Canada High-resolution Urban Morphology to be used for WRF application

This file provides information on how to extract Urban Morphology parameters for the province of British Columbia (BC), Canada. These high-resolution parameters are needed to successfully run the Weather Research and Forecasting (WRF) model coupled with Single Layer (option 1 in WRF urban physics) and Multi-layer (option 2 and 3 in WRF model, urban physics) Urban Canopy Models. This Shapefile dataset has been generated using the Housing dataset from the Canada data warehouse. The shapefiles were collected and combined to serve as input for the Python script.

In this file, "domain" refers to the domain the WRF model will be run for (e.g., d01, or d01, d02, d03, etc. if a nested domain is used in the WRF model). 

"geo_em.{domain}.nc" refers to the output of the geogrid.exe program which is one of the core programs of WPS program. 

For example, for a 3 nested domain setup, geo_em files will be annotated as follows: 'geo_em.d01.nc', 'geo_em.d02.nc', 'geo_em.d03.nc'.

This folder contains all the documents necessary for extracting the Urban Morphology parameters. Refer to Section 1 for a description of the folder structure. 

You will first need to extract the shapefile of the domain considered using the VERDI program (https://www.cmascenter.org/verdi/). For more details, refer to Section 2. Make sure to save this shapefile under the SHP subfolder described in Section 1. 

To utilize the dataset, ensure the geo_em files are placed in the 'Main' folder mentioned in Section 1, which should also contain the Python script named 'Canada_UrbanMorphology.py', the folder with the dataset (geo_em.{domain}.nc), and the folder with the shapefiles (Home_BC and domain).

The Python script 'Canada_UrbanMorphology.py' necessitates three libraries: NetCDF4, geopandas, and numpy, all of which can be installed via pip or Anaconda.  Follow the specific instructions provided in Section 3 to adjust the script, then proceed to running the program using the following command:

python3 Canada_UrbanMorphology.py

_____________________________________________________________________________________________________________________________________

Section 1. Folder structure

Main Folder

|- Canada_UrbanMorphology.py

|- geo_em.{domain}.nc

|- INT

    |- Home_BC.cpg

    |- Home_BC.dbf

    |- Home_BC.prj

    |- Home_BC.sbn

    |- Home_BC.sbx

    |- Home_BC.shp

    |- Home_BC.shx

|- SHP

    |- domain.{domain}.dbf

    |- domain.{domain}.prj

    |- domain.{domain}.shp

    |- domain.{domain}.shx

    |- domain.{domain}.fix

_____________________________________________________________________________________________________________________________________

Section 2. Creating a domain shapefile

0. If you have not yet installed the VERDI program, follow the instructions on https://www.cmascenter.org/verdi/ to obtain the source

code and install it.

1. Open the VERDI program.

2. On the 'Datasets' panel, click on 'add local dataset'.

3. Open the desired geo_em.{domain}.nc file for the desired domain.

4. Double-click on a 2d variable (e.g., LU_INDEX) on the 'Variables' panel.

5. Click on 'Tile Plot.'

6. Click on File > Export as Image/GIS.

7. Change the 'Files of Type' to 'Shapefile (*.shp, *.shx, *.dbf)'.

8. Name the file accordingly (i.e., geo_em.{domain}).

9. Click on Save.

10. Move the saved files to the SHP folder.

For example, to process the first domain (d01), geo_em filename would be geo_em.d01.nc and the associated file names would be geo_em.d01 with GIS extentions (i.e., *.shp, *.shx, *.dbf)

_____________________________________________________________________________________________________________________________________

Section 3. Required modification to the Python script Canada_UrbanMorphology.py

1. Adjust the grid names based on the number of nested domains used in the modeling framework (e.g., [d01] for a single domain, or [d01, d02, d03] for a three nested domain framework) in the 'main' function, line 17.

2. Adjust the grid size in meters (e.g., 1000*1000 for a domain with 1 km resolution) in the 'gridsize' function.

3. Check the output in the geo_em file. Due to the conversion of geo_em file to shapefile in the VERDI program, the indexing of grid cells may not be consistent between geo_em files and shapefiles. Therefore, the output after executing the Python script may look flipped (e.g., the downtown which is located in the southeast of the domain may appear in the northwest of the domain). If the output is flipped, follow the instructions on lines 163 - 170. Follow the example in the Python script to adjust the frame output.

4. Use the new geo_em files instead of previous ones for processing metgrid.exe core program in the WPS program.

Credit: Clair lab (https://cleanairlab.cive.uvic.ca/)