• Avissar, R., , and Schmidt T. , 1998: An evaluation of the scale at which ground-surface heat flux patchiness affects the convective boundary layer using large-eddy simulations. J. Atmos. Sci., 55, 26662689, doi:10.1175/1520-0469(1998)055<2666:AEOTSA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Balaji, V., , and Clark T. L. , 1988: Scale selection in locally forced convective fields and the initiation of deep cumulus. J. Atmos. Sci., 45, 31883211, doi:10.1175/1520-0469(1988)045<3188:SSILFC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Baldwin, M. E., , and Mitchell K. E. , 1998: Progress on the NCEP hourly multi-sensor U.S. precipitation analysis for operations and GCIP research. Preprints, Second Symp. on Integrated Observing Systems, Phoenix, AZ, Amer. Meteor. Soc., 1011.

  • Balling, R. C., 1985: Warm season nocturnal precipitation in the Great Plains of the United States. J. Climate Appl. Meteor., 24, 13831387, doi:10.1175/1520-0450(1985)024<1383:WSNPIT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Barros, A. P., , and Hwu W. , 2002: A study of land–atmosphere interactions during summertime rainfall using a mesoscale model. J. Geophys. Res., 107, 17-1–17-18, doi:10.1029/2000JD000254.

    • Search Google Scholar
    • Export Citation
  • Barros, A. P., , Li W. , , Kang D.-H. , , Prat O. , , Shrestha P. , , Brunsell N. , , and Rhan T. A. , 2013: Observations of surface fluxes and boundary layer instability in the southern Great Plains during CLASIC. Rep. EPL-2013-CLASIC-1, Environmental Physics Laboratory, Duke University, 57 pp.

  • Betts, A. K., 1992: FIFE atmospheric boundary layer budget methods. J. Geophys. Res.,97, 18 52318 531, doi:10.1029/91JD03172.

  • Bright, D. R., , and Mullen S. L. , 2002: Short-range ensemble forecasts of precipitation during the southwest monsoon. Wea. Forecasting, 17, 10801100, doi:10.1175/1520-0434(2002)017<1080:SREFOP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Carbone, R. E., , and Tuttle J. D. , 2008: Rainfall occurrence in the U.S. warm season: The diurnal cycle. J. Climate, 21, 41324146, doi:10.1175/2008JCLI2275.1.

    • Search Google Scholar
    • Export Citation
  • Carbone, R. E., , Tuttle J. D. , , Ahijevych D. A. , , and Trier S. B. , 2002: Inferences of predictability associated with warm season precipitation episodes. J. Atmos. Sci., 59, 20332056, doi:10.1175/1520-0469(2002)059<2033:IOPAWW>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Denis, B., , Côté J. , , and Laprise R. , 2002: Spectral decomposition of two-dimensional atmospheric fields on limited-area domains using the discrete cosine transform (DCT). Mon. Wea. Rev., 130, 18121829, doi:10.1175/1520-0493(2002)130<1812:SDOTDA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Dirmeyer, P. A., and et al. , 2012: Simulating the diurnal cycle of rainfall in global climate models: Resolution versus parameterization. Climate Dyn., 39, 399418, doi:10.1007/s00382-011-1127-9.

    • Search Google Scholar
    • Export Citation
  • Dong, X., and et al. , 2011: Investigation of the 2006 drought and 2007 flood extremes at the southern Great Plains through an integrative analysis of observations. J. Geophys. Res.,116, D03204, doi:10.1029/2010JD014776.

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

    • Search Google Scholar
    • Export Citation
  • Gilmore, M. S., , and Wicker L. J. , 1998: The influence of midtropospheric dryness on supercell morphology and evolution. Mon. Wea. Rev., 126, 943958, doi:10.1175/1520-0493(1998)126<0943:TIOMDO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Hong, S.-Y., , and Lim J.-O. J. , 2006: The WRF Single-Moment 6-Class Microphysics Scheme (WSM6). J. Korean Meteor. Soc., 42, 121151.

  • Hong, S.-Y., , Noh Y. , , and Dudhia J. , 2006: A new vertical diffusion package with an explicit treatment of entrainment processes. Mon. Wea. Rev., 134, 23182341, doi:10.1175/MWR3199.1.

    • Search Google Scholar
    • Export Citation
  • James, R. P., , Markowski P. M. , , and Fritsch J. M. , 2006: Bow echo sensitivity to ambient moisture and cold pool strength. Mon. Wea. Rev., 134, 950964, doi:10.1175/MWR3109.1.

    • Search Google Scholar
    • Export Citation
  • Janjić, Z. I., 2002: Nonsingular implementation of the Mellor–Yamada level 2.5 scheme in the NCEP Meso model. NCEP Office Note 437, 61 pp. [Available online at www.emc.ncep.noaa.gov/officenotes/newernotes/on437.pdf.]

  • Kain, J. S., 2004: The Kain–Fritsch convective parameterization: An update. J. Appl. Meteor., 43, 170181, doi:10.1175/1520-0450(2004)043<0170:TKCPAU>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Koster, R. D., and et al. , 2004: Regions of strong coupling between soil moisture and precipitation. Science, 305, 11381140, doi:10.1126/science.1100217.

    • Search Google Scholar
    • Export Citation
  • Lamb, P., , Portis D. H. , , and Zangvil A. , 2012: Investigation of large-scale atmospheric moisture budget and land surface interactions over U.S. southern Great Plains including for CLASIC (June 2007). J. Hydrometeor., 13, 1719–1738, doi:10.1175/JHM-D-12-01.1.

    • Search Google Scholar
    • Export Citation
  • Lanicci, J. M., , Carlson T. N. , , and Warner T. T. , 1987: Sensitivity of the Great Plains severe-storm environment to soil-moisture distribution. Mon. Wea. Rev., 115, 26602673, doi:10.1175/1520-0493(1987)115<2660:SOTGPS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • LeMone, M. A., , and Pennell W. T. , 1976: The relationship of trade wind cumulus distribution to subcloud layer fluxes and structure. Mon. Wea. Rev., 104, 524539, doi:10.1175/1520-0493(1976)104<0524:TROTWC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Lin, Y., , and Mitchell K. E. , 2005: The NCEP Stage II/IV hourly precipitation analyses: Development and applications. Preprints, 19th Conf. on Hydrology, San Diego, CA, Amer. Meteor. Soc., 1.2. [Available online at https://ams.confex.com/ams/pdfpapers/83847.pdf.]

  • Luo, Y., , Berbery E. H. , , Mitchell K. E. , , and Betts A. K. , 2007: Relationships between land-surface and near-surface atmospheric variables in the NCEP North American Regional Reanalysis. J. Hydrometeor., 8, 11841203, doi:10.1175/2007JHM844.1.

    • Search Google Scholar
    • Export Citation
  • Lynn, B. H., , Tao W.-K. , , and Wetzel P. J. , 1998: A study of landscape-generated deep moist convection. Mon. Wea. Rev., 126, 928942, doi:10.1175/1520-0493(1998)126<0928:ASOLGD>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Mahrt, L., , Sun J. , , Vickers D. , , Macpherson J. I. , , Pederson J. R. , , and Desjardins R. L. , 1994: Observations of fluxes and inland breezes over a heterogeneous surface. J. Atmos. Sci., 51, 24842499, doi:10.1175/1520-0469(1994)051<2484:OOFAIB>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • McPherson, R. A., and et al. , 2007: Statewide monitoring of the mesoscale environment: A technical update on the Oklahoma Mesonet. J. Atmos. Oceanic Technol., 24, 301321, doi:10.1175/JTECH1976.1.

    • Search Google Scholar
    • Export Citation
  • Mesinger, F., and et al. , 2006: North American Regional Reanalysis. Bull. Amer. Meteor. Soc., 87, 343360, doi:10.1175/BAMS-87-3-343.

    • Search Google Scholar
    • Export Citation
  • Miller, M. A., and et al. , 2007: SGP Cloud and Land Surface Interaction Campaign (CLASIC): Science and Implementation Plan. Rep. DOE/SC-ARM-0703,Office of Biological and Environmental Research, U.S. Department of Energy, 14 pp. [Available online at www.arm.gov/publications/programdocs/doe-sc-arm-0703.pdf.]

  • Parker, M. D., , and Johnson R. H. , 2000: Organizational modes of midlatitude mesoscale convective systems. Mon. Wea. Rev., 128, 34133436, doi:10.1175/1520-0493(2001)129<3413:OMOMMC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Santanello, J. A., Jr., , Friedl M. A. , , and Ek M. B. , 2007: Convective planetary boundary layer interactions with the land surface at diurnal time scales: Diagnostics and feedbacks. J. Hydrometeor., 8, 10821097, doi:10.1175/JHM614.1.

    • Search Google Scholar
    • Export Citation
  • Santanello, J. A., Jr., , Peters-Lidard C. D. , , Kumar S. V. , , Alonge C. , , and Tao W.-K. , 2009: A modeling and observational framework for diagnosing local land–atmosphere coupling on diurnal time scales. J. Hydrometeor., 10, 577599, doi:10.1175/2009JHM1066.1.

    • Search Google Scholar
    • Export Citation
  • Santanello, J. A., Jr., , Peters-Lidard C. D. , , and Kumar S. V. , 2011: Diagnosing the sensitivity of local land–atmosphere coupling via the soil moisture–boundary layer interaction. J. Hydrometeor., 12, 766786, doi:10.1175/JHM-D-10-05014.1.

    • Search Google Scholar
    • Export Citation
  • Schär, C., , Lüthi D. , , Beyerle U. , , and Heise E. , 1999: The soil–precipitation feedback: A process study with a regional climate model. J. Climate, 12, 722741, doi:10.1175/1520-0442(1999)012<0722:TSPFAP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Seo, D. J., 1998: Real-time estimation of rainfall fields using radar rainfall and raingauge data. J. Hydrol., 209, 131147.

  • Shaw, B. L., , Pielke R. A. , , and Ziegler C. L. , 1997: A three-dimensional numerical simulation of a Great Plains dryline. Mon. Wea. Rev., 125, 14891506, doi:10.1175/1520-0493(1997)125<1489:ATDNSO>2.0.CO;2.

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

  • Stensrud, D. J., 2007: Parameterization Schemes: Keys to Understanding Numerical Weather Prediction Models. Cambridge University Press, 459 pp.

  • Sun, X., , and Barros A. P. , 2012: The impact of forcing dataset on the high resolution simulation of tropical storm Ivan (2004) in the southern Appalachians. Mon. Wea. Rev., 140, 3300–3326, doi:10.1175/MWR-D-11-00345.1.

    • Search Google Scholar
    • Export Citation
  • Sun, X., , and Barros A. P. , 2014: High resolution simulation of tropical storm Ivan (2004) in the Southern Appalachians: Role of planetary boundary-layer schemes and cumulus parameterization. Quart. J. Roy. Meteor. Soc., doi: 10.1002/qj.2255, in press.

    • Search Google Scholar
    • Export Citation
  • Tao, K., , and Barros A. P. , 2008: Metrics to describe the dynamical evolution of atmospheric moisture: Intercomparison of model (NARR) and observations (ISCCP). J. Geophys. Res., 113, D14125, doi:10.1029/2007JD009337.

    • Search Google Scholar
    • Export Citation
  • Trier, S. B., , Chen F. , , and Manning K. W. , 2004: A study of convection initiation in a mesoscale model using high-resolution land-surface initial conditions. Mon. Wea. Rev., 132, 29542976, doi:10.1175/MWR2839.1.

    • Search Google Scholar
    • Export Citation
  • Tuttle, J. D., , and Davis C. A. , 2006: Corridors of warm season precipitation in the central United States. Mon. Wea. Rev., 134, 22972317, doi:10.1175/MWR3188.1.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., 1975: Diurnal variations in precipitation and thunderstorm frequency over the conterminous United States. Mon. Wea. Rev., 103, 406419, doi:10.1175/1520-0493(1975)103<0406:DVIPAT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Weckwerth, T. M., , Wilson J. W. , , and Wakimoto R. M. , 1996: Thermodynamic variability within the convective boundary layer due to horizontal convective rolls. Mon. Wea. Rev., 124, 769784, doi:10.1175/1520-0493(1996)124<0769:TVWTCB>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Weckwerth, T. M., , Wilson J. W. , , Wakimoto R. M. , , and Crook N. A. , 1997: Horizontal convective rolls: Determining the environmental conditions supporting their existence and characteristics. Mon. Wea. Rev., 125, 505526, doi:10.1175/1520-0493(1997)125<0505:HCRDTE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wei, J., , Dickinson R. E. , , and Chen H. , 2008: A negative soil moisture–precipitation relationship and its causes. J. Hydrometeor., 9, 13641376, doi:10.1175/2008JHM955.1.

    • Search Google Scholar
    • Export Citation
  • Weisman, M. L., 1992: The role of convectively generated rear-inflow jets in the evolution of long-lived mesoconvective systems. J. Atmos. Sci., 49, 18261847, doi:10.1175/1520-0469(1992)049<1826:TROCGR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wu, D., , Dong X. , , Xi B. , , Feng Z. , , Kennedy A. , , Mullendore G. , , Gilmore M. , , and Tao W.-K. , 2013: Impacts of microphysical scheme on convective and stratiform characteristics in two high precipitation squall line events. J. Geophys. Res. Atmos., 118, 11 11911 135, doi:10.1002/jgrd.50798.

    • Search Google Scholar
    • Export Citation
  • Zeng, X., , Barlage M. , , Castro C. , , and Fling K. , 2010: Comparison of land–precipitation coupling strength using observations and models. J. Hydrometeor., 11, 979994, doi:10.1175/2010JHM1226.1.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Weather Surveillance Radar-1988 Doppler (WSR-88D) reflectivity plots from station KTLX for (a) 0127, (b) 0316, (c) 0435, and (d) 0612 UTC 20 Jun 2007. The location of the KTLX radar in Oklahoma City, OK, is marked (www.ncdc.noaa.gov/oa/radar).

  • View in gallery

    Time–lat (Hovmöller) plot of hourly stage IV precipitation (http://data.eol.ucar.edu) averaged over 100–95°W from 0000 UTC 13 Jun to 0000 UTC 22 Jun. Stage IV data have a horizontal grid spacing of 4 km.

  • View in gallery

    Region of study and model domains including topography (m). D01 (outermost grid) has 12-km horizontal grid spacing (122 × 122 grid cells), D02 (middle nest) has 4-km horizontal grid spacing (163 × 163 grid cells), and D03 (innermost nest) has 1.333-km horizontal grid spacing (124 × 124 grid cells).

  • View in gallery

    Land surface attributes in D03: (a) land use type and (b) soil texture for D03 from the USGS land cover and soil texture dataset available in the WRF preprocessing system. Marked lat cross sections A–A″, B–B″, C–C″, and D–D″ in (a) and northwest–southeast in (b) are referred to elsewhere in the manuscript, including Figs. 5, 6, 12, and 16.

  • View in gallery

    Hovmöller diagrams of sensible heat flux for the time period from 0900 UTC 19 Jun to 0900 UTC 20 Jun 2007: (a) YSU-HET along A–A″, (b) YSU-ADG along A–A″, (c) YSU-HET along D–D″, and (d) YSU-ADG along D–D″. The negative sign indicates that the flux is leaving the land surface. Cross sections A–A″ and D–D″ are marked in Fig. 4a.

  • View in gallery

    As in Fig. 5, but for latent heat fluxes.

  • View in gallery

    Pressure-weighted integrated vertical motion (shaded) fields for wavelengths less than 12 km in the lower (from the surface to 700 hPa) and midtroposphere (from 400 to 700 hPa, where the ±5 hPa m s−1 contour is shown and dashed contours indicate downward motion) at 1730 UTC 19 Jun 2007 for (a) the YSU-HET and (b) YSU-ADG and at 1840 UTC 19 Jun 2007 for (c) YSU-HET, (d) YSU-ADG, (e) MYJ-HET, and (f) MYJ-ADG. Shading shows the development of boundary layer eddies, while the contour lines (5 hPa m s−1) show that organized vertical motions at midlevels align perpendicular to the structures at low levels.

  • View in gallery

    D03-average atmospheric thermodynamic profile (temperature, red; dewpoint, blue) for (a) 1200, (b) 1500, (c) 1800, and (d) 2100 UTC 19 Jun 2007 and (e) 0000 and (f) 0300 UTC 20 Jun 2007 (MYJ-HET, solid lines; YSU-HET, dashed lines).

  • View in gallery

    LoCo mixing diagram for D03 pixels containing the following combinations of soil texture and land use: sand + grassland (black), silt loam + dryland/cropland/pasture (medium gray), and silt loam + grassland (light gray) for the YSU-HET simulation.

  • View in gallery

    (a) Scatterplot of 0000 UTC 20 Jun 2007 CAPE values and entrainment Bowen ratio for D03 pixels containing the following soil texture and land use combinations: sand + grassland pixels (black), silt loam + dryland/cropland/pasture pixels (medium gray), and silt loam + grassland pixels (light gray) for the MYJ-HET simulation. (b) Scatterplot of 0000 UTC 20 Jun 2007 CAPE values and surface Bowen ratio for sand + grassland pixels (black), silt loam + dryland/cropland/pasture pixels (medium gray), and silt loam + grassland pixels (light gray) for the YSU-HET simulation.

  • View in gallery

    YSU-HET D03-average sensible (solid) and latent (dashed) heat fluxes and soil saturation (triangles) for sandy (red) and silt loam pixels (blue) from 1200 UTC 19 Jun to 0000 UTC 20 Jun 2007.

  • View in gallery

    Along-storm cross-sectional view in D03 (grid resolution Δs = 1.333 km; cross section marked in Fig. 4b) of (a) vertical velocity, (b) equivalent potential temperature, (c) simulated radar reflectivity, and (d) CAPE for MYJ-HET at 0620 UTC 20 Jun 2007. The quantities on the x axis are grid points numbered from west to east.

  • View in gallery

    (a) D03-average rainfall for MYJ simulations MYJ-ADG (black), MYJ-D3G (dashed medium gray), MYJ-D3S (short dashed), and MYJ-HET (dashed light gray). Observations from stage IV precipitation data are shown in black asterisks. (b) D03-average rainfall for YSU-ADG (black), YSU-D3G (dashed medium gray), YSU-D3S (short dashed), and YSU-HET (dashed light gray). The timestamp corresponds to 20 Jun 2007.

  • View in gallery

    (d) Synthetic example of development of cold pool velocity Hovmöller diagram from (a)–(c) three velocity fields. Each line in the diagram was created by dividing the domain into bins one grid cell wide, aligned in the northwest–southeast direction, from the lower left corner (in the southwest) to the upper right corner (in the northeast). Along each northwest–southeast line (perpendicular to the storm motion), averages were computed for cells that contained a nonmissing value of cold pool speed. Objects moving east of a 45° angle appear to move toward increasing numbers on the x axis [in (d)] and objects moving south of a 45° angle appear to move to the left. Objects moving on a 45° angle, as in this example, appear to move straight in (d).

  • View in gallery

    Theoretical cold pool speed as in the Hovmöller diagram presented in Fig. 14 for (a) MYJ-D3G, (b) MYJ-D3S, (c) MYJ-HET, and (d) YSU-HET. The timestamp corresponds to 20 Jun 2007.

  • View in gallery

    Hovmöller plot of PBL height along D–D″ (marked in Fig. 4a) for (a) MYJ-D3G, (b) MYJ-D3S, and (c) YSU-HET. (d) The 0–3-km vertical wind shear vector at 0100 UTC for MYJ-D3G (black) and MYJ-D3S (purple). The difference in the magnitude of the shear vector MYJ (D3G − D3S) is shaded, with warm colors representing stronger shear in the case of the D3S simulation.

  • View in gallery

    Skew T–logp diagrams for (a) 0000 UTC MYJ-D3S (dashed) and MYJ-D3S-1way (solid), (b) 0200 UTC MYJ-D3S (dashed) and MYJ-D3S-1way (solid), (c) 0000 UTC MYJ-D3S (dashed) and MYJ-D3G (solid), and (d) 0200 UTC MYJ-D3S (dashed) and MYJ-D3G (solid) 20 Jun 2007. Profiles are averaged over the portion of D02 spanning from 36.25°N, 100°W to 36.75°N, 99.5°W. Each panel contains a trace of temperature and dewpoint temperature. Different dash patterns are used to differentiate between simulations.

  • View in gallery

    Temporal evolution of the spatial distribution of CAPE for a parcel with properties averaged over the lowest 500 m at the times indicated for (a)–(c) MYJ-D3S and (d)–(f) MYJ-D3S-1way. Also marked are the 10-m wind vectors.

  • View in gallery

    (a) D03-average rainfall for MYJ simulations MYJ-D3S-1way, MYJ-D3G, MYJ-D3S, and MYJ-D3G-1way as indicated in the legend. Observations from stage IV precipitation data are shown as black asterisks. The timestamp corresponds to 20 Jun 2007. (b) D02 cumulative rainfall for (left) MYJ-D3S and (right) MYJ-D3S-1way for wavelengths >12 km.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 28 28 1
PDF Downloads 19 19 0

A Study of the Role of Daytime Land–Atmosphere Interactions on Nocturnal Convective Activity in the Southern Great Plains during CLASIC

View More View Less
  • 1 Pratt School of Engineering, Duke University, Durham, North Carolina
© Get Permissions
Full access

Abstract

This study examines whether and how land–atmosphere interactions can have an impact on nocturnal convection over the southern Great Plains (SGP) through numerical simulations of an intense nocturnal mesoscale convective system (MCS) on 19–20 June 2007 with the Weather Research and Forecasting (WRF) Model. High-resolution nested simulations were conducted using realistic and idealized land surfaces and two planetary boundary layer (PBL) parameterizations (PBLp): Yonsei University (YSU) and Mellor–Yamada–Janjić (MYJ). Differences in timing and amount of MCS precipitation among observations and model results were examined in the light of daytime land–atmosphere interactions, nocturnal prestorm environment, and cold pool strength. At the meso-γ scale, land cover and soil type have as much of an effect on the simulated prestorm environment as the choice of PBLp: MYJ simulations exhibit strong sensitivity to changes in the land surface in contrast to negligible impact in the case of YSU. At the end of the afternoon, as the boundary layer collapses, a more homogeneous and deeper PBL (and stronger low-level shear) is evident for YSU as compared to MYJ when initial conditions and land surface properties are the same. At the meso-β scale, propagation speed is faster and organization (bow echo morphology) and cold pool strength are enhanced when nocturnal PBL heights are higher, and there is stronger low-level shear in the prestorm environment independent of the boundary layer parameterization for different land surface conditions. A comparison of one- and two-way nested MYJ results demonstrates how daytime land–atmosphere interactions modify the prestorm environment remotely through advection of low-level thermodynamic features. This remote feedback strongly impacts the MCS development phase as well as its spatial organization and propagation velocity and, consequently, nocturnal rainfall. These results indicate that synoptic- and meso-α-scale dynamics can play an important role in determining the spatial and temporal scales over which precipitation feedbacks of land–atmosphere interactions emerge regionally. Finally, this study demonstrates the high degree of uncertainty in defining the spatial and temporal scales of land–atmosphere interactions where and when organized convection is dominant.

Corresponding author address: Ana P. Barros, Duke University, Box 90287, 2447 CIEMAS Fitzpatrick Bldg., Durham, NC 27708. E-mail: barros@duke.edu

Abstract

This study examines whether and how land–atmosphere interactions can have an impact on nocturnal convection over the southern Great Plains (SGP) through numerical simulations of an intense nocturnal mesoscale convective system (MCS) on 19–20 June 2007 with the Weather Research and Forecasting (WRF) Model. High-resolution nested simulations were conducted using realistic and idealized land surfaces and two planetary boundary layer (PBL) parameterizations (PBLp): Yonsei University (YSU) and Mellor–Yamada–Janjić (MYJ). Differences in timing and amount of MCS precipitation among observations and model results were examined in the light of daytime land–atmosphere interactions, nocturnal prestorm environment, and cold pool strength. At the meso-γ scale, land cover and soil type have as much of an effect on the simulated prestorm environment as the choice of PBLp: MYJ simulations exhibit strong sensitivity to changes in the land surface in contrast to negligible impact in the case of YSU. At the end of the afternoon, as the boundary layer collapses, a more homogeneous and deeper PBL (and stronger low-level shear) is evident for YSU as compared to MYJ when initial conditions and land surface properties are the same. At the meso-β scale, propagation speed is faster and organization (bow echo morphology) and cold pool strength are enhanced when nocturnal PBL heights are higher, and there is stronger low-level shear in the prestorm environment independent of the boundary layer parameterization for different land surface conditions. A comparison of one- and two-way nested MYJ results demonstrates how daytime land–atmosphere interactions modify the prestorm environment remotely through advection of low-level thermodynamic features. This remote feedback strongly impacts the MCS development phase as well as its spatial organization and propagation velocity and, consequently, nocturnal rainfall. These results indicate that synoptic- and meso-α-scale dynamics can play an important role in determining the spatial and temporal scales over which precipitation feedbacks of land–atmosphere interactions emerge regionally. Finally, this study demonstrates the high degree of uncertainty in defining the spatial and temporal scales of land–atmosphere interactions where and when organized convection is dominant.

Corresponding author address: Ana P. Barros, Duke University, Box 90287, 2447 CIEMAS Fitzpatrick Bldg., Durham, NC 27708. E-mail: barros@duke.edu

1. Introduction

Previously, the southern Great Plains (SGP) region was identified as a maximum “hot spot” for land–atmosphere interactions on time and spatial scales relevant for climate studies (Koster et al. 2004), though the coupling mechanism proper, in particular the seasonality and spatial scales of soil moisture S and precipitation P feedbacks (e.g., the SP relationship) and the role of evapotranspiration, remains the subject of active research (Luo et al. 2007; Wei et al. 2008). Warm season precipitation in the SGP exhibits a strong diurnal cycle with a nocturnal maximum (Wallace 1975; Balling 1985; Carbone et al. 2002), which cannot be captured in global climate model simulations, and thus poses significant challenges to elucidating land–atmosphere interactions at the transition from weather to climate time scales (Dirmeyer et al. 2012).

The analysis of SP feedbacks in the context of climate studies is typically conducted at seasonal or at least monthly time scales, and the metrics used tend to be local, meaning that the expression of cause–effect relationships is investigated on a gridpoint by gridpoint basis (Koster et al. 2004; Zeng et al. 2010; Tao and Barros 2008), thus not accounting explicitly for lateral advection effects. At weather time scales, many idealized modeling studies have investigated the impact of the partitioning of latent and sensible heat fluxes on the evolution and organization of the boundary layer and convection. Balaji and Clark (1988) suggested that the initiation of deep convection through heterogeneous sensible heating in sheared environments can influence the spatial and temporal distribution of new convective cells in multicellular storm systems, leading to the development of internal gravity waves above the boundary layer that trigger the initiation of convective lifting and the formation of cumulus clouds. Using field observations and modeling experiments, Weckwerth et al. (1997) showed that there is a strong relationship between the magnitude of surface sensible fluxes and the circulation regime in a convective boundary layer, with rolls forming even in the presence of weak shear for high sensible heat fluxes before degenerating into disorganized convection. Avissar and Schmidt (1998) investigated the scale dependence of boundary layer rolls due to spatial differences in surface sensible fluxes and pointed out the strong co-organization of clouds and moisture convergence associated with boundary roll circulations consistent with field observations (e.g., Weckwerth et al. 1996; LeMone and Pennell 1976). Mahrt et al. (1994) and Lynn et al. (1998) showed that discontinuities in turbulent fluxes and soil moisture, respectively, can generate inland sea breeze–like fronts that are associated with the initiation of moist convection. These fronts are dependent upon patch size (strong circulations require large patches on the order of 100 km) and background wind, which govern the type of clouds that form. Furthermore, the importance of correct initialization of soil moisture to improve the estimation of surface energy fluxes has been widely recognized in different types of studies (e.g., Lanicci et al. 1987; Trier et al. 2004; Shaw et al. 1997; Schär et al. 1999; Santanello et al. 2007).

Overall, previous modeling studies using idealized forcing have shown that the triggering of gravity waves and their persistence in both space and time are linked to differential heating of the land surface, the environmental directional shear, and the relationships between the two. They also show that convective initiation is localized in regions of moisture convergence and high instability in the boundary layer. If and how local-scale effects, such as the effects of heterogeneity in land cover and soil moisture and texture, are able to affect nonlocal convective processes such as the initiation and propagation of mesoscale convective systems (MCSs) has not yet been established.

Carbone and Tuttle (2008) describe three dominant mechanisms of nocturnal rainfall in the SGP: 1) the eastward propagation of rainfall systems originating along the Continental Divide, 2) the nocturnal reversal of the mountain–plains circulations, and 3) moisture convergence and instability associated with the Great Plains low-level jet. The purpose of this study is to address the question of whether and how daytime land–atmosphere interactions can influence the eastward propagation of nocturnal deep convection in the warm season. The focus is on the day-to-night evolution of environmental conditions in the lower troposphere for the case of an MCS (Fig. 1) that propagated over Oklahoma (OK) on 19 and 20 June 2007 during the Cloud and Land Surface Interaction Campaign (CLASIC; Miller et al. 2007) that is illustrative of eastward-propagating rainfall systems.

Fig. 1.
Fig. 1.

Weather Surveillance Radar-1988 Doppler (WSR-88D) reflectivity plots from station KTLX for (a) 0127, (b) 0316, (c) 0435, and (d) 0612 UTC 20 Jun 2007. The location of the KTLX radar in Oklahoma City, OK, is marked (www.ncdc.noaa.gov/oa/radar).

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

Lamb et al. (2012) conducted a large-scale analysis of the moisture budget in the SGP during CLASIC and reported that the contribution of regional moisture recycling to the above-normal precipitation was small, which could be interpreted as indicative of weak SP coupling, albeit from a moisture budget perspective only. Here, we are interested in characterizing the physical mechanisms by which land surface heterogeneities modulate the surface energy balance and boundary layer processes during daytime and how these mechanisms impact storm propagation and nocturnal rainfall through model simulations. In particular, we investigate the hypothesis that the nighttime release of residual low-level instability is tied, both dynamically and thermodynamically, to land surface conditions in the prestorm environment and helps fuel convective propagation. A complementary goal of this research is to assess the impact of model configuration and, in particular, boundary layer parameterizations on emergent simulated physics. Because eastward-propagating MCSs play a critical role in SGP hydrometeorology, this case study can provide insight to understand land–atmosphere interactions at climate scales.

June 2007 was a historically wet month for the state of Oklahoma, and the heavy rainfall during the night of 19 June and early morning of 20 June followed a 1-week-long period of daily late afternoon rainfall, with the exception of the preceding day. A detailed study of the large-scale dynamical conditions associated with the extremely wet conditions in 2007 was conducted by Dong et al. (2011). Specifically, they point out anomalously low 500-hPa geopotential heights over Texas and Oklahoma, a strong low-level jet, and enhanced southerly moisture transport (Dong et al. 2011). Figure 2 shows the Hovmöller diagram of hourly stage IV rainfall (nominally 4-km horizontal grid spacing) for the 10-day period 12–22 June 2007 and the longitudinal band (95°–100°W) studied by Tuttle and Davis (2006). Stage IV is a national rainfall product that combines the regional hourly and 6-hourly multisensor (radar and rain gauge) precipitation analysis produced by the National Weather Service (NWS) River Forecast Centers over the continental United States (Baldwin and Mitchell 1998; Lin and Mitchell 2005; Seo 1998). The break in storm activity after 13 June and the fact that late afternoon rainfall was constrained to Texas before the event allowed for redistribution of soil moisture and temperature, thus enhancing the signatures of soil and land use heterogeneities (as seen in Mesonet observations) on the spatial distribution of surface heat fluxes. Afternoon rainfall in southern Oklahoma on 18 June established a strong gradient in initial soil moisture conditions within the state.

Fig. 2.
Fig. 2.

Time–lat (Hovmöller) plot of hourly stage IV precipitation (http://data.eol.ucar.edu) averaged over 100–95°W from 0000 UTC 13 Jun to 0000 UTC 22 Jun. Stage IV data have a horizontal grid spacing of 4 km.

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

The manuscript is organized as follows. The numerical simulations are described in section 2, and results are presented and discussed in sections 3, 4, and 5. Section 3 is devoted to the analysis of boundary layer development, section 4 is dedicated to the evolution of the prestorm environment and interactions with convection, and storm propagation conditions are analyzed in section 5. A summary of the research and conclusions are presented in section 6.

2. Numerical simulations: Experimental setup

The model used in this study is the Advanced Research Weather Research and Forecasting (ARW; Skamarock et al. 2008) Model, version 3.3, coupled to the Noah land surface model (Ek et al. 2003). Numerical simulations were conducted using three nested domains (see Fig. 3) with horizontal grid spacing of 12 km for the outermost domain (Domain 1; D01), 4 km for the intermediate domain (Domain 2; D02), and 1.333 km for the inner domain (Domain 3; D03). The dynamical core of the ARW Model solves the nonhydrostatic and compressible Euler equations and uses a horizontal Arakawa C grid. One set of simulations used the Yonsei University (YSU) planetary boundary layer (PBL) parameterization (PBLp; Hong et al. 2006). The other used the Mellor–Yamada–Janjić (MYJ) PBLp (Janjić 2002). For all numerical experiments, the model configuration included the Weather Research and Forecasting (WRF) single-moment (WSM) 6-class microphysics parameterization (Hong and Lim 2006) for all domains and the new Kain–Fritsch cumulus parameterization (Kain 2004) for the outermost (and coarsest) domain only.

Fig. 3.
Fig. 3.

Region of study and model domains including topography (m). D01 (outermost grid) has 12-km horizontal grid spacing (122 × 122 grid cells), D02 (middle nest) has 4-km horizontal grid spacing (163 × 163 grid cells), and D03 (innermost nest) has 1.333-km horizontal grid spacing (124 × 124 grid cells).

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

The simulations were two-way nested, in which the higher-resolution domains receive boundary conditions from the outer nests and subsequently feed back the higher-resolution calculations to the corresponding area in the coarser parent domains. A sensitivity analysis using one-way nesting was also conducted. The model used 60 stretched vertical levels, with seven layers below 1 km and the lowest level located at approximately 28 m above ground level. The model top was fixed at 100 hPa. North American Regional Reanalysis (NARR; Mesinger et al. 2006) 3-hourly data at 32-km horizontal grid spacing were used to initialize the land and atmospheric models and to update the boundary conditions at 3-h intervals. The simulations were performed for 19–20 June 2007 during CLASIC (Miller et al. 2007), when fair-weather cumulus clouds were observed during the daytime followed by a severe weather event later in the evening and overnight. On 19 June, convection with a multicellular mode initiated during the afternoon along a stationary boundary in central Kansas and also in the Texas Panhandle. During the early evening, these clusters of storms merged, grew upscale, and eventually produced a leading convective line–trailing stratiform-type (Parker and Johnson 2000) mesoscale convective system that propagated south-southeast across the region, eventually dissipating in central Texas. Across the SGP, reports of severe weather for this event (using NWS criteria from 2007) included 82 severe wind reports, 5 tornado reports, and 103 severe hail reports. Most notably, softball-sized hail (10.8 cm) fell near Goltry, Oklahoma, and several gusts of over 70 mph (31.29 m s−1) were recorded on 19–20 June by Oklahoma Mesonet stations (McPherson et al. 2007). The model was run for 24 h from 0900 UTC 19 June to 0900 UTC 20 June 2007. Similar to previous work (e.g., Sun and Barros 2012, 2014), short-duration simulations (not shown) to investigate model spinup behavior show that for simulations initialized with NARR and starting in the early morning [0400 central daylight time (CDT; CDT = UTC − 5 h)] before sunrise, the time required is less than 3 h. The simulations terminate after the mature MCS propagates through and leaves the innermost domain (D03).

The dominant land cover in the innermost nest is grassland, with dryland/cropland/pasture and savannah comprising the next largest contributions (Fig. 4a). The dominant soil texture is silt loam, although there is also the presence of other types of loam and sandy soils (Fig. 4b). The land cover and soil texture were obtained from the U.S. Geological Survey (USGS) dataset, part of the WRF preprocessing system. Additional sensitivity experiments were conducted with the model using homogeneous land cover for the same domains. One homogeneous simulation for each PBLp used a land surface consisting of silt loam soil, grassland land cover, leaf area index (LAI) of 2.0, and fractional vegetation cover (FVC) of 0.7 (MYJ-D3G and YSU-D3G) in the innermost nest. In another experiment, the realistic land cover in all three domains was replaced with grassland (MYJ-ADG). Soil moisture conditions were initialized to be uniform as well, with a volumetric soil moisture value of 0.3 due to the wet conditions preceding the nocturnal MCS. These values for initial conditions were based on the land surface conditions present in the NARR data. Other simulations (MYJ-D3S and YSU-D3S) used land cover with dry, bare, sandy soil (LAI of 0.0 and volumetric soil moisture of 0.1). The purpose of the homogeneous land cover simulations is to separate large-scale effects from those driven by local land surface heterogeneities. A summary list of simulations and the naming conventions used throughout this manuscript is provided in Table 1.

Fig. 4.
Fig. 4.

Land surface attributes in D03: (a) land use type and (b) soil texture for D03 from the USGS land cover and soil texture dataset available in the WRF preprocessing system. Marked lat cross sections A–A″, B–B″, C–C″, and D–D″ in (a) and northwest–southeast in (b) are referred to elsewhere in the manuscript, including Figs. 5, 6, 12, and 16.

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

Table 1.

WRF simulations performed for this experiment. The first column indicates the naming convention with the form PBLp–land cover, while the second and third columns indicate the PBL parameterization used and the land cover modifications made.

Table 1.

3. Results: Daytime boundary layer development

To monitor the temporal development of boundary layer eddies, time–longitude Hovmöller diagrams of the diurnal cycle of surface fluxes in the innermost nest were constructed for four latitudinal cross sections in the north–south direction (A–A″, B–B″, C–C″, and D–D″, marked in Fig. 4a). The sensitivity of the partitioning of surface fluxes to land cover and soil texture can be assessed by comparing the Hovmöller diagrams of sensible and latent heat fluxes in Fig. 5 and Fig. 6, respectively, for the same cross sections in the YSU-HET and YSU-ADG simulations. The broken patterns and changes in magnitude with longitude between 1200 and 2300 UTC that can be clearly seen for YSU-HET along cross sections A–A″ and D–D″ in Figs. 5a and 5c and Figs. 6a and 6c, respectively, correspond to spatial differences in land cover and soil texture as shown in Fig. 4a. Note the uniform variation of YSU-ADG surface fluxes with longitude in Figs. 5b and 5d and Figs. 6b and 6d. The uniformity of land cover across all domains imposed in YSU-ADG reduces the signature of soil texture differences, especially along D–D″ (cf. Figs. 6c and 6d for latent heat fluxes).

Fig. 5.
Fig. 5.

Hovmöller diagrams of sensible heat flux for the time period from 0900 UTC 19 Jun to 0900 UTC 20 Jun 2007: (a) YSU-HET along A–A″, (b) YSU-ADG along A–A″, (c) YSU-HET along D–D″, and (d) YSU-ADG along D–D″. The negative sign indicates that the flux is leaving the land surface. Cross sections A–A″ and D–D″ are marked in Fig. 4a.

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

Fig. 6.
Fig. 6.

As in Fig. 5, but for latent heat fluxes.

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

The heat fluxes simulated by the model are in good agreement with eddy-correlation data from a tower collocated at Fort Cobb, OK (35.10°N, 98.44°W), during CLASIC and from concurrent tethersonde measurements (Barros et al. 2013). The daytime amplitude of sensible heat fluxes does not exceed 250 W m−2, and tethersonde-based estimates of latent heat fluxes using the Bowen ratio method and net radiation from the tower were as high as 388 W m−2 at 1500 CDT, consistent with values simulated by the model at the same location. Note that the tethersonde was launched only four times during the afternoon of 19 June because of safety concerns on account of very strong low-level winds, which also impacted the latent heat flux estimates from the tower (Barros et al. 2013).

The differences between land cover in the western and eastern portions of D03 and between A–A″ and D–D″ explain the daytime variability in the magnitude of latent and sensible heat fluxes with longitude in the Hovmöller diagrams. As expected, when sensible heat increases, latent heat decreases and vice versa. The magnitude of sensible and latent heat fluxes is not generally very different from each other for YSU-HET, except in the early afternoon and at specific locations, as illustrated by Figs. 5a and 6a. However, there is a very large difference (~200 W m−2) between latent and sensible heat fluxes for YSU-ADG at peak times during the early and midafternoon, thus yielding Bowen ratios <1, consistent with higher evapotranspiration, and therefore a higher influx of moisture to the PBL. This difference is split nearly equally between the sensible and latent heat fluxes for YSU-HET: YSU-HET sensible heat fluxes are higher than YSU-ADG by about 100 W m−2, and YSU-HET latent heat fluxes are lower than YSU-ADG by the same quantity. As the YSU-HET simulation progresses, the sensible heat flux gradually increases to 250 W m−2 around 1515 UTC (1015 CDT). At this point, turbulent mixing begins to develop in the lower troposphere, producing boundary layer eddies. A spatial filter was applied to separate the wavelengths of vertical motion smaller than 12 km following the method proposed by Denis et al. (2002). Figure 7a shows that these eddies are present throughout the domain by 1730 UTC, and they maintain a northwest–southeast orientation, perpendicular to the organization of vertical motion at midlevels (contour lines) and continue to increase in the early afternoon at 1840 UTC (Fig. 7c). The efficiency of this mechanism in exciting the lower troposphere is greater in the northern than in the southern sectors of the innermost domain, as revealed by the spatial density of the 5 hPa m s−1 contour line of midtropospheric vertical motion, reflecting the latitudinal gradient in mean sensible heat flux from south to north (see results for A–A″ and D–D″ in Fig. 5). Boundary layer rolls form as well in the homogeneous YSU-ADG simulation (Figs. 7b,d) because of large-scale forcing interactions with Rayleigh–Bénard convection and differential sensible heating patterns associated with the east–west diurnal march of solar forcing. The vertical motion of the rolls in the homogeneous simulation, however, is significantly weaker in the northern half of the domain (Fig. 7b), where sensible heating is higher in the heterogeneous simulations and low-level shear is stronger. Figures 7e and 7f show the vertical motion fields at 1840 UTC for the MYJ-HET and MYJ-ADG, respectively. In both MYJ and YSU simulations, the initial rolls form over the southern half of D03, consistent with higher soil moisture initial conditions due to the late afternoon rainfall on 18 June (Fig. 2). However, note the dramatic difference in the spatial structure [stronger intensity, difference in wavelength (much shorter for MYJ, ~5 km as per grid spacing constraints), and space filling characteristics] of eddy vertical motion between MYJ and YSU at 1840 UTC. The effect of differences in low-level shear on storm propagation is discussed in section 4 below.

Fig. 7.
Fig. 7.

Pressure-weighted integrated vertical motion (shaded) fields for wavelengths less than 12 km in the lower (from the surface to 700 hPa) and midtroposphere (from 400 to 700 hPa, where the ±5 hPa m s−1 contour is shown and dashed contours indicate downward motion) at 1730 UTC 19 Jun 2007 for (a) the YSU-HET and (b) YSU-ADG and at 1840 UTC 19 Jun 2007 for (c) YSU-HET, (d) YSU-ADG, (e) MYJ-HET, and (f) MYJ-ADG. Shading shows the development of boundary layer eddies, while the contour lines (5 hPa m s−1) show that organized vertical motions at midlevels align perpendicular to the structures at low levels.

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

4. Results: Evolution of the prestorm environment and interactions with convection

Overall, the YSU-HET simulation captures well the vertical profile of the temperature of the atmosphere, especially in the boundary layer (not shown), when compared against radiosonde profiles from Chickasha, OK (35.05°N, 97.94°W), during CLASIC at the same stage of storm evolution. The timing of storm propagation is delayed in the model (i.e., simulated storm moves at slower velocity than observed). Therefore, the stages of storm evolution were determined based on simulated radar reflectivity patterns consistent with the leading line–trailing stratiform MCS that was observed by Next Generation Weather Radar (NEXRAD). Atmospheric profiles averaged over D03 are presented in Fig. 8 for the heterogeneous simulations MYJ-HET and YSU-HET. These diagrams show the same synoptic-scale setup: a capping inversion around 825 hPa and a persistent dry layer around 700 hPa. Low levels, however, are very moist, and if the weak capping inversion can be overcome, these profiles are conducive to deep convection. As the sun rises, turbulent mixing in the boundary layer develops and begins to erode the inversion. The incorporation of inversion-level air also results in dry and warm air entrainment into the boundary layer. By midafternoon (2100 UTC, 1600 CDT), the inversion has been mostly mixed out, and air from the warm, dry layer above can be entrained into the boundary layer. Throughout the day, parcels initiating from the lowest 2 km have very high values of convective available potential energy (CAPE), exceeding 6000 J kg−1 in some areas, consistent with CAPE calculated from CLASIC radiosonde data at Chickasha, OK. The dry layer, however, greatly reduces the buoyancy of parcels originating around 2 km, which can be seen in the reduced CAPE values and large convective inhibition (CIN) values for parcels lifted from this level. Because the lifting mechanism in this case is expected to come from the surface, it is more important to consider the air parcels from lower levels. Average profiles over the innermost domain (D03) for the homogeneous simulations show very little difference when compared with those from the heterogeneous simulations on larger scales, such as in the domain mean.

Fig. 8.
Fig. 8.

D03-average atmospheric thermodynamic profile (temperature, red; dewpoint, blue) for (a) 1200, (b) 1500, (c) 1800, and (d) 2100 UTC 19 Jun 2007 and (e) 0000 and (f) 0300 UTC 20 Jun 2007 (MYJ-HET, solid lines; YSU-HET, dashed lines).

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

Interestingly, a comparative analysis of the simulations with realistic land cover (YSU-HET and MYJ-HET) and the simulations with homogeneous land surface properties in D03 suggests that the choice of PBLp is equally as influential in the evolution of the simulated prestorm environment as land surface heterogeneity. The more vigorous mixing inherent in nonlocal closure parameterizations like YSU (Fig. 8, dashed line; Bright and Mullen 2002, Stensrud 2007) produces a deeper and drier PBL than the MYJ PBLp (a local closure method within the convective PBL), which retains a moister, shallower, and somewhat more unstable profile (~700 J kg−1 greater CAPE by the end of the afternoon for parcels originating near the surface). The homogeneous grassland (D3G) simulations also have lower levels of free convection and higher values of mostly unstable CAPE at the time of storm entry, when compared to the dry soil simulations (D3S) and heterogeneous simulations (HET), as expected. Because of the differences in the representation of mixing processes between the PBL parameterizations used in this study, the YSU-HET and YSU-D3G simulations evolve much more similarly than their MYJ counterparts as the differences in properties due to the land surface are mixed throughout a deeper PBL in the YSU (not shown). The magnitude of the difference in CAPE between MYJ-HET and MYJ-ADG are similar to the differences between YSU-HET and MYJ-HET. Likewise, the difference in the lower-level potential temperature, water vapor mixing ratio, and CAPE between the MYJ-D3G and MYJ-D3S runs is of similar magnitude as the difference between the MYJ-HET and YSU-HET simulations. That is, a change in PBLp (YSU-HET versus MYJ-HET) has as much of an effect on the low-level thermodynamic environment as dramatic changes in the land surface have on the MYJ simulations.

To diagnose the contributions from surface energy fluxes and entrainment from aloft, we use the local land–atmosphere coupling (LoCo) framework presented in Santanello et al. (2009, 2011). This approach uses the diurnal coevolution of 2-m specific humidity and temperature plotted in energy space (Betts 1992) to quantify the effects of heat and moisture fluxes from the land surface and from the top of the PBL (Fig. 9). The slopes of the vectors in Fig. 9 are equal to the Bowen ratio, and their components are proportional to sensible and latent heat fluxes as
e1
for sensible heat flux and
e2
for latent heat flux. The overbars denote time-averaged fluxes, Δt is the time over which they are averaged, cp is the specific heat of dry air at constant pressure, Lυ is the latent heat of vaporization, PBLH is the average PBL height, and ρm is the average air density. By comparing magnitudes of these vector components, one can also derive the daytime entrainment ratios for heat and moisture at the top of the boundary layer, a metric that is not easily quantifiable using traditional measurements alone. For a more thorough description of this framework, the reader is referred to Santanello et al. (2009).
Fig. 9.
Fig. 9.

LoCo mixing diagram for D03 pixels containing the following combinations of soil texture and land use: sand + grassland (black), silt loam + dryland/cropland/pasture (medium gray), and silt loam + grassland (light gray) for the YSU-HET simulation.

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

Among different types of land use and land cover (LULC), there do not appear to be differences in terms of range in the LoCo parameters when comparing these parameters to diagnostics useful for anticipating where convection is likely or unlikely. Figure 10a shows two orthogonal regimes in the relationship between CAPE and entrainment Bowen ratio above and below 5500 J kg−1 that are generally independent of LULC. Sandy soils, though, do tend to exhibit narrower ranges of variability: small changes in the entrainment Bowen ratio for all CAPE values below the high CAPE threshold and a wide variation above the threshold. Based on CLASIC tethersonde and radiosonde data for 19 June, values of CAPE above 5500 J kg−1 occurred in the SGP in the early afternoon, between 1200 and 1600 LST, when the PBL is well developed (Barros et al. 2013).

Fig. 10.
Fig. 10.

(a) Scatterplot of 0000 UTC 20 Jun 2007 CAPE values and entrainment Bowen ratio for D03 pixels containing the following soil texture and land use combinations: sand + grassland pixels (black), silt loam + dryland/cropland/pasture pixels (medium gray), and silt loam + grassland pixels (light gray) for the MYJ-HET simulation. (b) Scatterplot of 0000 UTC 20 Jun 2007 CAPE values and surface Bowen ratio for sand + grassland pixels (black), silt loam + dryland/cropland/pasture pixels (medium gray), and silt loam + grassland pixels (light gray) for the YSU-HET simulation.

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

However, the evolution of the PBL in energy space varies spatially as a function of soil texture. In particular, we highlight a differentiation between sand and silt loam soil textures when considering the surface energy budget expressed by the surface Bowen ratio in Fig. 10b. Figure 10b shows that soil texture and LULC determine the range of variability of the Bowen ratio as pointed out by Barros and Hwu (2002), with a broader range of variability corresponding to conditions that support a higher amplitude in the diurnal cycle of surface latent heat fluxes, that is, the silt loam and grassland combination. Figure 11 shows that sandy soils start near 70% saturation and dry out more quickly throughout the day when compared to the silt loam soils, as expected. This results in a higher mean latent heat flux for sandy pixels and also a difference in the relative contribution of the surface and entrainment fluxes (Fig. 9). Although the values of surface-based 0000 UTC instability for the two soil types span roughly the same range of values (Fig. 10), regardless of land cover, the sandy pixels rely more heavily on surface evaporation (with a significantly lower surface Bowen ratio) and subsequent dry and warm air entrainment to achieve the same values. However, there is no significant dependence on the range of CAPE variability itself and LULC.

Fig. 11.
Fig. 11.

YSU-HET D03-average sensible (solid) and latent (dashed) heat fluxes and soil saturation (triangles) for sandy (red) and silt loam pixels (blue) from 1200 UTC 19 Jun to 0000 UTC 20 Jun 2007.

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

5. Results: Storm propagation

In the late afternoon, convection initiates over Kansas (D02) along a stationary boundary and begins to propagate southeastward. The timing difference in the initiation and early development stages of the MCS is about 40 min in D02 (MYJ PBLp faster) for simulations with a heterogeneous D02. Cross sections of the storm evolution in Fig. 12 show lifting along the cold pool front (along the diagonal cross section marked in Fig. 4b) as the storm moves forward and ingests the layer that was modified by the surface earlier in the day. Each simulation produced a bow-shaped MCS that propagated southeastward across the domain, and these vertical cross sections show that the model reasonably reproduces the vertical structure of the leading line–trailing stratiform MCS that was observed. However, the results of the various simulations, when compared, show large differences in timing and amount of rainfall with respect to stage IV observations and among model runs (see Fig. 13 for D03 integrated values as a function of time). Most interestingly, for the MYJ PBLp simulations, those with the artificially imposed grassland land cover have a timing of the MCS propagation through the innermost domain most similar to the observations. The MCS enters the innermost domain in the MYJ-ADG simulation at 0300 UTC, the MYJ-D3G simulation enters at 0335 UTC, the MYJ-HET simulation enters at 0425 UTC, and the MYJ-D3S simulation enters at 0535 UTC. For the cases using the YSU PBLp, the timing of the MCS remains nearly the same for all of the simulations independently of land cover, with the storm entering D03 around 0530 UTC. The timing of the simulated MCSs in the YSU runs most closely resembles the timing of the storm in the MYJ-D3S simulation.

Fig. 12.
Fig. 12.

Along-storm cross-sectional view in D03 (grid resolution Δs = 1.333 km; cross section marked in Fig. 4b) of (a) vertical velocity, (b) equivalent potential temperature, (c) simulated radar reflectivity, and (d) CAPE for MYJ-HET at 0620 UTC 20 Jun 2007. The quantities on the x axis are grid points numbered from west to east.

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

Fig. 13.
Fig. 13.

(a) D03-average rainfall for MYJ simulations MYJ-ADG (black), MYJ-D3G (dashed medium gray), MYJ-D3S (short dashed), and MYJ-HET (dashed light gray). Observations from stage IV precipitation data are shown in black asterisks. (b) D03-average rainfall for YSU-ADG (black), YSU-D3G (dashed medium gray), YSU-D3S (short dashed), and YSU-HET (dashed light gray). The timestamp corresponds to 20 Jun 2007.

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

The thermodynamic structure of the atmosphere within the storm environment, specifically the presence of dry and moist layers, has been investigated previously to determine its relationship to storm propagation, downdraft strength, and near-surface wind speeds. In midlatitudes, air with low relative humidity in the lower troposphere has been thought to increase the evaporative cooling and subsequently induce strong downdrafts (e.g., Gilmore and Wicker 1998). However, idealized numerical simulations of quasi-linear convective systems (James et al. 2006) have shown that dry air aloft reduced total rainfall and condensate and reduced the updraft and downdraft mass fluxes, except in high CAPE environments. The strength of the cold pool along its leading edge was also found to remain unchanged or weakened when dry air was present aloft. The downdraft mass flux was only strengthened in the stratiform region of the storm. James et al. (2006) determined that bowing segments were found for moisture ranges that promote intermediate cold pool strengths, enhancing convergence locally when the cold pool is strong enough to overwhelm the low-level wind shear. Stronger cold pools tend to favor large, slablike mesoscale convective systems, while weaker cold pools typically favor more a more cellular mode.

The strong sensitivity to land surface conditions in the case of the MYJ simulations, and the lack thereof in the case of YSU, begs further questions into how land cover may influence the speed and strength of the cold pool, which in turn will affect the timing of the storm. A measure of cold pool strength c2, an integrated measure of negative buoyancy, was used here to compare cold pool speeds from simulation to simulation (Weisman 1992; James et al. 2006):
e3
where D is the depth of the cold pool, defined as the −1 K potential temperature perturbation, and g is the gravitational acceleration. The overbar in Eq. (3) denotes the base state of the model, and θρ is the density potential temperature calculated by
e4
where qυ is the water vapor mixing ratio, qt is the total mixing ratio, and ε is the ratio of the dry to moist gas constants. Since the storm propagates from northwest to southeast across the third domain (and not along any specific parallel or meridian), Hovmöller diagrams are constructed that are approximately orthogonal to the storm motion. Specifically, averages are computed along cross sections drawn at a 45° angle (northwest–southeast) across the third domain, as illustrated for a synthetic example in Fig. 14. Values that appear to the left of center of the x axis indicate points west with a southward component, while values to the right of center represent value points east of a due southeast trajectory.
Fig. 14.
Fig. 14.

(d) Synthetic example of development of cold pool velocity Hovmöller diagram from (a)–(c) three velocity fields. Each line in the diagram was created by dividing the domain into bins one grid cell wide, aligned in the northwest–southeast direction, from the lower left corner (in the southwest) to the upper right corner (in the northeast). Along each northwest–southeast line (perpendicular to the storm motion), averages were computed for cells that contained a nonmissing value of cold pool speed. Objects moving east of a 45° angle appear to move toward increasing numbers on the x axis [in (d)] and objects moving south of a 45° angle appear to move to the left. Objects moving on a 45° angle, as in this example, appear to move straight in (d).

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

Plots of theoretical cold pool speed are shown in Fig. 15 for the MYJ-HET, the MYJ-D3G, the MYJ-D3S, and the YSU-HET simulations. The direction of propagation is most similar in the MYJ-D3G, the MYJ-D3S, and the YSU-HET. In these simulations, the MCS propagates in a more easterly direction, whereas in the MYJ-HET simulation, the segment of the storm takes a more southern route across the innermost domain. Note the space–time patterns of the Hovmöller diagrams of cold pool speeds for MYJ-D3S and YSU-HET, which are representative of all other YSU simulations. There are very small differences in terms of timing (5 min) and negligible differences in terms of storm rainfall (Figs. 13 and 15), and the storm trajectory and theoretical cold pool speed are very similar among the YSU runs and MYJ-D3S. An examination of the storm structure shown along the same cross section in Fig. 12 suggests that the PBL structures as described by CAPE and vertical wind shear are very similar, which is consistent with the significantly deeper PBL height (and stronger shear) in MYJ-D3S when compared to other MYJ simulations, as illustrated in Fig. 16. Thus, the storm tends to propagate along the more easterly track into the unstable environment with more favorable vertical wind shear in these circumstances, with the storm arriving earlier at Domain 3 for MYJ-D3G with weaker shear compared to MYJ-D3S (Fig. 16d).

Fig. 15.
Fig. 15.

Theoretical cold pool speed as in the Hovmöller diagram presented in Fig. 14 for (a) MYJ-D3G, (b) MYJ-D3S, (c) MYJ-HET, and (d) YSU-HET. The timestamp corresponds to 20 Jun 2007.

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

Fig. 16.
Fig. 16.

Hovmöller plot of PBL height along D–D″ (marked in Fig. 4a) for (a) MYJ-D3G, (b) MYJ-D3S, and (c) YSU-HET. (d) The 0–3-km vertical wind shear vector at 0100 UTC for MYJ-D3G (black) and MYJ-D3S (purple). The difference in the magnitude of the shear vector MYJ (D3G − D3S) is shaded, with warm colors representing stronger shear in the case of the D3S simulation.

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

Finally, another simulation was performed in order to investigate the changes in the MYJ simulations that show stronger sensitivity to land cover change by switching off two-way coupling for the MYJ-D3G and MYJ-D3S cases, with all else remaining the same. These simulations are referred to as one way (Table 1). The results highlight the important role of upwind advection of low-level thermodynamic features in modifying the remote prestorm environment where the MCS originates. Over the course of the afternoon, the low-level flow over Oklahoma shifts to a predominantly southeasterly direction. Downstream, in the area of D02 to the northwest of D03, the effects of a homogeneous innermost nest can be seen in modified temperature and moisture profiles along an axis that coincides with the tracks of the simulated MCSs (Fig. 17). In the case of the MYJ-D3S simulation, the advection of warm and dry air leads to a reduction in CAPE by nearly 1500 J kg−1 in the area of D02 just downstream of D03 for the two-way nested simulation (Fig. 18). At the same time, southeasterly air from D02 that was not subjected to a land cover change is transported into D03, which shifts the area of interest, from a thermodynamic perspective, out of the highest-resolution domain. As the MCS approaches D03, it interacts with the air mass that was modified by the land surface in D03 and transported downstream earlier in the day, that is, a remote land–atmosphere interaction. In the case of the dry, barren land cover simulations, the reduced low-level moisture and instability appear to inhibit convection first, and next slow down the development phase of the MCS downstream of D03 (though convection still develops readily along the edges of the axis of reduced instability). The moister low-level conditions in the grassland simulations allow the storm to develop more quickly, resulting in an earlier time of entry into D03. However, when the third nest (D03) is not allowed to interact with its parent nest (D02), the impact of downstream low-level advection is not transferred to D02, and the remote feedback is lost: the D03-averaged timing and magnitude of rainfall is nearly the same in the MYJ-D3S-1way and the MYJ-D3G-1way simulations (Fig. 19a). The differences in rainfall amount and spatial organization in D02 between the two- and one-way nested MYJ-D3S simulations are shown in Fig. 19b for wavelengths greater than 12 km (at smaller scales the differences are negligible, not shown). Note the agreement between the region where daytime downstream impacts on CAPE are strongest in Fig. 18 (centered approximately at 36.5°N and 99.5°W) and the cumulative precipitation pattern in Fig. 19b (left) for MYJ-D3S. Thus, the largest observable effect of the land surface at this resolution is downwind of the core land–atmosphere interactions region in D03 during daytime. This is an important result that shows remote feedbacks of daytime land–atmosphere interactions on nocturnal rainfall in the SGP at the meso-β scale (20–200 km). The precipitation amount simulated in the one-way simulations is significantly lower than in the two-way simulations, which should be interpreted in the context that precipitation processes are first being resolved in the coarsest domain, with no feedback from higher-resolution nests that may better capture these processes.

Fig. 17.
Fig. 17.

Skew T–logp diagrams for (a) 0000 UTC MYJ-D3S (dashed) and MYJ-D3S-1way (solid), (b) 0200 UTC MYJ-D3S (dashed) and MYJ-D3S-1way (solid), (c) 0000 UTC MYJ-D3S (dashed) and MYJ-D3G (solid), and (d) 0200 UTC MYJ-D3S (dashed) and MYJ-D3G (solid) 20 Jun 2007. Profiles are averaged over the portion of D02 spanning from 36.25°N, 100°W to 36.75°N, 99.5°W. Each panel contains a trace of temperature and dewpoint temperature. Different dash patterns are used to differentiate between simulations.

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

Fig. 18.
Fig. 18.

Temporal evolution of the spatial distribution of CAPE for a parcel with properties averaged over the lowest 500 m at the times indicated for (a)–(c) MYJ-D3S and (d)–(f) MYJ-D3S-1way. Also marked are the 10-m wind vectors.

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

Fig. 19.
Fig. 19.

(a) D03-average rainfall for MYJ simulations MYJ-D3S-1way, MYJ-D3G, MYJ-D3S, and MYJ-D3G-1way as indicated in the legend. Observations from stage IV precipitation data are shown as black asterisks. The timestamp corresponds to 20 Jun 2007. (b) D02 cumulative rainfall for (left) MYJ-D3S and (right) MYJ-D3S-1way for wavelengths >12 km.

Citation: Journal of Hydrometeorology 15, 5; 10.1175/JHM-D-14-0016.1

Nevertheless, the effect of remote land–atmosphere interactions is significantly reduced in the case of YSU simulations (not shown), which only show small sensitivity to total storm rainfall and no sensitivity to timing. That is, the strong nonlocal mixing characteristics of the YSU PBL parameterization effectively diffuse the low-level thermodynamic heterogeneities associated with land surface conditions, and consequently, the remote impact of land–atmosphere interactions on convective propagation is the same for all surface conditions.

6. Summary and conclusions

The overarching goal of this study was to investigate the question of SP feedbacks in the SGP, specifically focusing on the strong nocturnal peak in the regional diurnal cycle of rainfall. A series of 10 numerical experiments was conducted to investigate land–atmosphere interactions on nighttime convection using the WRF Model to simulate an MCS that originated in Kansas and moved southeastward on 19–20 June 2007 during CLASIC. By the time the storm reached central Oklahoma, it exhibited a mature leading line–trailing stratiform structure, and this event included severe weather reports of strong winds and hail. In this study, the specific research objective was to examine daytime land–atmosphere feedbacks and day-to-night transition processes associated with land surface conditions and their impact on propagating convection in its mature stage rather than at the initial stages.

To this end, 10 simulations using various combinations of heterogeneous and homogeneous land cover, soil texture, soil moisture, temperature states (in all domains and then only in the innermost nest) and two different PBL parameterizations (MYJ and YSU) were conducted. One set of experimental configurations used land cover with a moist, homogeneous grassland and the other used dry, barren soil to benchmark the effect of the land surface at two extremes. The purpose of using a homogeneous land surface was to separate the effects of the land surface from the effects of larger-scale forcing. In the prestorm environment, an analysis of the thermodynamic profiles shows an atmosphere primed for convection at the regional scale. In the late afternoon, convection initiated well to the north of the innermost domain and began to propagate over the study region after nightfall, at which point differences in the contributions from land surface heat fluxes were minimal. All simulations exhibited a time delay in precipitation as compared to observations, with rainfall timing errors varying between 1 and 3 h for MYJ simulations and 3 h for all YSU-based simulations. Timing discrepancies were ultimately attributed to the following mechanisms: 1) daytime remote feedback through low-level advection and 2) feedbacks having to do with differences in mixing processes that affect PBL height and shear locally. Differences in how the prestorm environment evolves were evident over different soil textures for the case of the MYJ and YSU simulations with realistic land covers. Both simulations overestimated the total storm rainfall, but the timing error was smaller for MYJ. The MYJ simulations exhibited strong sensitivity to land cover, with significant timing differences between the homogeneous grassland simulations and the homogeneous simulations with barren sandy soil, the former most closely resembling the observations. The YSU-based simulations showed some sensitivity to land surface conditions in terms of rainfall amount, but negligible sensitivity in terms of timing (Fig. 13). Overall, more differences were seen between simulations with the two different boundary layer parameterizations and same surface conditions than between simulations that used a realistic homogenized land surface. This underscores the importance of making appropriate choices of PBLp when studying the coupled land–atmosphere system since drastic, unrealistic changes in land cover impact the simulated storm nearly as much as changes in the representation of PBL mixing processes. Recently, Sun and Barros (2014) also reported on the importance of the representation of mixing processes in both MYJ and YSU, including sensitivity to stability conditions in the context of their simulations of the terrestrial evolution of hurricane Ivan. For the MCS case reported here, the difference between local closure (MYJ) and nonlocal closure (YSU) strategies and how this affects PBL mixing appears to be the dominant control in term of location-specific feedbacks (meso-γ scale, <20 km). Clearly, even though different types of vegetation are associated with different surface roughness (and consequently surface friction velocities) and eddy dispersion, these differences are quickly mixed in the YSU simulations that have deeper PBLs and strong low-level vertical shear.

The thermodynamic prestorm environment where the MCS originates shows changes consistent with advection by the low-level southeasterly afternoon winds of thermodynamic features associated with daytime land–atmosphere interactions for the various MYJ configurations. These changes affect the rates of convective development after sundown (e.g., MYJ-D3G accelerates and MYJ-D3S slows down). Thus, this study shows that, depending on the PBL parameterization, remote (downwind) daytime feedback effects of local processes on nocturnal development and propagation of the MCS can be significant and that they are controlled by meso-α- (>200 km) and synoptic-scale dynamical controls on regional storm activity. On the other hand, differences between simulations using different PBL parameterizations but same land surface conditions are as large as differences among simulations for the same PBL parameterization, namely MYJ, using different land surface conditions. This ambiguity prevents establishing a definite (and general) cause–effect relationship independent of the model physics. Nevertheless, this finding, compounded with the storm timing error for all simulations and the fact that the total precipitation error was the smallest when all three nest domains are covered by grass, suggests that the representation of surface frictional effects in both parameterizations needs to be revisited. Note that replacing large areas with a homogeneous land surface is a dramatic step when exploring the effect of the land surface on convection, and actual impacts of realistic land cover changes are likely to be more difficult to trace. Finally, independent of the remote effects of daytime land–atmosphere interactions and PBL parameterization, analysis of storm structure shows that simulations with the same timing error and similar storm propagation characteristics as described by cold pool dynamics tend to present deeper boundary layers and stronger low-level vertical shear in the prestorm environment ahead of the cold pool.

To elucidate the distinct roles of mesoscale transport and redistribution of low-level instability (daytime remote feedbacks) and low-level shear in the downwind prestorm environment (nighttime local feedbacks), which is to separate the nonlinear land–atmosphere physical processes from PBLp-specific effects on simulated storm dynamics, requires addressing the (phase) delay of several hours in storm development and propagation between the observed and the simulated MCS (Figs. 13 and 19). This can be tentatively attributed in part to model forcing [imperfect initial and boundary conditions as in Sun and Barros (2012)] and the representation of turbulence and mixing processes in the PBLp. Wu et al. (2013) report large differences in rainfall amounts and storm vertical structure for the simulations of two squall lines in a sensitivity study to microphysical parameterizations in WRF. A study looking at the joint sensitivity to PBL and microphysical parameterizations should provide further insight on storm evolution and propagation.

The effects of the land surface on the momentum budget of an MCS may also be relevant, though perhaps on smaller scales than the grid spacing (Δs = 1.333 km) used in this study. It is also important to note that the final result of each simulation represents the total effect of the land surface as well as how different parameterizations process the differences originating as the lower boundary condition for the atmosphere. This means that, although the land surface was initialized to be homogeneous, heterogeneities in the temperature and moisture fields were allowed to develop and evolve because of internal nonlinear dynamics.

Although the MCS examined here is illustrative of SGP hydrometeorology, how to statistically upscale the local and nonlocal (i.e., remote) feedbacks found in this study to the climate scale, and thus elucidate the sign and seasonality of SP feedbacks in the region, requires additional research for different types of storm systems. The contribution of this study is to identify two mechanisms by which the land surface can affect preexisting nocturnal convection in the SGP locally (PBL mixing; Figs. 7 and 10) and remotely (daytime low-level advection; Figs. 13 and 19). In particular, teleconnections between land–atmosphere interactions and upstream convection due to daytime advection of low-level unstable air were identified (Fig. 18), which point to the need for further work on the detection and attribution of remote feedbacks with a focus on spatial scales from the local to the mesoscale.

Overall, this study demonstrates that there is large uncertainty in defining the spatial and temporal scales of land–atmosphere interactions, particularly with regard to organized convection. This result is especially important as the grids in climate models begin to approach the meso-β scale, and therefore, remote SP feedbacks can be resolved explicitly. Specifically, it would be very useful to repeat the study for a large number of MCSs during one or various summer seasons and to examine the model climatology to determine the sensitivity of these conclusions over a wide range of initial and boundary conditions.

Acknowledgments

The authors thank Prabhakar Shrestha for his work on the early stages of this research and on an earlier version of this manuscript and Yajuan Duan and Miguel Nogueira for their help with the figures. The first author has been supported in part by an American Meteorological Society graduate fellowship, a Pratt–Gardner graduate fellowship from the Pratt School of Engineering at Duke University, and a J.B. Duke graduate fellowship. This study was also supported by NSF Award EAR-0711430 and by NASA Grant NNX07AP81G to participate in the CLASIC experiment. NCAR Command Language (NCL) software version 6.1.0-beta was used to produce figures (doi:10.5065/D6WD3XH5).

REFERENCES

  • Avissar, R., , and Schmidt T. , 1998: An evaluation of the scale at which ground-surface heat flux patchiness affects the convective boundary layer using large-eddy simulations. J. Atmos. Sci., 55, 26662689, doi:10.1175/1520-0469(1998)055<2666:AEOTSA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Balaji, V., , and Clark T. L. , 1988: Scale selection in locally forced convective fields and the initiation of deep cumulus. J. Atmos. Sci., 45, 31883211, doi:10.1175/1520-0469(1988)045<3188:SSILFC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Baldwin, M. E., , and Mitchell K. E. , 1998: Progress on the NCEP hourly multi-sensor U.S. precipitation analysis for operations and GCIP research. Preprints, Second Symp. on Integrated Observing Systems, Phoenix, AZ, Amer. Meteor. Soc., 1011.

  • Balling, R. C., 1985: Warm season nocturnal precipitation in the Great Plains of the United States. J. Climate Appl. Meteor., 24, 13831387, doi:10.1175/1520-0450(1985)024<1383:WSNPIT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Barros, A. P., , and Hwu W. , 2002: A study of land–atmosphere interactions during summertime rainfall using a mesoscale model. J. Geophys. Res., 107, 17-1–17-18, doi:10.1029/2000JD000254.

    • Search Google Scholar
    • Export Citation
  • Barros, A. P., , Li W. , , Kang D.-H. , , Prat O. , , Shrestha P. , , Brunsell N. , , and Rhan T. A. , 2013: Observations of surface fluxes and boundary layer instability in the southern Great Plains during CLASIC. Rep. EPL-2013-CLASIC-1, Environmental Physics Laboratory, Duke University, 57 pp.

  • Betts, A. K., 1992: FIFE atmospheric boundary layer budget methods. J. Geophys. Res.,97, 18 52318 531, doi:10.1029/91JD03172.

  • Bright, D. R., , and Mullen S. L. , 2002: Short-range ensemble forecasts of precipitation during the southwest monsoon. Wea. Forecasting, 17, 10801100, doi:10.1175/1520-0434(2002)017<1080:SREFOP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Carbone, R. E., , and Tuttle J. D. , 2008: Rainfall occurrence in the U.S. warm season: The diurnal cycle. J. Climate, 21, 41324146, doi:10.1175/2008JCLI2275.1.

    • Search Google Scholar
    • Export Citation
  • Carbone, R. E., , Tuttle J. D. , , Ahijevych D. A. , , and Trier S. B. , 2002: Inferences of predictability associated with warm season precipitation episodes. J. Atmos. Sci., 59, 20332056, doi:10.1175/1520-0469(2002)059<2033:IOPAWW>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Denis, B., , Côté J. , , and Laprise R. , 2002: Spectral decomposition of two-dimensional atmospheric fields on limited-area domains using the discrete cosine transform (DCT). Mon. Wea. Rev., 130, 18121829, doi:10.1175/1520-0493(2002)130<1812:SDOTDA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Dirmeyer, P. A., and et al. , 2012: Simulating the diurnal cycle of rainfall in global climate models: Resolution versus parameterization. Climate Dyn., 39, 399418, doi:10.1007/s00382-011-1127-9.

    • Search Google Scholar
    • Export Citation
  • Dong, X., and et al. , 2011: Investigation of the 2006 drought and 2007 flood extremes at the southern Great Plains through an integrative analysis of observations. J. Geophys. Res.,116, D03204, doi:10.1029/2010JD014776.

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

    • Search Google Scholar
    • Export Citation
  • Gilmore, M. S., , and Wicker L. J. , 1998: The influence of midtropospheric dryness on supercell morphology and evolution. Mon. Wea. Rev., 126, 943958, doi:10.1175/1520-0493(1998)126<0943:TIOMDO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Hong, S.-Y., , and Lim J.-O. J. , 2006: The WRF Single-Moment 6-Class Microphysics Scheme (WSM6). J. Korean Meteor. Soc., 42, 121151.

  • Hong, S.-Y., , Noh Y. , , and Dudhia J. , 2006: A new vertical diffusion package with an explicit treatment of entrainment processes. Mon. Wea. Rev., 134, 23182341, doi:10.1175/MWR3199.1.

    • Search Google Scholar
    • Export Citation
  • James, R. P., , Markowski P. M. , , and Fritsch J. M. , 2006: Bow echo sensitivity to ambient moisture and cold pool strength. Mon. Wea. Rev., 134, 950964, doi:10.1175/MWR3109.1.

    • Search Google Scholar
    • Export Citation
  • Janjić, Z. I., 2002: Nonsingular implementation of the Mellor–Yamada level 2.5 scheme in the NCEP Meso model. NCEP Office Note 437, 61 pp. [Available online at www.emc.ncep.noaa.gov/officenotes/newernotes/on437.pdf.]

  • Kain, J. S., 2004: The Kain–Fritsch convective parameterization: An update. J. Appl. Meteor., 43, 170181, doi:10.1175/1520-0450(2004)043<0170:TKCPAU>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Koster, R. D., and et al. , 2004: Regions of strong coupling between soil moisture and precipitation. Science, 305, 11381140, doi:10.1126/science.1100217.

    • Search Google Scholar
    • Export Citation
  • Lamb, P., , Portis D. H. , , and Zangvil A. , 2012: Investigation of large-scale atmospheric moisture budget and land surface interactions over U.S. southern Great Plains including for CLASIC (June 2007). J. Hydrometeor., 13, 1719–1738, doi:10.1175/JHM-D-12-01.1.

    • Search Google Scholar
    • Export Citation
  • Lanicci, J. M., , Carlson T. N. , , and Warner T. T. , 1987: Sensitivity of the Great Plains severe-storm environment to soil-moisture distribution. Mon. Wea. Rev., 115, 26602673, doi:10.1175/1520-0493(1987)115<2660:SOTGPS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • LeMone, M. A., , and Pennell W. T. , 1976: The relationship of trade wind cumulus distribution to subcloud layer fluxes and structure. Mon. Wea. Rev., 104, 524539, doi:10.1175/1520-0493(1976)104<0524:TROTWC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Lin, Y., , and Mitchell K. E. , 2005: The NCEP Stage II/IV hourly precipitation analyses: Development and applications. Preprints, 19th Conf. on Hydrology, San Diego, CA, Amer. Meteor. Soc., 1.2. [Available online at https://ams.confex.com/ams/pdfpapers/83847.pdf.]

  • Luo, Y., , Berbery E. H. , , Mitchell K. E. , , and Betts A. K. , 2007: Relationships between land-surface and near-surface atmospheric variables in the NCEP North American Regional Reanalysis. J. Hydrometeor., 8, 11841203, doi:10.1175/2007JHM844.1.

    • Search Google Scholar
    • Export Citation
  • Lynn, B. H., , Tao W.-K. , , and Wetzel P. J. , 1998: A study of landscape-generated deep moist convection. Mon. Wea. Rev., 126, 928942, doi:10.1175/1520-0493(1998)126<0928:ASOLGD>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Mahrt, L., , Sun J. , , Vickers D. , , Macpherson J. I. , , Pederson J. R. , , and Desjardins R. L. , 1994: Observations of fluxes and inland breezes over a heterogeneous surface. J. Atmos. Sci., 51, 24842499, doi:10.1175/1520-0469(1994)051<2484:OOFAIB>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • McPherson, R. A., and et al. , 2007: Statewide monitoring of the mesoscale environment: A technical update on the Oklahoma Mesonet. J. Atmos. Oceanic Technol., 24, 301321, doi:10.1175/JTECH1976.1.

    • Search Google Scholar
    • Export Citation
  • Mesinger, F., and et al. , 2006: North American Regional Reanalysis. Bull. Amer. Meteor. Soc., 87, 343360, doi:10.1175/BAMS-87-3-343.

    • Search Google Scholar
    • Export Citation
  • Miller, M. A., and et al. , 2007: SGP Cloud and Land Surface Interaction Campaign (CLASIC): Science and Implementation Plan. Rep. DOE/SC-ARM-0703,Office of Biological and Environmental Research, U.S. Department of Energy, 14 pp. [Available online at www.arm.gov/publications/programdocs/doe-sc-arm-0703.pdf.]

  • Parker, M. D., , and Johnson R. H. , 2000: Organizational modes of midlatitude mesoscale convective systems. Mon. Wea. Rev., 128, 34133436, doi:10.1175/1520-0493(2001)129<3413:OMOMMC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Santanello, J. A., Jr., , Friedl M. A. , , and Ek M. B. , 2007: Convective planetary boundary layer interactions with the land surface at diurnal time scales: Diagnostics and feedbacks. J. Hydrometeor., 8, 10821097, doi:10.1175/JHM614.1.

    • Search Google Scholar
    • Export Citation
  • Santanello, J. A., Jr., , Peters-Lidard C. D. , , Kumar S. V. , , Alonge C. , , and Tao W.-K. , 2009: A modeling and observational framework for diagnosing local land–atmosphere coupling on diurnal time scales. J. Hydrometeor., 10, 577599, doi:10.1175/2009JHM1066.1.

    • Search Google Scholar
    • Export Citation
  • Santanello, J. A., Jr., , Peters-Lidard C. D. , , and Kumar S. V. , 2011: Diagnosing the sensitivity of local land–atmosphere coupling via the soil moisture–boundary layer interaction. J. Hydrometeor., 12, 766786, doi:10.1175/JHM-D-10-05014.1.

    • Search Google Scholar
    • Export Citation
  • Schär, C., , Lüthi D. , , Beyerle U. , , and Heise E. , 1999: The soil–precipitation feedback: A process study with a regional climate model. J. Climate, 12, 722741, doi:10.1175/1520-0442(1999)012<0722:TSPFAP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Seo, D. J., 1998: Real-time estimation of rainfall fields using radar rainfall and raingauge data. J. Hydrol., 209, 131147.

  • Shaw, B. L., , Pielke R. A. , , and Ziegler C. L. , 1997: A three-dimensional numerical simulation of a Great Plains dryline. Mon. Wea. Rev., 125, 14891506, doi:10.1175/1520-0493(1997)125<1489:ATDNSO>2.0.CO;2.

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

  • Stensrud, D. J., 2007: Parameterization Schemes: Keys to Understanding Numerical Weather Prediction Models. Cambridge University Press, 459 pp.

  • Sun, X., , and Barros A. P. , 2012: The impact of forcing dataset on the high resolution simulation of tropical storm Ivan (2004) in the southern Appalachians. Mon. Wea. Rev., 140, 3300–3326, doi:10.1175/MWR-D-11-00345.1.

    • Search Google Scholar
    • Export Citation
  • Sun, X., , and Barros A. P. , 2014: High resolution simulation of tropical storm Ivan (2004) in the Southern Appalachians: Role of planetary boundary-layer schemes and cumulus parameterization. Quart. J. Roy. Meteor. Soc., doi: 10.1002/qj.2255, in press.

    • Search Google Scholar
    • Export Citation
  • Tao, K., , and Barros A. P. , 2008: Metrics to describe the dynamical evolution of atmospheric moisture: Intercomparison of model (NARR) and observations (ISCCP). J. Geophys. Res., 113, D14125, doi:10.1029/2007JD009337.

    • Search Google Scholar
    • Export Citation
  • Trier, S. B., , Chen F. , , and Manning K. W. , 2004: A study of convection initiation in a mesoscale model using high-resolution land-surface initial conditions. Mon. Wea. Rev., 132, 29542976, doi:10.1175/MWR2839.1.

    • Search Google Scholar
    • Export Citation
  • Tuttle, J. D., , and Davis C. A. , 2006: Corridors of warm season precipitation in the central United States. Mon. Wea. Rev., 134, 22972317, doi:10.1175/MWR3188.1.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., 1975: Diurnal variations in precipitation and thunderstorm frequency over the conterminous United States. Mon. Wea. Rev., 103, 406419, doi:10.1175/1520-0493(1975)103<0406:DVIPAT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Weckwerth, T. M., , Wilson J. W. , , and Wakimoto R. M. , 1996: Thermodynamic variability within the convective boundary layer due to horizontal convective rolls. Mon. Wea. Rev., 124, 769784, doi:10.1175/1520-0493(1996)124<0769:TVWTCB>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Weckwerth, T. M., , Wilson J. W. , , Wakimoto R. M. , , and Crook N. A. , 1997: Horizontal convective rolls: Determining the environmental conditions supporting their existence and characteristics. Mon. Wea. Rev., 125, 505526, doi:10.1175/1520-0493(1997)125<0505:HCRDTE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wei, J., , Dickinson R. E. , , and Chen H. , 2008: A negative soil moisture–precipitation relationship and its causes. J. Hydrometeor., 9, 13641376, doi:10.1175/2008JHM955.1.

    • Search Google Scholar
    • Export Citation
  • Weisman, M. L., 1992: The role of convectively generated rear-inflow jets in the evolution of long-lived mesoconvective systems. J. Atmos. Sci., 49, 18261847, doi:10.1175/1520-0469(1992)049<1826:TROCGR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wu, D., , Dong X. , , Xi B. , , Feng Z. , , Kennedy A. , , Mullendore G. , , Gilmore M. , , and Tao W.-K. , 2013: Impacts of microphysical scheme on convective and stratiform characteristics in two high precipitation squall line events. J. Geophys. Res. Atmos., 118, 11 11911 135, doi:10.1002/jgrd.50798.

    • Search Google Scholar
    • Export Citation
  • Zeng, X., , Barlage M. , , Castro C. , , and Fling K. , 2010: Comparison of land–precipitation coupling strength using observations and models. J. Hydrometeor., 11, 979994, doi:10.1175/2010JHM1226.1.

    • Search Google Scholar
    • Export Citation
Save