These files are the source code of CLIGEN V-4.2, recoded to conform to the
"Water Erosion Prediction Project (WEPP) Fortran-77 Coding Convention".
The work was performed by Charles R. Meyer, NSERL-USDA-ARS, from August to
November 1999, in order to facilitate the maintenance, modification, and
extension of the CLIGEN code.  The extensive help of David Hall, USFS, Moscow,
Idaho in providing many valuable definitions is gratefully acknowledged.
C. R. Meyer may be contacted at: 1196 Bldg. SOIL, Purdue University, West
Lafayette, IN 47907; ph: 765/494-8695; e-mail: meyerc@ecn.purdue.edu.

  Approach:
Code complexity was reduced by breaking sections of the code out into eight
additional modules.  Common blocks were removed into "include" files, resulting
in a single copy of each.  Unused variables were removed -- 66 from the common
blocks alone.  The extremely unstructured condition of original code was
greatly improved, largely through the addition of two varibles, MOVETO a
global variable, and NDLOOP a local variable -- and though careful selection
of code to be isolated as a module, based on its functional structure.  This
permitted removal of a large number of GOTO's that made spaghetti of the code.
Where a GOTO seemed necessary, appropriate choice of module boundaries and
use of the MOVETO variable was substituted.

  Results:
The software metrics supplied by the McCabe software analysis tool indicate
the number of unstructured constructs declined from 46 to 2.  (The 2 are in
the main module.)  The complexity of the most complex module declined from
89 linearly independent paths, to 13.

  Additional Note:
Code that seemed wrong was commented with the string "XXX".  Meyer is not
a climate/weather expert, so it is not obvious to him what should be done
-- only that the code appears incorrect.  The user is left to his/her own
interpretation.

  Between the 11/15/99 and 12/15/99 posting:
Some explanatory comments were added to the section where the daily climate
outputs are generated.  The logic remains the same.

  What's New:
An error was discovered in function DSTG.  It generates an arithmetic
underflow in long runs.  To correct the problem, add the following line
to DSTG in your existing code:
      double precision xx,fu


----Charles. R. ("Chuck") Meyer, 1/06/2000.