Updated 2019 version - diagnostic output fields for the Rapid Refresh and HRRR
RAP and HRRR both use the common NCEP post-processing program, Unipost, used now for all NCEP models. GRIB table identifiers for HRRRv2 2-d fields

Some of the RUC-derived diagnostics (described in http://ruc.noaa.gov/ruc/vartxt.html ) were added to Unipost and are retained in the RAP. All of the fields shown are output for the 3km HRRR also. (Sept 2016 - specific HRRR-only fields are now added.)

HRRR-specific variables
  • reflectivity
  • lightning diagnostic
  • updraft helicity
  • vertical velocity
  • VIL - vertically integrated liquid
  • Echo-top level
  • hourly maximum fields

    Relative humidity - Defined with respect to saturation over water in the RAP/Unipost isobaric fields and in the surface relative humidity field. (as done also for NAM but not for GFS as of April 2012).

    Relative humidity with respect to precipitable water

  • RH with respect to precipitable water - calculated as ratio between precipitable water and PW if full column was saturated (described more in upcoming manuscript by Benjamin).

    Diagnosis of 2m temp, dewpoint temp, 10m wind - 2m temperature and dewpoint temperatures in RAP are no longer diagnosed using a more detailed "minimum topography" field as in the RUC.

    Skin temperature - Temperature of top level in 9-level soil model (Smirnova et al, MWR, 2016) over land, and the sea-surface (or lake-surface) temperature over water.

    Sea-level pressure using MAPS reduction (MAPS SLP) - This reduction is the one used in previous versions of RUC/MAPS using the 700 hPa temperature to minimize unrepresentative local variations caused by local surface temperature variations. This reduction is described in Benjamin and Miller (1990, October, Monthly Weather Review, pp. 2099-2116. PDF ) This method has improvement over the standard reduction method in mountainous areas and gives geostrophic winds that are more consistent with observed surface winds.

    Precipitation accumulation - All precipitation values, including the model run total, are liquid equivalents, regardless of whether the precipitation is rain, snow, or frozen.
    - 1h accumulations. 1h accumulation is over last 1h period in model forecast. RAP does not output 3-h accumulated precip, only in 1h buckets.

    Instantaneous precipitation rate - Total precipitation (resolved and sub-grid-scale) in last physics time step is written in mm/s.

    Resolvable and sub-grid scale precipitation - RAP uses some version of the Grell scheme ( Grell-Freitas, 2014, ACPD ) as of 2015 as its convective parameterization scheme, replacing the Grell-Devenyi (2002, GRL) scheme.
    As in most other convective parameterization schemes used at similar horizontal grid spacings, this scheme is not designed to completely eliminate grid-scale saturation in its feedback to temperature and moisture fields. One result of this is that the precipitation from weather systems that might be considered to be largely convective will nevertheless be reflected in the RAP model with the Grell 3-d scheme with a substantial proportion of resolvable-scale precipitation. Thus, the sub-grid scale precipitation from RAP should not be considered equivalent to "convective precipitation".

    Snow accumulation for RAP and HRRR

    - Starting 3/3/2015, snow accumulation in ESRL experimental versions of RAP and HRRR now use only actual snow accumulation and do not use graupel/sleet accumulation any longer. This change was propagated to the NCEP versions of RAP and HRRR (RAPv3, HRRRv2) implemented in Aug 2016. The Thompson microphysics used in RAP and HRRR calculates explicitly the fall of snow mixing ratio, graupel mixing ratio, and rain mixing ratio reaching the surface, using separate fall speeds for each. This allows separate diagnosis of accumulation for each variable.

    Snow/sleet accumulation (in web product) - For NCEP RAP/HRRR forecasts until summer 2015, snow and sleet are combined into a single product called "snow accumulation". This fixed-density snow accumulation is calculated using a 10:1 ratio from the accumulated snow water equivalent with both snow and graupel/sleet combined. This ratio varies in reality, but the ratio used for this product was set at this constant value so that users will know the water equivalent exactly. The snow accumulation (through the snow liquid water equivalent) is explicitly forecast through the mixed-phase cloud microphysics in the model and specifically from snow mixing ratio and graupel mixing ratio fall out to the surface. The graupel field means that this snow/sleet accumulation field includes both sleet and even graupel from convective storms, especially in the 3-km HRRR model. This addition of sleet/graupel into snow accumulation was ended for RAP/HRRR runs by ESRL as of 3/3/2015 and was ended at NCEP for RAPv3/HRRRv2 in Aug 2016.

    Frozen precip percentage - This field uses the explicit precipitation (rain, snow or graupel) produced from the multi-species Thompson cloud microphysics. It is calculated as (snow-accumulated + graupel-accumulated) divided by (snow-accumulated + graupel-accumulated + rain-accumulated). No rime factor is used in this explicit calculation.

    Graupel accumulation - Graupel accumulation is defined as the water equivalent of the combination of sleet (frozen rain droplets, i.e., ice pellets) and heavily rimed snow.

    Snow depth - This field is the current estimated snow depth using the latest snow density, which is also an evolving variable. (Snow water equivalent cycles internally within the RAP 1-h cycle.)
    The 10:1 ratio is kept only for fresh snow falling on the ground surface when 2-m air temperature is below -15 C. When 2-m temperature is above -15 C the density of falling snow is computed using an exponential dependency on 2-m temperature, and usually the ratio will be less than 10:1, but not less than 2.5:1. The density of snow pack is computed as the weighed average of old and fresh snow, and it changes with time due to compaction, temperature changes, melted water held within the snow pack and addition of more fresh snow. (See Koren et al., 1999, J. Geophys. Res., for snow density formulations.
    Snow density was provided in the RUC grib output (but not in RAP) together with snow water equivalent and snow depth. Snow density in RAP (Kg/m**3) = Snow-water equivalent [kg/m**2] / snow depth [m].

    RAP uses 2011 version of RUC land-surface model with 2-level snow model and cold-season effects (freezing and thawing of moisture in soil). The RAP cycles snow depth/cover, as well as snow temperature in the top 5 cm and below that top snow layer.

    Categorical precipitation types - rain/snow/ice pellets/freezing rain - These yes/no indicators are calculated from the 3-d hydrometeor mixing ratios calculated in the explicit cloud microphysics parameterization (references: Thompson 2008, Mon. Wea. Rev., Thompson/Eidhammer 2014 MWR) in the HRRR and RAP models.

    2016 Wea. Forecasting article on precip-type diagnostic used in HRRR/RAP: Explicit precipitation-type diagnosis from a model using a mixed-phase bulk cloud-precipitation microphysics parameterization Authors: Benjamin, Brown, Smirnova, NOAA Earth System Research Laboratory, Boulder, CO

    These p-type values from the post-processing are not mutually exclusive. More than one value can be yes (1) at a grid point. Here is how the diagnostics are done (all precipitation rates below are at the ground and in liquid-water equivalent):

    Freezing levels - Two sets of freezing levels are output from RAP, one searching from the bottom up, and one searching from the top down. Of course, these two sets may be equivalent under many situations, but they may sometimes identify multiple freezing levels. The bottom-up algorithm will return the surface as the freezing level if any of the bottom 3 native levels (up to about 80 m above the surface) are below freezing (per instructions from Aviation Weather Center, which uses this product). The top-down freezing level returns the first level at which the temperature goes above freezing searching from the top downward. For both the top-down and bottom-up algorithms, the freezing level is actually interpolated between native levels to estimate the level at which the temperature goes above or below freezing.

    3-h surface pressure change - These fields are determined by differencing surface pressure fields at valid times separated by 3 h. Since altimeter setting values (surface pressure) are used in the RAP analyses, this field reflects the observed 3-h pressure change fairly closely over areas with surface observations. It is based on the forecast in data-void regions.
    -The 3-h pressure change field during the first 3 h of a model forecast often shows some non-physical features resulting from gravity wave sloshing in the model, despite use of digital filter initialization (DFI) in RAP/WRF model. After 3 h, the pressure change field are better behaved. The smaller-scale features in this field appear to be very useful for seeing predicted movement of lows, surges, etc. despite the slosh at the beginning of the forecast.

    CAPE - convective available potential energy - RAP uses standard Unipost definition of CAPE including use of virtual temperature. RAP CAPE vs. RUC CAPE differences are described in http://ruc.noaa.gov/pdf/RUC_cape_vs_RAP_Unipost_cape_apr12.pdf . CAPE values are provided for surface-based CAPE, most unstable CAPE (MUCAPE) in lowest 300 hPa, and mixed-layer (lowest ~50 hPa mixed) CAPE (MLCAPE). The calculation of CAPE considers only positively-buoyant contributions of the ascending air parcel, starting at the parcel's Lifted Condensation Level and ending at the Equilibrium Level.

    CIN - convective inhibition - indicates accumulated negative buoyancy contributions for the ascending parcel, starting at the parcel's Lifted Condensation Level (LCL) and ending at it's Equilibrium Level. By this definition, CIN is mainly accumulated between the LCL and the Level of Free Convection (LFC), and represents the negative bouyant energy that must be overcome in order for the parcel to become positively buoyant once it reaches its LCL. This is also the standard Unipost definition.

    EL - equilibrium level - the highest positively buoyant level. This is also the standard Unipost definition. The EL provied is associated with the most unstable CAPE parcel (MUCAPE; using the parcel with highest theta-e in the lowest 300hPa).

    Lifted index / Best lifted index - Lifted index uses the surface parcel, and best lifted index uses buoyant parcel from native level with maximum buoyancy within 300 hPa of surface (also standard Unipost definition).

    Precipitable water - Integrated precipitable water vapor from surface to top level (10 hPa).

    Helicity and storm motion

    Standard Unipost definition - uses Bunkers et al. 2000, Weather and Forecasting.

    Reference: Bunkers, M. J., and co-authors, 2000: Predicting supercell motion using a new hodograph technique. Wea. Forecasting, 15, 61-79.

    What about the high values of helicity?

    The units of helicity are m^2/s^2. The value of 150 is generally considered to be the low threshold for tornado formation. Helicity is basically a measure of the low-level shear, so in high shear situations, such as behind strong cold fronts or ahead of warm fronts, the values will be very large maybe as high as 1500. High negative values are also possible in reverse shear situations.

    Lightning / Thunder This thunder parameter is from David Bright (NOAA).

    At any point where convective precip is forecast to occur (i.e., where the convective parameterization scheme is active), thunder is predicted if all of the following are true:

  • The temp of the LCL is greater than -10 C
  • The temp of the EL is less than -18 C
  • The CAPE in the layer between 0 and -20 C exceeds 75 J/kg (this threshold from David Bright via personal communication to Geoff Manikin)

    Additional information is available at http://www.spc.noaa.gov/publications/bright/ltgparam.pdf

    Soil moisture - cycled continuously in the RAP model/assimilation cycle without resetting from external models. There are 6 levels in the RUC land-surface model used in the RAP configuration of WRF, extending down to 3 m deep, but the field shown are for only the top 3 levels, at the surface, 5 and 20 cm depth. The surface value is indicitive of the top 2 cm of soil only, so this field responds quickly to recent precipitation or surface drying. In general, the deeper in the soil, the more slowly do soil conditions change..

    Tropopause Pressure -

    Vertical velocity - Following NCEP unipost convention, vertical velocity in m/s is converted to omega in Pa/s using the formula omega = -rho*g*w, where rho is air density and g = 9.80665 m/s*s.

    (The RAP vertical motion is at a given time step and is not time-averaged.)

    PBL depth

    gust wind speed potential - The diagnostic should be understood as a gust potential, reflecting an estimate of maximum wind gusts with an area at a given time. This gust potential diagnostic will overestimate gust values at many stations and gives values higher than the mean gust of stations in a given area, but it will also provide a "relatively good" probability of detection of the highest observed wind gusts.


    Note: Gust wind speed was temporarily larger at night in RAPv3/HRRRv2 due to use of a hybrid PBL diagnostic.
    13 Oct 2016 - ESRL experimental RAPv3/HRRRv2 changed back to use theta-v-profile PBL depth diagnostic for wind gust calculation.
    2 Nov 2016 - NCEP operational RAPv3/HRRRv2 also changed back to use theta-v-profile PBL depth for the wind gust diagnostic.

    cloud base height (or ceiling - note differences)

    best combined with 3 other conditions listed further below).
    Units - meters above sea level (ASL).
    Horizontal grid points without any cloud layer are indicated with -99999. Note that the RAP/HRRR graphics show cloud base in above GROUND level (AGL) -- the RAP/HRRR terrain elevation height is subtracted first. But in the actual GRIB files, cloud base height is in ASL.
    Note that for RAP/HRRR fields using this diagnostic, there are differences: cloud-base height for RAPv1/v2 and HRRRv1 issued by NCEP used conditions 1-4 (more cloud/ceiling detection, while the ceiling field used only condition 1). Starting with RAPv3 and HRRRv2 at NCEP, the ceiling field is the diagnostic with all 4 conditions applied, while the cloud base uses only condition #1. Do NOT use "cloud base" field for ceiling and instead, always use GRIB "ceiling" field from RAP and HRRR grids.

    cloud top height - Top level at which combined cloud and ice mixing ratio exceeds 10**-6 g/g. Units - meters above sea level. Horizontal grid points without any cloud layer are indicated with -99999.

    cloud fraction
    Following NWS convention (defined for general forecasting and obviously not for aviation purposes), "low" cloud fraction is defined as surface-642 hPa, medium as 642-350 hPa, and "high" as 350 hPa to top.

    visibility - RUC/RAP extension of Stoelinga-Warner (JAM, 1999) algorithm

    Special HRRR output fields

  • Reflectivity
    Reflectivity is computed for each model grid point based on rain, snow, graupel/hail, and temperature at that grid point. The temperature is used to determine if melting snow is present (“bright band”).
  • Lightning diagnostic
    Hourly maximum lightning threat is a measure of total lightning (cloud-to-ground and in-cloud). It is calculated as the sum of 2 different diagnostics for each model column. The first is based vertical graupel flux (vertical motion and graupel) at -15C. The second is based on the vertically integrated ice (cloud ice, snow, graupel). (McCaul et al. 2009, Wea. Forecasting) . The units are flashes per square km every 5 mins. No horizontal smoothing or broadening is applied, so the lightning threat is dependent on full-resolution cloud forecasts (similar to the detail of the deterministic HRRR reflectivity forecasts). It attempts to capture both lower frequency, broad anvil lightning and higher frequency lightning near updrafts.
  • Updraft helicity
    Hourly maximum updraft helicity is an hourly maximum valid at the end of each hour. Updraft helicity is derived from upward vertical velocity and cyclonic vertical vorticity between 2 and 5 km AGL. It indicates updraft rotation in forecasted convection. It can imply a threat for tornadoes but does not explicitly predict tornadoes. It will not identify anticyclonic rotation and associated hazards. Since updraft helicity depends partially on updraft strength, it can be small in low CAPE, highly sheared environments. It does not discriminate between elevated and surface based convection.
  • Vertical velocity
    Hourly maximum updraft velocity / downdraft velocity are the maximum upward/downward vertical velocity (m s-1) between the surface and 400 mb. They do not indicate where in the vertical column the maximum occurred or when during the hour. Hourly mean vertical velocity is the average vertical velocity (m s-1) between sigma level 0.8 and 0.5 (approximately 800 mb and 470 mb)
  • VIL - vertical integrated liquid
    Calculated from reflectivity to produce vertically integrated liquid in kg/m**2. With an average vertical profile within a convective storm, 12 kg/m**2 VIL is very roughly equivalent to a 50 dBZ reflectivity although VIL is, by definition a vertically integrated quantity.
  • Echo-top level
    Maximum height (in m above sea level) at which reflectivity is 18 dBZ. Calculated from vertical profile of reflectivity.
  • Hourly maximum fields Maximum hourly fields contain the maximum value across every model time-step (20 seconds for HRRR model) at each grid point during that hour.
    Care must be taken to interpret these fields.
    (Back to RAP homepage)

    Prepared by Stan Benjamin stan.benjamin@noaa.gov, 303-497-6387