Scheduled Downtime
On Friday 21 April 2023 @ 5pm MT, this website will be down for maintenance and expected to return online the morning of 24 April 2023 at the latest

wrf fails run with ECMWF model forcing data


Dear Colleges,

I am using ECMWF high-resolution analysis downloaded from to run the WRF model. I had difficulties with preprocessing the ECMWF by ungrib for some of surface variables which is a known issue ( So, I had to replace Sea-Surface Temperature, Sea-Ice Fraction, lsm and Water Equivalent of Accumulated Snow Depth by ERA5 reanalysis data. After that I succesfully ran the metgrid.exe and real.exe. However, when I run wrf.exe it stops after few minutes. Please upload my wrfinput*, wrfbdy_d01, namelist and error.log files using this link:

I would much appreciate your help.

I believe this issue is caused by different landmasks in your input data. The ERA5 landmask is different to that for the ECMWF IFS CY41r2 High-Resolution data. Therefore, you must specify different landmasks for the two input data.
Please take a look at the file Slides 53-57 of this file gives detailed instruction how to specify landmask if your input data are different.
Hi Ming,

Thanks for your reply! I am looking the presentation suggested by you. I would like to clarify that I am using single landmask which is the ERA5 landmask since ungrib did not allow me to preprocess ECMWF IFS CY41r2 High-Resolution landmask data. This means that my intermediate files produced by ungrib contain single landmask field. After that I successfully ran metgrig.exe and real.exe, but it fails to run wrf.exe. So, my question is whether the issue of using multiple landmasks described in the presentation (Slides 53-57) is relevant for my case?

Those slides describe in detail how landsea mask works and how to include specific landsea mask from your input data. Now that you cannot obtain landsea mask from your ECMWF data, i suppose you need to find this data (and related information) from other dataset such as ERA-I. When using different surface input data, correct specification of landsea is important. Otherwise some variables related to soil, seaice, sst etc. will not be processed correctly. Note that WPS handles interpolation and won't involve any physics, there may not have any issue when running WPS. But once wrf.exe starts, those unreasonable data will lead to model crash, just as what you hav sen in your case.