MESO, Inc. MM5 Real Time Forecast System Change Log
08 April 1999: Changed main grid cu scheme to Grell.
March 1999: Changed some of land use characteristics applicable to east coast.
12 February 1999: Changed land/water calculation so that final distribution
is more consistent with grid scale. Land/water data is 10 min resolution, so land/water
distribution on the 45 km grid is now an average of all the 10 min data points within
a 45 grid box. This may affect some site forecasts along the East coast. The grid 2
land/water distribution is unchanged.
04 February 1999: Changed cape calculation to mix lowest 5 model levels
(approximately lowest 50 mb) and use that for parcel, instead of using
lowest level parcel as previously. Fixed bug in eta verification that
affected eta wind direction statistics. Modified terrain and land/water
generation software and surface characteristic fields plotting software
to be more consistent with differences between mm5 dot and cross points.
15 January 1999: Altered MM5 software to eliminate noise generation
on north and south nest boundaries.
28 December 1998: Added in point site tables of forecast output.
18 December 1998: Added specific soundings to web site graphics output.
14 December 1998: Changed all runs so that they use microphysics
look-up table option. Seems to save about 5% in user time, but increases needed
memory. Change effective with 98121500 run. Changed parallel 1-grid run from
initial analysis experiment to more
general experiment run. Right now that run is set to use the mixed phase
microphysics scheme (vs. simple ice in the other runs).
13 December 1998: Changed eta verification to verify against obs within
mm5 domain (not all obs in eta domain). Change effective with 98121300 verifications.
2 December 1998: Changed initialization for all runs to use enhanced
analysis (multiquadratic option, uses raobs and METAR surface obs). Change
starts in 3 December 0000 UTC runs. Added verification statistics online.
November 1998: Fixed many errors in RAWINS (used in enhanced initial
analysis runs) related to modifications
for incorporation of other data sets.
3 November 1998: Added in initialization of soil moisture availability
and snow cover. The SMA is initialized from the Eta 12 hr forecast SMA from the
previous Eta run (the 12Z run, so the 12 hr forecast SMA is valid at 00Z,
the MM5 start time). The snow cover is initialized from the latest available
NESDIS IMS snow cover analysis.
30 October 1998: Added Jom Bresch's modifications to MRF PBL to change
soil moisture availability during the model run. Added in usage of IFSNOW option
to model run to include snow cover effects during simulation. Display snow cover
and soil moisture availability during run for coarse and fine grids.
29 October 1998: Added option to use AVN as first-guess when Eta is unavailable.
Added explicit date/time write to main html file. Added 925 mb convergence and
1000-850 mb thickness plots to grid 2 display. Reversed color indices for RH plots so
blue is moistest.
04 October 1998: Switched cu scheme in main 2-grid run from Grell to K-F. Still
doing cu scheme comparison runs, so extra 1-grid run is now Grell. Rawins test run
uses K-F and is compared to coarse grid of 2-grid run.
23 September 1998: Changed source of SST data to NCEP weekly 1 degree
resolution data set. This data set does have Great Lake temperatures.
22 September 1998: Changed to version 2.9 of MM5v2. Switched PBL scheme from
Blackadar to MRF. For the 22 September case, the changes resulted in a cpu time
reduction of 11%, from 20320 secs to 18114 secs for the 2 grid 36 hr run. The
reduction for the 1 grid run (with the K-F parameterization) was 15%, from 6875 secs
to 5817 secs.
2 September 1998: Changed SST to September climatology (monthly
climatology from Shea, Trenberth, Reynolds 2 degree monthly climatology dataset).
28 August 1998: Added in parallel coarse grid run with Kain-Fritsch
cu scheme, display table to compare forecasts with Kain-Fritsch vs. Grell.
10 August 1998: Changed grid 2 topography and land use to finer scale
(interpolated directly from data to 15 km grid 2, rather being interpolated
from coarse grid). Changed nest start time to time 0. instead of 12 hrs into
the simulation. Switched cumulus parameterization from Kain-Fritsch to Grell.
6 August 1998: Changed cumulus parameterization from Grell to
Kain-Fritsch. Modified land-use to reflect data in 10 min resolution land-water
data set. Moved nest a few points to the east. Added CAPE and CIN to display, added
4-panel displays, added color bar to plots. Fixed a few errors in displays.
21 June 1998: Started daily real-time MM5 runs with version 2.7. Parameterizations
include simple ice, Grell cu, Blackadar PBL, cloud-scheme radiation. The coarse grid
is 93*78*25 with 45 km resolution. A nest with 15 km resolution is initialized
12 hrs into the simulation. The coarse grid is initialized at 0000 UTC from the
previous 12Z Eta 12 hr forecast - no additional data is used. The 12Z Eta forecasts
are also used for the lateral bc.