CM1 version 1 release 14:  23 October 2009.

Summary of changes.

-------------------------------------------------------------
1.  New options/features.

 - When netcdf output format is requested (output_format = 2), ALL data
   files are written in netcdf format.  (Now including stats and parcel
   trajectory data files.)

 - The Thompson microphysics scheme (ptype=3) and the Morrison microphysics
   scheme (ptype=5) now both work with the energy-conserving equation set 
   (neweqts >= 1).  Both schemes still use original formulation for neweqts=0.

 - Added the ability to use a vertically implicit integration scheme for
   turubulent diffusion.  (See vimpl in README.namelist)  This scheme is
   absolutely stable, and is useful when vertical grid spacing is very
   small (dz < 100 m).

 - Added a new stat output:  radius of maximum azimuthal velocity for 
   axisymmetric model simulations.  To use, set stat_rmw in namelist.input.

 - New documentation is available:  "The Governing Equations for CM1"
   Available at the CM1 website:  http://www.mmm.ucar.edu/people/bryan/cm1/


-------------------------------------------------------------
2.  Changes, fixes, modifications, etc.

 - Made some minor changes and upgrades to the Thompson microphysics scheme
   (ptype=3).  The scheme is now identical to the version in WRFV3.1.1.  
   (thompson.F)

 - Made some minor fixes and upgrades to the Morrison microphysics scheme
   (ptype=5).  Reflectivity is no longer cut off at ~15 dBZ.  (morrison.F)

 - Made some small changes to the pressure equation to better account for 
   mass conservation.  Small affect on results unless simulations are very 
   long (> 10 days).  Thanks to Hamish Ramsey (Columbia Univ) for identifying 
   this problem.  (solve.F)

 - Fixed a minor bug to ensure that Rayleigh damping near horizontal
   boundaries (hrdamp=1) is applied even when upper-level Rayleigh damping 
   is not used (irdamp=0).  (solve.F)

 - Changed loop indices for calculation of some base-state soundings to 
   prevent floating-point error.  Thanks to Adam French (NCSU) for 
   identifying this problem.  (base.F)

 - Modified the parameterized turbulence scheme (iturb=3) so that it works
   with 3d simulations.  (Previously, iturb=3 only worked with axisymmetric
   simulations.)  (turbtke.F, solve.F)

 - Made some minor changes to correct mass/energy budgets.  Only affects 
   stats output.  Does not change results of simulations.  (multiple files)

 - Made a slight change in calculation of moist Brunt-Vaisala frequency to 
   prevent aggressive generation of subgrid turbulence in some ice-phase
   environments.  Thanks to Dan Kirshbaum (U. Reading) for identifying 
   this problem.  (turbtke.F)

 - Made a minor change to the upper and lower boundary conditions of the 
   weno advection scheme to prevent excessive dissipation near those 
   boundaries.  (weno.F)

 - Updated the netcdf writeout routine.  Removed obsolete netcdf commands.
   These changes should address problems that some users have reported.  
   (writeout_nc.F)

 - Made a small change to better conserve energy during fallout when 
   efall=1 is selected.  (A forthcoming article will explain this option
   in more detail.)  (kessler.F)

 - Made some relatively minor bug fixes to the parcel driver (iprcl=1).
   It should work properly now for axisymmetric simulations, and for
   simulations without moisture.  (parcel.F)

 - Checked that certain analytic soundings and 3d initial conditions work 
   properly in dry environments (i.e., when imoist=0).  (base.F, init3d.F)

 - Added compiler flags and settings in the Makefile for different kinds 
   of linux clusters (including Intel compiler and PathScale compiler).
   (Makefile)

 - Updated the skew-T/log-p plotting program for GrADS that is available
   on the CM1 webpage.  Among other changes, CAPE and CIN are now calculated 
   more accurately.  Thanks to Bob Hart (Florida State Univ) for updating 
   his skew-T script.  (plotskew.gs)