Remove IS_FCST and clean up associated logic in replay TYPE

[sanAkel]

Testing by @rtodling has found that these lines of code:

GEOSgcm_GridComp/GEOSagcm_GridComp/GEOS_AgcmGridComp.F90

Lines 1730 to 1734 in 8faac49

	if (ISFCST/=0 ) then
	TYPE = FORECAST
	else if(.not. DasMode) then
	TYPE = FREERUN
	else

are unnecessary, cause confusion and (potential) bugs. @atrayano is also aware of these lines of code.

@rtodling please add your notes as needed.

This IS_FCST must be removed also the above mentioned if block cleaned up.

IF
...
ENDIF

[tclune]

@sanAkel This is "assigned" to me. I'm assuming that this is just meant to keep me informed rather than actually having me implement the fix?

[sanAkel]

@tclune, perhaps you would be able to (re-) assign someone to it?

[mathomp4]

I mean, I could do the coding if I knew what is the "right" way to fix it. Could you and/or @rtodling let me know?

[rtodling]

Guys, before we change anything I suggest we do a little investigation on who is (if anyone) using this flag set to 1. If nobody is using the flag the simplest thing to do is to wipe out the implementation associated w/ the option 1.

I also strongly suggest talking to Larry Takacs, He might recall what the motivation might have been for distinguishing between a forecast and a "freerun" ... I sincerely cannot see the difference from an intellectual point of view, but clearly in the code lots of things happen that differ the two cases ... so not sure ... I suggest being careful w/ this.

I'll also add that there are other fish to fry w/ higher priority so I would leave this here as a note for us to remember, but would not spend too much time on this.

[atrayano]

Ricardo, I agree that talking Larry Takacs is the right thing to do. I am more or less aware of the initial motivation: this has to do with the so-called bias correction done in Agcm. But again, the whole bias correction was Larry's idea, and the implementation is done also by him.

[sanAkel]

Ricardo, I agree that talking Larry Takacs is the right thing to do. I am more or less aware of the initial motivation: this has to do with the so-called bias correction done in Agcm. But again, the whole bias correction was Larry's idea, and the implementation is done also by him.

Thanks @mathomp4, @rtodling, @atrayano

Sometime - not "soon", I plan to have a chat with @lltakacs
Reg replay. Will add this to my list of questions.