Scheduled Downtime
On Friday 21 April 2023 @ 5pm MT, this website will be down for maintenance and expected to return online the morning of 24 April 2023 at the latest

process mode for the resolution of new geographical static data

This post was from a previous version of the WRF&MPAS-A Support Forum. New replies have been disabled and if you have follow up questions related to this post, then please start a new thread from the forum home page.

sd16121

Member
Dear Colleagues,

In our team meeting, we discussed an interesting question about the model configuration. As we know, the spatial resolution of WRF normally is 1 km. If we use the local climate zone as an urban landscape map, with 30 m spatial resolution, how does WRF handle the input data that are with higher resolution than its own? We supposed that WRF may average the 30 m map into 1 km but would like to confirm it with you.

Looking forward to hearing from you as soon as possible. Many thanks.

Best wishes,
Stella
 
Hi Stella,
When you have input data that are a much higher resolution that the resolution of the domain of interest, we use nesting for this, using a resolution ratio between nests, and we have some best practices that should be followed for that. For e.g., we recommend using either a 3:1 or 5:1 grid size ratio between nests. So if, for e.g., you were using GFS 0.25 degree input data to drive the model, that is about 27 km resolution and let's say you were ultimately interested in a domain with a resolution of 3 km. It would be reasonable for your outer (parent) domain to be 9 km, and then have a nested domain inside that is 3 km (and this would satisfy the 3:1 ratio). You can find additional information about nesting in this presentation. And you may also be interested in some of the best practices for setting up your namelist and domain. I would also advise you to refer to our WRF Users' Guide.
 
Dear Colleague,

Many thanks for your detailed reply. Based on my previous question, here are three other things I am curious about:

1. When we interpolate the map with different resolutions (30 m or 1 km) as the geographical static data, what are the differences for the WRF processing?
2. Could you please kindly introduce how the WRF handles when I interpolate 30 m map as the geographical static data into the 1 km grid?
3. I am interested in WRF Urban, and I would like to know which model (land surface model or land terrain model) is used in WRF under this circumstance?

Looking forward to hearing from you. Your early reply will be highly appreciated.

Best regards,
Stella
 
Hi Stella,
Thank you for your patience. I have reached out to a colleague who may be better to explain this. They will either respond to you, or I'll keep you posted when I receive a reply from them.

P.S. I see that you posted the same question in different areas of the forum. To keep the forum organized, and to keep from having two people working on the same topic simultaneously, I'm going to delete those posts and we will just work with this topic here.
 
Hi,

Thank you very much for your reply. Here is a supplementary question:

After interpolating the geographical static data generated by WUDAPT, namely the Local Climate Zone map of a city, it divides the urban areas into 17 classifications, how the WRF handles the map that has different fractions for the 17 zones (please note that the map has a higher resolution than that of the grid)?

We have two assumptions:
1. according to the proportions of 17 zones in the high-resolution map to distribute them in the large grid of WRF.
2. only keep the local climate zone that occupies the most ratio.

Please kindly provide your answer as well. :D

Best wishes,
Stella
 
Hi kwerner,

Just a kind reminder about the answers to my four questions. If there are any progresses, please kindly let me know. Many thanks.

All the best,
Stella
 
Stella,
I am still waiting on a reply from a colleague. If I don't hear anything in the next few days, I'll try to ask someone else. Thank you for your patience.
 
Dear Colleague,

Many thanks for your help! For your convenience, I put all the four questions together (see below):

1. When we interpolate the map with different resolutions (30 m or 1 km) as the geographical static data, what are the differences for the WRF processing?

2. Could you please kindly introduce how the WRF handles when I interpolate 30 m map as the geographical static data into the 1 km grid?

3. I am interested in WRF Urban, and I would like to know which model (land surface model or land terrain model) is used in WRF under this circumstance?

4. After interpolating the geographical static data generated by WUDAPT, namely the Local Climate Zone map of a city, it divides the urban areas into 17 classifications, how the WRF handles the map that has different fractions for the 17 zones (please note that the map has a higher resolution than that of the grid)?
We have two assumptions:
a. according to the proportions of 17 zones in the high-resolution map to distribute them in the large grid of WRF.
b. only keep the local climate zone that occupies the most ratio.
 
Again, I'd like to apologize for the long delay, and to thank you so much for your patience. I was able to receive some answers to your questions.

1 & 2: There are two types of static data that WPS/geogrid handles, continuous (e.g. terrain) and categorical (e.g. land use). For continuous data, if the resolution of the model grid is lower than the input data (for example, using 30 second resolution terrain data for 10 km model grid), then data points inside the model grid box are averaged to obtain the value at that grid point. Afterwards, we use a smoother-desmoother to remove high frequency noises in the gridded field. For categorical data such as landuse, one can provide the data in the original resolution. If you provide data at higher resolution than model grid you'd use (e.g. input data at 30 meter resolution, but you would like to run model at 1 km), geogrid will sum up all different categories in the 1 km model grid cell, and compute the percentage of landuse by each categories. It will then determine the dominant category (the WRF model typically uses dominant categories only) by taking the category that has the largest percentage in the model grid cell. As a user you can have some control over this, and this can be done by modifying the inter_option in geogrid/GEOGRID.TBL. The options are explained in Chapter 3 of the user's guide.

3. The urban options in WRF can be used with either the Noah LSM or NoahMP LSM (sf_surface_physics = 2 or 4).

4. The LCZ handling will be included in WRF V4.3 release, which should happen before the end of April. You can view the brief document here.
 
Dear Colleague,

Many thanks for your detailed reply. I think maybe my expression for question 3 is inappropriate, I may need to re-ask this question:

I would like to know which model is handled by WRF, Digital Surface Model or Digital Terrain Model (specifically Digital Elevation Model)?

Best wishes,
Stella
 
Hi Stella,
I'm not actually sure of the answer to your question regarding the urban model. That part of the model was written by a different group and therefore I'll point you to their resource page regarding the different urban options. If you aren't able to find what you need there, there is contact information there so you can contact someone to ask.
 
Top