By default, WRF comes with its own terrestrial fields dataset for static initialization, although it is allowed for one to use his/her own datasets. For doing so, however, WRF users should pay attention to creating the static fields of their interest in the appropriate format for WPS (i.e. WRF’s suite of programs for static initialization). In this post, I will present an almost ready-to-use method for generating a high-resolution terrain dataset for integration in WRF. The presented method incorporates procedures that have been already referenced in the official WRF forum, as well as in other sites/blogs coping with WRF. Hence, it is my aim to sum up all this information in a single step-by-step guide for easier reference.
Resampled 250 m SRTM terrain elevation, kindly provided by Andy Jarvis at https://hc.app.box.com/shared/1yidaheouv. You are allowed to download the data tile of your interest (in my case, I will be working with the NE tile).
Python (https://www.python.org) with GDAL bindings
QGIS (or any other GIS platform) **Optional, for preprocessing input datasets
1. (Optional) Load the source SRTM tiff to QGIS and extract the particular area of interest. This may save time by reducing the area that will be later processed.
2. Run convert_geotiff to get a WPS-compliant dataset based on the source SRTM tiff
./convert_geotiff -w 4 -t 1200 -u “meters MSL” -d “Topography height” sourceSRTM.tiff
convert_geotiff will create (within the directory from which you issued the above command) the necessary data tiles for WPS, along with the corresponding index file. The parameters that are required for correctly generating this file are set at the command line used for running convert_geotiff. More specifically, -w 4 defines the word length , -t 1200 defines that each tile will be contain 1200 x 1200 points, -u “meters MSL” defines the units, -d “Topography height” defines the description (the latter two may be ignored, although it is advisable to include them).
3. Within the directory you have the default WRF terrestrial dataset, create a new folder and name it topo_3s (or whatever you desire).
4. Move the data tiles convert_geotiff generated in the directory you previously created.
5. Move into your WPS directory and open /geogrid/GEOGRID.TBL.ARW (the TBL file actually linked to the GEOGRID.TBL in use). Look for variable HGT_M and add the following lines at the appropriate places:
interp_option= 3s: average_gcell(4.0)+four_pt+average_4pt
rel_path= 3s: topo_3s (or whatever the name of your directory (3) is).
6. To make use of the new dataset, edit the geog_data_res entry of your namelist.wps:
geog_data_res = “3s+30s” (this instructs geogrid.exe to look first for the new high-resolution data and then for 30s resolution data).