Npp

[ckcastillo@purdue_edu]

Hi! I would just like to confirm - is there a means by which NPP is altered after the DGVM modules are called (i.e. after DGVMEstablishmentMod.F90)? The calling sequence in both the manual and the code indicate that NPP is determined first, and does not seem to be re-calculated or scaled or normalized after all the DGVM modules and before being written out.

I ask just because I have this run branching off a 500-year spin-up in which modifications in DGVMEstablishment seem to cause a change in NPP in the very first time step compared to the control branch run.

 

[slevis]

I think you're right that the NPP should not change in the first time step of a branch run. You could try some tests to understand what's going on. E.g. you could write out the NPP from different parts of the code to find out when the value changes in the branch versus the control run. You could also try running the branch without your code modifications to see if your mods are responsible or not.

Debugging is a big part of our jobs as modelers, isn't it?

Sam

 

[ckcastillo@purdue_edu]

Thanks - I thought I may have just misunderstood the calling sequence! I've tried the last bit, which is to run the branch without the modifications and the NPP was fine so I know my mods are responsible in some way even if they come at the end of a year's run. I will try writing the NPP variable out at different stages of the DGVM, as you suggest.

As I understand it, photosynthesis is involved in NPP calculations, but curiously, the photosynthesis as plots by the land diagnostics package show both the control and modified branches correctly starting from the same point in the first year.

 

[ckcastillo@purdue_edu]

Hi, Dr. Levis,

I was able to solve the previous NPP problem - it was just a matter of matching this year's reported NPP with the previous year's pft distribution when summing up and plotting the NPP per pft.

I'm running the CLM with the fully-coupled mode now (CCSM3 B config) and something odd has occurred with the NPP. I have three runs branching from a spinup - control, deforestation A and deforestation B. The NPP in the first year of the B run correctly matches the NPP of the spinup a year after where the 3 runs branched off. However, the NPP of control and A do not match the spinup but match each other in the first year.

The 3 runs should all have the same NPP in the first year. I double-checked and all three runs branch from the same place and use the same input land surface files. The only difference was in the number of nodes I used, which should not affect the results. Would you have any idea how this could occur?

Thanks!

 

[slevis]

Sorry, I cannot think of a good reason...

 

[ckcastillo@purdue_edu]

Ok, thanks.

I actually repeated the control and A runs with the same load configuration and number of nodes as the spinup and this time the NPP matched...I did not think that the nodes could affect the smoothness of the branch/restart but I have not yet found anything else to explain.