Large change in 2 m Temperature along coastline

[Israt246]

Hello,
I am running WRF v4.2.2 forced by ERA5 reanalysis. I ran a 1 year simulation from 2001-08-01 00 UTC to 2002-09-01 00 UTC using restart. I randomly looked at some of my WRF output files and observed there are some weird large changes in 2 m temperature in adjacent grid cells along coastline. e.g., around 8 degree celsius between long island and ocean. I outlined this in the attached picture. Does anyone have any idea why this is happening and if this something I need to be concerned about?

I am using two-way nesting with two domains. The outer domain has 12 km resolution and the inner domain has 4 km resolution. Number of vertical levels is 51. I have attached a picture of temperature at 2m at a certain time step in the inner domain (northeast USA). I used sst_update=1 and sst_skin=1 in namelist.input and set SST update interval as 24 hours, but I did not use any external SST data other than what's in ERA5. These weird coarse looking large temperature changes are not always present at each time step, but occur frequently. I have also attached my namelist.wps and nameless.input files.
I would really appreciate any insights on this.
Best,
Israt

 

[kwerner]

Apologies for the long delay in response while our team tended to time-sensitive obligations. Thank you for your patience.

When you processed the ERA5 data, did you make sure to include the "invariant" data? See How to Process ERA5 Data.

 

[Israt246]

Apologies for the long delay in response while our team tended to time-sensitive obligations. Thank you for your patience.

When you processed the ERA5 data, did you make sure to include the "invariant" data? See How to Process ERA5 Data.

Hi Kelly,
I did not use ERA5 data from NCAR RDA dataset. I downloaded the ERA5 data from Copernicus Climate Data Store for single level and for pressure levels. I downloaded all the variables that are in VTable.ECMWF including Land-Sea mask. I downloaded all surface level variables including the land sea mask variable as a single grib file.
However, I did not add constants_name = 'ERA5_INVARIANT' in &metgrid section of the namelist like the way it is described in the link that you shared. Did I need to add it in my namelist.wps even though I have the Land-Sea mask variable with other surface level variables?