The meteorological background fields that are used for an MM5 simulation are extracted from a larger-scale (either global or regional) simulation that geographically covers the MM5 simulation domain and simulation period entirely. The background fields are used to generate MM5’s “first guess” and lateral boundary conditions. The first guess is improved with synoptic and mesoscale observations to generate an analysis that is transformed into an initial state for the model. The lateral boundary conditions provide forcing to the model throughout the simulation, and they can be especially influential in the later stages of the simulation.

Since CMAQ will be used to simulate very high-resolution domains, it is important to select a high-quality background dataset for the model. The MM5 system can ingest and process global analyses data generated by either the European Centre for Medium-Range Weather Forecasts (ECMWF) or the National Centers for Environmental Prediction (NCEP). With an account on the NCAR Cray or other arrangements with NCAR (see http://www.mmm.ucar.edu/mm5/faqMM5.html), you may retrieve the pre-formatted global analyses data for use in MM5. ECMWF global analyses data is twice daily at a 2.5-degree resolution from January 1980 to December 1989. A special dataset of the European Centre for Medium-Range Tropical Ocean and Global Atmospheric program (ECMTOGA) global analyses is available for restricted usage in MM5. This dataset (at a 2.5-degree resolution from January 1985 through December 1997) is available from NCAR for a fee. NCEP global tropospheric analyses data (twice daily at a 2.5-degree resolution from July 1976 to April 1997) is also available from NCAR. Background analyses data is also available from sources other than NCAR. If you acquire such data from a source other than NCAR, you may need to write software to format the data into a form that REGRID expects, or you may need to modify REGRID to accept the analyses data.

Standard meteorological observations of temperature, relative humidity, height, sea-level pressure, and wind are required to improve the background fields with synoptic and mesoscale details. The analysis model (RAWINS, discussed in Section 4.5.3) is set up to process surface observations (including ship reports and buoys) and upper-air observations (including radiosonde observations and wind profiles). Other types of data may be included in the analysis if the appropriate modifications are made to the software.

With an account on the NCAR Cray or other arrangements with NCAR (see http://www.mmm.ucar.edu/mm5/faqMM5.html), you may retrieve pre-formatted surface and upper-air observations that were acquired by NCAR from the National Centers for Environmental Prediction (NCEP). Daily surface observations (including both land and water observations) are available for the period from January 1978 through today. Some of the dates include two sets of files: six-hour, and three-hour. The six-hour files include data observed at 00, 06, 12, and 18 Universal Time Coordinates (UTC). The three-hour files include data observed at 03, 09, 15, and 21 UTC. The more current surface observation files have all observed surface data merged into a single file. The upper-air data are available for the period from January 1985 to April 1997 for 00 UTC and 12 UTC.

If you acquire your meteorological observation data from a source other than NCAR (e.g., from special field studies), you may need to write software to format the observations into a form that RAWINS expects.

MM5 requires analyses data for sea-surface temperature and snow to more accurately simulate the meteorological conditions. With an account on the NCAR Cray or other arrangements with NCAR (see http://www.mmm.ucar.edu/mm5/faqMM5.html), you may retrieve such data. If you acquire sea-surface temperature and snow analyses data from a source other than NCAR, you may need to write software to convert the data into a form that REGRID (discussed in Section 3.5.2) expects or modify REGRID to accept the data.

NCAR has sea-surface temperature data from the U.S. Navy and NCEP, as well as sea-surface temperature data generated from climatological data. The Navy sea-surface temperature data is 63 x 63 Northern Hemisphere polar data. It is available for cases between November 1961 and December 1993. More current Navy sea-surface temperatures may be available from NCAR but are not listed in the NCAR catalog. The NCEP sea-surface temperature data is included as part of the global analysis data and is available at 2.5-degree resolution. The climatological sea-surface temperature data available from NCAR is at 2-degree resolution. When possible, it is more advantageous to use the measured and analyzed sea-surface temperatures rather than the climatology data. The measured and analyzed sea-surface temperatures are generally more representative of the actual field during the simulation. The use of measured sea-surface temperatures can be particularly important (e.g., in an El Niño year) when the actual temperatures deviate strongly from “normal” (e.g., climatology).

The snow data base at NCAR includes Northern Hemisphere analyses from NCEP. The snow analysis is updated about once a week, or following a significant snow event. If you are simulating a summer episode, you should be able to bypass the need for a snow analysis, and MM5 will assume that there is no snow cover in the simulation domain.