Divide by zero error

[chrimerss]

Hello,

I got the system error - floating divide by zero and was wondering if any of you have some insights. I was able to pin down the location of the error by printing the variables. I found my soil depths were all initialized to zeros for the first grid cell. Attached are my config files and lis output.

Appreciate your help!

Allen

ldt.txt
lis.txt
lislog.txt

[dmocko]

Hi @chrimerss

From where did you obtain the parameters in?: PARAMS/CONUS_1km_HYMAP

Can you please also attach your ldt/make/configure.ldt and lis/make/configure.lis files?

I am wondering if you configured LDT to use little-endian binary files; however, the files you are using are big-endian.

[agetiran]

Hi @chrimerss
Something is off here. the lis.config file shows that HyMAP is running in offline mode using NLDAS runoff as forcing:
Routing model: "HYMAP2 router"
External runoff data source: "NLDAS2 runoff data"
HYMAP2 routing model time step: "30mn"
TEMPLATE model timestep: "30mn"

And the LSM is set as template:
Number of surface model types: 2
Surface model types: "LSM" "Openwater"
Surface model output interval: "1da"
Land surface model: "none"
TEMPLATE model timestep: 1hr
Open water model: "template open water"

It's not clear to me why the model is crashing in NoahMP401. Could you please double check the config files used in this simulation?

[chrimerss]

@agetiran -
Sorry for the confusion. I've tried many ways such as online mode and offline mode and forgot to update the config file. Attached are the latest ones. At this time, I got seg error.

@dmocko -
I can confirm I configured using big-endian and it works perfectly when I crop a small domain instead of using the whole state.

ldt.txt
lisconfig.txt
lislog.txt

[agetiran]

Hi @chrimerss. The error is likely related to inconsistencies in the parameters you are using. @dmocko previously asked where you obtained the HyMAP parameters at 1km. Would you mind responding to that question?

Also:
In your config file, NoahMP and HyMAP have different time steps
HYMAP2 routing model time step: "30mn"
Noah-MP.4.0.1 model timestep: "1hr"
They must be the same. I suggest using 15mn for both.

I suggest turning off the ensemble mode in HyMAP
HYMAP2 run in ensemble mode: 0

[chrimerss]

Hi @agetiran @dmocko,

@agetiran you sent me the CONUS 1km parameter about a year ago.

I tried setting both timesteps to 15 mn but still didn't work. Same error.

[agetiran]

Just for a closure to this thread:
I worked offline with @chrimerss on the generation of a new customized parameter set for the domain, and that resolved the issue.