diabatic heating) at midlevels (Emanuel 1986; Gray 1998; Houze 2010). These characteristics are all favorable for TC formation in EPH.
In summary, the moisture underestimation does not affect TC formation simulation in favorable environments (type 1). For vorticity-sensitive TCs (type 2), the moisture patterns improved by the assimilation of GPS RO data are not sustained over the entire TC formation process. The increased midlevel moisture at the initial time and its nonlinear effects (e.g., moisture-dynamics interaction) are not sufficient to result in successful TC formation simulation. For moisture-sensitive TCs (type 3), the initial improved moisture and its nonlinear effects in EPH maintain the pre-TC disturbance and increase midlevel moisture within the center during the entire TC formation process. This is favorable for successful TC formation simulation.
b. Sensitivity experiments
Two ETCs and two MTCs in type 3 are selected to further explore the moisture importance to TC formation in the sensitivity experiment. These TCs form in four different environments: easterly west (Kajiki) and southwest (Lingling) of the subtropical high, monsoon confluence (Fitow), and north of monsoon trough (Mirinae). The experiment procedures of GTS-E and EPH-G for each TC are shown and defined in section 2c and Table 2. These sensitivity experiments examine the importance of the increased moisture alone isolated from the effects of GPS RO DA on other parameters (e.g., dynamics and temperature) during the assimilation period.
Figure 11 shows the average moisture magnitudes exchanged between GTS and EPH for the four sensitivity experiment cases. EPH has higher relative humidity around the pre-TC disturbance center than GTS, especially at mid- and high levels. Based on the GPI definition (i.e., 600-hPa relative humidity; Emanuel and Nolan 2004), the pre-TC disturbance in EPH has a more favorable environmental condition than GTS at the initial time of free simulations. In the sensitivity experiments with exchanged moisture fields, TC formation can be detected in all GTS-E experiments for four TCs, while EPH-G cannot (Table 4). This indicates that the midlevel moisture change at the initial time of free simulations is more important than other parameters changed by GPS RO DA for the formation simulation of these TCs. The disturbances with improved moisture fields (i.e., the dry bias is reduced) at midlevels can develop into TCs.
Difference in relative humidity (colored shading; %) between EPH and GTS at the initial time of free simulations for four sensitivity experiment cases in (a) 600-hPa composite average and (b) azimuthal average. The origin in each panel represents the disturbance location. Due to the experiment sample size, the significance in difference is not shown.
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
Simulation performance for sensitivity experiments and correlation coefficients (R) between diabatic PV and wind speed at each level during TC formation process. The circles (O) indicate the TC formation is detected in the simulation, while the crosses (X) are not. R950, R850, R700, R500, and R300 indicate the correlations at 950, 850, 700, 500, and 300 hPa, respectively. Asterisks (*) indicate the correlation passes the Pearson’s correlation test at a 95% confidence level.
To explain the moisture contribution to dynamic process of TC formation, the diabatic PV around the disturbance center is further analyzed. Because the moisture within 2° radius is significantly increased by the assimilation of GPS RO data (Fig. 8c), the diabatic PV within 2° radius is calculated. As Table 4 shows, the diabatic PV magnitudes are significantly correlated with the disturbance intensity at midlevels (700 and 500 hPa) during the TC formation process, and the intensity of MTCs is more sensitive to the diabatic PV than ETCs. The sensitivity to midlevel moisture for disturbance intensification is greater than low- or high-level moisture.
Figure 12 shows the time–height section of difference in diabatic PV averaged over four cases at the time relative to the initial time of free simulations (t = 0 h in the figure) between GTS-E and GTS and between EPH and EPH-G. GTS-E and EPH have higher diabatic PV than GTS and EPH-G below 500 hPa at the initial time of free simulations. These positive differences persist at mid- and low levels for the next 48 h. Furthermore, changing the time reference to the time relative to TC formation (t = 0 h) shows that stronger PV contributed by diabatic heating is observed in GTS-E and EPH before TC formation (Fig. 13). The main contribution is distributed over the period 48–8 h before TC formation for GTS-E and over the periods 48–36 h and 20–0 h before TC formation for EPH. This results in the performance difference in TC formation simulation between the sensitivity and original experiments.
Difference in time–height section of diabatic PV averaged within a 2° radius of the disturbance center (colored shading; PVU; 1 PVU = 10−6 m2 K s−1 kg−1) after 3-day cycling DA (t = 0 h) between (a) GTS-E and GTS and (b) EPH and EPH-G. Due to the experiment sample size, the significance in difference is not shown.
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
As in Fig. 12, but for the time before TC formation (t = 0 h).
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
The simulation performance of TC formation is highly sensitive to midlevel moisture patterns for these four type-3 TCs. When the midlevel moisture is increased (EPH and GTS-E), stronger PV is produced by diabatic heating below the midlevel during the formation process, favoring TC formation. This result addresses the physical contribution and importance of midlevel moisture to TC formation. A larger difference in diabatic PV between GTS-E and GTS than that between EPH and EPH-G may be attributable to that the 3-day cycling DA in EPH also improves parameters other than moisture (e.g., wind and temperature) and causes a smaller difference in diabatic PV between EPH and EPH-G. Although the results shown in this section are only based on the experiments for the four TCs, 73% of type-3 TCs have a consistent result for the moisture-exchanged experiment (not shown). The moisture and diabatic PV are the main factors affecting type-3 TC formation, but other parameters improved by GPS RO DA (e.g., wind, temperature, or nonlinear interactions) may also affect type-3 TC formation.
5. Discussion and conclusions
a. Discussion
Model resolution can affect the performance of TC formation simulation (Tory et al. 2006; Montgomery et al. 2010; Blake and Kotal 2018). Simulations with a higher resolution (e.g., grid spacings of 5 km or less) can better capture the low-level mesoscale convergence and updrafts around the pre-TC disturbance center, favoring deep convection formation, maintenance, and successful TC formation simulation. This is one of the potential explanations for dry biases shown in Figs. 6 and 9. For instance, if the experiments of typhoon Kajiki (2007) use two nested domains (grid spacings of 15 and 5 km), their TC formations can be detected whether in GTS or EPH (not shown). The midlevel moisture underestimation around the pre-TC disturbance center is overcome by the stronger low-level moisture convergence and updraft caused by the resolution increase, leading to TC formation in the simulation. However, the science perspective of this study, which statistically analyzes the TC formation characteristics in 15-km-resolution simulations for many cases, is different from studies using high-resolution simulations for a small number of selected cases. Due to computational cost and timeliness, most synoptic-scale operational TC forecasts use models with grid spacing larger than 15 km. The purpose of this study is therefore to understand the moisture importance for TC formation at synoptic scales and find the potential factors that can improve TC formation forecast in operational models and DA products.
Previous studies used regional models with 12–35-km resolution to analyze the influence of synoptic-scale environment and interannual variability on TC formation (Jourdain et al. 2011; Suzuki-Parker 2012; Hsieh et al. 2017; Chen et al. 2020). In their studies, TC detection rates without additional assimilation of satellite data were 20%–40% in simulations of 48–120-h lead time. Similar detection rates were also found in operational global models (Halperin et al. 2013). Average TC detection rates do not differ between previous studies and this study (GTS), but this study clearly shows different TC detection rate in different environments. Moreover, Hsieh et al. (2017) and Chen et al. (2020) noticed that models had a low simulation skill for some TC formations, regardless of initial data, physical parameterizations, and lead time. This characteristic was explained by a stronger influence of stochastic process in these TC formations, causing a lower probability of TC formation being detected in models (Ooyama 1982). In turn, TCs with a stronger influence of deterministic process had a higher detection rate. Corresponding to these previous studies, ETC formations are likely to have greater stochastic processes associated with a smaller area of favorable conditions than MTC formations, resulting in a lower simulation skill for ETCs, especially over southwest of the subtropical high (Table 3).
In addition, the direction of the “teardrop” streamline of ETCs (Dunkerton et al. 2009) is likely another mechanism contributing to the sensitivity of ETC formation to environmental moisture in simulations. As shown by the comoving streamlines in Fig. 1, the main streamlines directed inward to the disturbance center come from drier areas for ETCs, while they are from moister areas for MTCs. The monsoon environment has higher moisture and is easier to transport moisture into the disturbance center. Conversely, the high moisture area in the easterly environment is narrow and has a lower probability of being transported into the disturbance center. This may result in MTC formation being less sensitive but ETC formation being more sensitive to moisture accuracy in simulations. A more detailed moisture transport analysis is provided in supplement 2 in the online supplemental material.
Wang et al. (2018) analyzed practical predictability for TC formations in the North Atlantic, and found that TC formations within tropical transition pathways had lower predictability than the others. Based on the contingency table (WWRP/WGNE 2015), the simulation skill of TC formation analyzed in this study is different from the practical predictability of TC formation. First, only developing TCCs are analyzed in this study. The correct negatives are not considered (i.e., nonformation process of nondeveloping TCCs). Second, no false alarm is counted and discussed in the experiments. The model of this study is not set up for exploring the practical predictability. Therefore, although this study explains the difference in TC detection rate between monsoon and easterly environments, the complete predictability of TC formation in different environments deserves to be studied further.
The TC formation process analyzed in this study is the period from TCCs forming to developing into TCs. This is slightly different from previous studies (Emanuel 1986; Simpson et al. 1997; Hendricks et al. 2004; Dunkerton et al. 2009), which analyzed the process from a weak initial vortex or a wave pouch to TC formation. Compared to the initial vortex in their studies, TCCs have initial low-level cyclonic circulation, moisture condition, and deep convection around the center (Teng et al. 2019, 2020). To compare the statistical characteristics of TC formation between multiple ETCs and MTCs at a similar stage of their formation process, the disturbance with TCC intensity is regarded as the initial disturbance for the ETC and MTC formation process in this study. Specifically, MTCs have no pouch structure and form within a large background trough system in the WNP. These characteristics are significantly different from ETCs and TC formations in the North Atlantic. Although the disturbances with TCC intensity have a weak vorticity and convective center, the characteristics of their development into TCs can still represent the main process of TC formation.
b. Conclusions
In the WNP, the importance of midlevel moisture for TC formation in monsoon and easterly environments has remained unclear. This study first compares the climatological environmental characteristics between ETC and MTC formation, and then explores the moisture contribution to TC formation in simulations of 35 TCs with and without assimilation of GPS RO data. The simulations with additional assimilation of GPS RO data (EPH) enhance moisture closer to observations relative to the simulations without (GTS). The GPS RO data, during the period from 120 to 48 h prior to TC formation, are found to mainly affect moisture patterns above 850 hPa around the disturbance center, rather than affecting the vorticity.
In general, the monsoon system provides a larger area of higher moisture for TC formation, while the easterly system supplies higher vorticity around the TCC center. Because a dry bias exists in all simulations for ETCs and MTCs, the easterly environments may have insufficient moisture for TC formation in simulations and have a significantly lower TC detection rate than the monsoon environments. The numerical model has a higher simulation skill for TCs that form in environments with higher vorticity and moisture (type 1 TCs), most of which (92%) are MTCs, especially the monsoon trough environment. Other than those TCs, the simulation skill for TCs dominated by midlevel moisture (type 3 TCs) is increased if the midlevel moisture underestimation is partially improved by GPS RO data. The simulation performance of TCs dominated by midlevel vorticity (type 2 TCs) are not sensitive to the midlevel moisture improved by GPS RO data. Most ETCs (92%) belong to types 2 and 3. Because the averaged midlevel moisture magnitude around the observed TC formation location is low in easterly environments, its importance to TC formation is higher than that in monsoon environments.
For moisture-sensitive TCs (type 3), the moisture improved by the GPS RO data at the initial time of free simulations persists and affects midlevel moisture patterns during the TC formation process, leading to TC formation in EPH. To examine the moisture importance for TC formation, four TCs that form in different environments are selected to perform sensitivity experiments. If the lower moisture fields in GTS are replaced by higher moisture fields in EPH, TC formation can be detected in the simulations (GTS-E). In turn, the EPH simulations with GTS moisture fields (EPH-G) cannot successfully simulate TC formation. Specifically, higher moisture magnitudes at mid- and high levels (above 850 hPa) increase the diabatic heating release during the TC formation process. This difference results in stronger diabatic PV at midlevels and increases the disturbance intensity, leading to the pre-TC disturbance developing into TC in GTS-E and EPH.
This study analyzes all TCs that developed within a single consistent environmental category throughout their formation process during 2006–09 to explore the midlevel moisture importance for TC formation in monsoon and easterly environments. Although the simulation skill is not equal to the model predictability, the results advance our understanding of synoptic-scale moisture effects on TC formation in the regional model. The relationship between the midlevel moisture underestimation and TC formation under different environments can guide TC forecast improvement. The assimilation of GPS RO data can substantially improve the simulation performance of the formation of moisture-sensitive TCs in this class of models, particularly in easterly environments.
Acknowledgments
The authors thank the editor, Professor Dunkerton, and another reviewer who helped to improve this paper. This study is supported by the National Science Foundation under Cooperative Agreement 1522830 and the National Space Organization (NSPO) of the Republic of China (Taiwan) via the American Institute in Taiwan (AIT) under NSPO-UCAR (University Corporation for Atmospheric Research) AIT-TECRO (Taipei Economic and Cultural Representative Office) Agreement Implementing Arrangement 5. The National Center for Atmospheric Research is sponsored by the National Science Foundation. The computing resource and storage are supported by the Computational Information Systems Laboratory of UCAR.
APPENDIX
Environmental Characteristics Corresponding to TC Types with Different Simulation Skill
To explain the difference in TC detection rate in GTS and EPH between ETCs and MTCs, the environmental differences in the reanalysis data (NCEP CFSR) among TCs of types 1–3 are analyzed in this appendix. Figure A1 shows the average time–height section of relative humidity and vorticity for the 35 TCs during their formation process. The time–height section considers the parameter averaged within a 5° radius of the TCC center and is composited based on the time relative to the TC formation (t = 0 h). The 5°-radius environment following the TCC center is regarded as the environment that affects TCC development (Teng et al. 2019). In general, two maxima in the vertical section of relative humidity are observed before TC formation (950–900 and 800–700 hPa). When TCCs develop into TCs, the relative humidity over 550–350 hPa decreases. This is likely driven by the vertical motion and condensation around the TCC center during TC formation process (Fig. A1a). The mid- and low-level vorticity (below 400 hPa) increases associated with the TCC intensification, starting from the low levels (950–850 hPa).
Time–height section of the 35 TCs for (a) relative humidity (colored shading; %) and vertical velocity (contours; Pa s−1) and (b) relative vorticity (colored shading; 10−6 s−1) averaged within a 5° radius of the disturbance center before TC formation (t = 0 h).
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
Figures A2 and A3 show the difference in time–height section of relative humidity and vorticity between each TC type and the average of the 35 TCs. Type-1 TCs have significantly higher relative humidity and vorticity than the average during their formation process. Most type-1 TCs develop in monsoon environments (92%) associated with more favorable conditions. Their formations have a higher probability of being detected in simulations. In contrast, TC formation cannot be detected in GTS when TCCs develop into TCs within environments with lower relative humidity and vorticity (Figs. A2b and A3b). Although these environmental conditions can still support TC formation in the observations, the underestimation of relative humidity and vorticity in the analyses and model may cause no TC formation in GTS. Comparing environmental patterns between type-2 and -3 TCs shows that the importance of relative humidity and vorticity to these two types of TCs differ. The formation process of type-3 TCs is associated with relatively higher relative humidity and lower vorticity at midlevels than for type-2 TCs. Based on the GPI formulation, the favorable environmental condition for type-3 TC formation is mainly dominated by the midlevel relative humidity (i.e., moisture-sensitive TCs; Figure A2c), while by the vorticity for type-2 TCs (i.e., vorticity-sensitive TCs; Figure A3c). The environmental characteristics of type-2 TCs are similar to the general characteristics of ETCs (Fig. 4), while type-3 TCs differ.
Difference in time–height section of relative humidity (colored shading; %) averaged within a 5° radius of the disturbance center before TC formation (t = 0 h) between (a) type 1 and the average of 35 TCs, (b) types 2–3 and the average of 35 TCs, and (c) type 3 and type 2. Black dots indicate the differences that pass the t test at a 90% confidence level. The average of 35 TCs is shown in Fig. A1.
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
As in Fig. A2, but for relative vorticity (colored shading; 10−6 s−1).
Citation: Monthly Weather Review 149, 7; 10.1175/MWR-D-20-0313.1
REFERENCES
Anthes, R. A., C. Rocken, and Y. H. Kuo, 2000: Applications of COSMIC to meteorology and climate. Terr. Atmos. Oceanic. Sci., 11, 115–156, https://doi.org/10.3319/TAO.2000.11.1.115(COSMIC).
Anthes, R. A., and Coauthors, 2008: The COSMiC/FORMOSAT-3—Mission early results. Bull. Amer. Meteor. Soc., 89, 313–333, https://doi.org/10.1175/BAMS-89-3-313.
Barker, D. M., W. Huang, Y.-R. Guo, A. J. Bourgeois, and Q. N. Xiao, 2004: A three-dimensional variational data assimilation system for MM5: Implementation and initial results. Mon. Wea. Rev., 132, 897–914, https://doi.org/10.1175/1520-0493(2004)132<0897:ATVDAS>2.0.CO;2.
Barker, D. M., and Coauthors, 2012: The Weather Research and Forecasting Model’s Community Variational/Ensemble Data Assimilation System: WRFDA. Bull. Amer. Meteor. Soc., 93, 831–843, https://doi.org/10.1175/BAMS-D-11-00167.1.
Beljaars, A. C. M., and A. A. M. Holtslag, 1991: Flux parametrization over land surfaces for atmospherics models. J. Appl. Meteor., 30, 327–341, https://doi.org/10.1175/1520-0450(1991)030<0327:FPOLSF>2.0.CO;2.
Bister, M., and K. A. Emanuel, 2002: Low frequency variability of tropical cyclone potential intensity. 1. Interannual to interdecadal variability. J. Geophys. Res., 107, 4801, https://doi.org/10.1029/2001JD000776.
Blake, E., and S. D. Kotal, 2018: Cyclogenesis: Operational forecasting perspective. Ninth Int. Workshop on Tropical Cyclones Rapporteur Report, WMO/WWRP/TCP, 28 pp.
Bruyère, C. L., G. J. Holland, and E. Towler, 2012: Investigating the use of a genesis potential index for tropical cyclones in the North Atlantic basin. J. Climate, 25, 8611–8626, https://doi.org/10.1175/JCLI-D-11-00619.1.
Camargo, S. J., K. A. Emanuel, and A. H. Sobel, 2007: Use of a genesis potential index to diagnose ENSO effects on tropical cyclone genesis. J. Climate, 20, 4819–4834, https://doi.org/10.1175/JCLI4282.1.
Chen, D. Y. C., K. K. W. Cheung, and C. S. Lee, 2012: A study on the synoptic-dynamical characteristics of compact tropical cyclones in the western North Pacific. Mon. Wea. Rev., 140, 4046–4065, https://doi.org/10.1175/MWR-D-11-00317.1.
Chen, F., and J. Dudhia, 2001: Coupling an advanced land surface-hydrology model with the Penn State–NCAR MM5 modeling system. Part I: Model implementation and sensitivity. Mon. Wea. Rev., 129, 569–585, https://doi.org/10.1175/1520-0493(2001)129<0569:CAALSH>2.0.CO;2.
Chen, S.-Y., C.-Y. Huang, Y.-H. Kuo, Y.-R. Guo, and S. Sokolovskiy, 2009: Assimilation of GPS refractivity from FORMOSAT-3/COSMIC using a nonlocal operator with WRF 3DVAR and its impact on the prediction of a typhoon event. Terr. Atmos. Oceanic Sci., 20, 133–154, https://doi.org/10.3319/TAO.2007.11.29.01(F3C).
Chen, S.-Y., Y. H. Kuo, and C. Y. Huang, 2020: The impact of GPS RO data on the prediction of tropical cyclogenesis using a nonlocal observation operator: An initial assessment. Mon. Wea. Rev., 148, 2701–2717, https://doi.org/10.1175/MWR-D-19-0286.1.
Chen, T. C., S. Y. Wang, M. C. Yen, and A. J. Clark, 2008: Are tropical cyclones less effectively formed by easterly waves in the western North Pacific than in the North Atlantic? Mon. Wea. Rev., 136, 4527–4540, https://doi.org/10.1175/2008MWR2149.1.
Chen, Y. C., M. E. Hsieh, L. F. Hsiao, Y. H. Kuo, M. J. Yang, C. Y. Huang, and C. S. Lee, 2015: Systematic evaluation of the impacts of GPSRO data on the prediction of typhoons over the northwestern Pacific in 2008–2010. Atmos. Meas. Tech., 8, 2531–2542, https://doi.org/10.5194/amt-8-2531-2015.
Choi, Y., D.-H. Cha, M.-I. Lee, J. Kim, C.-S. Jin, S.-H. Park, and M.-S. Joh, 2017: Satellite radiance data assimilation for binary tropical cyclone cases over the western North Pacific. J. Adv. Model. Earth Syst., 9, 832–853, https://doi.org/10.1002/2016MS000826.
Chu, J.-H., C. R. Sampson, A. S. Levin, and E. Fukada, 2002: The Joint Typhoon Warning Center tropical cyclone best-tracks, 1945–2000. Joint Typhoon Warning Center Rep. NRL/MR/7540-02-16, https://www.metoc.navy.mil/jtwc/products/best-tracks/tc-bt-report.html.
Doyle, J. D., C. A. Reynolds, C. Amerault, and J. Moskaitis, 2012: Adjoint sensitivity and predictability of tropical cyclogenesis. J. Atmos. Sci., 69, 3535–3557, https://doi.org/10.1175/JAS-D-12-0110.1.
Dunkerton, T. J., M. T. Montgomery, and Z. Wang, 2009: Tropical cyclogenesis in a tropical wave critical layer: Easterly waves. Atmos. Chem. Phys., 9, 5587–5646, https://doi.org/10.5194/acp-9-5587-2009.
Emanuel, K. A., 1986: An air-sea interaction theory for tropical cyclones. Part I: Steady-state maintenance. J. Atmos. Sci., 43, 585–604, https://doi.org/10.1175/1520-0469(1986)043<0585:AASITF>2.0.CO;2.
Emanuel, K. A., and D. S. Nolan, 2004: Tropical cyclone activity and the global climate system. 26th Conf. on Hurricanes and Tropical Meteorology, Miami, FL, Amer. Meteor. Soc., 10A.2, https://ams.confex.com/ams/26HURR/techprogram/paper_75463.htm.
Fu, B., M. S. Peng, T. Li, and D. E. Stevens, 2012: Developing versus nondeveloping disturbances for tropical cyclone formation. Part II: Western North Pacific. Mon. Wea. Rev., 140, 1067–1080, https://doi.org/10.1175/2011MWR3618.1.
Fudeyasu, H., and R. Yoshida, 2018: Western North Pacific tropical cyclone characteristics stratified by genesis environment. Mon. Wea. Rev., 146, 435–446, https://doi.org/10.1175/MWR-D-17-0110.1.
Gray, W. M., 1968: Global view of origin of tropical disturbances and storms. Mon. Wea. Rev., 96, 669–700, https://doi.org/10.1175/1520-0493(1968)096<0669:GVOTOO>2.0.CO;2.
Gray, W. M., 1984a: Atlantic seasonal hurricane frequency. Part I: El Niño and 30 mb quasi-biennial oscillation influences. Mon. Wea. Rev., 112, 1649–1668, https://doi.org/10.1175/1520-0493(1984)112<1649:ASHFPI>2.0.CO;2.
Gray, W. M., 1984b: Atlantic seasonal hurricane frequency. Part II: Forecasting its variability. Mon. Wea. Rev., 112, 1669–1683, https://doi.org/10.1175/1520-0493(1984)112<1669:ASHFPI>2.0.CO;2.
Gray, W. M., 1998: The formation of tropical cyclones. Meteor. Atmos. Phys., 67, 37–69, https://doi.org/10.1007/BF01277501.
Greeshma, M. M., C. V. Srinivas, V. Yesubabu, C. V. Naidu, R. Baskaran, and B. Venkatraman, 2015: Impact of local data assimilation on tropical cyclone predictions over the Bay of Bengal using the ARW model. Ann. Geophys., 33, 805–828, https://doi.org/10.5194/angeo-33-805-2015.
Guard, C. P., L. E. Carr, F. H. Wells, R. A. Jeffries, N. D. Gural, and D. K. Edson, 1992: Joint Typhoon Warning Center and the challenges of multibasin tropical cyclone forecasting. Wea. Forecasting, 7, 328–352, https://doi.org/10.1175/1520-0434(1992)007<0328:JTWCAT>2.0.CO;2.
Halperin, D. J., H. E. Fuelberg, R. E. Hart, J. H. Cossuth, P. Sura, and R. J. Pasch, 2013: An evaluation of tropical cyclone genesis forecasts from global numerical models. Wea. Forecasting, 28, 1423–1445, https://doi.org/10.1175/WAF-D-13-00008.1.
Hendricks, E. A., M. T. Montgomery, and C. A. Davis, 2004: The role of “vortical” hot towers in the formation of Tropical Cyclone Diana (1984). J. Atmos. Sci., 61, 1209–1232, https://doi.org/10.1175/1520-0469(2004)061<1209:TROVHT>2.0.CO;2.
Hennon, C. C., C. N. Helms, K. R. Knapp, and A. R. Bowen, 2011: An objective algorithm for detecting and tracking tropical cloud clusters: Implications for tropical cyclogenesis prediction. J. Atmos. Oceanic Technol., 28, 1007–1018, https://doi.org/10.1175/2010JTECHA1522.1.
Hill, K. A., and G. M. Lackmann, 2009: Influence of environmental humidity on tropical cyclone size. Mon. Wea. Rev., 137, 3294–3315, https://doi.org/10.1175/2009MWR2679.1.
Holland, G. J., 1995: Scale interaction in the western Pacific monsoon. Meteor. Atmos. Phys., 56, 57–79, https://doi.org/10.1007/BF01022521.
Hong, S.-Y., Y. Noh, and J. Dudhia, 2006: A new vertical diffusion package with an explicit treatment of entrainment processes. Mon. Wea. Rev., 134, 2318–2341, https://doi.org/10.1175/MWR3199.1.
Houze, R. A., 2010: Clouds in tropical cyclones. Mon. Wea. Rev., 138, 293–344, https://doi.org/10.1175/2009MWR2989.1.
Hsieh, Y.-H., C.-S. Lee, and C.-H. Sui, 2017: A study on the influences of low-frequency vorticity on tropical cyclone formation in the western North Pacific. Mon. Wea. Rev., 145, 4151–4169, https://doi.org/10.1175/MWR-D-17-0085.1.
Huang, C. Y., and Coauthors, 2010: Impact of GPS radio occultation data assimilation on regional weather predictions. GPS Solut., 14, 35–49, https://doi.org/10.1007/s10291-009-0144-1.
Huang, C. Y., Y. H. Kuo, S. H. Chen, and F. Vandenberghe, 2005: Improvements in typhoon forecasts with assimilated GPS occultation refractivity. Wea. Forecasting, 20, 931–953, https://doi.org/10.1175/WAF874.1.
Huang, C. Y., S. Y. Chen, S. K. Rao Anisetty, S. C. Yang, and L. F. Hsiao, 2016: An impact study of GPS radio occultation observations on frontal rainfall prediction with a local bending angle operator. Wea. Forecasting, 31, 129–150, https://doi.org/10.1175/WAF-D-15-0085.1.
Iacono, M. J., J. S. Delamere, E. J. Mlawer, M. W. Shephard, S. A. Clough, and W. D. Collins, 2008: Radiative forcing by long-lived greenhouse gases: Calculations with the AER radiative transfer models. J. Geophys. Res., 113, D13103, https://doi.org/10.1029/2008JD009944.
Jourdain, N. C., P. Marchesiello, C. E. Menkes, J. Lefèvre, E. M. Vincent, M. Lengaigne, and F. Chauvin, 2011: Mesoscale simulation of tropical cyclones in the South Pacific: Climatology and interannual variability. J. Climate, 24, 3–25, https://doi.org/10.1175/2010JCLI3559.1.
Kerns, B. W., and S. Y. S. Chen, 2013: Cloud clusters and tropical cyclogenesis: Developing and nondeveloping systems and their large-scale environment. Mon. Wea. Rev., 141, 192–210, https://doi.org/10.1175/MWR-D-11-00239.1.
Kueh, M. T., C. Y. Huang, S. Y. Chen, S. H. Chen, and C. J. Wang, 2009: Impact of GPS radio occultation refractivity soundings on a simulation of Typhoon Bilis (2006) upon landfall. Terr. Atmos. Oceanic Sci., 20, 115–131, https://doi.org/10.3319/TAO.2008.01.21.03(F3C).
Kukkonen, J., and Coauthors, 2012: A review of operational, regional-scale, chemical weather forecasting models in Europe. Atmos. Chem. Phys., 12, 1–87, https://doi.org/10.5194/acp-12-1-2012.
Kuo, H. C., J. H. Chen, R. T. Williams, and C. P. Chang, 2001: Rossby waves in zonally opposing mean flow: Behavior in Northwest Pacific summer monsoon. J. Atmos. Sci., 58, 1035–1050, https://doi.org/10.1175/1520-0469(2001)058<1035:RWIZOM>2.0.CO;2.
Kuo, Y. H., S. V. Sokolovskiy, R. A. Anthes, and F. Vandenberghe, 2000: Assimilation of GPS radio occultation data for numerical weather prediction. Terr. Atmos. Oceanic Sci., 11, 157–186, https://doi.org/10.3319/TAO.2000.11.1.157(COSMIC).
Kuo, Y. H., H. Zhang, H. Liu, C. Y. Huang, S. Y. Chen, L. F. Hsiao, M. E. Shieh, and Y. C. Chen, 2018: Impact of GPS RO data on the prediction of tropical cyclones. Fourth Int. Conf. on GPS Radio Occultation, Taipei, Taiwan, NSPO, 40 pp., http://w3.nspo.narl.org.tw/ICGPSRO2018/download/pdf/S1.3_Kuo_ICGPSRO_2018_Final.pdf.
Kuo, Y. H., S. Y. Chen, H. F. Teng, and C. Y. Huang, 2020: Impact of GPS radio occultation data on the prediction of tropical cyclogenesis. 24th Conf. on Integrated Observing and Assimilation Systems for the Atmosphere, Oceans, and Land Surface (IOAS-AOLS), Boston, MA, Amer. Meteor. Soc., 11.1, https://ams.confex.com/ams/2020Annual/meetingapp.cgi/Paper/367643.
Kursinski, E. R., G. A. Hajj, S. S. Leroy, and B. Herman, 2000: The GPS radio occultation technique. Terr. Atmos. Oceanic Sci., 11, 53–114, https://doi.org/10.3319/TAO.2000.11.1.53(COSMIC).
Lander, M. A., 1994: Description of a monsoon gyre and its effects on the tropical cyclones in the western North Pacific during August 1991. Wea. Forecasting, 9, 640–654, https://doi.org/10.1175/1520-0434(1994)009<0640:DOAMGA>2.0.CO;2.
Li, Z., and Z. Pu, 2014: Numerical simulations of the genesis of Typhoon Nuri (2008): Sensitivity to initial conditions and implications for the roles of intense convection and moisture conditions. Wea. Forecasting, 29, 1402–1424, https://doi.org/10.1175/WAF-D-14-00003.1.
Lim, A. H., J. A. Jung, S. E. Nebuda, J. M. Daniels, W. Bresky, M. Tong, and V. Tallapragada, 2019: Tropical cyclone forecasts impact assessment from the assimilation of hourly visible, shortwave, and clear-air water vapor atmospheric motion vectors in HWRF. Wea. Forecasting, 34, 177–198, https://doi.org/10.1175/WAF-D-18-0072.1.
Lim, K. S., and S. Hong, 2010: Development of an effective double-moment cloud microphysics scheme with prognostic cloud condensation nuclei (CCN) for weather and climate models. Mon. Wea. Rev., 138, 1587–1612, https://doi.org/10.1175/2009MWR2968.1.
Lin, P.-H., and M.-J. Yang, 2012: The potential vorticity budget of Typhoon Nari (2001). Atmos. Sci., 40, 1–27.
Lin, Y., M. Zhao, and M. Zhang, 2015: Tropical cyclone rainfall area controlled by relative sea surface temperature. Nat. Commun., 6, 6591, https://doi.org/10.1038/ncomms7591.
Liu, H., J. Anderson, and Y.-H. Kuo, 2012: Improved analyses and forecasts of Hurricane Ernesto’s genesis using radio occultation data in an ensemble filter assimilation system. Mon. Wea. Rev., 140, 151–166, https://doi.org/10.1175/MWR-D-11-00024.1.
MacQueen, J., 1967: Some methods for classification and analysis of multivariate observations. Fifth Berkeley Symp. on Mathematical Statistics and Probability, Berkeley, CA, University of California, 281–297.
Minamide, M., and F. Zhang, 2018: Assimilation of all-sky infrared radiances from Himawari-8 and impacts of moisture and hydrometer initialization on convection-permitting tropical cyclone prediction. Mon. Wea. Rev., 146, 3241–3258, https://doi.org/10.1175/MWR-D-17-0367.1.
Molinari, J., and D. Vollaro, 2013: What percentage of western North Pacific tropical cyclones form within the monsoon trough? Mon. Wea. Rev., 141, 499–505, https://doi.org/10.1175/MWR-D-12-00165.1.
Monin, A. S., and A. M. Obukhov, 1954: Basic laws of turulent mixing in the atmosphere near the ground. Tr. Inst. Teor. Geofiz. Akad. Nauk SSSR, 24, 1963–1987.
Montgomery, M. T., Z. Wang, and T. J. Dunkerton, 2010: Coarse, intermediate and high resolution numerical simulations of the transition of a tropical wave critical layer to a tropical storm. Atmos. Chem. Phys., 10, 10 803–10 827, https://doi.org/10.5194/acp-10-10803-2010.
NCEP, 2000: NCEP FNL operational model global tropospheric analyses, continuing from July 1999. Research Data Archive at the National Center for Atmospheric Research, Computational and Information Systems Laboratory, Boulder, CO, accessed 18 April 2020, https://doi.org/10.5065/D6M043C6.
NCEP, 2015: Observations and gridded forcing products input into the NCEP climate forecast system reanalysis. Research Data Archive at the National Center for Atmospheric Research, Computational and Information Systems Laboratory, Boulder, CO, accessed 18 April 2020, https://doi.org/10.5065/D6HM56HW.
Ooyama, K. V., 1982: Conceptual evolution of the theory and modeling of the tropical cyclone. J. Meteor. Soc. Japan, 60, 369–380, https://doi.org/10.2151/jmsj1965.60.1_369.
Peng, X. D., J. F. Fei, X. G. Huang, and X. P. Cheng, 2017: Evaluation and error analysis of official forecasts of tropical cyclones during 2005–14 over the western North Pacific. Part I: Storm tracks. Wea. Forecasting, 32, 689–712, https://doi.org/10.1175/WAF-D-16-0043.1.
Saha, S., and Coauthors, 2010: The NCEP Climate Forecast System Reanalysis. Bull. Amer. Meteor. Soc., 91, 1015–1057, https://doi.org/10.1175/2010BAMS3001.1.
Schreiner, W., C. Rocken, S. Sokolovskiy, and D. Hunt, 2010: Quality assessment of COSMIC/FORMOSAT-3 GPS radio occultation data derived from single- and double-difference atmospheric excess phase processing. GPS Solut., 14, 13–22, https://doi.org/10.1007/s10291-009-0132-5.
Schubert, W. H., S. A. Hausman, M. Garcia, K. V. Ooyama, and H. C. Kuo, 2001: Potential vorticity in a moist atmosphere. J. Atmos. Sci., 58, 3148–3157, https://doi.org/10.1175/1520-0469(2001)058<3148:PVIAMA>2.0.CO;2.
Shapiro, L. J., 1977: Tropical storm formation from easterly waves: A criterion for development. J. Atmos. Sci., 34, 1007–1022, https://doi.org/10.1175/1520-0469(1977)034<1007:TSFFEW>2.0.CO;2.
Simpson, J., E. Ritchie, G. J. Holland, J. Halverson, and S. Stewart, 1997: Mesoscale interactions in tropical cyclone genesis. Mon. Wea. Rev., 125, 2643–2661, https://doi.org/10.1175/1520-0493(1997)125<2643:MIITCG>2.0.CO;2.
Sippel, J. A., and F. Zhang, 2008: A probabilistic analysis of the dynamics and predictability of typhoon cyclogenesis. J. Atmos. Sci., 65, 3440–3459, https://doi.org/10.1175/2008JAS2597.1.
Skamarock, W. C., and Coauthors, 2019: A description of the Advanced Research WRF Model version 4. NCAR Tech. Note NCAR/TN-556+STR, 145 pp., https://doi.org/10.5065/1dfh-6p97.
Sokolovskiy, S., Y.-H. Kuo, and W. Wang, 2005: Assessing the accuracy of a linearized observation operator for assimilation of radio occultation data: Case simulations with a high-resolution weather model. Mon. Wea. Rev., 133, 2200–2212, https://doi.org/10.1175/MWR2948.1.
Suzuki-Parker, A., 2012: Uncertainties and Limitations in Simulating Tropical Cyclones. Springer, 78 pp.
Teng, H. F., C. S. Lee, and H. H. Hsu, 2014: Influence of ENSO on formation of tropical cloud clusters and their development into tropical cyclones in the western North Pacific. Geophys. Res. Lett., 41, 9120–9126, https://doi.org/10.1002/2014GL061823.
Teng, H. F., C. S. Lee, H. H. Hsu, J. M. Done, and G. J. Holland, 2019: Tropical cloud cluster environments and their importance for tropical cyclone formation. J. Climate, 32, 4069–4088, https://doi.org/10.1175/JCLI-D-18-0679.1.
Teng, H. F., J. M. Done, C. S. Lee, H. H. Hsu, and Y. H. Kuo, 2020: Large-scale environmental influences on tropical Cyclone formation processes and development time. J. Climate, 33, 9763–9782, https://doi.org/10.1175/JCLI-D-19-0709.1.
Teng, H. F., J. M. Done, and Y. H. Kuo, 2021: Landfalling tropical cyclone characteristics and their multi-timescale variability connected to monsoon and easterly formation environments over the western North Pacific. Quart. J. Roy. Meteor. Soc., https://doi.org/10.1002/qj.4108, in press.
Tory, K. J., M. T. Montgomery, and N. E. Davidson, 2006: Prediction and diagnosis of tropical cyclone formation in an NWP system. Part I: The critical role of vortex enhancement in deep convection. J. Atmos. Sci., 63, 3077–3090, https://doi.org/10.1175/JAS3764.1.
Wang, B., R. G. Wu, and K. M. Lau, 2001: Interannual variability of the Asian summer monsoon: Contrasts between the Indian and the western North Pacific–East Asian monsoons. J. Climate, 14, 4073–4090, https://doi.org/10.1175/1520-0442(2001)014<4073:IVOTAS>2.0.CO;2.
Wang, P. K., 2013: Physics and Dynamics of Clouds and Precipitation. Cambridge University Press, 467 pp.
Wang, Z., W. Li, M. S. Peng, X. Jiang, R. McTaggart-Cowan, and C. A. Davis, 2018: Predictive skill and predictability of North Atlantic tropical cyclogenesis in different synoptic flow regimes. J. Atmos. Sci., 75, 361–378, https://doi.org/10.1175/JAS-D-17-0094.1.
Wu, L., and Coauthors, 2012a: Relationship of environmental relative humidity with North Atlantic tropical cyclone intensity and intensification rate. Geophys. Res. Lett., 39, L20809, https://doi.org/10.1029/2012GL053546.
Wu, L., Z. P. Wen, R. H. Huang, and R. G. Wu, 2012b: Possible linkage between the monsoon trough variability and the tropical cyclone activity over the western North Pacific. Mon. Wea. Rev., 140, 140–150, https://doi.org/10.1175/MWR-D-11-00078.1.
Wu, L. G., H. J. Zong, and J. Liang, 2013: Observational analysis of tropical cyclone formation associated with monsoon gyres. J. Atmos. Sci., 70, 1023–1034, https://doi.org/10.1175/JAS-D-12-0117.1.
WWRP/WGNE, 2015: WWRP/WCRP Joint Working Group on Forecast Verification Research. WWRP/WCRP, accessed 14 June 2020, http://www.cawcr.gov.au/projects/verification/.
Zhang, F., and J. A. Sippel, 2009: Effects of moist convection on hurricane predictability. J. Atmos. Sci., 66, 1944–1961, https://doi.org/10.1175/2009JAS2824.1.
Zheng, Y., K. Alapaty, J. A. Herwehe, A. D. Del Genio, and D. Niyogi, 2016: Improving high-resolution weather forecasts using the Weather Research and Forecasting (WRF) Model with an updated Kain–Fritsch scheme. Mon. Wea. Rev., 144, 833–860, https://doi.org/10.1175/MWR-D-15-0005.1.
Zong, H. J., and L. G. Wu, 2015: Synoptic-scale influences on tropical cyclone formation within the western North Pacific monsoon trough. Mon. Wea. Rev., 143, 3421–3433, https://doi.org/10.1175/MWR-D-14-00321.1.
