|
RUC-2 - The Rapid Update Cycle Version 2 Technical Procedures Bulletin - draft |
ABSTRACT - Summary of RUC-2 vs. RUC-1
differences
Horizontal and vertical domain
Data assimilated
Assimilation frequency
Model physics
Surface analyses/forecasts
Variables available
Output formats available
Verification stats for 40km MAPS vs. 60km MAPS
and 48km Eta
Change log for operational RUC-2 running in real-time at NCEP
To RUC/MAPS homepage, including real-time images
1. INTRODUCTION AND HISTORY
A number of significant weather forecasting problems exist in the 0-12 hour range, including severe weather in all seasons (tornadoes, severe thunderstorms, crippling snow and ice storms) and hazards to aviation (clear air turbulence, icing, downbursts). Accurate short-term forecasts are clearly indispensable for the protection of life and property, but they also have tremendous economic value even under non- threatening conditions, for example, for agriculture and the recreation and power-generation industries. The Rapid Update Cycle (RUC) was designed to provide accurate numerical forecast guidance for weather-sensitive users for the next 12 hour period. The RUC runs at the highest frequency of any forecast model at the National Centers for Environmental Prediction (NCEP), assimilating recent observations aloft and at the surface to provide very high frequency updates of current conditions and short-range forecasts using a sophisticated mesoscale model.
This bulletin describes a new version of the RUC, called RUC-2, implemented at NCEP in April 1998. The RUC-2 produces new 3-d analyses and short-range forecasts every hour, compared to the 3-h updating in RUC-1. The original Rapid Update Cycle (RUC-1, Benjamin et al. 1994a (TPB) ,1994b ) was implemented in September 1994 at NCEP. Some number of smaller changes were made to RUC-1 over the 1995-1996 period, but the RUC-2 is a significant advance over RUC-1, not just in assimilation frequency, but also in resolution, types of data assimilated, and model physics. These changes (summarized in the TPB Abstract) allow the RUC-2 to more accurately represent significant weather systems across the United States in all seasons.
Summary
of characteristics in RUC-2 compared to RUC-1. Also see the
TPB Abstract.
Uses of the RUC
Some key users of the RUC
2. DOMAIN, RESOLUTION, TERRAIN FOR RUC-2
2.a. Domain
The RUC-2 domain (
Fig.1a) is on a Lambert conformal projection
matching that used for the AWIPS 212 NWS distribution grid. The mesh is
rectangular on this projection, and its size is 151 by 113 grid points
(compared to 81x62 for RUC-1
(Fig.1b)
). The grid length
is 40.635 km at 35 deg N. Due to the varying map-scale factor from the
projection, the actual grid length in RUC-2
varies from about 40.6 km at 35 deg N
to 33 km at the north boundary. The grid length is about 38 km at
43 deg N.
The RUC-2 domain was designed to:
The 40-km RUC-2 domain covers about 50% more area than the 60-km RUC-1 domain.
It extends farther than the RUC-1 domain in all directions, but especially in
the southeast, owing to use of the Lambert conformal projection and the need
to cover slightly east of the state of
Maine. It covers considerably more oceanic areas than the RUC-1 domain.
The RUC-2 latitude/longitude at each point in an ASCII file can be
downloaded from
http://maps.fsl.noaa.gov/MAPS.domain.html.
The lower left corner point is (1,1), and the upper right corner point is
(151,113),
as shown in the table below.
| RUC-2 point | AWIPS-212 point | Latitude | Longitude |
|---|---|---|---|
| (1,1) | (23,7) | 16.2810 N | 126.1378 W |
| (1,113) | (23,119) | 54.1731 N | 139.8563 W |
| (151,1) | (173,7) | 17.3400 N | 69.0371 W |
| (151,113) | (173,119) | 55.4818 N | 57.3794 W |
2.b. Horizontal resolution
Horizontal resolution in RUC-2 is 40 km compared to 60 km for the RUC-1. The higher resolution allows considerable improvement in resolution of topography and also in shapes of coasts and lakes. These improvement improve the ability of the RUC-2 to resolve local circulations and orographic precipitation patterns. Because the RUC-2 model has less internal smoothing than the Eta or NCAR/Penn State MM5 models, these features tend to be fairly well depicted considering its 40-km resolution.
2.c. Vertical resolution
The RUC-2 has 40 vertical levels compared to 25 levels in RUC-1. The RUC-2 continues to use a generalized vertical coordinate configured as a hybrid isentropic-sigma coordinate in both the analysis and model. This coordinate has proven to be very advantageous in RUC-1 in providing sharper resolution near fronts and the tropopause (e.g., Benjamin 1989, Johnson et al. 1993, Zapotocny et al. 1994). Some of the other advantages include:
A sample cross section of RUC-2 native levels is displayed in Fig. 2. The cross-section is across the United States, passing south of San Francisco, through Boulder (where a downslope windstorm occurred that morning) and through southern Virginia on the East Coast. The cross section is for a 12-h forecast valid at 1200 UTC 30 November 1995.
The typical RUC-2 resolution near fronts is apparent in this figure, as well as the tendency for more terrain-following levels to "pile up" in warmer regions (the eastern part of the cross section, in this case). The hybrid isentropic-sigma coordinate is defined by a 20-line section of code in both the analysis and forecast model. The rest of the code treats the analysis/model processes as a generalized vertical coordinate. The 20-line section of code can be changed to define a pure sigma terrain-following coordinate and has been tested in this mode.
In the RUC-2 (as well as in RUC-1), analysis/model levels which are isentropic in part of the domain can become terrain-following in other parts, as shown in Fig. 2. A reference potential temperature is assigned to each of the 40 levels (Table 1).
| Table 1. RUC-2 reference potential temperatures | |||||||
|---|---|---|---|---|---|---|---|
| 224. | 232. | 240. | 245. | 250. | 255. | 260. | 265. |
| 270. | 274. | 278. | 282. | 286. | 290. | 294. | 297. |
| 300. | 302. | 304. | 306. | 308. | 310. | 312. | 314. |
| 316. | 318. | 320. | 322. | 325. | 328. | 331. | 334. |
| 337. | 341. | 347. | 355. | 364. | 375. | 400. | 450. |
The prespecified pressure spacing in RUC-2, starting from the ground is 2, 5, 8, and 10 mb, followed by as many 15-mb layers as are needed. (Near-surface pressure spacing in RUC-1 was 20 mb.) This terrain-following spacing compacts somewhat as the terrain elevation increases. Excellent resolution of the boundary layer is provided in all locations, including over higher terrain. The lowest atmospheric level in RUC-2 is set at 5 m above the model terrain height. The effects of this choice are discussed in the analysis and model sections, but since the RUC-2 has an explicit level actually at the surface, no extrapolation from higher levels is necessary to diagnose values at the surface. The minimum potential temperature spacing occurs through much of the troposphere and is 2 K instead of 4 K as in RUC-1. The top level in RUC-2 is at 450 K as opposed to 410 K in RUC-1. Overall, the vertical resolution is somewhat higher both in the boundary layer and free atmosphere, and the domain extends farther into the stratosphere.
2.d. Terrain
The most obvious difference, of course, between terrain in RUC-2 (
Fig. 3a) and that in RUC-1
(Fig. 3b) is that finer-scale topographical
features are distinct in the RUC-2 terrain. RUC-2 analyses and forecasts
can depict many significant topographically induced features, including
mountain/valley circulations, mountain waves, sea breezes, and orographic
precipitation patterns.
The surface elevation of the RUC-2 is
defined by a "slope envelope" topography instead of the previous full envelope
topography used in RUC-1. The envelope topography is defined by adding the
sub-grid-scale terrain standard deviation (calculated from a 10-km terrain
field)
to the mean value over the grid box. In the slope envelope topography,
the terrain
standard
deviation is calculated with respect to a plane fit to the
high-resolution topography within each grid box. This gives more accurate
terrain
values, especially in sloping areas at the edge of high-terrain regions. It
also avoids
a tendency of the standard envelope topography to project the edge of plateaus
too far laterally onto low terrain regions. Using the slope envelope topography
gives lower terrain elevation at locations such as Denver and Salt Lake City
which
are located close to mountain ranges.
The RUC-2 topography at each point in an ASCII file can be
downloaded from
http://maps.fsl.noaa.gov/MAPS.domain.html.
3. ANALYSIS METHOD FOR RUC-2
3.a. Data assimilated
3.b. Optimum interpolation analysis
The optimal interpolation multivariate analysis used in RUC-1
has been substantially modified for the RUC-2, providing,
among other things, closer fit to observations, better use of aircraft
ascent/descent winds and temperatures, and greater efficiency.
A discussion of optimal interpolation analysis in isentropic
coordinates is provided in Benjamin (1989).
Sequence in RUC-2 analysis
This background field is
defined at the (x,y,p) points of the hybrid coordinate surfaces
for the background field. The quality control and analysis steps
below are carried out at these (x,y,p) points. This will result
in changes to virtual potential temperature at these points. For the RUC-2, the
next-to-last step (bullet) is in effect a repositioning of the coordinate
surfaces to correspond to the results of the analysis steps that preceed
it.
The RUC-2 analysis provides de facto analyses of cloud variables
and soil variables by using the previous 1-h forecast of these variables as
initial conditions for the next run. Although use of observations will later
provide improved fields for these variables (e.g., Kim and Nychka 1998),
this "cycling" provides substantial improvement over zero initial clouds and
climatology for soil variables.
3.c. Incorporation of the surface analysis within the 3-d analysis
With the 1-h assimilation cycle, the RUC-2 integrates into one system
the RUC-1 and RSAS (RUC Surface Analysis System) from the 60-km era.
The RUC-2 surface analyses are improved over those from RUC-1 due to
the use of a forecast background combined
with new design features to ensure that the 40km surface analyses
not only draw more closely
to the data, but also have better consistency and reliability.
Specific advantages of RUC-2 surface analyses over those from RUC-1 are:
The following new features are added to ensure that the 40km surface
analyses
draw very closely to surface data:
Comparison of 60km RUC-1 and
40km RUC-2 surface temperature analyses
for 1200 UTC 19 February 1997.
3.d. Quality control
As in RUC-1, quality control in RUC-2 involves a buddy check. The buddy check
is of observation residuals, the differences betwen the observation and
the background field interpolated to the observation point, and not of the
observations alone. This is an important distinction, since it means that
any known anomaly in the previous forecast has already been subtracted out,
improving the sensitivity of the QC procedure to actual errors.
At each
observation point, the parameter in question is estimated via optimum
interpolation of values from surrounding observation points. If the estimated
and measured values differ by more than a prescribed amount, further checks
determine whether the central observation or one of its neighbors is erroneous.
Bird contamination for radar/profiler winds
Checks are made for bird contamination for both VAD and profiler winds in
RUC-2. A careful check for bird (and other) contamination in profiler winds
is made at the Profiler Hub in Boulder, CO. This check includes use of
second-moment data to examine for likelihood of bird contamination.
If the quality control
flag produced by this check
indicates suspicious data, the profiler data at that level is not used
by RUC-2. For VAD winds, no second-moment data is available, so a cruder
and more conservative check is made. A solar angle is calculated, and if
the sun is down and the temperature is warmer than -2 deg C, VAD winds
are not used if they have a northerly component between 15 August and 15 November
or a southerly component between 15 February and 15 June.
3.e. Future improvements
A new 3-d variational analysis (Devenyi and Benjamin 1998) is nearing
completion for the RUC-2 and will follow the rest of the RUC-2 into
operations with a lag time of several months.
4. ONE-HOUR ASSIMILATION CYCLE FOR RUC-2
See also information on
1-h cycle and data used for more differences.
Grids from RUC-2 will be available almost 1 hour earlier
than those from RUC-1. The data cut-off
time from the 40km system is 20 min after the analysis valid time.
A catch-up cycle at 0000 and 1200 UTC runs at +55 min
to catch late-arriving rawinsonde data. Twelve-hour forecasts from
these times are run from the catch-up cycle analysis, rather than the
"early look" analysis at 20 min after.
5. RUC-2 FORECAST MODEL
The RUC-2 forecast model is an updated version of the generalized
vertical coordinate model described by Bleck and Benjamin (1993).
Modifications to a 20-line section of code in the model are sufficient
to modify it from the hybrid isentropic-sigma coordinate described
in section 2.c to either a pure sigma or pure isentropic model.
The RUC-2 model is considerably different from the RUC-1 model in
its parameterizations of physical processes such as cloud microphysics
(stable precipitation),
turbulent mixing, radiation, and convective precipitation. To some
extent, this was made possible by changing the RUC-2 model to use
the code structure of the NCAR/Penn State Mesoscale Model version 5 (MM5,
Grell et al. 1994). This allowed relatively easy transfer of MM5
parameterizations (cloud microphysics, radiation) into the RUC-2
model, and will continue to do so in the future as new MM5 parameterizations
are developed.
5.a. Basic dynamics/numerics
Here are some of the basic numerical characteristics of the RUC-2 model:
Explicit cloud/moisture processes. The explicit
microphysics from the
NCAR/Penn State mesoscale model MM5 (level 4, Reisner et al. 1997)
is used, with five
hydrometeor species -- cloud water, rain water, snow, ice, graupel (mixing
ratios
for each) and also with an explicit prediction of ice particle number
concentration. This improvement was made to provide improved
forecasts of clouds, icing, and precipitation from RUC-2.
In the 60km RUC-1, stable precipitation was defined simply by
supersaturation removal,
with no knowledge of cloud or ice processes. All of the cloud variables are
cycled
in the 40-km MAPS, meaning that there are initial cloud fields for each run.
In the RUC-2 model, all six cloud variables are advected using the
positive definite scheme of Smolarkiewicz (1983) on the isentropic-sigma
levels with adaptive vertical resolution.
The incorporation of this scheme into RUC-2 is described in detail by Brown
et al. (1998).
Improved surface physics.
The RUC-2 includes a 6-layer soil/vegetation/snow model
(schematic)
to improve forecasts of low-level conditions. Surface (shelter/anemometer
level) forecasts are often critically
dependent on accurate estimates of surface fluxes, and in turn,
on reasonably accurate soil moisture and temperature estimates.
The RUC-2
soil model contains heat and moisture transfer equations solved at 6 levels
at each grid point
together with the energy and moisture budget equations for the ground surface
(Smirnova et al. 1997a,b). The heat and moisture budgets are applied to a thin
layer spanning the ground surface and including both the soil and the
atmosphere with corresponding heat capacities and densities.
A treatment of the evapotranspiration process, developed by
Pan and Mahrt (1987), is implemented in the MAPS/RUC soil/vegetation scheme.
In the presence of snow cover, snow is considered to be an additional
upper layer of soil that interacts with the atmosphere, significantly affecting
its characteristics (Smirnova et al 1998).
The snow model contains a heat-transfer equation within the snow layer
together with the energy and moisture budget equations on the surface of
the snow pack. This budget is applied to the entire snow layer if the snow depth
is less than a threshold value, currently set equal to 7.5 cm, or to the
top 7.5 cm layer of snow if the snow pack is thicker. Snow evaporates at a
potential rate unless the snow layer would all evaporate before the end of
the time step. In this case the evaporation rate is reduced to that
which would just evaporate all the existing snow during the current time step.
A heat budget is also calculated at the boundary between the snow pack and
the soil, allowing melting from the bottom of the snow layer. Melting at the
top or bottom of the snow layer occurs if energy budgets produce temperatures
higher than the freezing temperature (0 deg C). In this case the snow temperature
is set equal to the freezing point, and the residual from the energy budget
is spent on melting snow. Water from melting snow infiltrates into the soil,
and if the infiltration rate exceeds the maximum possible value for the given
soil type, then the excess water becomes surface runoff.
The accumulation of snow on the ground surface is provided by the
microphysics algorithm of the MAPS/RUC forecast scheme (Reisner et al. 1997,
Brown et al. 1998). It predicts the total amount of precipitation and also
the distribution of precipitation between the solid and liquid phase.
The subgrid-scale ("convective") parameterization scheme also contributes
to the liquid precipitation. With or without snow cover, the liquid phase
is infiltrated into the soil at a rate not exceeding maximum infiltration rate,
and the excess goes into surface runoff. The solid phase in the form of snow
or graupel is accumulated on the ground/snow surface and is unavailable for
the soil until melting begins.
The RUC-2 surface
package provides surface temperature and dewpoint forecasts that are
clearly superior to those of the RUC-1 due to improved surface heat and
moisture fluxes.
The soil temperature and moisture has been evolving since 29 April 1996,
leading to fairly accurate estimates of these fields, certainly much
better than climatology
(Smirnova et al.
1997b).
Real-time estimates of soil moisture in the top 2 cm are
available on the MAPS/RUC home page,
http://maps.fsl.noaa.gov.
Atmospheric radiation.
The MM5 atmospheric radiation package (Dudhia 1989, Grell et al. 1994)
is used in the RUC-2, with additions
for attenuation
and scattering by all hydrometeor types. This scheme is a broadband
scheme with separate components for longwave and shortwave radiation.
In the RUC-1, there was no atmospheric
radiation at all, only a surface radiation budget with clouds diagnosed from
relative humidity.
The solar flux at the top of the atmosphere is now
variable, taking
into account the elliptical orbit of the earth around the sun.
Turbulent mixing.
In RUC-2, turbulent mixing at all levels, including the boundary layer,
is prescribed the explicit turbulence scheme of Burk and Thompson (1989).
This scheme is a level-3.0 scheme, with explicit forecast of turbulent
kinetic energy and three other turbulence variables, replacing the
Mellor-Yamada level-2.0 scheme in RUC-1. The surface layer mixing continues
to be prescribed by Monin-Obukhov similarity theory, specifically
the three-layer scheme described in Pan et al. (1994).
With the Burk-Thompson scheme, the RUC-2 forecasts TKE amounts of
5-20 J/kg in the boundary layer, and also forecasts TKE maxima aloft,
typically localized
in frontal zones, corresponding to estimated
areas of clear-air turbulence potential.
Convective parameterization.
A version of the Grell (1993) convective parameterization is used, updated from
that used in
the RUC-1, including fixes to downdraft detrainment, calculation
of cloud top, minimum cloud depth, and capping criteria.
This version gives somewhat larger amounts of precipitation
and more coherent patterns in convective areas than the RUC-1 version.
The inclusion of downdrafts in the Grell scheme results in smaller-scale
details in RUC-2 warm season precipitation patterns than may be evident
in that from the Eta model, which currently uses the Betts-Miller-Janjic
convective parameterization.
This same difference in character of
precipitation forecasts is also evident in NCEP/NSSL
experiments comparing the Kain-Fritsch (which also
includes downdrafts) and Betts-Miller-Janjic schemes
both within the MesoEta model (Kain et al. 1998).
5.c. Lateral boundary conditions
Lateral boundary conditions for the RUC-2 model are provided by the
early Eta run output at 3-h intervals. The Eta model forecasts are
interpolated to the 40-km RUC-2 domain on its hybrid coordinate levels.
Values of pressure thickness, virtual potential temperature, and
horizontal winds at the edge of the RUC-2 domain (up to 5 grid points
from the boundary) are nudged (Davies 1976)
toward the Eta values at each time step
in a model run.
The accuracy of RUC-2 forecasts is driven to some extent by the
time availability of Eta forecasts. RUC-2 forecasts initialized at
0000 or 1200 UTC are forced to use boundary conditions from Eta runs
already 12 h old, since the RUC-2 runs before the Early Eta.
This means that the 12-h RUC-2 forecasts valid at 0000 or 1200 UTC
are nudging toward 24-h Eta forecasts at these times.
Typically, the skill of RUC-2 forecasts jumps near the western
boundary by the 0300 or 1500 UTC runs, respectively, since the newer
Early Eta runs are available by those times.
5.d. Fields for surface boundary conditions
6. VERIFICATION OF RUC-2 ANALYSES AND FORECASTS
Verification statistics for the RUC-1, RUC-2, and
the 48-km Eta models are available at this link.
7. RUC-2 OUTPUT VARIABLES AND FILES
Variables available
In RUC-2, water vapor mixing ratio replaces condensation pressure (output
in RUC-1) as
the water vapor moisture variable in
the 3-d grids.
Height also replaces Montgomery stream function. The use of height
and water-vapor mixing ratio instead of the variables they replace
facilitates calculation of derived quantities involving these variables.
Height also stores more efficiently in GRIB.
40km RUC/MAPS output formats
From the 40km RUC at NCEP, five different formats will be available:
Click here for more
information on 40km RUC-2 variables.
RUC-2 GRIB table -
parameters for all RUC-2 variables
Use of GRIB gridded input and output. The MAPS database routines used in RUC-1
will no longer be used for grid input and output.
8. DIAGNOSTIC FIELDS FROM RUC-2
Explanation of diagnosed RUC-2 variables
, including surface fields, stability indices, precipitation
products, special level fields (tropopause, freezing level).
9. PLANS FOR THE FUTURE
10. REFERENCES
Last updated: 2 Feb 98
The new data sets assimilated in the 40-km RUC-2 include:
These new data sets already assimilated in the experimental
40-km RUC/MAPS at FSL will also be ingested into the RUC-2
as soon as available at NCEP:
Satellite-based precipitable water retrievals and cloud-drift
winds are currently used only over water points. Satellite
observations over land will be ingested in the near future.
Wind profiler, rawinsonde, aircraft, and surface (land and buoy)
observations continue to be utilized in the RUC-2, as they were
for RUC-1, except that the wind profiler and rawinsonde data
are used with higher vertical resolution due to the 40 levels
in RUC-2.
Here is a summary of the actual measurements used from other data sets:
More information on the use of wind data in RUC-2 is available in
Smith and Benjamin (1998).
1-h assimilation cycle. The background for each analysis is
the previous 1-h forecast. The 1-h cycle allows much more complete use of profiler,
surface,
and VAD data, which are all available at least hourly. The time window for
aircraft data is now -1h to 0h instead of -2h to +1h, meaning that aircraft
data are now applied closer to the time that they are actually reported.
The atmospheric prognostic variables of the RUC-2 forecast model are:
The soil prognostic variables (at six levels) of the RUC-2 forecast model are:
Other surface-related prognostic variables are snow water equivalent moisture
and snow temperature.
(~6 MB in GRIB format per output time)
(~10 MB in GRIB format per output time)
(0.35 MB per output time)
Naming convention
isobaric main:
RUC/ruc2.yymmdd/ruc2.TXXZ.pgrbanl
or
RUC/ruc2.yymmdd/ruc2.TXXZ.pgrbfzz
native:
RUC/ruc2.yymmdd/ruc2.TXXZ.bgrbanl
or
RUC/ruc2.yymmdd/ruc2.TXXZ.bgrbfzz
XX = cycle time (01,02,03..)
- zz = forecast hour (00,01,02...)
surface:
RUC/ruc2.yymmdd/ruc2.TXXZ.sgrbanl
or
RUC/ruc2.yymmdd/ruc2.TXXZ.sgrbfzz
BUFR sounding:
RUC/ruc2.yymmdd/ruc2.TXXZ.bufranl
or
RUC/ruc2.yymmdd/ruc2.TXXZ.bufrfzz
(~2.5 MB for each hour for a given run, soundings for all sites)
This page prepared by Stan Benjamin
