• Bae, S.-Y., S.-Y. Hong, and K.-S. Lim, 2016: Coupling WRF double-moment 6-class microphysics schemes to RRTMG radiation scheme in weather research forecasting model. Adv. Meteor., 2016, 111, https://doi.org/10.1155/2016/5070154.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Baek, S. H., 2017: A revised radiation package of G-packed McICA and two-stream approximation: Performance evaluation in a global weather forecasting model. J. Adv. Model. Earth Syst., 9, 16281640, https://doi.org/10.1002/2017MS000994.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bauer, P., and Coauthors, 2013: Model cycle 38r2: Components and performance. ECMWF Tech. Memo. 704, ECMWF, Reading, United Kingdom, 58 pp., https://www.ecmwf.int/sites/default/files/elibrary/2013/7986-model-cycle-38r2-components-and-performance.pdf.

  • Bélair, S., M. Roch, A.-M. Leduc, P. A. Vaillancourt, S. Laroche, and J. Mailhot, 2009: Medium-range quantitative precipitation forecasts from Canada’s new 33-km deterministic global operational system. Wea. Forecasting, 24, 690708, https://doi.org/10.1175/2008WAF2222175.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bhaskar Rao, D. V., D. Hari Prasad, D. Srinivas, and Y. Anjaneyulu, 2010: Role of vertical resolution in numerical models towards the intensification, structure and track of tropical cyclones. Mar. Geod., 33, 338355, https://doi.org/10.1080/01490419.2010.518066.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Charron, M., and Coauthors, 2012: The stratospheric extension of the Canadian global deterministic medium-range weather forecasting system and its impact on tropospheric forecasts. Mon. Wea. Rev., 140, 19241944, https://doi.org/10.1175/MWR-D-11-00097.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 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, 569585, https://doi.org/10.1175/1520-0493(2001)129<0569:CAALSH>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Choi, H.-J., and S.-Y. Hong, 2015: An updated subgrid orographic parameterization for global atmospheric forecast. J. Geophys. Res. Atmos., 120, 12 44512 457, https://doi.org/10.1002/2015JD024230.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Côté, J., J.-G. Desmarais, S. Gravel, A. Méthot, A. Patoine, M. Roch, and A. Staniforth, 1998: The operational CMC–MRB global environmental multiscale (GEM) model. Part II: Results. Mon.Wea. Rev., 126, 13971418, https://doi.org/10.1175/1520-0493(1998)126<1397:TOCMGE>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ek, M. B., K. E. Mitchell, Y. Lin, E. Rogers, P. Grunmann, V. Koren, G. Gayno, and J. D. Tarpley, 2003: Implementation of Noah land surface model advances in the National Centers for Environmental Prediction operational mesoscale Eta model. J. Geophys. Res., 108, 8851, https://doi.org/10.1029/2002JD003296.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Han, J., and H.-L. Pan, 2011: Revision of convection and vertical diffusion schemes in the NCEP Global Forecast System. Wea. Forecasting, 26, 520533, https://doi.org/10.1175/WAF-D-10-05038.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Han, J.-Y., S.-Y. Hong, K.-S. S. Lim, and J. Han, 2016: Sensitivity of a cumulus parameterization scheme to precipitation production representation and its impact on a heavy rain event over Korea. Mon. Wea. Rev., 144, 21252135, https://doi.org/10.1175/MWR-D-15-0255.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Han, J.-Y., S.-Y. Hong, and Y.-C. Kwon, 2017: Recent progress on cumulus parameterization in KIAPS Integrated Model (KIM). Preprints, Meteorology and Climate—Modeling for Air Quality Conf., Davis, CA, California Environmental Protection Agency’s Air Resources Board, https://www2.eventsxd.com/event/4129/meteorologyandclimatemodelingforairquality/sessions.

  • Holtslag, A. A., and B. A. Boville, 1993: Local versus nonlocal boundary-layer diffusion in a global climate model. J. Climate, 6, 18251842, https://doi.org/10.1175/1520-0442(1993)006<1825:LVNBLD>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hong, S.-Y., and J. Jang, 2018: Impacts of shallow convection processes on a simulated boreal summer climatology in a global atmospheric model. Asia-Pac. J. Atmos. Sci., 54, 361370, https://doi.org/10.1007/S13143-018-0013-3.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hong, S.-Y., J. Dudhia, and S.-H. Chen, 2004: A revised approach to ice microphysical processes for the bulk parameterization of clouds and precipitation. Mon. Wea. Rev., 132, 103120, https://doi.org/10.1175/1520-0493(2004)132<0103:ARATIM>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
  • Hong, S.-Y., and Coauthors, 2013: The Global/Regional Integrated Model System (GRIMs). Asia-Pac. J. Atmos. Sci., 49, 219243, https://doi.org/10.1007/s13143-013-0023-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hong, S.-Y., and Coauthors, 2018: The Korean Integrated Model (KIM) System for global weather forecasting. Asia-Pac. J. Atmos. Sci., 54, 267292, https://doi.org/10.1007/S13143-018-0028-9.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Huffman, G. J., and Coauthors, 2007: The TRMM Multisatellite Precipitation Analysis (TMPA): Quasi-global, multiyear, combined-sensor precipitation estimates at fine scales. J. Hydrometeor., 8, 3855, https://doi.org/10.1175/JHM560.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Iacono, M.-J., J. S. Delamere, E. J. Mlawer, M. W. Shepherd, S. A. Clough, and W. D. Collins, 2008: Radiative forcing by longlived greenhouse gases: Calculation with the AER radiative transfer models. J. Geophys. Res., 113, D13103, https://doi.org/10.1029/2008JD009944.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Inness, P. M., J. M. Slingo, S. J. Woolnough, R. B. Neale, and V. D. Pope, 2001: Organization of tropical convection in a GCM with varying vertical resolution; Implications for the simulation of the Madden–Julian Oscillation. Climate Dyn., 17, 777793, https://doi.org/10.1007/s003820000148.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kim, E.-J., and S.-Y. Hong, 2010: Impact of air-sea interaction on East Asian summer monsoon climate in WRF. J. Geophys. Res., 115, D19118, https://doi.org/10.1029/2009JD013253.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Koo, M.-S., S. Baek, K.-H. Seol, and K. Cho, 2017: Advances in land surface modeling of KIAPS based on the Noah land surface model. Asia-Pac. J. Atmos. Sci., 53, 361373, https://doi.org/10.1007/s13143-017-0043-2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kwon, Y. C., and S.-Y. Hong, 2017: A mass-flux cumulus parameterization scheme across gray-zone resolutions. Mon. Wea. Rev., 145, 583598, https://doi.org/10.1175/MWR-D-16-0034.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, E., and S.-Y. Hong, 2019: Impact of the sea surface salinity on simulated precipitation in a global numerical weather prediction model. J. Geophys. Res. Atmos., 124, 719730, https://doi.org/10.1029/2018JD029591.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, E.-H., E. Lee, R. Park, Y.-C. Kwon, and S.-Y. Hong, 2018: Impact of turbulent mixing in the stratocumulus-topped boundary layer on numerical weather prediction. Asia-Pac. J. Atmos. Sci., 54, 371384, https://doi.org/10.1007/S13143-018-0024-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mass, C. F., D. Ovens, K. Westrick, and B. A. Colle, 2002: Does increasing horizontal resolution produce more skillful forecasts? Bull. Amer. Meteor. Soc., 83, 407430, https://doi.org/10.1175/1520-0477(2002)083<0407:DIHRPM>2.3.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • McFarlane, S. A., J. H. Mather, and T. P. Ackerman, 2007: Analysis of tropical radiative heating profiles: A comparison of models and observations. J. Geophys. Res., 112, D14218, https://doi.org/10.1029/2006JD008290.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pan, H.-L., and W.-S. Wu, 1995: Implementing a mass flux convective parameterization package for the NMC medium-range forecast model. NMC Office Note 409, 40 pp.

  • Park, R.-S., J.-H. Chae, and S.-Y. Hong, 2016: A revised prognostic cloud fraction scheme in a global forecasting system. Mon. Wea. Rev., 144, 12191229, https://doi.org/10.1175/MWR-D-15-0273.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Philip, A., T. Bergot, Y. Bouteloup, and F. Bouyssel, 2016: The impact of vertical resolution on fog forecasting in the kilometric-scale model AROME: A case study and statistics. Wea. Forecasting, 31, 16551671, https://doi.org/10.1175/WAF-D-16-0074.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Retsch, M. H., C. Hohenegger, and B. Stevens, 2017: Vertical resolution refinement in an aqua-planet and its effect on the ITCZ. J. Adv. Model. Earth Syst., 9, 24252436, https://doi.org/10.1002/2017MS001010.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Richardson, D., J. Bidlot, R. Buizza, L. Ferranti, A. Ghelli, G. van der Grijn, F. Vitart, and E. Zsoter, 2006: Verification statistics and evaluations of ECMWF forecasts in 2005-2006. ECMWF Tech. Memo. 504. ECMWF, Reading, United Kingdom, 44 pp., https://www.ecmwf.int/sites/default/files/elibrary/2006/11902-verification-statistics-and-evaluations-ecmwf-forecasts-2005-2006.pdf.

  • Roeckner, E., and Coauthors, 2006: Sensitivity of simulated climate to horizontal and vertical resolution in the ECHAM5 atmosphere model. J. Climate, 19, 37713791, https://doi.org/10.1175/JCLI3824.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ruti, P. M., D. D. Rocco, and S. Gualdi, 2006: Impact of increased vertical resolution on simulation of tropical climate. Theor. Appl. Climatol., 85, 6180, https://doi.org/10.1007/s00704-005-0174-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shin, H. H., and S.-Y. Hong, 2015: Representation of the subgrid-scale turbulent transport in convective boundary layers at gray-zone resolutions. Mon. Wea. Rev., 143, 250271, https://doi.org/10.1175/MWR-D-14-00116.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shin, H. H., S.-Y. Hong, and J. Dudhia, 2012: Impacts of the lowest model level height on the performance of planetary boundary layer parameterizations. Mon. Wea. Rev., 140, 664682, https://doi.org/10.1175/MWR-D-11-00027.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Simmons, A. J., D. M. Burridge, M. Jarraud, C. Girard, and W. Wergen, 1989: The ECMWF medium-range prediction models: Development of the numerical formulations and the impact of increased resolution. Meteor. Atmos. Phys., 40, 2860, https://doi.org/10.1007/BF01027467.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Takeuchi, Y., and Coauthors, 2013: Outline of the Operational Numerical Weather Prediction at the Japan Meteorological Agency: Appendix to WMO Technical Progress Report on the Global Data-processing and Forecasting System (GDPFS) and Numerical Weather Prediction (NWP) Research. Japan Meteorological Agency, 188 pp.

  • Tardif, R., 2007: The impact of vertical resolution in the explicit numerical forecasting of radiation fog: A case study. Pure Appl. Geophys., 164, 12211240, https://doi.org/10.1007/s00024-007-0216-5.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tompkins, A. M., and K. A. Emanuel, 2000: The vertical resolution sensitivity of simulated equilibrium tropical temperature and water-vapour profiles. Quart. J. Roy. Meteor. Soc., 126, 12191238, https://doi.org/10.1002/qj.49712656502.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhang, D.-L., and X. Wang, 2003: Dependence of hurricane intensity and structures on vertical resolution and time-step size. Adv. Atmos. Sci., 20, 711725, https://doi.org/10.1007/BF02915397.

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

    (a) Layer depth as a function of height (both in hPa) for CTL and L91. (b) The information of the full level near the surface (0–200 m), including L65, CTL, and L91.

  • View in gallery

    Latitude–pressure cross sections of differences in zonally averaged daily mean temperature (K) and specific humidity (g kg−1), between experiments for 48–72-h forecasts during the month of July 2016: (a) L91 minus CTL, (b) L65 minus CTL, and (c) L91 minus L65. The black solid (dotted) line indicates positive (negative) change with a 0.05 (K, g kg−1) interval.

  • View in gallery

    Daily mean (a) PBL height (m), (b) latent heat flux (W m−2), and (c) sensible heat flux (W m−2) for 48–72-h forecasts in CTL during the month of July 2016. (d)–(f) Differences between CTL and L65 (L65 minus CTL).

  • View in gallery

    (a) Total precipitation, (b) convective precipitation, and (c) large-scale precipitation (mm day−1) accumulated from 48- to 72-h forecasts in CTL during the month of July 2016. (d)–(f) Differences between CTL and L65 (L65 minus CTL).

  • View in gallery

    Vertical profiles of (a) differences in relative humidity between CTL and L65 (L65 minus CTL). Vertical profiles of daily mean heating rates (K day−1) by (b) cumulus convection and (c) gridscale condensation in CTL (black solid line), L65 (red dashed line), and the differences between CTL and L65 (gray short-dashed line). All analyses were taken from the daily mean for 48–72-h forecasts during the month of July 2016 over (left) the tropics (TROP: 20°S–20°N), (middle) western Pacific Ocean (WP: 0°–20°N, 120°–150°E), and (right) eastern Pacific Ocean (EP: 0°–20°N, 240°–260°E).

  • View in gallery

    Vertical profiles of daily mean (a) moistening rate (g kg−1 day−1) by cumulus convection and (b) cloud ice mixing ratio (g kg−1) in 65 (black solid line) and L91 (red dashed line). Gray short-dashed line indicates the differences between L65 and L91. The averaged period and domains are the same as Fig. 5.

  • View in gallery

    Vertical profile of daily mean heating rates (K day−1) by the (a) shortwave and (b) longwave radiation in L65 (black solid line) and L91 (red dashed line). The averaged period and domains are the same as Fig. 5.

  • View in gallery

    Vertical profiles of daily mean (a) moistening rate (g kg−1 day−1) by the cumulus convection and (b) cloud ice mixing ratio (g kg−1) over the tropics in L65 (black solid line) and L91 (red dashed line) following the modified treatment for the specific humidity of a buoyant parcel.

  • View in gallery

    Time–pressure cross sections of biases (shaded) and RMSEs (contour) for temperature (K) in (a) CTL against the NCEP GFS-FNL data and (b) their differences between CTL and L91 (L91 minus CTL), averaged over (top) the tropics (20°S–20°N) and (bottom) Northern Hemisphere extratropics (20°–90°N), during the month of July 2016. Solid and dashed lines indicate positive and negative values, respectively.

  • View in gallery

    Time–pressure cross sections of biases (shaded) and RMSEs (contour) for specific humidity (g kg−1) in (a) CTL against the NCEP GFS-FNL data and (b) their differences between CTL and L91 (L91 minus CTL), averaged over (top) the tropics (20°S–20°N) and (bottom) Northern Hemisphere extratropics (20°–90°N), during the month of July 2016. Solid and dashed lines indicate positive and negative values, respectively.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 4 4 4
PDF Downloads 4 4 4

Impact of Increased Vertical Resolution on Medium-Range Forecasts in a Global Atmospheric Model

View More View Less
  • 1 Korea Institute of Atmospheric Prediction Systems, Seoul, South Korea
Free access

Abstract

This study aims to investigate the impact of increased vertical resolution on global medium-range forecasts. For this purpose, the dependencies of simulated atmospheric temperature and specific humidity on the lowest model level height and vertical grid spacing are investigated. The reduced first model level increases vertical turbulent mixing by determining a higher planetary boundary layer (PBL) height and associated turbulent diffusivities and velocity scales. This contributes to warming/drying within the PBL and cooling/moistening above the PBL. Resulting dryness near the surface enhances an increase in surface moisture flux from the ocean, which results in the apparent moistening of the troposphere. Consequently, large-scale precipitation increases due to more humid atmospheric conditions, while convective precipitation decreases because of drier conditions near the starting level of convection. Meanwhile, the reduced vertical grid spacing resolves the overshooting layer well in the cumulus convection process, which decreases the detrained moisture at the convective cloud top. This leads to a noticeable downward shift of ice clouds in the upper troposphere, and further contributes to the enhancement of longwave cooling via cloud–radiation processes. In medium-range forecasts, the increased vertical resolution exerts a significant impact on the simulated features of tropospheric temperature and humidity, while changes in the prediction accuracy precipitation are negligible owing to compensation between convective and large-scale precipitation. Finally, the discussion of possible methods to minimize the sensitivities of model’s physics to vertical resolution is presented.

© 2019 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: Eun-Hee Lee, eh.lee@kiaps.org

Abstract

This study aims to investigate the impact of increased vertical resolution on global medium-range forecasts. For this purpose, the dependencies of simulated atmospheric temperature and specific humidity on the lowest model level height and vertical grid spacing are investigated. The reduced first model level increases vertical turbulent mixing by determining a higher planetary boundary layer (PBL) height and associated turbulent diffusivities and velocity scales. This contributes to warming/drying within the PBL and cooling/moistening above the PBL. Resulting dryness near the surface enhances an increase in surface moisture flux from the ocean, which results in the apparent moistening of the troposphere. Consequently, large-scale precipitation increases due to more humid atmospheric conditions, while convective precipitation decreases because of drier conditions near the starting level of convection. Meanwhile, the reduced vertical grid spacing resolves the overshooting layer well in the cumulus convection process, which decreases the detrained moisture at the convective cloud top. This leads to a noticeable downward shift of ice clouds in the upper troposphere, and further contributes to the enhancement of longwave cooling via cloud–radiation processes. In medium-range forecasts, the increased vertical resolution exerts a significant impact on the simulated features of tropospheric temperature and humidity, while changes in the prediction accuracy precipitation are negligible owing to compensation between convective and large-scale precipitation. Finally, the discussion of possible methods to minimize the sensitivities of model’s physics to vertical resolution is presented.

© 2019 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: Eun-Hee Lee, eh.lee@kiaps.org

1. Introduction

The governing equations in numerical weather prediction (NWP) models describe virtually all dynamical and physical processes of the atmosphere, and both horizontal and vertical resolutions modulate the uncertainties in these processes because fine model grid spacing can resolve small-scale features (Mass et al. 2002). Over the past few decades, horizontal resolution has been increased in major NWP models. Many previous studies have indicated that increasing horizontal resolution alone does not always guarantee a better solution and that the simultaneous implementation of finer vertical resolution is also required (Zhang and Wang 2003). As global simulations with fine vertical resolution have become possible with recent advances in available computation power, increased vertical resolutions have been attempted in the global prediction system at multiple operational centers (Table 1). With these efforts, attention has been focused on the impact of increasing vertical resolution on improving forecasting ability in NWP models.

Table 1.

Vertical resolution updates in the global prediction system at various operation centers.

Table 1.

Previous studies have reported that increased vertical resolution is capable of improving reproduction of weather phenomena such as typhoon and fog. Tardif (2007) and Philip et al. (2016) showed that the cooling rate before the appearance and formation of fog is better represented when increased vertical resolution is used in both one-dimensional and three-dimensional models. Bhaskar Rao et al. (2010) reported that high vertical resolution throughout the troposphere improves the simulation of intensity and tracking of tropical cyclones as well as the prediction of vertical shear of horizontal winds. In addition, the impact of fine vertical resolution is known to be important for producing the initial conditions of NWP models. For example, Simmons et al. (1989) suggested that improved initial analyses taken from data assimilation with increased vertical resolution including the stratosphere, improve forecast skill scores.

These previously mentioned studies robustly indicate that the impact of increased vertical resolution on reproducing specific weather phenomena and data assimilations is positive. This is in line with the theoretical expectation that increased vertical resolution can improve the reproduction of weather events. However, the advantage of increased vertical resolution has been questioned with regard to improving the skills of global weather forecasting in NWP models. As demonstrated by Roeckner et al. (2006) and Ruti et al. (2006), higher vertical resolution leads to a marked redistribution of temperature, humidity and clouds due to the change in physical processes. Retsch et al. (2017) reported that high vertical resolution yields an equatorward shift of the intertropical convergence zone by stronger mixing between the updraft and its environment as simulated in the convection scheme. Bauer et al. (2013) also documented systematic changes caused by the increased vertical resolution, which are likely to be a result of multiple different sensitivities to vertical resolution in both the dynamics and parameterization. These studies suggest that systematic changes due to the vertical resolution in specific schemes of model physics and dynamics can also affect medium-range forecasts. Thus far, little effort has been made to include a clear and general determination of how simulated atmospheric fields are affected by the change of vertical resolution in NWP models.

In this study, we assess the impact of increased vertical resolution on predictable variables, including temperature and specific humidity, in a global atmospheric forecasting model along with related physical processes, including turbulent transport, precipitation, cloud and radiation processes. A statistical evaluation of medium-range forecasts with increased vertical resolution against an analysis dataset is also performed. Increases in vertical resolution are tested separately in two approaches. The first involves only the impact of reducing the lowest model level height, and the second requires reducing the vertical grid spacing without imposing changes to the lowest model level.

This paper is organized as follows: section 2 describes the model and experimental setup with various vertical resolutions. Changes in globally simulated temperature and specific humidity due to increased vertical resolution are investigated in section 3. The relevant physical processes are also discussed in detail in this section. In addition, impacts of increased vertical resolution on the biases and RMSE of the medium-range forecasts against an analysis dataset are presented. Finally, this study is concluded in section 4.

2. Model and experimental design

a. Model description

The atmospheric model used in this study is the Global/Regional Integrated Model system (GRIMs) (Hong et al. 2013) global model program (GMP), which is a multiscale atmospheric modeling system with unified physics, developed for use in NWP, seasonal simulations, and climate research at global and regional scales. The selected dynamic core is spherical harmonics. The advanced physics parameterization employed in this study include the revised Rapid Radiative Transfer Model for general circulation models (RRTMG; Iacono et al. 2008; RRTMK; Baek 2017), scale-aware Yonsei University (YSU) planetary boundary layer (PBL) scheme (Hong et al. 2006; Shin and Hong 2015) with top-down turbulent mixing by the stratocumulus topped boundary layer (Lee et al. 2018), the simplified Arakawa–Schubert (SAS) deep convection scheme (Pan and Wu 1995; Han and Pan 2011; Han et al. 2016; Kwon and Hong 2017), the Weather Research and Forecasting (WRF) single-moment 5-class (WSM5) microphysics scheme (Hong et al. 2004; Bae et al. 2016), the GRIMs shallow convection scheme with adjusted method (Hong and Jang 2018), the modified Noah land surface model (Chen and Dudhia 2001; Ek et al. 2003; Koo et al. 2017), subgrid-scale orographic drag parameterization scheme (Choi and Hong 2015), the ocean mixed layer model (Kim and Hong 2010), and the prognostic cloudiness parameterization scheme (Park et al. 2016). This physics package developed for the operational NWP models has been tested by Hong et al. (2018), in which the detailed performances of the physics packages are described.

b. Experimental setup

To examine the impacts of increased vertical resolution on simulated atmospheric fields, we set up numerical simulations using three different vertical grid configurations with a fixed horizontal resolution. The horizontal resolution of T254 to represent a triangular truncation at wavenumber 254 is selected, corresponding to a grid spacing of approximately 50 km. With this protocol, the control experiment uses 64 hybrid sigma-pressure vertical levels with a model top at 0.3 hPa, which is the default setting of the GRIMs-GMP T254 resolution, as suggested by Hong et al. (2013) (hereafter, CTL). The impact of the increased vertical resolution is examined by an experiment employing 91 levels of the hybrid sigma-pressure vertical coordinate with an expanded model top up to 0.01 hPa, as introduced by the European Centre for Medium-Range Weather Forecasts (ECMWF) and the high-resolution Integrated Forecasting System cycle 30r1 (Richardson et al. 2006) (hereafter L91). (Specific level information is available at https://www.ecmwf.int/en/forecasts/documentation-and-support/91-model-levels.)

The vertical coordinate of L91 with increased vertical resolution has the reduced the vertical grid spacing (Δz) throughout the atmosphere. More levels near the surface compared with CTL in L91, gives the reduced lowest model level height (z1) of around 10 m than the approximately 18 m in CTL. Consequently, it is necessary to determine the individual impact of reducing z1 and Δz. Accordingly, we add another experiment with 65 vertical levels (hereafter L65) which is almost identical to the vertical profile of CTL, except that z1 is reduced from approximately 18 to 10 m. To produce L65, one extra layer is added to the CTL profile near the surface, which provides the same full and interface level heights at the bottom of layers of L91 by modulating the second level. Figure 1 illustrates the profile shape of vertical levels and grid spacing in CTL, L65, and L91.

Fig. 1.
Fig. 1.

(a) Layer depth as a function of height (both in hPa) for CTL and L91. (b) The information of the full level near the surface (0–200 m), including L65, CTL, and L91.

Citation: Monthly Weather Review 147, 11; 10.1175/MWR-D-18-0387.1

All of the simulations are carried out for 10-day forecasts initiated at every 0000 UTC from 1 to 31 July 2016 with identical physics packages. Initial and surface boundary conditions are taken from the operational analysis data of the National Centers for Environmental Prediction (NCEP) Global Forecast System (GFS) with the resolution of T1534L64; the observed sea surface temperature (SST) and snow data are also obtained from NCEP GFS. Simulated SST is initialized every 24 h in order to avoid the excessive bias of SST during the integration, based on the assumption that the variation of daily mean SST is not large during the 10-day forecasts (Lee and Hong 2019). To validate simulated atmospheric fields, the NCEP GFS Final Analysis (FNL) dataset is used, which provides products every 6 h with a 1° × 1° spatial resolution and 31 pressure levels from 1000 to 10 hPa. For the evaluation, model outputs are interpolated to the time and location of NCEP GFS FNL dataset.

3. Results

a. Changes in temperature and specific humidity due to increased vertical resolution

To investigate the impact of increased vertical resolution on simulated global features, zonally averaged differences of daily mean temperature and specific humidity forecasts are presented in Fig. 2. The responses of primary variables with respect to increased vertical resolutions are consistent over the course of the forecast. Thus, we describe the changes based on the 48–72-h forecasts. First, the differences between CTL and L91 (i.e., L91 minus CTL) are depicted in Fig. 2a. It is apparent that L91 simulates a colder and moister atmosphere compared to CTL. The difference in temperature is an overall cooling throughout the atmosphere, except for model top and bottom layers, and the cores of cooling occurs over the tropics at 850, 250, and 100 hPa. In addition to the cooling that predominantly occurs over the tropics, slight warming occurs near the surface and above the tropopause. A systematic increase in the specific humidity is simulated in low- and midlatitudes regions by enhancing the vertical resolution from CTL to L91, which is apparent as an increase between 925 and 500 hPa. However, the simulation also exhibits an exceptional decrease in specific humidity below 925 hPa. These changes in simulated moisture are also centered over the tropics, but extend to midlatitudes. Furthermore, these responses are also seen over high latitudes; however, these changes are insignificant compared to those over the tropics. These remarkable changes over the tropics caused by increased vertical resolution are also noted out by Tompkins and Emanuel (2000) and Inness et al. (2001). It could be understood that a possible effect due to the extension of model top from the middle mesosphere to the mesopause is not related to the changes seen in the stratosphere because an initial condition is taken from independent data and the experimental period is limited to a medium-range forecast.

Fig. 2.
Fig. 2.

Latitude–pressure cross sections of differences in zonally averaged daily mean temperature (K) and specific humidity (g kg−1), between experiments for 48–72-h forecasts during the month of July 2016: (a) L91 minus CTL, (b) L65 minus CTL, and (c) L91 minus L65. The black solid (dotted) line indicates positive (negative) change with a 0.05 (K, g kg−1) interval.

Citation: Monthly Weather Review 147, 11; 10.1175/MWR-D-18-0387.1

Figures 2b and 2c show the individual impacts of the reduced z1 (the difference between L65 and CTL) and Δz (the difference between L91 and L65). The difference between L65 and CTL shows that the lower z1 contributes to changes in temperature and moisture in the lower troposphere, especially over the tropics (Fig. 2b). The differences clearly divide near 925 hPa, which can be considered a typical ocean boundary layer height. It exhibits a cooling and moistening above the boundary layer, and warming and drying within the boundary layer. In contrast, the difference between L65 and L91 represents that the reduced Δz results in changes in temperature of the entire troposphere, especially in the upper troposphere (top panel of Fig. 2c). Meanwhile, the impact of the reduced Δz on specific humidity appears to be weaker, compared to the response to the reduced z1 (bottom panel of Fig. 2c). According to these results, it is suggested that the changes in simulated temperature and specific humidity in L91, compared to changes in CTL (Fig. 2a) are caused by the combined effects of reduced z1 and Δz (Figs. 2b,c). The related physical processes arising from each change in z1 and Δz are discussed in the next sections.

b. Effects of the lowest model level height

The differences in simulated temperature and humidity between two experiments with varying heights of z1 (i.e., CTL minus L65) are noticeable in the lower troposphere. To examine the effects of z1 on simulated features in the lower troposphere, the PBL height and surface turbulent fluxes are analyzed. Figure 3 compares the distributions of daily mean PBL height and surface fluxes in L65 with CTL, and shows that the reduced z1 leads to an increase in the simulated global boundary layer height and latent heat flux, except in high-latitude regions. These changes are more intense over most tropical ocean regions. In contrast, it appears that the simulated global sensible heat flux is reduced, especially over the land regions. To analyze what causes these changes, the sensitivity of physical processes in the boundary layer and land/ocean surface to z1 is further investigated.

Fig. 3.
Fig. 3.

Daily mean (a) PBL height (m), (b) latent heat flux (W m−2), and (c) sensible heat flux (W m−2) for 48–72-h forecasts in CTL during the month of July 2016. (d)–(f) Differences between CTL and L65 (L65 minus CTL).

Citation: Monthly Weather Review 147, 11; 10.1175/MWR-D-18-0387.1

The YSU PBL scheme is a first-order closure scheme, where turbulent diffusion in the convective boundary layer is parameterized by using a K-profile approach with nonlocal countergradient mixing and explicit treatment of entrainment. The estimation of the PBL height in the YSU scheme is based on the bulk stability comparison with the critical value using the bulk Richardson number from the surface (Holtslag and Boville 1993). The bulk Richardson number Rib is expressed as
Rib(z)=gz[θυ(z)θs]θυa|U(z)|2,
θs=θυa+θT[=b(wθυ)¯0ws],
where g is gravitational acceleration, θυ(z) and U(z) represent the virtual potential temperature and horizontal wind speed at each model level, and θυa is the virtual temperature at z1 (=θυ,1). The appropriate temperature near the surface, θs, is obtained by considering surface buoyant thermal (θT) in Eq. (2). Here, (wθυ)¯0 is the virtual heat flux from the surface, and the proportionality factor b is a constant, and ws=u*ϕm1 is the mixed-layer velocity scale, where u* is the surface friction velocity scale, and ϕm is the wind profile function evaluated at the top of the surface layer. The critical Richardson number, Rib,c, is set to zero for the convective boundary layer and 0.25 for the stable boundary layer.

To apply these to the numerical simulation, a two-step calculation of PBL height is made because the mixed-layer velocity scale should also be determined according to the PBL height. The procedure, which is called the first guess of PBL height, uses θυ,1, instead of θs. Consequently, this first guess of the PBL height h exhibits sensitivity to z1. In the second step, the PBL height is revised by a thermal access θT, which indicates surface flux-dependent perturbation to the lowest level temperature, and considers the most buoyant eddies. Since the change in θT due to reduced z1 is weaker than that of θυ,1, the updated PBL is still determined at a higher level in the experiment with lower z1. Following this, representative velocity scales, such as the convective velocity scale and the entrainment velocity scale, are determined using the estimated PBL height, which results in enhanced velocity scales. As a result, within the enhanced PBL, the scheme simulates the larger eddy diffusivity by the K-profile function and turbulent velocity scale depending on heights as well as the enhanced entrainment flux. Consequently, the enhanced vertical mixing in L65 with the lower z1 results in cooling/moistening above the PBL and warming/drying within the PBL, as shown in Fig. 2b.

The sensitivity in bulk stability to z1, could be resolved by using surface temperature instead of the temperature at z1 and thermal access in the determination of Rib. We found that the sensitivity is not apparent when applied (not shown); however, this method has some risk because of the high uncertainty of simulated surface temperature in the NWP model.

The PBL height increase in L65 dominates over tropical oceans, with a mean difference of 35.54 m for the tropical ocean between 20°S and 20°N, where the strong convective ocean boundary layer height appears (Figs. 3a,d). However, the change in PBL is relatively small over land regions, with several exceptions in low- and midlatitude continental dry regions. These different changes in PBL heights over the ocean and land due to the reduced z1 are ascribed to distinct responses of surface temperature to the atmospheric characteristics changed by increased subgrid turbulent mixing. It is related to changes in the surface fluxes, which is further described below.

Surface fluxes are diagnosed using the momentum, temperature and moisture at the surface and z1. The bulk formula for sensible and latent heat fluxes in the model are expressed as
H0=ρCpθw¯=ρCpCHU1Δθ,
E0=ρLqw¯=ρLCqU1Δq,
where ρ is air density, Cp is the specific heat capacity of air, CH is the bulk transfer coefficient for heat, Cq is the bulk transfer coefficient for moisture, L is the latent heat of evaporation, and U1 is wind speed at z1. Between the surface and z1, Δθ and Δq are the difference of potential temperature and specific humidity, respectively. In the model, the vertical coordinate associated with a lower z1 directly leads to larger bulk transfer coefficients, CH and Cq, and also estimates smaller Δθ and Δq. Therefore, few changes in the sensible and latent heat fluxes are expected (Shin et al. 2012).

Unexpectedly, the increase in latent heat flux and the decrease in sensible heat flux are globally apparent (Figs. 3b,c,e,f), and are affected by changes near the surface due to enhanced PBL vertical mixing. Drier conditions within the PBL induced by enhanced vertical turbulent mixing in L65 increases the surface latent heat flux by increasing the surface-air gradient. This also contributes to the larger supply of moisture from the surface to the atmosphere, and consequently the amount of atmospheric water vapor increases, as previously discussed (and shown in Fig. 2b). Meanwhile warmer condition within the PBL in L65 reduces the temperature gradient between the surface and the lowest model level, which in turn decreases the sensible heat flux.

The land surface is also strongly affected by atmospheric changes. Clear decreases in sensible heat flux over most of the land regions is related to the decrease in land surface temperature, which is balanced with a decrease in downward shortwave radiation arising from increased cloud cover. The overall increase in low-level specific humidity in L65 results in an increase in low-level cloud cover, which contributes to the weakening of downward shortwave radiation that reaches the surface. Over land, the decrease in surface temperature due to reduced incoming solar radiation, in conjunction with the warmer boundary layer, enhances the reduction in sensible heat flux. However, the change in heat flux over the ocean is relatively weak because SST is almost invariant in our experiment, being constrained to the initial values at every 0000 UTC. Again, the different responses of surface temperature and heat flux between land and ocean is connected to the growth of the PBL, with changes in the PBL height dominantly over the ocean.

It is expected that the simulated cloud and precipitation processes would be strongly affected by changes in the surface flux and enhanced vertical mixing of moisture. Figure 4 presents the distribution of total, convective, and large-scale precipitation in CTL accumulated for a 48–72-h forecast and the differences between CTL and L65 (i.e., L65 minus CTL). In CTL, the strong precipitation occurs mainly over the tropics, and this is caused by convective precipitation rather than large-scale precipitation (Figs. 4a–c). The enhanced surface flux and vertical transport of moisture increases the large-scale precipitation over tropical oceans, but decreases the convective precipitation (Figs. 4e,f). These changes lead to slight increase in the amount of total precipitation (0.012 mm day−1, 0.4%) in the global mean (Fig. 4d).

Fig. 4.
Fig. 4.

(a) Total precipitation, (b) convective precipitation, and (c) large-scale precipitation (mm day−1) accumulated from 48- to 72-h forecasts in CTL during the month of July 2016. (d)–(f) Differences between CTL and L65 (L65 minus CTL).

Citation: Monthly Weather Review 147, 11; 10.1175/MWR-D-18-0387.1

Significant changes in convective and large-scale precipitations, observed mainly over the tropics, are caused by the different responses of the physical processes governing precipitation in the model arising from changes in the simulated atmospheric environment. Figure 5 shows the vertical profiles of daily mean heating rates by cumulus convection and gridscale condensation processes in CTL and L65, during the month of July 2016, over the tropics (TROP: 20°S–20°N) together with changes in relative humidity (RH). The profiles over the western Pacific Ocean (WP: 0°–20°N, 120°–150°E) and eastern Pacific (EP: 0°–20°N, 240°–260°E) are also shown, where the precipitation in CTL, and the differences in convective and large-scale precipitations are most pronounced. The enhanced moisture flux and vertical mixing over the tropical ocean resulting from the reduced z1 in L65 produces higher RH throughout the tropical convective cloud layers from approximately 925 to 100 hPa (Fig. 5a). Furthermore, L65 simulates a reduced RH in the PBL and a slight decrease in humidity is detected above 100 hPa.

Fig. 5.
Fig. 5.

Vertical profiles of (a) differences in relative humidity between CTL and L65 (L65 minus CTL). Vertical profiles of daily mean heating rates (K day−1) by (b) cumulus convection and (c) gridscale condensation in CTL (black solid line), L65 (red dashed line), and the differences between CTL and L65 (gray short-dashed line). All analyses were taken from the daily mean for 48–72-h forecasts during the month of July 2016 over (left) the tropics (TROP: 20°S–20°N), (middle) western Pacific Ocean (WP: 0°–20°N, 120°–150°E), and (right) eastern Pacific Ocean (EP: 0°–20°N, 240°–260°E).

Citation: Monthly Weather Review 147, 11; 10.1175/MWR-D-18-0387.1

Changes in daily mean heating rates arising from cumulus convection and gridscale condensation processes in CTL and L65 (Figs. 5b,c) reflect their responses to changes in RH in each physical process. The convection trigger in the SAS cumulus convection scheme is determined by RHs between the starting level of a buoyant parcel and the level of free convection (LFC), known as “relative humidity-based triggering” (Han et al. 2017; Hong et al. 2018). According to this method, the occurrence of cumulus convection is suppressed under the condition of lower RH below the LFC in L65, and results in decreased adiabatic warming by cumulus convection (Fig. 5b) and decreased convective precipitation (Fig. 4e). In contrast, environments with higher RH above the PBL provide favorable atmospheric conditions for gridscale condensation in L65 (Fig. 5c).

c. Effects of the vertical grid spacing

The comparison between L65 and L91 is conducted in order to examine the impact of Δz. As discussed in section 3a, the reduced vertical grid spacing modifies simulated temperature mainly in the upper atmosphere over the tropics (Fig. 2c). Therefore, we focus on changes in relevant convective process, which are vigorous over the tropics, in conjunction with radiative responses to simulated hydrometeors.

Figure 6 shows the vertical profiles of daily mean moistening rates by cumulus convection physics in the model (i.e., the SAS cumulus convection scheme) and cloud ice mixing ratio for the L65 and L91 experiments, which are regionally averaged for the TROP, WP, and EP regions. The L91 results in similar tendencies below 250 hPa and above 100 hPa but also a remarkable decrease in moistening rate near 100–250 hPa, compared to L65. These changes are more prominent for the specified convective regions of the WP and EP.

Fig. 6.
Fig. 6.

Vertical profiles of daily mean (a) moistening rate (g kg−1 day−1) by cumulus convection and (b) cloud ice mixing ratio (g kg−1) in 65 (black solid line) and L91 (red dashed line). Gray short-dashed line indicates the differences between L65 and L91. The averaged period and domains are the same as Fig. 5.

Citation: Monthly Weather Review 147, 11; 10.1175/MWR-D-18-0387.1

The moistening rate near the convective cloud top is strongly connected to the ability to capture the overshooting layer, simulated in the cumulus convection scheme. That is, use of reduced Δz (L91) provides the capability not only to detect the overshooting layer but also to resolve more levels for the assigned layer. In the SAS cumulus convection scheme, the detrained moisture at the convective cloud top is predominantly determined by the amount of moisture of a buoyant parcel, which represents the remaining water vapor in the conversion process of convective precipitation. The amount of moisture is parameterized by the conversion parameter for convective rain (C0), which is given a constant value for the overshooting layer. Consequently, more resolved levels contribute to reduce the specific humidity of the updraft parcel in the overshooting layer, because water vapor is converted to convective precipitation efficiently. As a result, it is evident that the detrained moisture at the convective cloud top is reduced in L91 (Fig. 6a) and the L91 simulates a decrease in cloud ice above 250 hPa, which contributes to a downward shift of ice cloud layers (Fig. 6b). This hydrometeor response to the reduced Δz, caused by the fact that drier conditions at the convective cloud top, weakens the formation of cloud ice near the tropopause in the microphysics scheme.

This change in the atmospheric hydrometeor strongly modulates atmospheric radiation. Vertical profiles of heating rate by shortwave and longwave radiative fluxes in L65 and L91 are displayed in Fig. 7. Decreased absorption of shortwave radiative flux at the cloud top is shown in L91. However, the differences in shortwave radiative heating between the two experiments are less noticeable than those in longwave radiative heating/cooling. The responses of longwave radiative heating/cooling are apparent from the middle to the upper troposphere, with major changes occurring around 200 hPa. This is related to the response of longwave radiative flux to ice clouds, being cooling near the cloud top and warming below the layer (McFarlane et al. 2007). Decrease in cloud ice near the top results in downward shift of ice clouds and leads to the redistribution of cooling/warming by the ice clouds, which is apparent as cooling in the ice cloud layer (Fig. 7b). This change in longwave cooling contributes to a decrease in temperature in the upper troposphere (Fig. 2c).

Fig. 7.
Fig. 7.

Vertical profile of daily mean heating rates (K day−1) by the (a) shortwave and (b) longwave radiation in L65 (black solid line) and L91 (red dashed line). The averaged period and domains are the same as Fig. 5.

Citation: Monthly Weather Review 147, 11; 10.1175/MWR-D-18-0387.1

We performed an additional test in order to eliminate the sensitivity to reduced Δz. To avoid the rapid decrease in the amount of moisture of buoyant parcels of air in the overshooting layer, C0 was modified as a function of temperature, as suggested by Han et al. (2016):
C0(z)=aexp{b[T(z)T0]}forTT0,
where a (=2 × 10−3 m−1) and b (=0.07°C−1) are constants, and T0 is the freezing temperature. C0 declines exponentially in the overshooting layer as a function of environmental temperature and this results in the amount of moisture of buoyant parcels being independent of vertical grid spacing. This modification leads to the vertical continuity of the amount of moisture of buoyant parcels in the overshooting and cloud layers. Consequently, differences in detrained moisture and cloud ice between L65 and L91 revised by modified C0 are much reduced (Fig. 8); that is, the modified method exhibits little dependency.
Fig. 8.
Fig. 8.

Vertical profiles of daily mean (a) moistening rate (g kg−1 day−1) by the cumulus convection and (b) cloud ice mixing ratio (g kg−1) over the tropics in L65 (black solid line) and L91 (red dashed line) following the modified treatment for the specific humidity of a buoyant parcel.

Citation: Monthly Weather Review 147, 11; 10.1175/MWR-D-18-0387.1

d. Impacts on the medium-range skill scores

In this section, the impacts of an increase in vertical resolution on medium-range forecasts are statistically verified against analysis data and observation. First, model performances of simulated temperature and specific humidity in July 2016 are compared to the NCEP GFS-FNL analysis dataset. Figures 9 and 10 depict the time series of vertical profiles of biases and root-mean-square errors (RMSEs) of simulated temperature and specific humidity in CTL and their differences between CTL and L91 (i.e., L91 minus CTL), averaged over the tropics and the Northern Hemisphere.

Fig. 9.
Fig. 9.

Time–pressure cross sections of biases (shaded) and RMSEs (contour) for temperature (K) in (a) CTL against the NCEP GFS-FNL data and (b) their differences between CTL and L91 (L91 minus CTL), averaged over (top) the tropics (20°S–20°N) and (bottom) Northern Hemisphere extratropics (20°–90°N), during the month of July 2016. Solid and dashed lines indicate positive and negative values, respectively.

Citation: Monthly Weather Review 147, 11; 10.1175/MWR-D-18-0387.1

Fig. 10.
Fig. 10.

Time–pressure cross sections of biases (shaded) and RMSEs (contour) for specific humidity (g kg−1) in (a) CTL against the NCEP GFS-FNL data and (b) their differences between CTL and L91 (L91 minus CTL), averaged over (top) the tropics (20°S–20°N) and (bottom) Northern Hemisphere extratropics (20°–90°N), during the month of July 2016. Solid and dashed lines indicate positive and negative values, respectively.

Citation: Monthly Weather Review 147, 11; 10.1175/MWR-D-18-0387.1

In CTL, over the tropics, a warm bias dominates the middle troposphere between 850 and 300 hPa for all 10-day forecasts, with increases apparent across the forecast time. Cold biases appear above 300 hPa and near the surface and the bias around 200 hPa become weaker over the forecast time (top panel in Fig. 9a). The temperature biases averaged over the Northern Hemisphere are similar to those over the tropics, but with a reduced warm bias and more persistent cold bias near the surface and tropopause (bottom panel in Fig. 9a). The increased vertical resolution in L91 exerts an effect on such biases throughout the troposphere by contributing toward a decreasing warm bias in the middle troposphere and cold biases below 850 hPa and above 50 hPa over the tropics and Northern Hemisphere, and increasing the cold bias between 300 and 100 hPa (Fig. 9b).

CTL simulates a moist bias below 925 hPa, but simulates a weak dry bias for most layers of the troposphere above the PBL, and these biases are predominant, especially in tropical regions (Fig. 10a). Both moist and dry biases in the lower and middle troposphere are significantly modified by the increase in the total amount and enhanced upward transport of atmospheric water vapor in L91, and this change reduces the RMSE during the period of the forecasts (Fig. 10b). This suggests that the changes in temperature and moisture arising from increased vertical resolution from CTL to L91 modulate the overall forecast accuracy in most layers of the atmosphere by decreasing the conventional biases of temperature and moisture.

The sensitivity of precipitation processes to increased vertical resolution can affect the spatial distribution of precipitation. In L65, with the reduced z1, convective precipitation is weakened, especially in the tropics, whereas large-scale precipitation is intensified. Reduced Δz also affects the ratio of convective to large-scale precipitation; however, the changes in total precipitation are minimal. The L91 consistently demonstrates a tendency to reduce convective precipitation over the ocean when compared with CTL. However, a part of the reduced convective precipitation is compensated for through enhanced large-scale precipitation. We investigated the precipitation skill scores using the Tropical Rainfall Measuring Mission (TRMM) Multisatellite Precipitation Analysis (TMPA) data (Huffman et al. 2007). It is found that the overall impact of increased vertical resolution on precipitation skill scores is insignificant despite the changes in spatial distribution of convective/nonconvective precipitation (not shown).

4. Conclusions

This study investigates the impact of increased vertical resolution on simulated temperature and specific humidity in the global atmospheric forecasting model, GRIMs. For this purpose, each effect of reduced z1 and Δz is examined by comparing three experiments; the comparison between CTL and L65 is performed to identify the effects of z1, and the comparison between L65 and L91 is performed to determine the effects of Δz. Furthermore, impacts of increased vertical resolution on medium-range forecasts are also investigated.

The simulation with reduced z1 (L65) produces enhanced vertical turbulent mixing by determining a higher PBL height and associated turbulent velocity scales, which results in cooling/moistening above the PBL and warming/drying within the PBL. The resulting dryness near the surface contributes to the increase in latent heat flux and moistening in the troposphere. Changes in simulated temperature and moisture environments intensify large-scale precipitation due to higher relative humidity; however, convective precipitation is suppressed because the drier conditions near the surface inhibit convection triggering over the tropics. The increase in large-scale precipitation occurs in the areas with reduced convective precipitation.

The simulation with reduced Δz (L91) results in decreased detrained moisture at the convective cloud top, which is caused by effectively resolving the overshooting layer and reducing the amount of moisture in a buoyant parcel due to the constant value of C0 in this layer. The decrease in detrained moisture at the convective cloud top induces drier conditions which in turn, reduces cloud ice. The effect of ice clouds on longwave radiation flux, which is represented by cooling at the cloud top and warming in the cloud layer, is suppressed by the downward shift of ice clouds, and this significantly affects a cooling in the upper troposphere in L91.

The sensitivities of the physical processes due to increased vertical resolution from the CTL to L91 cause changes in temperature and moisture of the whole troposphere. Consequently, these changes modulate the biases of temperature and specific humidity; the cold/moist biases observed near the surface and warm/dry biases observed in the lower and midtroposphere are reduced, in contrast, the cold bias between 300 and 100 hPa is intensified. However, the increased vertical resolution does not cause significant changes in total precipitation skill, despite the individual changes in convective and large-scale precipitation.

Sensitivities are undesirable in numerical models that contain many assumptions in order to reproduce the real atmosphere. We could reduce model sensitivity by fine-tuning of the model without knowing where the sensitivities come from and how each process responds to them, but this is not the optimal method. This study improves our understanding of all processes related to the sensitivity to vertical resolution and allows us to propose methods to reduce the sensitivities. To reduce the sensitivity to z1, for example, using surface temperature or the modification of the formula for thermal access, θT could be proposed in the determination of Rib. Using surface temperature requires caution; however, since there is some risk in using surface temperature because of its high uncertainty in the NWP model. Modification for thermal access also requires more experiments in order to optimize the empirical parameter. To reduce the sensitivity to Δz, more precise parameterization is needed for experiments using fine grid spacing. As an example of possible solutions to eliminate the sensitivity, it is helpful to use a revised C0, which contributes to the vertical continuity of the amount of moisture of a buoyant parcel throughout the convective cloud and overshooting layers.

Acknowledgments

This work has been carried out through the R&D Project on the Development of the Global Numerical Weather Prediction Systems of the Korea Institute of Atmospheric Prediction Systems (KIAPS) funded by the Korea Meteorological Administration (KMA).

REFERENCES

  • Bae, S.-Y., S.-Y. Hong, and K.-S. Lim, 2016: Coupling WRF double-moment 6-class microphysics schemes to RRTMG radiation scheme in weather research forecasting model. Adv. Meteor., 2016, 111, https://doi.org/10.1155/2016/5070154.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Baek, S. H., 2017: A revised radiation package of G-packed McICA and two-stream approximation: Performance evaluation in a global weather forecasting model. J. Adv. Model. Earth Syst., 9, 16281640, https://doi.org/10.1002/2017MS000994.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bauer, P., and Coauthors, 2013: Model cycle 38r2: Components and performance. ECMWF Tech. Memo. 704, ECMWF, Reading, United Kingdom, 58 pp., https://www.ecmwf.int/sites/default/files/elibrary/2013/7986-model-cycle-38r2-components-and-performance.pdf.

  • Bélair, S., M. Roch, A.-M. Leduc, P. A. Vaillancourt, S. Laroche, and J. Mailhot, 2009: Medium-range quantitative precipitation forecasts from Canada’s new 33-km deterministic global operational system. Wea. Forecasting, 24, 690708, https://doi.org/10.1175/2008WAF2222175.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bhaskar Rao, D. V., D. Hari Prasad, D. Srinivas, and Y. Anjaneyulu, 2010: Role of vertical resolution in numerical models towards the intensification, structure and track of tropical cyclones. Mar. Geod., 33, 338355, https://doi.org/10.1080/01490419.2010.518066.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Charron, M., and Coauthors, 2012: The stratospheric extension of the Canadian global deterministic medium-range weather forecasting system and its impact on tropospheric forecasts. Mon. Wea. Rev., 140, 19241944, https://doi.org/10.1175/MWR-D-11-00097.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 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, 569585, https://doi.org/10.1175/1520-0493(2001)129<0569:CAALSH>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Choi, H.-J., and S.-Y. Hong, 2015: An updated subgrid orographic parameterization for global atmospheric forecast. J. Geophys. Res. Atmos., 120, 12 44512 457, https://doi.org/10.1002/2015JD024230.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Côté, J., J.-G. Desmarais, S. Gravel, A. Méthot, A. Patoine, M. Roch, and A. Staniforth, 1998: The operational CMC–MRB global environmental multiscale (GEM) model. Part II: Results. Mon.Wea. Rev., 126, 13971418, https://doi.org/10.1175/1520-0493(1998)126<1397:TOCMGE>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ek, M. B., K. E. Mitchell, Y. Lin, E. Rogers, P. Grunmann, V. Koren, G. Gayno, and J. D. Tarpley, 2003: Implementation of Noah land surface model advances in the National Centers for Environmental Prediction operational mesoscale Eta model. J. Geophys. Res., 108, 8851, https://doi.org/10.1029/2002JD003296.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Han, J., and H.-L. Pan, 2011: Revision of convection and vertical diffusion schemes in the NCEP Global Forecast System. Wea. Forecasting, 26, 520533, https://doi.org/10.1175/WAF-D-10-05038.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Han, J.-Y., S.-Y. Hong, K.-S. S. Lim, and J. Han, 2016: Sensitivity of a cumulus parameterization scheme to precipitation production representation and its impact on a heavy rain event over Korea. Mon. Wea. Rev., 144, 21252135, https://doi.org/10.1175/MWR-D-15-0255.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Han, J.-Y., S.-Y. Hong, and Y.-C. Kwon, 2017: Recent progress on cumulus parameterization in KIAPS Integrated Model (KIM). Preprints, Meteorology and Climate—Modeling for Air Quality Conf., Davis, CA, California Environmental Protection Agency’s Air Resources Board, https://www2.eventsxd.com/event/4129/meteorologyandclimatemodelingforairquality/sessions.

  • Holtslag, A. A., and B. A. Boville, 1993: Local versus nonlocal boundary-layer diffusion in a global climate model. J. Climate, 6, 18251842, https://doi.org/10.1175/1520-0442(1993)006<1825:LVNBLD>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hong, S.-Y., and J. Jang, 2018: Impacts of shallow convection processes on a simulated boreal summer climatology in a global atmospheric model. Asia-Pac. J. Atmos. Sci., 54, 361370, https://doi.org/10.1007/S13143-018-0013-3.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hong, S.-Y., J. Dudhia, and S.-H. Chen, 2004: A revised approach to ice microphysical processes for the bulk parameterization of clouds and precipitation. Mon. Wea. Rev., 132, 103120, https://doi.org/10.1175/1520-0493(2004)132<0103:ARATIM>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
  • Hong, S.-Y., and Coauthors, 2013: The Global/Regional Integrated Model System (GRIMs). Asia-Pac. J. Atmos. Sci., 49, 219243, https://doi.org/10.1007/s13143-013-0023-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hong, S.-Y., and Coauthors, 2018: The Korean Integrated Model (KIM) System for global weather forecasting. Asia-Pac. J. Atmos. Sci., 54, 267292, https://doi.org/10.1007/S13143-018-0028-9.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Huffman, G. J., and Coauthors, 2007: The TRMM Multisatellite Precipitation Analysis (TMPA): Quasi-global, multiyear, combined-sensor precipitation estimates at fine scales. J. Hydrometeor., 8, 3855, https://doi.org/10.1175/JHM560.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Iacono, M.-J., J. S. Delamere, E. J. Mlawer, M. W. Shepherd, S. A. Clough, and W. D. Collins, 2008: Radiative forcing by longlived greenhouse gases: Calculation with the AER radiative transfer models. J. Geophys. Res., 113, D13103, https://doi.org/10.1029/2008JD009944.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Inness, P. M., J. M. Slingo, S. J. Woolnough, R. B. Neale, and V. D. Pope, 2001: Organization of tropical convection in a GCM with varying vertical resolution; Implications for the simulation of the Madden–Julian Oscillation. Climate Dyn., 17, 777793, https://doi.org/10.1007/s003820000148.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kim, E.-J., and S.-Y. Hong, 2010: Impact of air-sea interaction on East Asian summer monsoon climate in WRF. J. Geophys. Res., 115, D19118, https://doi.org/10.1029/2009JD013253.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Koo, M.-S., S. Baek, K.-H. Seol, and K. Cho, 2017: Advances in land surface modeling of KIAPS based on the Noah land surface model. Asia-Pac. J. Atmos. Sci., 53, 361373, https://doi.org/10.1007/s13143-017-0043-2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kwon, Y. C., and S.-Y. Hong, 2017: A mass-flux cumulus parameterization scheme across gray-zone resolutions. Mon. Wea. Rev., 145, 583598, https://doi.org/10.1175/MWR-D-16-0034.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, E., and S.-Y. Hong, 2019: Impact of the sea surface salinity on simulated precipitation in a global numerical weather prediction model. J. Geophys. Res. Atmos., 124, 719730, https://doi.org/10.1029/2018JD029591.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, E.-H., E. Lee, R. Park, Y.-C. Kwon, and S.-Y. Hong, 2018: Impact of turbulent mixing in the stratocumulus-topped boundary layer on numerical weather prediction. Asia-Pac. J. Atmos. Sci., 54, 371384, https://doi.org/10.1007/S13143-018-0024-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mass, C. F., D. Ovens, K. Westrick, and B. A. Colle, 2002: Does increasing horizontal resolution produce more skillful forecasts? Bull. Amer. Meteor. Soc., 83, 407430, https://doi.org/10.1175/1520-0477(2002)083<0407:DIHRPM>2.3.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • McFarlane, S. A., J. H. Mather, and T. P. Ackerman, 2007: Analysis of tropical radiative heating profiles: A comparison of models and observations. J. Geophys. Res., 112, D14218, https://doi.org/10.1029/2006JD008290.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pan, H.-L., and W.-S. Wu, 1995: Implementing a mass flux convective parameterization package for the NMC medium-range forecast model. NMC Office Note 409, 40 pp.

  • Park, R.-S., J.-H. Chae, and S.-Y. Hong, 2016: A revised prognostic cloud fraction scheme in a global forecasting system. Mon. Wea. Rev., 144, 12191229, https://doi.org/10.1175/MWR-D-15-0273.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Philip, A., T. Bergot, Y. Bouteloup, and F. Bouyssel, 2016: The impact of vertical resolution on fog forecasting in the kilometric-scale model AROME: A case study and statistics. Wea. Forecasting, 31, 16551671, https://doi.org/10.1175/WAF-D-16-0074.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Retsch, M. H., C. Hohenegger, and B. Stevens, 2017: Vertical resolution refinement in an aqua-planet and its effect on the ITCZ. J. Adv. Model. Earth Syst., 9, 24252436, https://doi.org/10.1002/2017MS001010.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Richardson, D., J. Bidlot, R. Buizza, L. Ferranti, A. Ghelli, G. van der Grijn, F. Vitart, and E. Zsoter, 2006: Verification statistics and evaluations of ECMWF forecasts in 2005-2006. ECMWF Tech. Memo. 504. ECMWF, Reading, United Kingdom, 44 pp., https://www.ecmwf.int/sites/default/files/elibrary/2006/11902-verification-statistics-and-evaluations-ecmwf-forecasts-2005-2006.pdf.

  • Roeckner, E., and Coauthors, 2006: Sensitivity of simulated climate to horizontal and vertical resolution in the ECHAM5 atmosphere model. J. Climate, 19, 37713791, https://doi.org/10.1175/JCLI3824.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ruti, P. M., D. D. Rocco, and S. Gualdi, 2006: Impact of increased vertical resolution on simulation of tropical climate. Theor. Appl. Climatol., 85, 6180, https://doi.org/10.1007/s00704-005-0174-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shin, H. H., and S.-Y. Hong, 2015: Representation of the subgrid-scale turbulent transport in convective boundary layers at gray-zone resolutions. Mon. Wea. Rev., 143, 250271, https://doi.org/10.1175/MWR-D-14-00116.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shin, H. H., S.-Y. Hong, and J. Dudhia, 2012: Impacts of the lowest model level height on the performance of planetary boundary layer parameterizations. Mon. Wea. Rev., 140, 664682, https://doi.org/10.1175/MWR-D-11-00027.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Simmons, A. J., D. M. Burridge, M. Jarraud, C. Girard, and W. Wergen, 1989: The ECMWF medium-range prediction models: Development of the numerical formulations and the impact of increased resolution. Meteor. Atmos. Phys., 40, 2860, https://doi.org/10.1007/BF01027467.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Takeuchi, Y., and Coauthors, 2013: Outline of the Operational Numerical Weather Prediction at the Japan Meteorological Agency: Appendix to WMO Technical Progress Report on the Global Data-processing and Forecasting System (GDPFS) and Numerical Weather Prediction (NWP) Research. Japan Meteorological Agency, 188 pp.

  • Tardif, R., 2007: The impact of vertical resolution in the explicit numerical forecasting of radiation fog: A case study. Pure Appl. Geophys., 164, 12211240, https://doi.org/10.1007/s00024-007-0216-5.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tompkins, A. M., and K. A. Emanuel, 2000: The vertical resolution sensitivity of simulated equilibrium tropical temperature and water-vapour profiles. Quart. J. Roy. Meteor. Soc., 126, 12191238, https://doi.org/10.1002/qj.49712656502.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhang, D.-L., and X. Wang, 2003: Dependence of hurricane intensity and structures on vertical resolution and time-step size. Adv. Atmos. Sci., 20, 711725, https://doi.org/10.1007/BF02915397.

    • Crossref
    • Search Google Scholar
    • Export Citation
Save