Unmatched ERA5 Coastlines Causing Irregular Temperature Edges Over Lake Michigan

[Henry18]

Hi,

I'm using ERA5 pressure-level and surface data as forcing for a WRF simulation over Chicago. Initially, I used the recently published era5_to_int.py tool to convert ERA5 netCDF files to intermediate and then metgrid files, following the workflow described here:

Question about Required ERA5 Data (ml and Invariant) for era5_to_int.py

Hi, I'm using the era5_to_int.py tool to process downloaded ERA5 pressure-level (pl) and surface (sfc) NetCDF (.nc) data into intermediate files for subsequent metgrid preprocessing. According to my colleague’s experience using ERA5 GRIB (.grb) data, only the pl and sfc datasets are required...

forum.mmm.ucar.edu

However, the simulation driven by these metgrid files produced problematic output—temperature-related fields show irregular, blocky edges along Lake Michigan, as the issue discussed here:

Using era5_to_int.py for ERA5 Pressure-Level Input but Getting Requests for ML Files

Hi, I’m using the era5_to_int.py tool to process ERA5 data into the intermediate files required for met_grid. I would like to use pressure-level (pl) data as the input. I have collected the pl, surface (sfc), and invariant datasets in my data directory, and I’m running the following command...

forum.mmm.ucar.edu

Because of this, I switched to the traditional GRIB-based workflow for ERA5. Here are the steps I took:

1. Used the cdsapi package and Python scripts to download:

• Pressure-level data from “ERA5 hourly data on pressure levels from 1940 to present”
• Surface data from “ERA5 hourly data on single levels from 1940 to present”

    These were saved in:

    /glade/derecho/scratch/hhou/TestData/GribData_ERA5

    as pl.grib and sfc.grib. The Python scripts used for downloading are in the same directory.

1. Followed the method described in this post to ungrib pl.grib and sfc.grib separately, producing intermediate files with distinct prefixes.

1. Ran metgrid using these intermediate files to create met_em* files, which are located in:

    /glade/derecho/scratch/hhou/Test_ERA5/WPS

1. Linked the met_em* files to my WRF directory and ran real and wrf from:

    /glade/derecho/scratch/hhou/Test_ERA5/WRF/test/em_real

Unfortunately, even using the traditional GRIB workflow, the resulting temperature fields still show irregular edges over Lake Michigan.

T2


TMN


After discussing with colleagues and examining the met_em* and wrflowinp* files, our current hypothesis is that the issue may be related to unmatched coastlines of SST field:

sst in met_em*


sst in wrflowinp*


My WPS version is 4.6, and my WRF version is 4.7.1. Looks like they are incapable of ingesting ERA5 SST fields through the WPS/real process. However, this is still only a hypothesis.

Given that many WRF users work with ERA5 forcing—often over coastal domains—I would like to ask:

Is this type of coastal edge artifact a known issue when using ERA5, particularly for domains that include coastlines? If so, is there a recommended solution or workaround?

Any insights would be appreciated.

Cheers,
Henry

 

[Ming Chen]

Hi Henry,

This is a landmask issue. Let's stay with era5_to_int.py, which we have well tested and we know for sure it works just fine. The issue you have is derived from incorrect input data. Would you please let me know where you download ERA5 data and what data did you have? Thanks.

 

[Henry18]

Hi Henry,

This is a landmask issue. Let's stay with era5_to_int.py, which we have well tested and we know for sure it works just fine. The issue you have is derived from incorrect input data. Would you please let me know where you download ERA5 data and what data did you have? Thanks.

Hi Ming,

Thank you for your prompt reply.

When using ERA5 NetCDF (.nc) data with the era5_to_int.py tool, I copy the required variables (pl, ml, sfc and invariant) from the following directories to my working directory on Derecho:

/glade/campaign/collections/rda/data/d633000/e5.oper.an.pl/
/glade/campaign/collections/rda/data/d633000/e5.oper.an.sfc/
/glade/campaign/collections/rda/data/d633006/e5.oper.an.ml/
/glade/campaign/collections/rda/data/d633006/e5.oper.invariant/

When using ERA5 GRIB data, I download the files (only pl and sfc) to my directory on Derecho using Python scripts from the following CDS webpages:

ERA5 hourly data on pressure levels from 1940 to present

ERA5 is the fifth generation ECMWF reanalysis for the global climate and weather for the past 8 decades. Data is available from 1940 onwards. ERA5 replaces the ERA-Interim reanalysis. Reanalysis combines model data with observations from across the world into a globally complete and consistent...

cds.climate.copernicus.eu

ERA5 Land hourly time-series data from 1950 to present

ERA5-Land is a reanalysis dataset providing a consistent view of the evolution of land variables over several decades at an enhanced resolution compared to the Fifth Generation of the European Centre for Medium-Range Weather Forecasts (ECMWF) Reanalysis (ERA5). Produced by replaying only the...

cds.climate.copernicus.eu

Thank you for your help.

Henry

 

[Ming Chen]

Hi Henry,

If you run era5_to_init.py in derecho, basically you don't need to download the data. This is because the script can automatically retrieve data from /glade/campaign. You only need to issue the command:

era5_to_int.py -i yyyy-mm-dd_00

and the intermediate file will be created. More options can be found in README.md

Please try and let me know if you still cannot make it work.

 

[Henry18]

Hi Henry,

If you run era5_to_init.py in derecho, basically you don't need to download the data. This is because the script can automatically retrieve data from /glade/campaign. You only need to issue the command:

era5_to_int.py -i yyyy-mm-dd_00

and the intermediate file will be created. More options can be found in README.md

Please try and let me know if you still cannot make it work.

Hi Ming,

Thank you for the suggestion. I tried just specifying the simulation period in the command and let it automatically retrieve the necessary ERA5 input data. The WPS processing and the simulation ran successfully, but the WRF output still shows irregular edges along the coastal lines.

I'm now focusing on resolving this coastal edge issue. I came across this thread discussing the problem:

Issue with ERA5 SST fields

Hello all, I'm making a WRF run forced wit ERA5 reanalysis, and I noted that ERA5 SST fields look weird... First, the mask values over land is 0ºC, which when WPS will interpolate the fields in metgrid will give irrealistic surface temperature fields over coastal areas. obviously, this will in...

forum.mmm.ucar.edu

It recommends editing the SST field in METGRID.TBL.ARW to include a four-point weighted average in the interpolation options and to mask the field over land. Do you think this would be an effective solution?

Best,
Henry

 

[Ming Chen]

Henry,
Can you upload a figure that shows irregular edges along the coastal lines?
Please also upload your namelist.wps and namelist.input for me to take a look. Thanks.
Ming

 

[Henry18]

Henry,
Can you upload a figure that shows irregular edges along the coastal lines?
Please also upload your namelist.wps and namelist.input for me to take a look. Thanks.
Ming

Thank you, Ming.

I've attached two pictures of the wrfout to show the edges over Lake Michigan, they are T2 and TMN. I've also attached my namelist.wps and namelist.input files for reference.

Best,
Henry