[Extract from /l_data/model_parameters/SOIL_DATA/README]

%================================================================================================

- HWSDv1.21-STATSGO2/
oc_sub_30sec.dat
oc_top_30sec.dat
clay_sub_30sec.dat
clay_top_30sec.dat
sand_sub_30sec.dat
sand_top_30sec.dat

HWSDv1.21 and STATSGO2 are merged as follows:
1) - use STATSGO2 if either (or both) its top- or subsoil information is available
   - if only one of both layers in STATSGO2 is available:
     keep that information, set other layer to NaN
2) - obtain consistent vertical profiles:
     extrapolate top(sub)soil information to sub(top)soil information
     where only one of both layers is given in the original data
3) - fill holes in space:
     areas where HWSDv1.21 nor STATSGO2 have any data are filled
     with neighbouring information
Note that this merged dataset covers the land area determined by GMAO's CLSM landmask.

#FORMAT - fortran binary files:
#-----------------

- little-endian
- global, 30'' resolution
- N columns  = 43200
- N rows     = 21600

- first row = south pole; last row = north pole
- first column = west; last column = east

- western and southern edges of the data array are along the dateline and the south pole:
center of LL corner pixel is at (-180. + 1/240, -90 + 1/240)

- Fortan read example:

integer, parameter :: nc = 43200, nr = 21600
real :: data(nc,nr)
do j =1,nr                         ! Reading rows from the south pole to the north pole
    read (10) (data (i), i = 1,nc) ! reading columns from the dateline to the dateline
end do

%=========================================EOF====================================================