We have noted some issues in the generated land-sea mask using topo_30s datasets.

1. Some lakes are in the land-sea mask.
2. Some single point islands have also generated in the land-sea mask due to the Gibbs effect.

We have not seen above given issues in the land-sea mask when we have used the landuse_30s_with_lakes datasets instead of topo_30s datasets for the creation of the land-sea mask. These both datasets are available at /bdd/WRF_Forcing/GEOG/geog/.

After this, we have also done some modification in the land-sea mask. What we have modified in the land-sea mask is given below.

1. We have transformed the single land grids (those grids which are surrounded by the ocean grids and also those grids whose corners are connected to the other land grids) to the ocean grids and also assigned the value of HGT_M is to zero at those points. [Line 979-996 in the process_tile_module.F]
2. Similarly, We have transformed the single ocean grids (only those grids which are surrounded by the land grids) to the land grids and also assigned the value of HGT_M is to averaged value of HGT_M over surrounded land grids at those points. [Line 999-1007 in the process_tile_module.F]
3. We have also opened the estuaries (for example Tajo close to Lisbonne). We have followed the given below steps for opening the estuaries.

• Chosen the longitude and latitude of the start and end points by using the google map, and stored in the text file. Please see the attached text file (OCEANPTS.txt).
• Read the longitude and latitude of both points from the text file by using the READOCEANPT subroutine. [Line 1353-1392 in the process_tile_module.F]
• Converted the longitude and latitude of the start and end points in the (i1, j1) and (i2, j2) by using the LOLA2IJ subroutine. [Line 1394-1426 in the process_tile_module.F]
• Finally, Opened the path in between (i1, j1) and (i2, j2) by using the FILLOP subroutine. [Line 1428-1456 in the process_tile_module.F]

To use, update the process_tile_module.F in …/WPS/geogrid/src and put OCEANPTS.txt in …/WPS/

and after that run the ./geogrid.exe

The files are available here : code by Namendra

I have used the 3 hourly MEDCORDEX-B simulated meteorological datasets as forcing datasets in the current WRF high-resolution simulation. The MEDCORDEX-B dataset was in .NC format and as we know that WRF UNGRIB program only reads the GRIB format datasets and generate the intermediate format meteorological datasets. Therefore, I have used the FORTRAN program to generate the intermediate format meteorological datasets because converting the NC file datasets into a GRIB format datasets for a longer duration would be a time-consuming process. After that, I have used the WRF METGRID program to horizontally interpolates the intermediate format meteorological datasets onto the simulation domains defined by the geogrid program. After that, I provided the interpolated datasets to the WRF real.exe.

In the simple word, I have not used the WRF UNGRIB program to generate the intermediate format meteorological datasets.

The program is written in the following sequence:

1. First, I opened the 3H hourly .NC file.
2. Getting the size of Time, Latitude, Longitude, and Level and allocating memory
3. Reading the Time, Latitude, Longitude, and Level
4. Allocating the memory to all the meteorological fields
5. Reading the 3D (Time, Latitude, Longitude) and 4D (Time, Latitude, Longitude, and Level) meteorological fields.
6. Some levels of the 4D variables have undefined values, I filled up using the FILLBLNK subroutine.
7. Then I have converted the TIME in actual format (YY:MM:DD HH:MM:SS). (for this I have used subroutines calendar.f90, errioipsl.f90, met_data_module.f90, stringop.f90)
8. Then I have wrote the data in intermediate format.
9. and then deallocating the memory.

For the compile and run I have used the Makefile

make netcdf2wps

./netcdf2wps

The files are available here : code by Namendra