Negative RADSH

[palderman]

@chporter and @kthorp ,

I'm getting floating point errors with the forage model related to the fact that RADSH is negative due to ADDF being negative in CANABS() in ETPHOT.for. When I uncomment the check for ADDF < 0.0 (see below), the problem goes away. The issue seems to be limited to situations with very low LAI.

Is there a reason why the checks for negative numbers are commented out in CANABS() in ETPHOT.for? Is there any problem with allowing ADDF, ADDFSL, etc to be set to zero when the original calculation is negative? See here:

dssat-csm-os/SPAM/ETPHOT.for

Lines 1793 to 1800 in a20a2b5

	C-KRT****************************
	C-KRT ADDF = ADIR - ADDR
	!-CHP ADDF = MAX(0.0,ADIR-ADDR)
	ADDF = ADIR - ADDR
	! IF (ADDF < 0.0) THEN
	! ADDF = 0.0
	! ENDIF
	C-KRT****************************

dssat-csm-os/SPAM/ETPHOT.for

Lines 1814 to 1827 in a20a2b5

	C-KRT************************************
	C-KRT ADDFSL = ADIRSL - ADDRSL
	C-KRT ADDFSH = ADDF - ADDFSL
	!-CHP ADDFSL = MAX(0.0,ADIRSL - ADDRSL)
	!-CHP ADDFSH = MAX(0.0,ADDF - ADDFSL)
	ADDFSL = ADIRSL - ADDRSL
	! IF (ADDFSL < 0.0) THEN
	! ADDFSL = 0.0
	! ENDIF
	ADDFSH = ADDF - ADDFSL
	! IF (ADDFSH < 0.0) THEN
	! ADDFSH = 0.0
	! ENDIF
	C-KRT************************************

dssat-csm-os/SPAM/ETPHOT.for

Lines 1862 to 1869 in a20a2b5

	C-KRT********************************
	C-KRT ADIFSH = ADIF - ADIFSL
	!-CHP ADIFSH = MAX(0.0,ADIF - ADIFSL)
	ADIFSH = ADIF - ADIFSL
	! IF (ADIFSH < 0.0) THEN
	! ADIFSH = 0.0
	! ENDIF
	C-KRT********************************

dssat-csm-os/SPAM/ETPHOT.for

Lines 1883 to 1890 in a20a2b5

	C-KRT********************************
	C-KRT AREFSL = AREF - AREFSH
	!-CHP AREFSL = MAX(0.0,AREF - AREFSH)
	AREFSL = AREF - AREFSH
	! IF (AREFSL < 0.0) THEN
	! AREFSL = 0.0
	! ENDIF
	C-KRT********************************

[fabiooliveira72]

Hi Dr. Alderman!

Yes, indeed. We have this problem right now. We had some problems with Canola experiments from Dr. White are having the same issue.

We are about the make a pull request to fix those issues. We need to check with a final testing if did not broke anything else.

Thank you so much,

Fabio Oliveira

[fabiooliveira72]

This issue was solved by merge Protections for Negative LAI and radiation absorbed by sunlit and shaded leaves

https://github.com/DSSAT/dssat-csm-os/tree/d0c9faa89c7dc3b276197e3ea7cdca64d41bdc5b