| North American Regional Climate Change Assessment Program |
Home
Data
Output Archive
| Output Archive | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Updated 2010/4/30 to reflect changes in filename conventions and 3-D data organization. General Information Output will be reported in NetCDF using CF conventions. Variable names and units should be consistent with IPCC. Please consult with the NARCCAP group if inconsistencies are found. More on CF conventions can be found at http://badc.nerc.ac.uk/help/formats/netcdf/index_cf.html. Output will be on the model’s native horizontal grid. A supplemental archive is planned to provide some variables on a common grid. File name File names should follow this structure: Where: "VariableName" is the IPCC/CF-convention Variable Name in the tables below. "ModelName" is a 4-character, all-caps identifier chosen by the modeling group (please clear it with the archival group beforehand to avoid conflicts) "DriverName" is either "ncep" or the name of the driving GCM, in all-lowercase. (For the timeslice experiments, the driver is "slice".) "Time" is the starting time of the file, in the form "YearMonthDayUTC", e.g., 1979010100 for 00 UTC on 1 Jan 1979. File sizes (a) two-dimensional fields: files covering 5 years (half a decade). For consistency with climatological averaging standards, all 2-D files should end on years ending in zero or five (e.g., 1981-1985, 1986-1990, etc.). There may be some odd file sizes at the beginning and end (e.g., 1979-1980), but there is merit in having output from all models cover the same set of periods in their files. The ending time in a file should be 00 UTC on 1 January, since fluxes are for the 3-hour period preceding that time. For instantaneous fields, the ending time is less important, but for consistency should be the same time. (b) three-dimensional fields: Note that the definition of a month may vary between GCMs and may also differ from the real calendar. The "month" should follow usage in the source of the boundary conditions. The ending time in a file should be 00 UTC on day 1 of the following month (i.e., at midnight UTC separating one month from the next). Three-dimensional fields are separated by pressure level into 2-D files structured as above. Following CF convention, the pressure level is recorded as a scalar coordinate variable. The pressure level is included in the filename, prefaced by a 'p', between the driver and the time. For example: ta_RCM3_ncep_p750_1981010103.nc is the file containing temperature data at the 750 hPa level for the ncep-driven run of RegCM3 starting Jan 1st 1981. Further details regarding NetCDF conventions to follow, especially for metadata, appear in Requirements for Standard Output Contributed to the NARCCAP Archive. Two-Dimensional Fields Table 1: Daily fields The "day" for minimum and maximum temperatures should be from 06 UTC - 06 UTC, with the date stamp referring to the beginning time. This follows NARCCAP Exp. 0 practice. Thus, for 06 UTC 30 June 1979 - 06 UTC 1 July 1979, the date of the record is 30 June 1979, the calendar day corresponding to the period recorded for North America. The min/max should be recorded from instantaneous “screen height” temperatures sampled every time step during the course of the day. The “screen height” or “reference height” should be in the range 1.5 – 2 m above the surface.
Table 2: Primary three-hourly surface fields needed for impacts assessment Each day, values reported at 03 UTC, 06 UTC, … 24 UTC. Solar radiation is averaged over the 3-hr period prior to the reporting time. Precipitation is the 3-hour average prior to the reporting time. Other quantities are instantaneous values as noted. Winds should be the true west-east and south-north winds (i.e., rotated from the model’s grid as needed to the earth’s longitudinal and latitudinal directions). The surface air temperature and specific humidity should be recorded from “screen height” values. The “screen height” or “reference height” should be in the range 1.5 – 2 m above the surface. Surface winds should be recorded from “anemometer level” values. The “anemometer level” should be 10 meters above the surface. Note units for precipitation and surface pressure (MKS – following CF standards).
Table 3: Additional three-hourly fields, guided by GCM output archived for IPCC Fourth Assessment Report See https://esg.llnl.gov:8443/about/ipccTables.do for more on IPCC Fourth Assessment Report. For each day, values reported at 03 UTC, 06 UTC, … 24 UTC, on the model’s native grid. Quantities that can be used for storage terms in budgets are instantaneous values. Quantities that are fluxes in budgets should be average values for the 3-hr period prior to the reporting time. Snow depth is snow-water equivalent and is the average depth over the entire grid box, even if the model allows snow to cover a fraction of the grid box.
Table 4: Fixed surface fields For land-cover type, modelers will supply a vegetation index for each grid point along with a table linking each index value to a land-cover description. Models using a mosaic of multiple surface types in a grid box should supply sufficient information to describe the mosaic. On the models’ native grids.
Three-Dimensional Fields Table 5: Three-dimensional atmospheric fields Reported every three hours (03 UTC, 06 UTC, 09 UTC, 12 UTC, 15 UTC, 18 UTC, 21 UTC, 24 UTC). Instantaneous values every 25 hPa from 1050 hPa to Note that according to the CF conventions, the ice cloud water and the liquid cloud water are each reported as the ratio (cloud-water mass)/(mass of gas+condensed constituents), i.e., more like specific humidity than mixing ratio.
For more information on the output archive, contact William Gutowski, Jr. (gutowski@iastate.edu) or Seth McGinnis (mcginnis@ucar.edu). |