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

[Tanja]

I am running CLM5.0 in single-point mode.

I am using my own forcing data on longwave down. In user_nl_datm I set tintalgo to “coszen” for the solar stream. I then get the error message:

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

I have checked my longwave down data, and there is no negative or zero values. If I set tintalgo to nearest, I don´t get the error.

In CLM5.0 user guide "coszen" is recommended for solar data. However, as my data i hourly, "nearest" should be okay, but I would like to understand why I get the error when using "coszen".

/Tanja

 

[oleson]

If you are running the model at the same time interval as the forcing data, then all of the forcing data can be in the same stream and tintalgo should be nearest.
If the model time step is different than the forcing data (e.g., as it is for the GSWP3 data that is usually used to force the model in global simulations), then the forcing data should be split into three streams (one containing solar radiation only), one containing precipitation only, and one containing the other forcing variables (including longwave radiation). The tintalgo for those streams should be coszen, nearest, and linear respectively.
If that doesn't resolve your question, please attach your datm_in and your stream text files (e.g., datm.streams.txt.CLMGSWP3v1.Solar, datm.streams.txt.CLMGSWP3v1.Precip, datm.streams.txt.CLMGSWP3v1.TPQW, or whatever you are using for these files) and I can take a closer look.

 

[erik]

Tanja based on what you are saying I think the problem is that you are using "coszen" mode for LW down data. The coszen" mode should ONLY be used for the solar streams like FSDS and NOT the LW down. LW down never goes to zero, while the solar goes to zero at night. The coszen mode is going to make the data go to zero at night which is part of what it does. So it makes sense that you'd get that error. LW down doesn't follow the cosine of the solar zenith angle like SW does, LW is fairly constant day to night and independent of a diurnal cycle because it most changes with cloud cover and atmosphere water content and Temperature. So using "coszen" mode for LW down is totally inappropriate to do.

But, like Keith above agree that using "nearest" makes the most sense for hourly data. That is the usual way we do it.