lnd_import ERROR: Longwave down sent from the atmosphere model is negative or zero

[BoHuang]

Dear all,
recently, I am working on CAM6 with variable-resolution simulation. I want to run the simulation with FHIST_BGC component set. I build the model and can successfully complete the the simulation for 5 days (STOP_N=5). However, the model always crash when I change the STOP_N to a longer time. The error occurs at time-step 1241, as "lnd_import ERROR: Longwave down sent from the atmosphere model is negative or zero". I have no idea how this error comes? Is there any one have idea how to fix it? Enclosed the log files. Thank you.

Greetings
Bo

 

[oleson]

You might try posting in the CAM Forum since the problem seems likely to be with the atmosphere. I personally don't have any ideas from the land side.

 

[BoHuang]

Hi Oleson,

thank you for your suggestion. I can pass the error by changing the nsplit.

However, I got another error in nstep=1256

-----
Warning: incorrect ice nucleation number (nuci reset =0)
15974.8806117785 211.094338257501 0.573135485888197
3.272696509873641E-002 0.000000000000000E+000 0.000000000000000E+000
0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000
1.524968712475104E-028 32.6903636583803
# of NaNs = 1
Which are NaNs = F F F F F F T F F F F F F F F F F F F F F F F F F F F F F F F
F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F
Sl_t
gridcell index = 17636
ENDRUN:
ERROR:
lnd_export ERROR: One or more of the output from CLM to the coupler are NaN
----

enclosed the log files. How can I track or fix this error? Thank you very much.

Greetings,
Bo

 

[yajun si]

Hi Bo Huang,
Sorry to bother you, BUT I met the same problem. How did you solve it

Thank you very much.

Greetings,
Si

 

[dannymleung@ucla_edu]

Hi all,
I have also seen similar messages in CAM6/CLM5 recently. I saw in my runs the error 1) emitted LW radiation being -ve or zero, and I also saw the warning 2) incorrect ice nucleation number (nuci reset =0). 1 is from CLM (lnd_import_export.F90), and 2 is from CAM (nucleate_ice.F90). Both messages usually happen together and 2 sometimes leads to 1, but 2 also appears alone without triggering 1. I am not familiar with the CAM ice nucleation module, but my guess is that there are not an adequate amount of aerosols in the atmosphere that act as ice nuclei, which caused the radiation calculation to misbehave and output negative values, and then CLM receives negative radiation and call the endrun function. It's usually coarse aerosol particles such as dust and sea salt that are responsible for ice nucleation. I was working on CLM dust emission parameterization and both 1 and 2 appear a lot, and I think it means in my CAM dust particles were too few/too many. I usually scale up/down my dust emissions in CLM by a factor of 2-10 to make sure CAM has enough dust, and that will usually solve the problem (or delay the time the error occurs again). One needs to guarantee that every gridcell around the globe has an adequate number of dust/sea salt aerosols, not too few and not too many. It seems that the ice nucleation scheme is very sensitive to the number of ice nuclei, and CESM seems heavily tuned to maintain stability, and only a narrow range of dust emissions can avoid the model from stopping the run. I am not sure if you are working on dust or sea salt, or you are just running into this error without changing the aerosol scheme, but my thought is that you make sure the dust or sea salt levels are adequate by scaling them up/down. You may need to try a few times to get a good scaling factor that avoids the error.
Another trick I use is that I manually set the LW radiation to be non-negative in CLM whenever the LW radiation is -ve, to avoid trigger CLM's endrun function. But this is not a good practice and is better not used.
Danny

 

[balaji]

Hi Bo Huang,
Sorry to bother you, BUT I met the same problem. How did you solve it

Thank you very much.

Greetings,
Si

Hi Oleson,

thank you for your suggestion. I can pass the error by changing the nsplit.

However, I got another error in nstep=1256

-----
Warning: incorrect ice nucleation number (nuci reset =0)
15974.8806117785 211.094338257501 0.573135485888197
3.272696509873641E-002 0.000000000000000E+000 0.000000000000000E+000
0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000
1.524968712475104E-028 32.6903636583803
# of NaNs = 1
Which are NaNs = F F F F F F T F F F F F F F F F F F F F F F F F F F F F F F F
F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F
Sl_t
gridcell index = 17636
ENDRUN:
ERROR:
lnd_export ERROR: One or more of the output from CLM to the coupler are NaN
----

enclosed the log files. How can I track or fix this error? Thank you very much.

Greetings,
Bo

Hello. Did you solve this? I am getting the same error. Can anyone help me in this regard?

 

[JIAA]

Hi Oleson,

thank you for your suggestion. I can pass the error by changing the nsplit.

However, I got another error in nstep=1256

-----
Warning: incorrect ice nucleation number (nuci reset =0)
15974.8806117785 211.094338257501 0.573135485888197
3.272696509873641E-002 0.000000000000000E+000 0.000000000000000E+000
0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000
1.524968712475104E-028 32.6903636583803
# of NaNs = 1
Which are NaNs = F F F F F F T F F F F F F F F F F F F F F F F F F F F F F F F
F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F
Sl_t
gridcell index = 17636
ENDRUN:
ERROR:
lnd_export ERROR: One or more of the output from CLM to the coupler are NaN
----

enclosed the log files. How can I track or fix this error? Thank you very much.

Greetings,
Bo

Hello sir! Do you solve the question? May I ask how you solved it? Thank you

 

[slevis]

I will move to the atm forum.

 

[peverley]

Hello sir! Do you solve the question? May I ask how you solved it? Thank you

Have you tried any of the tips that Danny proposed? See below:

Hi all,
I have also seen similar messages in CAM6/CLM5 recently. I saw in my runs the error 1) emitted LW radiation being -ve or zero, and I also saw the warning 2) incorrect ice nucleation number (nuci reset =0). 1 is from CLM (lnd_import_export.F90), and 2 is from CAM (nucleate_ice.F90). Both messages usually happen together and 2 sometimes leads to 1, but 2 also appears alone without triggering 1. I am not familiar with the CAM ice nucleation module, but my guess is that there are not an adequate amount of aerosols in the atmosphere that act as ice nuclei, which caused the radiation calculation to misbehave and output negative values, and then CLM receives negative radiation and call the endrun function. It's usually coarse aerosol particles such as dust and sea salt that are responsible for ice nucleation. I was working on CLM dust emission parameterization and both 1 and 2 appear a lot, and I think it means in my CAM dust particles were too few/too many. I usually scale up/down my dust emissions in CLM by a factor of 2-10 to make sure CAM has enough dust, and that will usually solve the problem (or delay the time the error occurs again). One needs to guarantee that every gridcell around the globe has an adequate number of dust/sea salt aerosols, not too few and not too many. It seems that the ice nucleation scheme is very sensitive to the number of ice nuclei, and CESM seems heavily tuned to maintain stability, and only a narrow range of dust emissions can avoid the model from stopping the run. I am not sure if you are working on dust or sea salt, or you are just running into this error without changing the aerosol scheme, but my thought is that you make sure the dust or sea salt levels are adequate by scaling them up/down. You may need to try a few times to get a good scaling factor that avoids the error.
Another trick I use is that I manually set the LW radiation to be non-negative in CLM whenever the LW radiation is -ve, to avoid trigger CLM's endrun function. But this is not a good practice and is better not used.
Danny