Avoid double accounting of N in senescence #8985
Conversation
Models/PMF/Organs/SorghumLeaf.cs
Outdated
| @@ -891,13 +891,13 @@ private void ApplySenescence() | |||
| if (!MathUtilities.IsPositive(dltSenescedBiomass)) return; | |||
|
|
|||
| double slnToday = MathUtilities.Divide(Live.N, laiToday, 0.0); | |||
| DltSenescedN += DltSenescedLai * Math.Max(slnToday, 0.0); | |||
| double dltSenescedN = DltSenescedLai * Math.Max(slnToday, 0.0); | |||
There was a problem hiding this comment.
From memory, DltSenescedN is used as a reporting/tracking variable. It can also be senesced when providing N to another organ and is therefore cumulative. I should be confident that we weren't double counting it as it mirrored the outputs from Classic that were much easier to follow - but it's been a while since I was in that code.
The outcome here shouldn't be an exception though - it should set the senescence to the live amount and then kill the plant - normally the reduction in LAI would catch this, but I'm guessing that the amounts that we're talking about are very small. There is an issue with how the LAI checks work due to frost removing LAI but not killing the plant completely. There's probably also a case to be made about going below the structural minimum %?
|
retest this please jenkins |
|
@jbrider @peter-devoil The stats have changed for this one - mostly improved. Are you happy for this to be merged? |
|
I'll work it through with jason. From what I can tell, the offender is being called twice/day when it should be once. |
|
@jbrider, I've had another look at this and suspect it is double accounting. The "double calling" isn't happening. It is set first in ProvideNThroughSenescence() which calculates the N lost in senesced material in differing ways pre/post flowering It's again incremented in ApplySenescence(), which is using the same DltSenescedLai calculated in ProvideNThroughSenescence() earlier I think the test, and second increment is redundant. |
|
undo |
working on #8917