[MITgcm-cvs] MITgcm/pkg/gmredi CVS Commit

[Michael Bates,54-1517,6172535458,]

Update of /u/gcmpack/MITgcm/pkg/gmredi
In directory forge:/tmp/cvs-serv29711

Modified Files:
	GMREDI.h gmredi_calc_eigs.F gmredi_calc_tensor.F 
	gmredi_calc_urms.F gmredi_check.F gmredi_diagnostics_init.F 
	gmredi_init_varia.F gmredi_k3d.F gmredi_readparms.F 
Added Files:
	gmredi_read_pickup.F gmredi_write_pickup.F 
Log Message:
Changes associated with the PV eddy closure (GM_K3D) include:
o To improve efficiency the call to solve for the eigenvectors only happen once every GM_K3D_vecFreq seconds.  This required the following changes:
  - read and write pickup files for the eigenvectors and deformation radius (gmredi_read_pickup.F and gmredi_write_pickup.F)
  - making the number of modes (GM_K3D_NModes) a parameter which must be specified at compile time in GMREDI.h
  - A new namelist variable, GM_K3D_vecFreq
  - Added modesC, modesW, modesS and Rdef to the common block
o If the CPP option use_lapack is undefined, then a WKB approximation to the eigenvectors and deformation radius is now used (although, it seems unstable; so for the moment an error is raised in gmredi_check if GM_K3D is defined but use_lapack is not).
o Changed gmredi_calc_eigs returns the deformation radius rather than the deformation wavenumber
o Fixed bug in calculation of tfluxX and tfluxY for the instance where the surface layer is the depth of the water column.
o Added warning messages if there are problems with calculating eigenmodes and eigenvectors
o Cleaned up code
o Improved documentation
o Rationalised diagnostics
o Added some extra startup checks (gmredi_check)