• Chang, L.-Y., 2013: Tropical cyclone formation associated with trade-wind surges. Ph.D. thesis, National Taiwan University, 206 pp.

  • Chang, L.-Y., K. K. W. Cheung, and C.-S. Lee, 2010: The role of trade wind surges for tropical cyclone formations in the western North Pacific. Mon. Wea. Rev., 138, 41204134, https://doi.org/10.1175/2010MWR3152.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cheung, K. K. W., and R. L. Elsberry, 2006: Model sensitivities in numerical simulations of the formation of Typhoon Robyn (1993). Terr. Atmos. Oceanic Sci., 17, 5389, https://doi.org/10.3319/TAO.2006.17.1.53(SWS).

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ching, L., C.-H. Sui, and M.-J. Yang, 2010: An analysis of the multiscale nature of tropical cyclone activities in June 2004: Climate background. J. Geophys. Res., 115, D24108, https://doi.org/10.1029/2010JD013803.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Crosbie, E., and Y. L. Serra, 2014: Intraseasonal modulation of synoptic-scale disturbances and tropical cyclone genesis in the eastern North Pacific. J. Climate, 27, 57245745, https://doi.org/10.1175/JCLI-D-13-00399.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Duchon, C. E., 1979: Lanczos filtering in one and two dimensions. J. Appl. Meteor., 18, 10161022, https://doi.org/10.1175/1520-0450(1979)018<1016:LFIOAT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dudhia, J., 1989: Numerical study of convection observed during the Winter Monsoon Experiment using a mesoscale two-dimensional model. J. Atmos. Sci., 46, 30773107, https://doi.org/10.1175/1520-0469(1989)046<3077:NSOCOD>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dunkerton, T. J., M. T. Montgomery, and Z. Wang, 2009: Tropical cyclogenesis in a tropical wave critical layer: Easterly waves. Atmos. Chem. Phys., 9, 55875646, https://doi.org/10.5194/acp-9-5587-2009.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • ECMWF, 2010: WCRP and WWRP THORPEX YOTC (Year of Tropical Convection) Project. Research Data Archive, Computational and Information Systems Laboratory, National Center for Atmospheric Research, Boulder, CO, accessed 30 March 2017, https://doi.org/10.5065/D6R20ZDD.

    • Crossref
    • Export Citation
  • Elsberry, R. L., H.-C. Tsai, and M. S. Jordan, 2014: Extended-range forecasts of Atlantic tropical cyclone events during 2013 using the ECMWF 32-day ensemble predictions. Wea. Forecasting, 29, 271288, https://doi.org/10.1175/WAF-D-13-00104.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frank, W. M., and P. E. Roundy, 2006: The role of tropical waves in tropical cyclogenesis. Mon. Wea. Rev., 134, 23972417, https://doi.org/10.1175/MWR3204.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 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, 10671080, https://doi.org/10.1175/2011MWR3618.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gall, J. S., and W. M. Frank, 2010: The role of equatorial Rossby waves in tropical cyclogenesis. Part II: Idealized simulations in a monsoon trough environment. Mon. Wea. Rev., 138, 13831398, https://doi.org/10.1175/2009MWR3115.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gall, J. S., W. M. Frank, and M. C. Wheeler, 2010: The role of equatorial Rossby waves in tropical cyclogenesis. Part I: Idealized numerical simulations in an initially quiescent background environment. Mon. Wea. Rev., 138, 13681382, https://doi.org/10.1175/2009MWR3114.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Grell, G. A., and D. Dévényi, 2002: A generalized approach to parameterizing convection combining ensemble and data assimilation techniques. Geophys. Res. Lett., 29, 1693, https://doi.org/10.1029/2002GL015311.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hendricks, E. A., M. T. Montgomery, and C. A. Davis, 2004: On the role of “vortical” hot towers in the formation of Tropical Cyclone Diana (1984). J. Atmos. Sci., 61, 12091232, https://doi.org/10.1175/1520-0469(2004)061<1209:TROVHT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 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, 23182341, https://doi.org/10.1175/MWR3199.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Houze, R. A., Jr., and Coauthors, 2006: The Hurricane Rainband and Intensity Change Experiment: Observations and modeling of Hurricanes Katrina, Ophelia, and Rita. Bull. Amer. Meteor. Soc., 87, 15031521, https://doi.org/10.1175/BAMS-87-11-1503.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Houze, R. A., Jr., W.-C. Lee, and M. M. Bell, 2009: Convective contribution to the genesis of Hurricane Ophelia (2005). Mon. Wea. Rev., 137, 27782800, https://doi.org/10.1175/2009MWR2727.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Janjić, Z. I., 2000: Comments on “Development and evaluation of a convection scheme for use in climate models.” J. Atmos. Sci., 57, 36863686, https://doi.org/10.1175/1520-0469(2000)057<3686:CODAEO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jourdain, N. C., P. Marchesiello, C. E. Menkes, J. Lefévre, 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, 325, https://doi.org/10.1175/2010JCLI3559.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kain, J. S., and J. M. Fritsch, 1993: Convective parameterization for mesoscale models: The Kain–Fritsch scheme. The Representation of Cumulus Convection in Numerical Models, Meteor. Monogr., No. 46, Amer. Meteor. Soc., 165–170.

    • Crossref
    • Export Citation
  • Knapp, K. R., and Coauthors, 2011: Globally gridded satellite observations for climate studies. Bull. Amer. Meteor. Soc., 92, 893907, https://doi.org/10.1175/2011BAMS3039.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 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, 10351050, https://doi.org/10.1175/1520-0469(2001)058<1035:RWIZOM>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 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, 640654, https://doi.org/10.1175/1520-0434(1994)009<0640:DOAMGA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, C.-S., Y.-L. Lin, and K. K.-W. Cheung, 2006: Tropical cyclone formations in the South China Sea associated with the mei-yu front. Mon. Wea. Rev., 134, 26702687, https://doi.org/10.1175/MWR3221.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, C.-S., K. K. W. Cheung, J. S. N. Hui, and R. L. Elsberry, 2008: Mesoscale features associated with tropical cyclone formations in the western North Pacific. Mon. Wea. Rev., 136, 20062022, https://doi.org/10.1175/2007MWR2267.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 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, 14021424, https://doi.org/10.1175/WAF-D-14-00003.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Li, Z., Z. Pu, J. Sun, and W.-C. Lee, 2014: Impacts of 4DVAR assimilation of airborne Doppler radar observations on numerical simulations of the genesis of Typhoon Nuri (2008). J. Appl. Meteor. Climatol., 53, 23252343, https://doi.org/10.1175/JAMC-D-14-0046.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Liang, J., L. G. Wu, and H. J. Zhong, 2014: Idealized numerical simulations of tropical cyclone formation associated with monsoon gyres. Adv. Atmos. Sci., 31, 305315, https://doi.org/10.1007/s00376-013-2282-1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lim, K.-S. S., and S.-Y. 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, 15871612, https://doi.org/10.1175/2009MWR2968.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mlawer, E. J., S. J. Taubman, P. D. Brown, M. J. Iacono, and S. A. Clough, 1997: Radiative transfer for inhomogeneous atmospheres: RRTM, a validated correlated-k model for the longwave. J. Geophys. Res., 102, 16 66316 682, https://doi.org/10.1029/97JD00237.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Montgomery, M. T., M. E. Nicholls, T. A. Cram, and A. B. Saunders, 2006: A vortical hot tower route to tropical cyclogenesis. J. Atmos. Sci., 63, 355386, https://doi.org/10.1175/JAS3604.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Montgomery, M. T., L. L. Lussier III, R. W. Moore, and Z. Wang, 2010: The genesis of Typhoon Nuri as observed during the Tropical Cyclone Structure 2008 (TCS-08) field experiment—Part 1: The role of the easterly wave critical layer. Atmos. Chem. Phys., 10, 98799900, https://doi.org/10.5194/acp-10-9879-2010.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Montgomery, M. T., and Coauthors, 2012: The Pre-Depression Investigation of Cloud-Systems in the Tropics (PREDICT) Experiment: Scientific basis, new analysis tools, and some first results. Bull. Amer. Meteor. Soc., 93, 153172, https://doi.org/10.1175/BAMS-D-11-00046.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Nakano, M., M. Sawada, T. Nasuno, and M. Satoh, 2015: Intraseasonal variability and tropical cyclogenesis in the western North Pacific simulated by a global nonhydrostatic atmospheric model. Geophys. Res. Lett., 42, 565571, https://doi.org/10.1002/2014GL062479.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • NCEP, 2000: NCEP FNL Operational Model Global Tropospheric Analyses, continuing from July 1999. Research Data Archive, Computational and Information Systems Laboratory, National Center for Atmospheric Research, Boulder, CO, accessed 30 March 2017, https://doi.org/10.5065/D6M043C6.

    • Crossref
    • Export Citation
  • Park, M.-S., H.-S. Kim, C.-H. Ho, R. L. Elsberry, and M.-I. Lee, 2015: Tropical Cyclone Mekkhala’s (2008) formation over the South China Sea: Mesoscale, synoptic-scale, and large-scale contributions. Mon. Wea. Rev., 143, 88110, https://doi.org/10.1175/MWR-D-14-00119.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ritchie, E. A., and G. J. Holland, 1999: Large-scale patterns associated with tropical cyclogenesis in the western Pacific. Mon. Wea. Rev., 127, 20272043, https://doi.org/10.1175/1520-0493(1999)127<2027:LSPAWT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Skamarock, W., and Coauthors, 2008: A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, 113 pp., http://dx.doi.org/10.5065/D68S4MVH.

    • Crossref
    • Export Citation
  • Suzuki-Parker, A., 2012: Uncertainties and Limitations in Simulating Tropical Cyclones. Springer, 78 pp.

  • Teng, H.-F., 2016: Formation and development of tropical cloud cluster in the western North Pacific. Ph.D. thesis, National Taiwan University, 178 pp.

  • Thatcher, L., and Z. Pu, 2013: Evaluation of tropical cyclone genesis precursors with relative operating characteristics (ROC) in high-resolution ensemble forecasts: Hurricane Ernesto. Trop. Cyclone Res. Rev., 2, 131148.

    • Search Google Scholar
    • Export Citation
  • Tory, K. J., M. T. Montgomery, and N. E. Davidson, 2007: Prediction and diagnosis of tropical cyclone formation in an NWP system. Part III: Developing and nondeveloping storms. J. Atmos. Sci., 64, 31953213, https://doi.org/10.1175/JAS4023.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tsai, H.-C., R. L. Elsberry, M. S. Jordan, and F. Vitart, 2013: Objective verifications and false alarm analyses of western North Pacific tropical cyclone event forecasts by the ECMWF 32-day ensemble. Asia-Pac. J. Atmos. Sci., 49, 409420, https://doi.org/10.1007/s13143-013-0038-6.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, Z., M. T. Montgomery, and T. J. Dunkerton, 2010: Genesis of pre–Hurricane Felix (2007). Part I: The role of the easterly wave critical layer. J. Atmos. Sci., 67, 17111729, https://doi.org/10.1175/2009JAS3420.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wu, L., and J. Duan, 2015: Extended simulation of tropical cyclone formation in the western North Pacific monsoon trough. J. Atmos. Sci., 72, 44694485, https://doi.org/10.1175/JAS-D-14-0375.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wu, L., H. Zong, and J. Liang, 2013: Observational analysis of tropical cyclone formation associated with monsoon gyres. J. Atmos. Sci., 70, 10231034, https://doi.org/10.1175/JAS-D-12-0117.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Xu, Y., T. Li, and M. Peng, 2014: Roles of the synoptic-scale wave train, the intraseasonal oscillation, and high-frequency eddies in the genesis of Typhoon Manyi (2001). J. Atmos. Sci., 71, 37063722, https://doi.org/10.1175/JAS-D-13-0406.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhan, R. F., Y. Wang, and C.-C. Wu, 2011: Impact of SSTA in East Indian Ocean on the frequency of northwest Pacific tropical cyclones: A regional atmospheric model study. J. Climate, 24, 62276242, https://doi.org/10.1175/JCLI-D-10-05014.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zong, H., and L. Wu, 2015: Synoptic-scale influences on tropical cyclone formation within the western North Pacific monsoon trough. Mon. Wea. Rev., 143, 34213433, https://doi.org/10.1175/MWR-D-14-00321.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery

    Nested model domain for WRF simulation. The horizontal resolutions are 36 and 12 km for the larger and smaller domains, respectively.

  • View in gallery

    Scatter diagram of 10-day low-pass-filtered 850-hPa vorticity (×10−5 s−1, abscissa) vs 10-day high-pass-filtered 850-hPa vorticity (×10−5 s−1, ordinate) inside 5° radius at −24 to −48 h for 288 TCs during 2000–09 using NCEP-FNL data. Circles (black dots) indicate the TCs in 2008–09 (2000–07), gray dots indicate TCs Dujuan (2009) and Nuri (2008). Numbers in the corners indicate the TC number in 2000–09 for each quadrant, and the dashed line indicates the divide between HTC and LTC.

  • View in gallery

    Composites relative to the TC centers for the cloud-top temperature (K, shaded), 850-hPa vorticity (>5 × 10−5 s−1, blue contours with interval of 5 × 10−5 s−1), and 850-hPa wind vector (m s−1) for (a),(b) HTCs and (c),(d) LTCs at (left) −48 and (right) 0 h. Green contour indicates the wet area (averaged specific humidity at 1000–700 hPa > 0.014 kg kg−1). Distance from the TC center is in degrees latitude and longitude.

  • View in gallery

    Significant difference (pass 95% t test) between composite results of HTCs and LTCs at (left) −48 and (right) 0 h for cloud-top temperature (K, shaded) and wind vector (m s−1) at 850 hPa. Results are shown for (a),(b) unfiltered fields; (c),(d) 10-day high-pass-filtered fields; and (e),(f) 10-day low-pass-filtered fields. Unit for the abscissa (ordinate) is degrees longitude (latitude).

  • View in gallery

    Time series of the average track errors (km, solid line) and the median of the track errors (km, dashed lines) for simulations with different initial times (as labeled). Time 0 represents the time of TC formation.

  • View in gallery

    Stacked column chart (percentage of stack’s total) of three types of simulation results for (left) LTCs and (right) HTCs at different initial times.

  • View in gallery

    Stacked column chart (percentage of stack’s total) of three types of simulation results of cumulus-scheme experiments for (left) LLTCs and (right) HHTCs at different initial times.

  • View in gallery

    The 850-hPa vorticity (×10−5 s−1) and streamlines at 1800 UTC 16 Aug 2008 for the simulations started at (a) −48, (b) −72, (c) −96, and (d) −120 h for NURI (2008), with two initial conditions (shown on the right), and four cumulus schemes (from left to right are the Kain–Fritch, Betts–Miller–Janjić, Grell–Devenyi ensemble, and Grell 3D ensemble cumulus schemes). The red cross indicates the location of Nuri at 0 h in the ECMWF-YOTC analysis.

  • View in gallery

    The 850-hPa vorticity (×10−5 s−1) and streamlines at 1800 UTC 3 Sep 2009 for the simulations started at (a) −48, (b) −72, (c) −96, and (d) −120 h for Dujuan (2009), with two initial conditions (shown on the right), and four cumulus schemes (from left to right are the Kain–Fritch, Betts–Miller–Janjić, Grell–Devenyi ensemble, and Grell 3D ensemble cumulus schemes). Red rectangles mark the false TCs, and the red cross indicates the location of Dujuan at 0 h in the ECMWF-YOTC analysis.

  • View in gallery

    The 850-hPa wind vector (m s−1), 850-hPa vorticity (blue contours at 5, 10, and 50 × 10−5 s−1), and cloud-top temperature (K, shaded) based on GridSat data at (a) 0600 UTC 15 Aug and (d) 0600 UTC 16 Aug 2008. (b),(c),(e),(f) As in (d), but for model simulations with shading indicating the simulated reflectivity (the maximum reflectivity at grid column) using different cumulus schemes as labeled. All simulations are started at −48 h (1800 UTC 14 Aug 2008) using ECMWF-YOTC for the initial conditions.

  • View in gallery

    The 850-hPa wind vector (m s−1), 850-hPa vorticity (blue contours at 5, 10, and 50 × 10−5 s−1), and cloud-top temperature (K, shaded) based on GridSat data at (a) 0600 UTC 2 Sep and (d) 0600 UTC 3 Sep 2009. (b),(c),(e),(f) As in (d), but for model simulations with shading indicating the simulated reflectivity (the maximum reflectivity at grid column) using different cumulus schemes as labeled. All simulations are started at −48 h (1800 UTC 1 Sep 2009) using ECMWF-YOTC for the initial conditions.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 49 49 12
PDF Downloads 28 28 7

A Study on the Influences of Low-Frequency Vorticity on Tropical Cyclone Formation in the Western North Pacific

View More View Less
  • 1 Department of Atmospheric Sciences, National Taiwan University, Taipei, Taiwan
  • 2 Department of Atmospheric Sciences, National Taiwan University, and Taiwan Typhoon and Flood Research Institute, National Applied Research Laboratories, Taipei, Taiwan
  • 3 Department of Atmospheric Sciences, National Taiwan University, Taipei, Taiwan
© Get Permissions
Full access

Abstract

The WRF Model is used to simulate 52 tropical cyclones (TCs) that formed in the western North Pacific during 2008–09 to study the influence of the low-frequency mode of environmental vorticity on TC formation [Vmax ~ 25 kt (~13 m s−1)]. All simulations, using the same model setting, are repeated at four distinct initial times and with two different initial datasets. These TCs are classified into two groups based on the environmental 850-hPa low-frequency vorticity (using a 10-day low-pass filter) during the period 24–48 h prior to TC formation. Results show that the WRF Model is more capable of simulating the TC formation process, but with larger track errors for TCs formed in an environment with higher low-frequency vorticity (HTC). In contrast, the model is less capable of simulating the TC formation process for TCs formed in an environment with lower low-frequency vorticity (LTC), but with smaller track errors. Fourteen selected TCs are further simulated to examine the sensitivity of previous results to different cumulus parameterization schemes. Results show that the capability of the WRF Model to simulate HTC formation is not sensitive to the choice of cumulus scheme. However, for an LTC, the simulated convection pattern is very sensitive to the cumulus scheme used; therefore, model simulation capability for LTC depends on the cumulus scheme used. Results of this study reveal that the convection process is not a dominant factor in HTC formation, but is very important for LTC formation.

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

Corresponding author: Cheng-Shang Lee, cslee@ntu.edu.tw

Abstract

The WRF Model is used to simulate 52 tropical cyclones (TCs) that formed in the western North Pacific during 2008–09 to study the influence of the low-frequency mode of environmental vorticity on TC formation [Vmax ~ 25 kt (~13 m s−1)]. All simulations, using the same model setting, are repeated at four distinct initial times and with two different initial datasets. These TCs are classified into two groups based on the environmental 850-hPa low-frequency vorticity (using a 10-day low-pass filter) during the period 24–48 h prior to TC formation. Results show that the WRF Model is more capable of simulating the TC formation process, but with larger track errors for TCs formed in an environment with higher low-frequency vorticity (HTC). In contrast, the model is less capable of simulating the TC formation process for TCs formed in an environment with lower low-frequency vorticity (LTC), but with smaller track errors. Fourteen selected TCs are further simulated to examine the sensitivity of previous results to different cumulus parameterization schemes. Results show that the capability of the WRF Model to simulate HTC formation is not sensitive to the choice of cumulus scheme. However, for an LTC, the simulated convection pattern is very sensitive to the cumulus scheme used; therefore, model simulation capability for LTC depends on the cumulus scheme used. Results of this study reveal that the convection process is not a dominant factor in HTC formation, but is very important for LTC formation.

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

Corresponding author: Cheng-Shang Lee, cslee@ntu.edu.tw

1. Introduction

Tropical cyclone (TC) formation is a multiscale process during which a tropical disturbance over the ocean intensifies into a self-sustaining system. Previous studies have revealed that atmospheric conditions, including tropical waves (Frank and Roundy 2006; Ching et al. 2010; Gall et al. 2010; Gall and Frank 2010) and the synoptic environment (Ritchie and Holland 1999; Lee et al. 2006; Lee et al. 2008; Dunkerton et al. 2009; Chang et al. 2010), can play a significant role in TC formation. These conditions, either individually or combined, provide a favorable environment for the initial vortex to intensify to the stage of self-maintained convective vortex dynamics, that is, TC formation (Hendricks et al. 2004; Montgomery et al. 2006; Houze et al. 2009). However, it is difficult to identify the dominant processes from various possible TC formation mechanisms because of the complexity of multiple convection variables and tropical dynamics. Furthermore, TC formation generally occurs far offshore, inhibiting direct observations. Understanding the mechanisms driving TC formation in the western North Pacific (WNP) is particularly difficult because background flows in this region are more complicated.

Previous field campaigns studying TC formation, such as the Hurricane Rainband and Intensity Change Experiment (RAINEX; Houze et al. 2006), and the Pre-Depression Investigation of Cloud-Systems in the Tropics (PREDICT; Montgomery et al. 2012) experiment, have focused on the Atlantic. As such, observations over the WNP are scarce. Advanced numerical modeling is an effective approach for studying the physical mechanisms associated with TC formation. Detailed information regarding these mechanisms can be obtained using TC hindcasting and theoretical experiments (Liang et al. 2014; Xu et al. 2014). For example, Li and Pu (2014) used the nonhydrostatic Weather Research and Forecasting (WRF) Model to examine the sensitivity of Typhoon Nuri’s (2008) formation to different initial conditions. Their results showed that the WRF Model required appropriate data for the initial and lateral boundary conditions to accurately simulate TC formation. Furthermore, Li and Pu (2014) found that the replication of intense convection near the center of the wave pouch was necessary for the model to successfully simulate the formation of Typhoon Nuri, supporting the results of previous studies by Dunkerton et al. (2009) and Wang et al. (2010).

Results of previous numerical studies on TC formation have demonstrated that further research is required to fully understand the various driving mechanisms of TC formation. For example, Tsai et al. (2013) and Elsberry et al. (2014) showed that the European Centre for Medium-Range Weather Forecasts (ECMWF) 32-day ensemble forecast model resolved the formation of most TCs during 2009–10; however, some weak and short-lived TCs were not resolved. Nakano et al. (2015) also found that the formation of weak and short-lived TCs could be difficult to simulate or predict. However, they also determined that some systems could be successfully reproduced within 1–2 weeks of model integration, for example, Typhoon Songda (2004), which formed within an active monsoon shear line environment. These results suggest that the capability of a numerical model to accurately simulate the TC formation process is influenced by different atmospheric conditions. However, the predictability of TC formation using numerical simulations is complicated by the many natural and artificial factors involved, including model initial conditions, parameterization schemes, and settings used to perform the simulation.

In this study, we systematically simulate TC formation using the WRF Model and analyze the relationship between simulation results and environments to understand the dominant mechanisms for TC formation under certain conditions. To achieve this, TCs that formed in the WNP during 2008–09 are simulated using uniform model settings. The sensitivity of the WRF Model to the choice of cumulus scheme is then tested for a selection of TC cases during the 2008–09 seasons. The data used in this study, experimental design, and model settings are described in section 2. The classification of TCs is described in section 3, and the simulated results for each TC class are discussed in section 4. Results of sensitivity experiments using different cumulus parameterization schemes are presented in section 5. Finally, section 6 presents the discussion and our conclusions.

2. Data and experimental design

Four datasets were used in this study. First, TC data are taken from the Joint Typhoon Warning Center best track dataset (http://www.usno.navy.mil/NOOC/nmfc-ph/RSS/jtwc/best_tracks/wpindex.php). TC formation was defined as the time when the first 25-kt maximum wind speed was recorded (note that 1 kt = 0.5144 m s−1). Second, Gridded Satellite (GridSat-B1) infrared window (IR) channel data (Knapp et al. 2011) were used for determining cloud-top temperatures. Third, National Centers for Environmental Prediction (NCEP) Final Operational Global Analysis (FNL) data, which are available 6-hourly on 1° × 1° grids (NCEP 2000), were used as the source for the model initial conditions. Finally, ECMWF Year of Tropical Convection (YOTC) data on 0.25° × 0.25° grids (ECMWF 2010), available only for the period May 2008–April 2010, were used as a second source for the model initial conditions. During 2008–09, 55 TCs formed; however, 3 TCs were excluded from this analysis. Two TCs occurred before the YOTC data were available (prior to May 2008) and the third TC was excluded because the location of the circulation center determined using ECMWF-YOTC data varied unreasonably prior to TC formation.

The formation processes for the remaining 52 TCs during 2008 and 2009 were simulated using version 3.2 of the WRF-ARW Model (Skamarock et al. 2008) with two nested domains, and horizontal grid spacings of 36 and 12 km, respectively. The 36-km-resolution domain extended from eastern Africa to western North America (590 × 340 grid cells), and the inner domain covered the WNP (706 × 400 grid cells), as shown in Fig. 1. There are 31 levels in the vertical with the model top at 50 hPa. The average thicknesses are 0.07 and 1 km at the lower and upper levels, respectively.

Fig. 1.
Fig. 1.

Nested model domain for WRF simulation. The horizontal resolutions are 36 and 12 km for the larger and smaller domains, respectively.

Citation: Monthly Weather Review 145, 10; 10.1175/MWR-D-17-0085.1

All simulations (except for those in the cumulus sensitivity experiment) used the same physical parameterization schemes. These schemes were determined by referring to recently published studies of TC simulations using the WRF Model, including Crosbie and Serra (2014), Li et al. (2014), and Xu et al. (2014). They appear to be appropriate for the simulation of TC formation and include the Kain–Fritsch cumulus scheme (Kain and Fritsch 1993), the Yonsei University PBL scheme (Hong et al. 2006), WRF double-moment 6-class microphysics (Lim and Hong 2010), the Rapid Radiative Transfer Model longwave radiation scheme (Mlawer et al. 1997), and the Dudhia shortwave radiation scheme (Dudhia 1989). For each TC, two different sets of initial conditions and lateral boundary conditions (NCEP-FNL and ECMWF-YOTC data) were used and the simulation was initialized at four different times; 48, 72, 96, and 120 h before TC formation (for convenience, denoted as −48, −72, −96, and −120 h, respectively). Each simulation was run until 24 h (denoted as +24 h) after the observed TC formation time (denoted as 0 h). In summary, there were 8 simulations for each TC and 416 simulations in total.

3. Synoptic environments for TC formation

As discussed in section 1, differences in environmental conditions appear to affect the model performance when simulating TC formation. Therefore, TCs were classified into different groups based on the low-level atmospheric conditions during their formation. Previous studies have identified several synoptic patterns associated with TC formation, including monsoon gyres, monsoon shear lines, easterly waves, and trade wind surges (Lander 1994; Ritchie and Holland 1999; Kuo et al. 2001; Lee et al. 2008; Chang et al. 2010). Furthermore, TC formation is affected by the various time scales of atmospheric circulation systems, including the 10-day period of WNP monsoons (Wu et al. 2013) and the 3–8-day period of easterly waves (Fu et al. 2012). Rather than using the large-scale patterns associated with tropical cyclogenesis as discussed by Ritchie and Holland (1999), we applied a low-pass–high-pass filter with a 10-day cutoff period to both the NCEP-FNL and ECMWF-YOTC data to obtain the low- and high-frequency winds at 850 hPa (Duchon 1979; Wu et al. 2013). The average vorticity inside the 5° radius of the circulation center for each pre-TC disturbance during the period 24–48 h prior to TC formation is computed using the low- and high-frequency 850-hPa winds, and the results are termed the (relative) low- and high-frequency vorticities, respectively.

The scatter diagram of the computed low- versus high-frequency vorticity using NCEP-FNL data for 288 pre-TC disturbances during 2000–09 is shown in Fig. 2, in which circles represent systems that formed during 2008–09 and black dots are for those that formed during 2000–07. The mean value of low-frequency vorticity for all 2000–09 cases is 1.39 × 10−5 s−1, with a standard deviation of 0.84 × 10−5 s−1; while for the high-frequency vorticity, the mean value is 0.76 × 10−5 s−1, with a standard deviation of 0.68 × 10−5 s−1. Values range between −0.88 × 10−5 and 2.68 × 10−5 s−1 for high-frequency vorticity, and −1.12 × 10−5 and 4.32 × 10−5 s−1 for low-frequency vorticity. The values of low-frequency vorticity for pre-TC disturbances are spread over a much wider range when compared to those of high-frequency vorticity.

Fig. 2.
Fig. 2.

Scatter diagram of 10-day low-pass-filtered 850-hPa vorticity (×10−5 s−1, abscissa) vs 10-day high-pass-filtered 850-hPa vorticity (×10−5 s−1, ordinate) inside 5° radius at −24 to −48 h for 288 TCs during 2000–09 using NCEP-FNL data. Circles (black dots) indicate the TCs in 2008–09 (2000–07), gray dots indicate TCs Dujuan (2009) and Nuri (2008). Numbers in the corners indicate the TC number in 2000–09 for each quadrant, and the dashed line indicates the divide between HTC and LTC.

Citation: Monthly Weather Review 145, 10; 10.1175/MWR-D-17-0085.1

When the 55 pre-TC disturbances during 2008–09 are considered, results are similar (passing the F test at 95%, but not the t test at 95%). Therefore, for further analysis, TCs that formed during 2008–09 were classified into two groups: TCs formed in an environment with higher low-frequency vorticity (HTC) and TCs formed in an environment with lower low-frequency vorticity (LTC). A TC was classified as an HTC (LTC) if the value of its low-frequency vorticity was among the top (bottom) half of all cases. In addition to the three TCs previously excluded, a further eight TCs were excluded because they were classified into different groups when different analysis datasets were used (NCEP-FNL or ECMWF-YOTC). Therefore, there are 22 HTCs and 22 LTCs, with a low-frequency vorticity threshold of 1.15 × 10−5 s−1.

Composites of 6-hourly cloud-top temperatures and 850-hPa flow during −48 to 0 h over a domain of 30° × 30°, centered at the 850-hPa circulation center, were computed for HTCs and LTCs, respectively. Results show that HTCs and LTCs are embedded in quite different environments during their formation stage, as shown in Fig. 3. At −48 h, the distribution of cloud-top temperature (Tb) for HTCs (Fig. 3a) shows that active convection (<260 K) is located to the south of the circulation center and spread over an area approximately 30° wide. For LTCs, convection is spread over a much smaller area (approximately 10° wide) or confined to the proximity of the circulation center (Fig. 3c), and is weaker as compared to HTC convection. At 0 h (Figs. 3b,d), convection becomes more concentrated for HTCs and stronger near the center for LTCs; however, the patterns of convection are similar to those at −48 h for both LTCs and HTCs. Active convection is still spread over a wider area for HTCs compared to that of LTCs. Results also show that the pre-TC disturbances for HTCs are embedded in an environment with high specific humidity and a broad cyclonic circulation at 850 hPa (Fig. 3a). For LTCs, the environment is much drier and the cyclonic circulation only occurs over a limited area (Fig. 3c).

Fig. 3.
Fig. 3.

Composites relative to the TC centers for the cloud-top temperature (K, shaded), 850-hPa vorticity (>5 × 10−5 s−1, blue contours with interval of 5 × 10−5 s−1), and 850-hPa wind vector (m s−1) for (a),(b) HTCs and (c),(d) LTCs at (left) −48 and (right) 0 h. Green contour indicates the wet area (averaged specific humidity at 1000–700 hPa > 0.014 kg kg−1). Distance from the TC center is in degrees latitude and longitude.

Citation: Monthly Weather Review 145, 10; 10.1175/MWR-D-17-0085.1

Composite results show that HTCs and LTCs form in distinct synoptic environments. HTCs form in a monsoonlike environment, while LTCs form in an easterly wave–like environment. The characteristics of these two synoptic environments are similar to the typical features of monsoon-related (with a broad westerly wind to the south side) and easterly wave (with an inverse trough structure in the easterly wind) environments associated with TC formations, as discussed by Ritchie and Holland (1999). Although the synoptic environment of each individual TC is somewhat different from that of the composite, the composite results effectively demonstrate the general low-level features of HTCs and LTCs. Specifically, the composites highlight the contrasts between these two classifications of TCs.

Figure 3 also shows that the 850-hPa flow to the south of the circulation center at −48 h is predominantly westerly for HTCs, but more easterly for LTCs. This difference is more evident in the difference field (Fig. 4a) and persists up to the time of TC formation (Fig. 4b). This difference is significant using a t test at the 95% confidence level and is mainly due to the difference in the 10-day low-pass-filtered 850-hPa wind field, as shown in Figs. 4c–f. The differences in the high-pass-filtered 850-hPa wind fields between HTCs and LTCs at −48 and 0 h (Figs. 4c,d) are much smaller than those in the low-pass-filtered 850-hPa wind fields (Figs. 4e,f). The differences in the low-pass-filtered 850-hPa wind fields between HTCs and LTCs are also significant using a t test at the 95% confidence level.

Fig. 4.
Fig. 4.

Significant difference (pass 95% t test) between composite results of HTCs and LTCs at (left) −48 and (right) 0 h for cloud-top temperature (K, shaded) and wind vector (m s−1) at 850 hPa. Results are shown for (a),(b) unfiltered fields; (c),(d) 10-day high-pass-filtered fields; and (e),(f) 10-day low-pass-filtered fields. Unit for the abscissa (ordinate) is degrees longitude (latitude).

Citation: Monthly Weather Review 145, 10; 10.1175/MWR-D-17-0085.1

To test the effectiveness of our classification method, TCs were then classified into two groups based on the high-frequency vorticity of the system, instead of the low-frequency vorticity used in the previous analysis. Four TCs were excluded from the analysis because they were classified into different groups (TCs with higher high-frequency vorticity versus TCs with lower high-frequency vorticity) when different analysis datasets were used (NCEP-FNL or ECMWF-YOTC). Results (not shown) indicate that the synoptic environments for the two groups of TCs are similar, and there is no significant difference in the synoptic environment between the two groups. These results demonstrate that the low-frequency component of the wind field has a more substantial impact on the environmental conditions important to TC formation as compared to the high-frequency component.

4. Results of systematic simulations for HTCs and LTCs

To quantify the performance of the numerical model in simulating the process of TC formation, the presence of a TC must first be determined in each simulation. Although the criteria used to identify TC formation varies for different studies, the common aim is to detect a large cyclonic circulation system with a strong, concentrated 850-hPa positive vorticity at the center and to identify the time at which the TC forms (Jourdain et al. 2011; Zhan et al. 2011). In this study, the 850-hPa area-averaged vorticity inside the 1.5°, 3°, and 5° radii at 0 h for all TC cases are computed using the ECMWF-YOTC data. The vorticity value of a simulated TC should be higher than the minimum value of all TCs in the analysis dataset; therefore, the mean value of the computed area-averaged vorticity for all cases minus one standard deviation is used as the threshold to determine whether a TC has formed in the 12-km-resolution grid of the simulation. A model run is considered to have successfully simulated TC formation if the area-averaged vorticity of the simulated vortex meets two criteria for more than 12 h during the period from −12 to +12 h. The first criterion is that the area-averaged vorticity is greater than the threshold of 7.87 × 10−5 s−1 inside the 1.5° radius, and the second criterion is that the area-averaged vorticity exceeds 3.79 × 10−5 s−1 inside the 3° radius or 1.50 × 10−5 s−1 inside the 5° radius. A simulation that does not meet the above criteria is defined as “no TC formation” (No_TC).

For those simulations with TC formations, the 6-hourly track error (with respect to the observed track from the ECMWF-YOTC data) of every simulated TC is computed following Tsai et al. (2013). Figure 5 shows the mean track error for all simulated TCs for simulations initialized at −48, −72, −96, and −120 h. The track error of a TC is considered unreasonable if its value is larger than the mean value; otherwise, it is considered reasonable. The mean track errors of all simulated TCs at 0 h are 249, 301, 441, and 600 km for simulations initialized at −48, −72, −96, and −120 h, respectively. All simulations with TC formations are further classified into two groups based on the track errors of the simulated TCs. The first group includes simulations with continuous unreasonable track errors exceeding 12 h (for three or more continuous 6-hourly time periods), or having unreasonable track errors during four or more 6-hourly time periods during the TC formation period (−48 to +12 h). These simulations are classified as “Large_error.” The remaining simulations are considered to have simulated TC formation (probably simulated) and are labeled as “Simulated_P.”

Fig. 5.
Fig. 5.

Time series of the average track errors (km, solid line) and the median of the track errors (km, dashed lines) for simulations with different initial times (as labeled). Time 0 represents the time of TC formation.

Citation: Monthly Weather Review 145, 10; 10.1175/MWR-D-17-0085.1

The 352 simulations for the 44 TCs are classified into three groups as shown in Table 1. Approximately 32% of the 88 simulations started at −48 h are classified as Simulated_P. It is reasonable that the percentage of Simulated_P decreases as the simulations are initialized at earlier times. Consequently, the percentages of simulations classified as Simulated_P are 28%, 26%, and 27% for simulations initialized at −72, −96, and −120 h, respectively. The percentages of simulations classified as No_TC are 25%, 25%, 28%, and 32% for simulations initialized at −48, −72, −96, and −120 h, respectively. As expected, these percentages increase as the integration time becomes longer.

Table 1.

Classification of simulation results for each TC: O, Simulated_P; L, Large_error; and —, No_TC.

Table 1.

Approximately two-thirds (65%) of all No_TC simulations use NCEP-FNL data for the initial conditions. These results indicate that the overall capability of the WRF Model to simulate TC formation is not only affected by differences in the initialization time, but also by the initial conditions. This is similar to the findings of Li and Pu (2014), who showed that simulation results using the WRF Model were sensitive to the initial data used in simulations of Typhoon Nuri (2008). This can be attributed to the fact that the NCEP-FNL dataset has a coarser grid compared to that of the ECMWF-YOTC dataset. Table 1 also shows that simulations classified as No_TC or Large_error tend to be for specific TCs. For example, all eight simulations of Nuri (2008) are classified as No_TC, and all eight simulations of Morakot (2009) are classified as Large_error. It is important to note that the low-frequency vorticities of Nuri are relatively smaller than those of Morakot, which are 0.37 × 10−5 and 3.65 × 10−5 s−1, respectively, suggesting that the performance of the numerical model in simulating TC formation is affected by the low-frequency vorticity associated with the pre-TC disturbance. It has to be noted that different thresholds in determining a simulated TC (i.e., the minimum value of all cases or the mean value plus one standard deviation) were tested, and the relative numbers of No_TC results were similar.

Among the 28 Simulated_P simulations initialized at −48 h, 11 are for LTC cases, accounting for 25% of all 44 simulations for LTC cases, and 17 are for HTC cases, accounting for 39% of all simulations, as shown in Fig. 6. These percentages vary only slightly for simulations with different initial times. For LTCs, they are 30%, 20%, and 25% for simulations initialized at −72, −96, and −120 h, respectively. For HTCs, they are 27%, 32%, and 27% for simulations initialized at −72, −96, and −120 h, respectively (as shown in Fig. 6). However, most No_TC simulations are for LTCs. For example, of the 88 simulations initialized at −48 h, 22 are No_TC simulations and 86% of those (19/22) are LTCs.

Fig. 6.
Fig. 6.

Stacked column chart (percentage of stack’s total) of three types of simulation results for (left) LTCs and (right) HTCs at different initial times.

Citation: Monthly Weather Review 145, 10; 10.1175/MWR-D-17-0085.1

For HTCs, 92% of the 176 simulations can simulate TC formation; however, this percentage varies slightly for simulations with different initial times (Fig. 6). For LTCs, this percentage is only 54%, which is substantially lower than that of HTCs. Moreover, among the 92% of simulations that can simulate the TC formation process for HTC, 66% are classified as Large_error. This percentage is higher than that of LTCs, which is at 54%. These results indicate that simulation results concerning TC formation are affected by the environment in which TC formation occurs. It is more likely that the WRF Model will successfully simulate TC formation in an environment with a larger low-frequency vorticity (similar to a monsoon system); however, the simulated TC tends to have a larger track error (Fig. 6). Conversely, the probability that the WRF Model will successfully simulate TC formation in an environment with smaller low-frequency vorticity (an easterly wave–like environment) is lower; however, the track error tends to be smaller. Furthermore, there is a positive correlation (r = 0.4) between the low-frequency vorticity in the analysis data (ECMWF-YOTC) and the mean vorticity errors (No_TC results are excluded) for all 52 TCs at 0 h. Therefore, the mean intensity errors for different areas (inside 1.5°, 3°, and 5° radii) are also larger for HTCs than for LTCs (pass t test at 95%). In other words, the overall simulation results using the intensity errors as the classification criteria might be similar to current results. However, using the intensity errors as the classifications criteria is likely not appropriate in this study because there are more complicated processes involved in the TC intensity changes (such as the effect of a nearby TC, the landmass and convection patterns, etc.), which deserve further study in the future.

Similar analyses, with TCs classified into two groups based on the high-frequency vorticity of the system, have also been applied to the simulation results. These results (not shown) reveal that there is almost no difference between the two classifications of TCs regarding the percentage of each type of simulated result (No_TC, Simulated_P, and Large_error). These results suggest that high-frequency vorticity only affects the accuracy of the simulated TC (location, intensity, and convection), but not the actual occurrence of TC formation. These results further support our classification of TCs according to low-frequency vorticity.

5. Cumulus scheme experiments

The results discussed in section 4 are based on simulations using the model parameterization options described in section 2; therefore, these results may differ if other parameterization schemes are used. Because simulation results are likely to be sensitive to the choice of cumulus scheme, an additional set of experiments (CU_EXP) is carried out for 14 selected TCs. These experiments vary the cumulus scheme while using the same model settings, initial conditions, and initial times discussed in section 4. In addition to the Kain–Fritsch scheme used previously, three other cumulus schemes are evaluated, namely the Betts–Miller–Janjić scheme (Janjić 2000), the Grell–Dévényi ensemble scheme, and the Grell 3D ensemble scheme (Grell and Dévényi 2002). Therefore, there are 24 more simulations for each of the 14 selected TCs in the CU_EXP experiments. Seven selected cases are HTCs with the highest low-frequency vorticity during TC formation, called HHTCs, and another seven selected cases are LTCs with the lowest low-frequency vorticity during TC formation, called LLTCs (Table 2).

Table 2.

Classifications and names of TCs in CU_EXP.

Table 2.

For the CU_EXP experiments, the simulation classification criteria and analysis procedures are the same as those used in section 4. Stacked column charts displaying the results for these 14 TCs are shown in Fig. 7. The overall classification proportions for the HHTCs are similar to those of the HTCs (Fig. 6). However, the overall classification proportions for the LLTCs differ from those of the LTCs. Specifically, the percentage of No_TC results is 75% for LLTCs and 46% for LTCs, and the percentage of Simulated_P results is 12% for LLTCs and 25% for LTCs. However, the percentage of Simulated_P results for HHTCs is only slightly larger than that for HTCs (33% compared to 31%), and the percentage of No_TC simulations is approximately the same for HHTCs and HTCs (9.0% and 8.0%). In addition, more than 81% of Large_error results are for HHTCs and 89% of No_TC results are for LLTCs.

Fig. 7.
Fig. 7.

Stacked column chart (percentage of stack’s total) of three types of simulation results of cumulus-scheme experiments for (left) LLTCs and (right) HHTCs at different initial times.

Citation: Monthly Weather Review 145, 10; 10.1175/MWR-D-17-0085.1

For simulations initialized at −48 and −72 h, the percentage of No_TC results for LLTCs (73% and 79%) is almost double that of LTCs (43% and 39%). These differences can be attributed to difficulties with the WRF Model in replicating TC formation for the extreme TCs (having very small low-frequency vorticities) simulated for CU_EXP. Consequently, the proportion of No_TC results increases significantly for LLTCs.

These results indicate that WRF can simulate the TC formation process in an environment with large low-frequency vorticity, using a range of cumulus schemes. Hence, the choice of cumulus scheme is not a key factor for WRF when simulating TC formation in an environment with large low-frequency vorticity. This suggests that, while it is a required process, convection is not a major forcing mechanism for this type of TC formation. However, for TCs that have formed in an environment with small low-frequency vorticity, the WRF Model is less capable of simulating TC formation. Furthermore, the capability of the model to replicate the formation of this type of TC is heavily dependent on the choice of cumulus scheme.

a. Simulation results for selected cases: Nuri (2008) and Dujuan (2009)

Simulations using different cumulus schemes for the same TC case are likely to share a similar synoptic-scale evolution pattern; however, the track and vorticity of the simulated disturbance can differ markedly. To account for this, simulation results for the formation times of Nuri (2008) and Dujuan (2009), which represent an LTC and HTC, respectively, are shown in Figs. 8 and 9 to explore the similarities and differences between simulations and TCs. Only 3 of the 32 CU_EXP simulations replicate TC formation for Nuri, while the majority (30 of 32) of CU_EXP simulations reproduce TC formation for Dujuan. It should be noted that Nuri and Dujuan are both extreme TC cases, located at the extremities of Fig. 2, with area-averaged low-frequency vorticities inside the 5° radius of 0.37 × 10−5 and 3.95 × 10−5 s−1, respectively, but with similar high-frequency vorticities inside the 5° radius (0.25 × 10−5 and 0.29 × 10−5 s−1).

Fig. 8.
Fig. 8.

The 850-hPa vorticity (×10−5 s−1) and streamlines at 1800 UTC 16 Aug 2008 for the simulations started at (a) −48, (b) −72, (c) −96, and (d) −120 h for NURI (2008), with two initial conditions (shown on the right), and four cumulus schemes (from left to right are the Kain–Fritch, Betts–Miller–Janjić, Grell–Devenyi ensemble, and Grell 3D ensemble cumulus schemes). The red cross indicates the location of Nuri at 0 h in the ECMWF-YOTC analysis.

Citation: Monthly Weather Review 145, 10; 10.1175/MWR-D-17-0085.1

Fig. 9.
Fig. 9.

The 850-hPa vorticity (×10−5 s−1) and streamlines at 1800 UTC 3 Sep 2009 for the simulations started at (a) −48, (b) −72, (c) −96, and (d) −120 h for Dujuan (2009), with two initial conditions (shown on the right), and four cumulus schemes (from left to right are the Kain–Fritch, Betts–Miller–Janjić, Grell–Devenyi ensemble, and Grell 3D ensemble cumulus schemes). Red rectangles mark the false TCs, and the red cross indicates the location of Dujuan at 0 h in the ECMWF-YOTC analysis.

Citation: Monthly Weather Review 145, 10; 10.1175/MWR-D-17-0085.1

For Nuri, the low-level (850 hPa) features of the 32 simulation results (Fig. 8) are diverse and there is no TC-like vortex in many of the simulations, especially those using the Kain–Fritsch scheme. Simulations using the Betts–Miller–Janjić cumulus scheme tend to have stronger vorticity, and two simulations that successfully replicate TC formation use this scheme (note that only three simulations successfully replicate TC formation for Nuri). All experiments have clear cyclonic circulation at 850 hPa for the Dujuan case, and the simulation results using the Betts–Miller–Janjić scheme produced the highest vorticity at 850 hPa (Fig. 9). The numbers of simulation results classified as Simulated_P, Large_error, and No_TC are 13, 17, and 2, respectively. Despite the success of the WRF Model in replicating TC formation in this case, the positions (and tracks) of the simulated TCs differ markedly between simulations.

Typhoon Nuri (2008) formed in an easterly wave environment (Montgomery et al. 2010). Thirty-six hours before TC formation (0600 UTC 15 August 2008), the pre-Nuri disturbance was a weak inverse trough with a small area of convection located near 16°N and 152°E, as shown in Fig. 10a. The system moved westward and intensified significantly during the following 24 h. An 850-hPa cyclonic circulation accompanied by convection was observed at 0600 UTC 16 August (Fig. 10d). On the other hand, Tropical Storm Dujuan (2009) formed in an environment with a large and long-lasting 850-hPa cyclonic circulation pattern prior to TC formation. In contrast to the pre-Nuri disturbance, a large low-level cyclonic circulation pattern with vorticity greater than 5 × 10−5 s−1 existed 36 h prior to the formation of Dujuan (0600 UTC 2 September 2009; Fig. 11a). Convection was spread over a broader region and was better organized than that of Nuri. During the following 24 h, strong convection became more concentrated on the southwestern side of the circulation center (Fig. 11d) with a more concentrated strong vorticity field. The environments of the pre-Nuri and pre-Dujuan disturbances are similar to the composites of LTCs and HTCs (shown in Fig. 3), indicating that they are typical LTC and HTC cases, respectively.

Fig. 10.
Fig. 10.

The 850-hPa wind vector (m s−1), 850-hPa vorticity (blue contours at 5, 10, and 50 × 10−5 s−1), and cloud-top temperature (K, shaded) based on GridSat data at (a) 0600 UTC 15 Aug and (d) 0600 UTC 16 Aug 2008. (b),(c),(e),(f) As in (d), but for model simulations with shading indicating the simulated reflectivity (the maximum reflectivity at grid column) using different cumulus schemes as labeled. All simulations are started at −48 h (1800 UTC 14 Aug 2008) using ECMWF-YOTC for the initial conditions.

Citation: Monthly Weather Review 145, 10; 10.1175/MWR-D-17-0085.1

Fig. 11.
Fig. 11.

The 850-hPa wind vector (m s−1), 850-hPa vorticity (blue contours at 5, 10, and 50 × 10−5 s−1), and cloud-top temperature (K, shaded) based on GridSat data at (a) 0600 UTC 2 Sep and (d) 0600 UTC 3 Sep 2009. (b),(c),(e),(f) As in (d), but for model simulations with shading indicating the simulated reflectivity (the maximum reflectivity at grid column) using different cumulus schemes as labeled. All simulations are started at −48 h (1800 UTC 1 Sep 2009) using ECMWF-YOTC for the initial conditions.

Citation: Monthly Weather Review 145, 10; 10.1175/MWR-D-17-0085.1

b. Comparisons between simulation results using different cumulus schemes

Results of CU_EXP experiments also show that the percentages of the 112 simulations (14 TCs, four initial times, and two datasets) that can simulate TC formation are 64%, 77%, 45%, and 47% for simulations using the Kain–Fritch, Betts–Miller–Janjić, and Grell–Dévényi ensemble, as well as the Grell 3D (G3D) ensemble scheme, respectively. Note that simulations using the Kain–Fritch and Betts–Miller–Janjić schemes tend to produce stronger convection; thus, the percentage is higher for simulations using these two schemes (especially for LLTCs, for which over 90% of the simulation results are classified as No_TC when Grell and G3D schemes are used). However, there are several simulations (especially for HTCs) using the Kain–Fritch and Betts–Miller–Janjić schemes that produce false TC formation as indicated in Fig. 9.

To explore the impact of the choice of cumulus scheme on the simulated convection and circulation patterns, results of simulations initialized at −48 h using different cumulus schemes are analyzed for Nuri and Dujuan. Simulations for both cases use ECMWF-YOTC initial conditions. The simulations start at 1800 UTC 16 August 2008 for Nuri and at 1800 UTC 3 September 2009 for Dujuan.

Results of 12-h forecasts (valid at 0600 UTC 15 August, or −36 h) show that all simulations for Nuri can replicate the inverse open trough observed near 16°N and 152°E (Fig. 10a). However, the convection pattern differs markedly between simulations using different cumulus schemes. For example, the simulation using the Kain–Fritsch scheme produces a reasonable convection pattern, while simulations using the Grell–Dévényi and Grell 3D ensemble schemes (Betts–Miller–Janjić scheme) tend to produce weaker (stronger) convection compared to the observations. At 0600 UTC 16 August 2008 (−12 h), the simulated convection and vorticity are still weaker than observations for simulations using the Grell–Dévényi and Grell 3D ensemble schemes; therefore, the vorticity of the simulated pre-TC disturbance does not reach the threshold value (Figs. 10e,f). For the simulation using the Kain–Fritch scheme, the overall pattern of simulated circulation and convection appears reasonable; however, active convection only occurs in a limited region to the north of the trough (Fig. 10b), which is not consistent with the observations. As a result, the simulated cyclonic circulation does not develop into a TC using the Kain–Fritsch scheme. TC formation only occurs in the simulation using the Betts–Miller–Janjić scheme (Fig. 10c).

Simulation results for Nuri are consistent with those of Li et al. (2014) and Li and Pu (2014). They indicate that the assimilation of airborne Doppler radar data and the finer initial conditions (especially the low-level water vapor) would lead to convection occurring near the center of the pre-TC disturbance and could help the formation of Nuri in the WRF simulations. Our results support this finding.

For Dujuan, the overall patterns of simulated convection appear similar (but with different intensity) for simulations using the different cumulus schemes and the patterns match the observations. The strongest convection is located on the southwestern side of the circulation center (17°N, 130°E) at −36 h. Convection either becomes stronger or spreads over a larger area during the following 24 h, and is better organized at −12 h, as shown in Figs. 11b,c,e,f. Similar to those in Nuri’s simulations, the simulated convection is stronger (weaker) than observed for simulations using the Betts–Miller–Janjić scheme (Grell–Dévényi and Grell 3D ensemble schemes). However, in this case, all simulations can simulate the TC formation. Aside from the differences in the intensity of simulated convection, the vorticities of the simulated TCs at −12 h (the 36-h forecast) for simulations using the Grell–Dévényi and Grell 3D ensemble schemes (Figs. 11e,f) are much weaker than the observations. Additionally, the centers for the simulated TCs for these two simulations are located far from the observed location; thus, these two simulations are classified as Large_error.

These results indicate that the location and strength of convection prior to TC formation is very important to enabling the model to simulate the formation of Nuri (an LTC) but is less important for the model to simulate the formation of Dujuan (an HTC). Results are similar for the other six LTCs and six HTCs, analyzed for CU_EXP. Therefore, it is reasonable to conclude that simulation results regarding TC formation are highly (less) sensitive to simulated convections for LTCs (HTCs). Furthermore, the WRF Model can successfully simulate the formation of a HTC using a variety of cumulus schemes; however, not all cumulus schemes facilitate the formation of an LTC. An appropriate cumulus scheme is required to simulate LTC formation.

6. Discussion and conclusions

Following previous studies (e.g., Ritchie and Holland 1999; Lee et al. 2008; Teng 2016), this study aims to advance our understanding regarding the environmental influences on TC formation in the WNP by means of systematic simulations of TC formation using the WRF Model. The simulations used fixed model domains and settings. Two different datasets, ECMWF-YOTC and NCEP-FNL, were used for the initial and boundary conditions. The simulations were initialized at four different times (48, 72, 96, and 120 h) prior to TC formation and were run for up to 24 h after formation. Therefore, eight simulations with four different integration times were produced for TCs that occurred during the period of 2008–09, totaling 416 simulations for 52 TCs. This approach minimized the impact of artificial factors on the simulation results and allowed for the analysis of statistical relationships between the simulation results and synoptic environments.

To highlight the important environmental factors influencing TC formation, all TCs were classified into two groups (HTC and LTC) based on the environmental low-frequency vorticity prior to TC formation. This approach follows the findings of Ching et al. (2010), Wu and Duan (2015), and Zong and Wu (2015), who identified the importance of low-frequency vorticity on TC formation. Analysis of the systematic simulation results shows that for HTCs that form in a monsoon-related system (with higher low-frequency vorticity) the WRF Model demonstrated a high degree of skill in simulating TC formation; however, the simulated TCs tended to have large track errors. In contrast, the WRF Model was less capable of simulating LTC formation; however, the track errors of simulated TCs tended to be smaller.

To address the effect of the cumulus parameterization scheme on the simulation results, cumulus experiments (CU_EXP) were conducted. Seven extreme HTCs and seven extreme LTCs (called HHTCs and LLTCs) were simulated using three alternate cumulus schemes (Betts–Miller–Janjić, Grell–Dévényi, and Grell 3D) in addition to the Kain–Fritsch scheme used in the previous systematic simulations. An additional 336 simulations (CU_EXP) were conducted to test the sensitivity of TC formation to the choice of cumulus scheme. The analysis procedures for CU_EXP followed those of the previous systematic simulations. The overall simulation results for HHTC were similar to those for HTCs, but the overall simulation results for LLTCs differed from those for LTCs. More than 80% of the simulations with large track errors were for HHTCs and 89% of No_TC simulations were for LLTCs. Furthermore, the percentage of No_TC simulation results for LLTCs was much greater than that for LTCs.

The CU_EXP simulation results for Nuri (2008) and Dujuan (2009), representing an LTC and HTC case, respectively, were also analyzed. Results showed that the intensity of simulated convection was very sensitive to the choice of the cumulus scheme. In the case of TC Nuri, which formed in an environment with smaller low-frequency vorticity, the choice of cumulus scheme is critical for simulating the increase in low-level vorticity. Hence, the cumulus scheme and initial conditions are critical for determining whether a disturbance can develop to TC intensity. Previous studies, including the work of Cheung and Elsberry (2006), Li et al. (2014), and Li and Pu (2014), have also determined that an appropriate cumulus scheme and initial conditions are vital for accurate TC simulation. However, for Dujuan, an HTC case that formed in an environment with large low-frequency vorticity, the overall pattern of convection was less sensitive to the choice of cumulus scheme. In this case, all simulations were able to simulate the formation of a TC, regardless of cumulus scheme; however, the simulated TCs displayed different intensities (vorticities) and track errors.

These results indicate that the choice of cumulus scheme is not a key factor for the WRF Model when simulating TC formation in an environment with large low-frequency vorticity; however, the TC track error may be large. In other words, convection does not play a dominant role in the formation of this type of TC. However, as found by Tory et al. (2007), it does affect the timing and location of TC center development. Based on our analysis and on previous studies (e.g., Zong and Wu 2015), we hypothesize that in a monsoon-related environment where the background low-frequency vorticity is large, organized convection can easily concentrate vorticity, leading to the formation of a TC. Thus, the WRF Model can more readily simulate TC formation in this environment. However, such an environment is also conducive to the initiation of convection (Lee et al. 2008; Chang 2013) and the stochastic nature of convection results in larger simulated TC track errors, similar to the argument of Tory et al. (2007). Furthermore, Wu and Duan (2015) found that TC formation only occurred when using the low-frequency background as the initial conditions. Therefore, interactions between the pre-TC disturbance and its environmental low-frequency background appear to be critical for the formation of this type of TC (Zong and Wu 2015). It is also important to understand the role of tropical waves, synoptic perturbations, and convection processes on HTC formation, as has been established by Park et al. (2015) for Tropical Storm Mekkhala (2008).

On the other hand, the WRF Model is less capable of simulating TC formation in an environment with small low-frequency vorticity. In this environment, the capability of the WRF Model to simulate TC formation is strongly influenced by the choice of cumulus scheme. This suggests that, for LTCs, convection might represent a critical low-level vorticity source, which is strongly affected by the choice of cumulus scheme. These results may explain why some weak and short-lived TCs are difficult to simulate or predict in global models during TC formation, as was found by Tsai et al. (2013), Elsberry et al. (2014), and Nakano et al. (2015). It is likely that convection processes are crucial for the formation of these weak systems, but not only in the WNP, because similar simulation results have been found for Atlantic easterly wave cases (Wang et al. 2010; Suzuki-Parker 2012; Thatcher and Pu 2013). However, the stochastic nature of convection makes it difficult for models to accurately simulate the convection process during TC formation, such that the model tends to easily miss the formation of these TCs. Therefore, it is important to investigate convective processes during TC formation to understand the key forcing mechanisms for the formation of this type of system.

Our results show that the percentage of simulations that can simulate TC formation, is higher for simulations using the Kain–Fritch or Betts–Miller–Janjić schemes than those using the Grell–Dévényi or Grell 3D ensemble schemes (especially for LTCs). However, the Kain–Fritsch and Betts–Miller–Janjić schemes are not necessarily more suitable for simulating TC formation because these schemes simulate several false TC formations owing to more active convection (especially for HTCs). Therefore, understanding the detailed evolution of simulated convection using different cumulus schemes, as well as their effects on TC formation, is important for determining the key factors and triggering mechanisms for TC formation. An understanding of these processes will form the basis of future investigations.

Acknowledgments

CS Lee is supported by the National Taiwan University and the Taiwan Typhoon Flood Research Institute, National Applied Research Laboratories, and CH Sui and YH Hsieh are supported by the National Taiwan University. This research is supported by the Ministry of Science and Technology (MOST) of the Republic of China (Taiwan) under Grants MOST 102-2111-M-002-001-, MOST 103-2111-M-002-003-, MOST 104-2111-M-002-003-, MOST 105-2111-M-002-004-, and MOST 106-2111-M-002-009-.

REFERENCES

  • Chang, L.-Y., 2013: Tropical cyclone formation associated with trade-wind surges. Ph.D. thesis, National Taiwan University, 206 pp.

  • Chang, L.-Y., K. K. W. Cheung, and C.-S. Lee, 2010: The role of trade wind surges for tropical cyclone formations in the western North Pacific. Mon. Wea. Rev., 138, 41204134, https://doi.org/10.1175/2010MWR3152.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cheung, K. K. W., and R. L. Elsberry, 2006: Model sensitivities in numerical simulations of the formation of Typhoon Robyn (1993). Terr. Atmos. Oceanic Sci., 17, 5389, https://doi.org/10.3319/TAO.2006.17.1.53(SWS).

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ching, L., C.-H. Sui, and M.-J. Yang, 2010: An analysis of the multiscale nature of tropical cyclone activities in June 2004: Climate background. J. Geophys. Res., 115, D24108, https://doi.org/10.1029/2010JD013803.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Crosbie, E., and Y. L. Serra, 2014: Intraseasonal modulation of synoptic-scale disturbances and tropical cyclone genesis in the eastern North Pacific. J. Climate, 27, 57245745, https://doi.org/10.1175/JCLI-D-13-00399.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Duchon, C. E., 1979: Lanczos filtering in one and two dimensions. J. Appl. Meteor., 18, 10161022, https://doi.org/10.1175/1520-0450(1979)018<1016:LFIOAT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dudhia, J., 1989: Numerical study of convection observed during the Winter Monsoon Experiment using a mesoscale two-dimensional model. J. Atmos. Sci., 46, 30773107, https://doi.org/10.1175/1520-0469(1989)046<3077:NSOCOD>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dunkerton, T. J., M. T. Montgomery, and Z. Wang, 2009: Tropical cyclogenesis in a tropical wave critical layer: Easterly waves. Atmos. Chem. Phys., 9, 55875646, https://doi.org/10.5194/acp-9-5587-2009.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • ECMWF, 2010: WCRP and WWRP THORPEX YOTC (Year of Tropical Convection) Project. Research Data Archive, Computational and Information Systems Laboratory, National Center for Atmospheric Research, Boulder, CO, accessed 30 March 2017, https://doi.org/10.5065/D6R20ZDD.

    • Crossref
    • Export Citation
  • Elsberry, R. L., H.-C. Tsai, and M. S. Jordan, 2014: Extended-range forecasts of Atlantic tropical cyclone events during 2013 using the ECMWF 32-day ensemble predictions. Wea. Forecasting, 29, 271288, https://doi.org/10.1175/WAF-D-13-00104.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frank, W. M., and P. E. Roundy, 2006: The role of tropical waves in tropical cyclogenesis. Mon. Wea. Rev., 134, 23972417, https://doi.org/10.1175/MWR3204.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 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, 10671080, https://doi.org/10.1175/2011MWR3618.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gall, J. S., and W. M. Frank, 2010: The role of equatorial Rossby waves in tropical cyclogenesis. Part II: Idealized simulations in a monsoon trough environment. Mon. Wea. Rev., 138, 13831398, https://doi.org/10.1175/2009MWR3115.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gall, J. S., W. M. Frank, and M. C. Wheeler, 2010: The role of equatorial Rossby waves in tropical cyclogenesis. Part I: Idealized numerical simulations in an initially quiescent background environment. Mon. Wea. Rev., 138, 13681382, https://doi.org/10.1175/2009MWR3114.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Grell, G. A., and D. Dévényi, 2002: A generalized approach to parameterizing convection combining ensemble and data assimilation techniques. Geophys. Res. Lett., 29, 1693, https://doi.org/10.1029/2002GL015311.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hendricks, E. A., M. T. Montgomery, and C. A. Davis, 2004: On the role of “vortical” hot towers in the formation of Tropical Cyclone Diana (1984). J. Atmos. Sci., 61, 12091232, https://doi.org/10.1175/1520-0469(2004)061<1209:TROVHT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 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, 23182341, https://doi.org/10.1175/MWR3199.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Houze, R. A., Jr., and Coauthors, 2006: The Hurricane Rainband and Intensity Change Experiment: Observations and modeling of Hurricanes Katrina, Ophelia, and Rita. Bull. Amer. Meteor. Soc., 87, 15031521, https://doi.org/10.1175/BAMS-87-11-1503.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Houze, R. A., Jr., W.-C. Lee, and M. M. Bell, 2009: Convective contribution to the genesis of Hurricane Ophelia (2005). Mon. Wea. Rev., 137, 27782800, https://doi.org/10.1175/2009MWR2727.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Janjić, Z. I., 2000: Comments on “Development and evaluation of a convection scheme for use in climate models.” J. Atmos. Sci., 57, 36863686, https://doi.org/10.1175/1520-0469(2000)057<3686:CODAEO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jourdain, N. C., P. Marchesiello, C. E. Menkes, J. Lefévre, 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, 325, https://doi.org/10.1175/2010JCLI3559.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kain, J. S., and J. M. Fritsch, 1993: Convective parameterization for mesoscale models: The Kain–Fritsch scheme. The Representation of Cumulus Convection in Numerical Models, Meteor. Monogr., No. 46, Amer. Meteor. Soc., 165–170.

    • Crossref
    • Export Citation
  • Knapp, K. R., and Coauthors, 2011: Globally gridded satellite observations for climate studies. Bull. Amer. Meteor. Soc., 92, 893907, https://doi.org/10.1175/2011BAMS3039.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 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, 10351050, https://doi.org/10.1175/1520-0469(2001)058<1035:RWIZOM>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 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, 640654, https://doi.org/10.1175/1520-0434(1994)009<0640:DOAMGA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, C.-S., Y.-L. Lin, and K. K.-W. Cheung, 2006: Tropical cyclone formations in the South China Sea associated with the mei-yu front. Mon. Wea. Rev., 134, 26702687, https://doi.org/10.1175/MWR3221.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, C.-S., K. K. W. Cheung, J. S. N. Hui, and R. L. Elsberry, 2008: Mesoscale features associated with tropical cyclone formations in the western North Pacific. Mon. Wea. Rev., 136, 20062022, https://doi.org/10.1175/2007MWR2267.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 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, 14021424, https://doi.org/10.1175/WAF-D-14-00003.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Li, Z., Z. Pu, J. Sun, and W.-C. Lee, 2014: Impacts of 4DVAR assimilation of airborne Doppler radar observations on numerical simulations of the genesis of Typhoon Nuri (2008). J. Appl. Meteor. Climatol., 53, 23252343, https://doi.org/10.1175/JAMC-D-14-0046.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Liang, J., L. G. Wu, and H. J. Zhong, 2014: Idealized numerical simulations of tropical cyclone formation associated with monsoon gyres. Adv. Atmos. Sci., 31, 305315, https://doi.org/10.1007/s00376-013-2282-1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lim, K.-S. S., and S.-Y. 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, 15871612, https://doi.org/10.1175/2009MWR2968.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mlawer, E. J., S. J. Taubman, P. D. Brown, M. J. Iacono, and S. A. Clough, 1997: Radiative transfer for inhomogeneous atmospheres: RRTM, a validated correlated-k model for the longwave. J. Geophys. Res., 102, 16 66316 682, https://doi.org/10.1029/97JD00237.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Montgomery, M. T., M. E. Nicholls, T. A. Cram, and A. B. Saunders, 2006: A vortical hot tower route to tropical cyclogenesis. J. Atmos. Sci., 63, 355386, https://doi.org/10.1175/JAS3604.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Montgomery, M. T., L. L. Lussier III, R. W. Moore, and Z. Wang, 2010: The genesis of Typhoon Nuri as observed during the Tropical Cyclone Structure 2008 (TCS-08) field experiment—Part 1: The role of the easterly wave critical layer. Atmos. Chem. Phys., 10, 98799900, https://doi.org/10.5194/acp-10-9879-2010.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Montgomery, M. T., and Coauthors, 2012: The Pre-Depression Investigation of Cloud-Systems in the Tropics (PREDICT) Experiment: Scientific basis, new analysis tools, and some first results. Bull. Amer. Meteor. Soc., 93, 153172, https://doi.org/10.1175/BAMS-D-11-00046.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Nakano, M., M. Sawada, T. Nasuno, and M. Satoh, 2015: Intraseasonal variability and tropical cyclogenesis in the western North Pacific simulated by a global nonhydrostatic atmospheric model. Geophys. Res. Lett., 42, 565571, https://doi.org/10.1002/2014GL062479.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • NCEP, 2000: NCEP FNL Operational Model Global Tropospheric Analyses, continuing from July 1999. Research Data Archive, Computational and Information Systems Laboratory, National Center for Atmospheric Research, Boulder, CO, accessed 30 March 2017, https://doi.org/10.5065/D6M043C6.

    • Crossref
    • Export Citation
  • Park, M.-S., H.-S. Kim, C.-H. Ho, R. L. Elsberry, and M.-I. Lee, 2015: Tropical Cyclone Mekkhala’s (2008) formation over the South China Sea: Mesoscale, synoptic-scale, and large-scale contributions. Mon. Wea. Rev., 143, 88110, https://doi.org/10.1175/MWR-D-14-00119.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ritchie, E. A., and G. J. Holland, 1999: Large-scale patterns associated with tropical cyclogenesis in the western Pacific. Mon. Wea. Rev., 127, 20272043, https://doi.org/10.1175/1520-0493(1999)127<2027:LSPAWT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Skamarock, W., and Coauthors, 2008: A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, 113 pp., http://dx.doi.org/10.5065/D68S4MVH.

    • Crossref
    • Export Citation
  • Suzuki-Parker, A., 2012: Uncertainties and Limitations in Simulating Tropical Cyclones. Springer, 78 pp.

  • Teng, H.-F., 2016: Formation and development of tropical cloud cluster in the western North Pacific. Ph.D. thesis, National Taiwan University, 178 pp.

  • Thatcher, L., and Z. Pu, 2013: Evaluation of tropical cyclone genesis precursors with relative operating characteristics (ROC) in high-resolution ensemble forecasts: Hurricane Ernesto. Trop. Cyclone Res. Rev., 2, 131148.

    • Search Google Scholar
    • Export Citation
  • Tory, K. J., M. T. Montgomery, and N. E. Davidson, 2007: Prediction and diagnosis of tropical cyclone formation in an NWP system. Part III: Developing and nondeveloping storms. J. Atmos. Sci., 64, 31953213, https://doi.org/10.1175/JAS4023.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tsai, H.-C., R. L. Elsberry, M. S. Jordan, and F. Vitart, 2013: Objective verifications and false alarm analyses of western North Pacific tropical cyclone event forecasts by the ECMWF 32-day ensemble. Asia-Pac. J. Atmos. Sci., 49, 409420, https://doi.org/10.1007/s13143-013-0038-6.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, Z., M. T. Montgomery, and T. J. Dunkerton, 2010: Genesis of pre–Hurricane Felix (2007). Part I: The role of the easterly wave critical layer. J. Atmos. Sci., 67, 17111729, https://doi.org/10.1175/2009JAS3420.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wu, L., and J. Duan, 2015: Extended simulation of tropical cyclone formation in the western North Pacific monsoon trough. J. Atmos. Sci., 72, 44694485, https://doi.org/10.1175/JAS-D-14-0375.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wu, L., H. Zong, and J. Liang, 2013: Observational analysis of tropical cyclone formation associated with monsoon gyres. J. Atmos. Sci., 70, 10231034, https://doi.org/10.1175/JAS-D-12-0117.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Xu, Y., T. Li, and M. Peng, 2014: Roles of the synoptic-scale wave train, the intraseasonal oscillation, and high-frequency eddies in the genesis of Typhoon Manyi (2001). J. Atmos. Sci., 71, 37063722, https://doi.org/10.1175/JAS-D-13-0406.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhan, R. F., Y. Wang, and C.-C. Wu, 2011: Impact of SSTA in East Indian Ocean on the frequency of northwest Pacific tropical cyclones: A regional atmospheric model study. J. Climate, 24, 62276242, https://doi.org/10.1175/JCLI-D-10-05014.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zong, H., and L. Wu, 2015: Synoptic-scale influences on tropical cyclone formation within the western North Pacific monsoon trough. Mon. Wea. Rev., 143, 34213433, https://doi.org/10.1175/MWR-D-14-00321.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
Save