User's Guide to NCAR CCM3.6 Search Page

3. CCM3.6 Internals

    3.5 Output Datasets

        3.5.1 History Files

3.5.1.2 History File Header Records

• It provides the information required by postprocessors to read and locate specific data fields.
• It serves an archival function, storing information such as date/time the input datasets used etcetera.

Contents of the three header records are shown below in Table 3.1  (integer), Table 3.2 (character), and Table 3.3 (real). Each variable is listed by name and word number within each header record, along with a short description and a typical value based on the T42, 18 vertical-layer case. Output fields in a CCM3 history file are located on the hybrid vertical coordinate grid. Thus, three arrays of hybrid coefficients are saved in the real header record along with the Gaussian latitudes and weights (see Table 3.3).

There are some differences in the format depending on what machine wrote the files out. On IEEE machines integer words are 4-bytes long, while floating point words are 8-bytes long. On Cray machines both integer and floating point words are 8-bytes long. Furthermore, numbers are written in native machine format. For example, on the Cray, floating point numbers are written in Cray floating point format, while on a Sun workstation floating point numbers will be written in IEEE floating point format. Also, some machines may have words that are "byte-swapped" (a different Endian representation) from other machines. FORTRAN adds "control-words" to the begining and end of each FORTRAN record. The control-words are a full integer word and represent the number of bytes written out on each record. A FORTRAN program that reads the file doesn't need to worry about these words, but a program written in another language (such as "C") will need to process them. The Operating system may add blocking control words as well (ie. the Cray adds COS-blocking control words). The offline program "ccm2nc" deals with most of these issues (other than byte-swapping) and as such is generally able to read CCM files produced on a non-native machine platform.

Table 3.1: History File Header Record 1-- Integer Variables
(Note: On IEEE machines integer words are 4-bytes long, on Cray machines integer words are 8-bytes long)

Variable Name	Word Number	Description	T42 Value
LENHDI	1	Length of header record 1 (in integer words, 4-bytes for IEEE machines and 8-bytes for Cray machines)	292
MFTYP	2	Format code for this history file	43
MFILH	3	Number of this "time sample" within this disk file or Mass Store tape.	1
MFILTH	4	Maximum number of time samples in this history file.	20
NRBD	5	Number of records before data records	3
MAXSIZ	6	Length of packed data record in words	33662
NDAVU	7	Length of the data record in words after unpacking	65282
	8	Unused	0
NLON	9	Number of longitude points in one latitude line	128
NLONW	10	Number of longitude data values written on the history file	128
NOREC	11	Number of latitude lines and number of data records	64
NLEV	12	Number of vertical levels written on the history file	18
NTRM	13	M spectral truncation parameter	42
NTRN	14	N spectral truncation parameter	42
NTRK	15	K spectral truncation parameter	42
NFLDH	16	Number of fields in this time sample	85
NSTEPH	17	Timestep on which this data was written	0
NSTPRH	18	Iteration number for the start of this run	0
NITSLF	19	Number of timesteps since last time sample was written	0
NDBASE	20	Base day number for this case	0
NSBASE	21	Base number of seconds for this case	0
NDCUR	22	Current day number corresponding to NSTEPH	0
NSCUR	23	Seconds of the current day NDCUR	0
NBDATE	24	Base date (yr mo day) as 6-digit integer	000901
NBSEC	25	Seconds to complete NBDATE date.  (Note: may be different from NSBASE.)	0
NCDATE	26	Current date (yr mo day) as 6-digit integer corresponding to NSTEPH	000901
NCSEC	27	Current seconds for date NCDATE	0
MDT	28	Model timestep in seconds	1200
MHISF	29	Frequency in timesteps at which time samples are written	72
MFSTRT	30	Flag to indicate type of run:  = 0 for initial run  = 1 for restart  = 3 for branch run	0
LENHDC	31	Length of header record 2	259
LENHDR	32	Length of header record 3	239
MPSIG	33	Position in header record 3 of first word of sigma value list.  sigapb(1) = REALHD(MPSIG)	1
MPLAT	34	Position in header record 3 of list of latitude lines.  hdlat(1) = REALHD(MPlat)	112
MPWTS	35	Position in header record 3 of list of Gaussian weights.  hdwt(1) = REALHD(MPwts)	176
MPFLDS	36	Position of header field information list in header record 1  (integer values)	38
MPCFLD	37	Position of field information list in header record 2 (character values), given as a number of 8-character items from the start of record 2	90
MFLDS	38	Integer field information, dimensioned (3, NFLDH). See description of MFLDS array in Table 3.4 .	--

 

Table 3.2: History File Header Record 2 -- Character Variables*

Variable Name	Word Number	Format	Description	Example T42 Value
MCASE	1	C*8	Case identifier	clsst01
MCSTIT	2-11	C*80	Case title	Test run
Current History File
LNHSTC	12-21	C*80	Pathname of this history file	/USER/csm/clsst01/ccm3/hist/h0001
LDHSTC	22	C*8	Date this header record was written (MM/DD/YY)	04/27/98
LTHSTC	23	C*8	Time this header record was written (HH:MM:SS)	00:28:25
LSHSTC	24	C*8	Unused
First History File of Case
LNHSTF	25-34	C*80	Pathname of first history file for this case	/USER/csm/clsst01/ccm3/hist/h0001
LDHSTF	35	C*8	Date this case started  (MM/DD/YY)	04/27/98
LTHSTF	36	C*8	Time this case started (HH:MM:SS)	00:28:22
LSHSTF	37	C*8	Unused
Initial Dataset File
LNHSTI	38-47	C*80	Pathname of initial dataset file	SEP1.T42.0198.nc
LDHSTI	48	C*8	Date initial dataset was created  (MM/DD/YYun)	02/06/98
LTHSTI	49	C*8	Time initial dataset created (HH:MM:SS)	06:26:35
LSHSTI	50	C*8	Unused
--	51-63	--	Words 51-63 are unused.
Time-variant SST Dataset File
LNHSTVS	64-73	C*80	Pathname of SST boundary dataset	STR-sst.T42.0596.nc
LDHSTVS	74	C*8	Date SST boundary dataset was created (MM/DD/YY)	05/17/97
LTHSTVS	75	C*8	Time SST boundary dataset was created (HH:MM:SS)	17:03:13
LSHSTVS	76	C*8	Unused
Ozone Dataset
LNHSTVO	77-86	C*80	Pathname of ozone boundary dataset	ozn.0596.r8.nc
LDHSTVO	87	C*8	Date ozone boundary dataset was created  (MM/DD/YY)	09/21/92
LTHSTVO09/21/92	88	C*8	Time ozone boundary dataset was created (HH:MM:SS)	13:34:26
LSHSTVO	89	C*8	Unused
MCFLDS	90	C*8	Character field information, dimensioned (2,NFLDH). See description of MCFLDS array in Table 3.4 .	--
*All character strings are left-justified, blank-filled, and multiples of 8 characters.

Table 3.3: History File Header Record 3 -- Real Variables
(Note: These are 8-byte native floating point numbers)

Variable Name*	Word Number	Description
REALHD(MPSIG)		Contains 3 arrays of (2 * NLEV +1) hybrid coefficients as follows:  sigapb = hybrid A+B coefficients  siga = hybrid A (pressure) coefficients  sigb = hybrid B (sigma) coefficients
REALHD(MPLAT)		Points to the first word, relative to the start of the header, of a list of NOREC Gaussian latitudes in degrees (south to north)
REALHD(MPWTS)		Points to the first word of a list of NOREC Gaussian weights corresponding to the Gaussian latitudes
*MPSIG, MPLAT, and MPWTS are from header record 1, words 33, 34, and 35, respectively.

Two arrays mflds (starting at word 38 of the integer header) and mcflds (starting at word 90 of the character header) contain integer and character information describing each field stored in a latitude data record. Table 3.4 shows details of the contents of these two arrays. Also see figure Figure 3.8 for more details.

Table 3.4: Description of Field Information List Arrays, MFLDS and MCFLDS

Variable Name	Description
MFLDS(1,IFld)	Flag for number of levels and field flag, packed in one integer word. The ones digit is for the level number flag; the tens digit states whether the field is an instantaneous value, a value averaged between history file writes, a time minimum between writes, or a time maximum between writes.  ones digit of MFLDS(1,ifld)   0 = single-level field,   1 = multilevel field at layer interfaces,   2 = multilevel field at layer midpoints.  tens digit:   0 = instantaneous field,   1 = time-averaged field.   2 = minimum field value since the last time sample.   3 = maximum field value since the last time sample.
MFLDS(2,IFld)	Pointer to the first field value in packed data record (in terms of 8-byte words, see Figure 3.8).
MFLDS(3,IFld)	Data-packing flag (set to 1 if not packed).
MCFLDS(1,IFld)	Field name for field number IFld. (8 characters)
MCFLDS(2,IFld)	Units of this field (SI units). See Table 4.1 . (8 characters)
The index ifld goes from 1 to nfldh (the number of fields in the history buffer).

  

 Search page

Questions on these pages can be sent to... erik@ucar.edu .