3-05-2009
	   Fixed bugs in CAM3 radiation:
            - do not allow fice to be 0 
            - output TOA SW is fsnt, not fsntoa
	    CliMT now replicates model output fsnt and flnt to within
	   ~0.03 W/m2 (remanining error to do with spatially-varying
	   stratospheric distribution of trace gases)

29-04-2009
	   - Added new input to CAM3 radiation: surface LW upwelling
	      radiation (FLUS)
	   - Corrected bugs in CAM3 radiation: 
	       - fice now correctly computed when ciwp input externally
	       - pstd set to 1000 mb instead of ps

16-04-2009 
	   - CCM3 and Chou schemes now accept explicit external
	      definition of cloud ice water path ciwp 
	   - CAM3, CCM3 and Chou schemes now expect drop radii r_liq
	      and r_ice as 3D fields, not 0D parameters as before
	   
15-04-2009 
	   - Added SwToa, LwToa, SwSrf and LwSrf outputs to CCM3 and Chou schemes
	   - swflx and lwflx outputs are now at level MIDPOINTS instead of
	      interfaces as before 
	   - Added check in cam3/src/radae.F90:initialize_radbuffer to see if buffers already
	      allocated (otherwise CliMT will crash if CAM3 radiation instantiated
	      more than once)
	   - Cloud ice fraction ciwp can now be passed explicitly to CAM3 
              (before, it was calculated internally as in CCM3; this
	      is still done if ciwp not passed explicitly)

09-04-2009 ++ Version 0.6.5
            - Added TOA and surface flux output  from CAM3 radiation
 	    - Corrected bug in CAM3 LW cloud forcing (sign was wrong!)
	    - Added check for negative or zero humidity and ozone
               input to CAM3, which give NaNs
	    - Disabled inf/nan flags in CAM3 code for portability
	    - Updated examples; all now use CAM3

07-04-2009 
	   Removed useless 'from numpy import *' in __init__.py which was
           causing namespace collision problems (__version__ is now correctly reported)

12-02-2009 ++ Version 0.6.4
	   Added CAM3 radiation scheme

05-02-2009 
	   setup.py reads version number from file 'Version'

13-03-2008 
	   io.py now compatible with both Scientific.IO and PyNGL.Nio
	   As of version 1.2, PyNGL doesn't work very well for NetCDF output

11-03-2008 
	   Small numpy compatibility fix in grid.py -- check that axis
	   is array, not scalar   

2-02-2008 ++ Version 0.6.0
       	  - Updated build system to use python distutils
          - CliMT now works *only* with numpy

10-01-2007 
	   Added functions thermodyn.CAPE and thermodyn.CINE 
	   to compute convective available potential energy and 
	   convective inhibition energy

27-11-2006 
	   Added function get_nlev(), get_nlat(), get_nlon() to
	   __init__py to retrieve compiled number of levels,
	   latitudes, longitudes

25-10-2006   
	     Added keywords to thermodyn functions qs, ws etc.
	     Corrected bug in qs: specific humidity is 
	     q = eps*e/(p+(1-eps)*e)  
	     and not
	     q = eps*e/p 
	     as before

24-10-2006   
	     Added function to make skew-T lnp plot.
 
19-10-2006  
	    Changed function to compute thetae to use definition in Bohren+Albrecht

18-10-2006  
	    Added function  pseudoadiab() to thermodyn module that
	    computes both pseudoadiabats by integrating the
	    pseudoadiabatic lapse rate equation

18-10-2006  
	    Added function z() to thermodyn module that computes height
	    given temp and press profiles, assuming hydrostatic 
            equilibrium and 

23-09-2006  
	    Bug fix: calday now written to ouput file each time
	    write() is called

15-09-2006  
	    Added CliMT version, run date and restart file name to 
            global attributes of output netcdf file

12-09-2206  
	    Bug fix: orbital parameters in insolation module are now
            correctly reset to values specified in kwargs when insolation module
            is federated 

12-09-2006  
	    Added check for Darwin version in CompilerFlags, needed to
            get the right link flags in xlf configure file 

08-08-2006  
	    Added flags for compilation using Intel ifort on Mac OS X

28-06-2006  
	    Added simplified Betts-Miller convection scheme 
	    (Dargan Frierson/Jonathan Mitchell)

28-06-2006  
	    Improvements to Monitor -- now printing max and min
	    values of 2D arrays, and using interpolation='nearest'