NCAR-CSM History global attribute.\\012 200000.0400000000 $Header: /cvsroot/nco/nco/data/in.cdl,v 1.55 2004/09/07 19:50:05 zender Exp $ time Monotonically increasing coordinate pressure hybrid_sigma_pressure down hyam hybm P0 PS ilev Monotonically decreasing coordinate pressure Homebrew level coordinate for LSM Soil depth meter Wavelength meter hybrid A coefficient at layer midpoints hybrid B coefficient at layer midpoints reference pressure pascal surface pressure pascal Character coordinate [chr] 3 base date as 6 or 8 digit integer (YYMMDD or YYYYMMDD) current date as 6 or 8 digit integer (YYMMDD or YYYYMMDD) Gaussian weights Gaussian weights which sum to two for n = 4. These weights are all have floor of 0.0 so should cause SIGFPE when applied to integer types in weighted average. partial mask, partial missing value example 9.99999962e+35 partial missing value example 9.99999962e+35 scalar missing value 9.99999962e+35 offset partial missing value example -999.000000 _FillValue example -999.000000 no missing value one regular value, one missing value 9.99999962e+35 Packed scalar variable Original packed value was 1s with scale_factor = 2.0d and add_offset = 1.0d. Unpacked value should be 3.0 = 2.0d*1s + 1.0d. NCO algorithms would pack this variable as scale_factor = 0.0d and add_offset = 3.0d. 2.000000000000000 1.000000000000000 Packed array Packed value is -32767s, 0s, 1s, 32767s, unpacked is same in double 1.000000000000000 0.000000000000000 Unpacked scalar variable Unpacked value is 3.0d0, upk=unpack(pck)= 2.0d0*1s + 1.0d0 = 3.0d0. Packing this variable should create an NC_SHORT scalar = 0s with packing attribute add_offset=3.0d and either no scale_factor (ncap) or scale_factor = 0.0d (ncpdq). Unpacked array Unpacked value is -32767.d, 0.d, 1.d, 32767.d, packed is same in short. Packing algorithm should yield an NC_SHORT array = [] with packing attributes scale_factor=1.0d, add_offset=0.0d scalar equal to 1 -99 1, 1 -99 17000, 17000 -99 1, mss_val -99 Scalar with value 0.5 9.99999962e+35 0.5,0.5 9.99999962e+35 all values are one all missing values example 9.99999962e+35 scalar variable Fraction float double double 1.000000000000000e+36 int short char 1 Character variable with whitespace on ends 1 Character variable containing one NUL 1 Character variable containing multiple NULs 3 Variable contains a file name 80 Variable that contains a short array of file names [sng] 80 Variable contains a one-dimensional array of characters that is not NUL-terminated [chr] 4 Variable contains a two-dimensional array of characters that are not NUL-terminated [chr] 4 byte zero one two e, base of the natural logarithm e, base of the natural logarithm three four negative one lev_var lev_wgt g Height meter Height stored with a monotonically increasing coordinate Height meter Height stored with a monotonically decreasing coordinate 0 Sentence one.\\012Sentence two.\\012 37 73 73.0000000 72.0000000 71.0000000 73.00000000000000 0 Sentence one.\\012Sentence two.\\012 37 73 73.0000000 73.00000000000000 three dimensional variable with CCM coordinate convention C=[lat,lev,lon], Fortran=(lon,lev,lat) Fraction three dimensional variable with COORDS coordinate convention C=[lev,lat,lon], Fortran=(lon,lat,lev) Fraction Surface pressure pascal two dimensional variable Fraction mask variable like ORO Fraction Orography Fraction Gaussian weight Fraction gw variable like gw Fraction gw variable like gw_T42 Fraction record variable, float record variable, double one dimensional record variable second Temperature kelvin Leave mispelled for NCO User's guide example record variable, double, with double missing values This variable is used to generate the packed variable rec_var_dbl_mss_val_dbl_pck, so its missing_value should not be too far out of range, i.e., it should be representable by a short. However, the missing_value should itself be the same type as the unpacked variable, NC_DOUBLE in this case. -999.0000000000000 record variable, double, with double missing values, packed Packed version of rec_var_dbl_mss_val_dbl_upk -999.0000000000000 -9.155413138017850e-05 5.000000000000000 record variable, float, with float missing values 9.99999962e+35 record variable, float, packed Demonstrate that rounding of means of packed data are handled correctly 0.100000001 100.000000 record variable, double, packed Demonstrate that rounding of means of packed data are handled correctly 0.1000000000000000 100.0000000000000 record variable, float, with double missing values 1.000000000000000e+36 record variable, float, with integer missing values -999 record variable, integer, with integer missing values -999 record variable, integer, with float missing values -999.000000 record variable, integer, with double missing values -999.0000000000000 record variable, double packed as long, with double missing values although not usual, packing doubles into longs (rather than shorts) is still a considerable cost savings -999.0000000000000 -9.155413138017850e-05 5.000000000000000 one dimensional record variable of string 10 Record variable of longitude coordinate two dimensional record variable of string 3 two dimensional record variable watt meter-2 three dimensional record variable watt meter-2 three dimensional record variable of type double watt meter-2 -99.00000000000000 three dimensional record variable of type int watt meter-2 -99 three dimensional record variable watt meter-2 -99 four dimensional record variable watt meter-2 hours since 1900-01-01 00:00:0.0 0000-00-00 06:00:0.0 Zonal wind speed meter second-1 Meridional wind speed meter second-1