Email:  Password:  
  Forgot Password?  Register Now
UCAR > NCAR > CISL
Search:  
  
FAQs | Contact Us | For Staff
 NCEP/NCAR Reanalysis > CRAY BINARYGo to Dataset: 
 
STILL UNDER CONSTRUCTION!

Several reanalysis file_types use this format. These file_types are:

File_Type: sanl, sges

This format is similar to the format in ds084.2 dataset. All values are the spherical harmonic representation, in 4-byte IBM words (must be converted with UNP360).
Record # Length (byte) Description
132 header
2848 fcst hr, init hr, mon, day, yr, sigma layer interfaces 1-29, sigma layer midlevel values 1-28 (4-byte words) 62,28,1,2,1,80 as the last four values
316128 orography
416128 ln(surface pressure)
516128 layer 1 virtual temperature
.
3216128 layer 28 virtual temperature
3316128 layer 1 divergence
3416128 layer 1 vorticity
3516128 layer 2 divergence
3616128 layer 2 vorticity
.
8716128 layer 28 divergence
8816128 layer 28 vorticity
8916128 layer 1 specific humidity
.
11616128 layer 28 specific humidity, sanl files end here
11772192 unknown, in sges files only

File_Type: sfcanl, bges

Here is an example of REANALYSIS sfcanl file type. I think I understand the first two records correctly. But I am not sure about the rest records (record 3 to 16). The bges type reanalysis files have the same variables as sfcanl. 
enter input COS-blocked file? 
sfcanl87072600
 record 1 fcst hour:   0.
 record 2 h,m,d,y:     0  7  26  87

                (192x94)
                 # of       grid      grid       grid       grid    ?variable?
                 grids  location   max value  location   min value
 GAUSS record: 3 18048 :   5893    318.911621   18048    206.802551   Temp
 GAUSS record: 4 36096 :   1020      0.550000   36096      0.384590
 GAUSS record: 5 18048 :  16478  55000.000000   18048  50000.148438
 GAUSS record: 6 36096 :   5786    313.072266   36096    207.820938
 GAUSS record: 7 18048 :   8736    301.599854   18048    207.819061
 GAUSS record: 8 18048 :   6196    270.000000   18048      2.000000
 GAUSS record: 9 18048 :  12259      0.787536   18048      0.000000
 GAUSS record:10 18048 :      1    100.000000   15878      5.000000
 GAUSS record:11 18048 :   7439     22.000000   18048      0.000000
 GAUSS record:12 18048 :      1      0.800000   18048      0.750000
 GAUSS record:13 18048 :      1      2.000000   18048      1.000000
 GAUSS record:14 18048 :   1020   1000.000000   18048    994.999756   sfc Pres
 GAUSS record:15 18048 :    732      1.999999   18048      1.773500
 GAUSS record:16 18048 :  12703      0.989515   18048      0.800689
END_OF_FILE during read, exit

The strange thing is that records 4 and 6 have twice 
the number of one full grids (192x94=18048).

According to the reanaysis paper, there are 16 variables in this file:
Number Description
1 earth surface temperature (K)
2 soil moisture level 1
3 soil moisture level 2
4 snow depth (m)
5 soil temperature level 1 (K)
6 soil temperature level 2 (K)
7 soil temperature level 3 (K)
8 surface roughness length (m)
9 convective cloud cover (%)
10 convective cloud bottom height (sigma)
11 convective cloud top height (sigma)
12 albedo (fraction)
13 snow/ice/land mask
14 minimum stomatal resistence
15 canopy water content
16 ratio of 10m and lowest sigma level winds (fraction)

File_type: znl.f00, znl.f06 files have size of 50456 bytes

Byte Description
1-32 header
33-56 date information
57- data of six zones (90S-60S, 60S-30S, 30S-30N, 30N-60N, 60N-90N, global average)