Hello everyone, could you please tell me how to run a local case using the WRF-CMAQ two-way coupled model? What data needs to be prepared in advance? Do I only need to modify the run_cctm_Bench_2018_12NE3.WRFCMAQ.csh script? What are the differences between running a local case with the two-way coupled model and the offline coupled model? I’ve successfully run the test case. Additionally, I noticed that the WPS preprocessing module is not included. Does this mean I need to run it in advance to generate the files, and then input them into the WRF-CMAQ two-way coupled model?
Hi Sher,
In short, offline mode uses the meteorology to predict the pollutants concentration where online two-way coupled model considers the influence of pollutants on meteorology (could formation, precipitation, etc.). So, using either one depends on your application and research question.
Depending on your case, resolution, study area, and study period, you might need to run WPS in addition to providing and preprocessing some additional datasets. You can get more help by providing additional details on your case.
This is only part of it: in the two-way coupled model, the meteorology acts on the pollutant concentration field at the internal time step of the meteorology model (which may well be only a few seconds).
In the offline mode, the meteorology is sampled at a much longer time step (typically, one hour), and those sampled values are used to act upon the concentration field. Except at very coarse modeling resolutions, that sampling is not adequate to represent the actual spatio-temporal feature-variations in the meteorology field.
At 4KM, for example, spectral analysis shows that the meteorology model resolves features with sizes less than 2000 seconds; basic information theory (the Shannon-Nyquist Theorem) says that a sampling time step larger than 1000 seconds cannot resolve these features. On the other hand, the met-model internal time step (perhaps 35 seconds?) does resolve them…
[And that is one reason why the original CMAQ systems architecture made the model operation essentially independent of the input time step of the various meteorology files. It was once quite easy to run the model with a meteorology input times step of, say, 5 or 10 minutes, provided you have sufficient disk space; and indeed I first did so in a project two decades ago for the California Air Resources Board.]
1 Like
Thank you very much.
I have successfully run the official test case. However, during practical operation, I found that the script run_cctm_wrf-cmaq.csh requires input data such as the initial condition, boundary condition, and grid files. I would like to ask how to obtain the grid files that are originally generated by MCIP, and how to prepare the ICON and BCON data.
Can the WRF-CMAQ bidirectional coupling model be run with multiple nested domains? I noticed that most published studies seem to use only a single domain.
In the WRF namelist settings, are the four-dimensional data assimilation (FDDA) and sea surface temperature (SST) options necessary?
When processing emission inventories, the CMAQ-MCIP output requires a GRIDDESC file. How can this file be obtained?
Thank you very much for your time and help.
Best regards,
Sher
Hi Sher,
Here are answers to your questions:
1. I would like to ask how to obtain the grid files that are originally generated by MCIP, and how to prepare the ICON and BCON data.
WRF-CMAQ coupled is no different than traditional offline CMAQ model. They both require ICON and BCON. In the CMAQ documentation section, it provides detailed procedure to guide you obtaining those two input files. The difference is the meteorological information. Traditional offline CMAQ model, the meteorological information is from the output of the MCIP process. Basically it gather information from wrfout files and put the information in CMAQ-ready format. On the other hand, WRF-CMAQ coupled, the meteorological information is directly from the WRF model, i.e. no MCIP process is needed.
2. Can the WRF-CMAQ bidirectional coupling model be run with multiple nested domains? I noticed that most published studies seem to use only a single domain.
No. If you want to run a nested domain case, you have to go through one domain at a time and sequentially. Recently I have made the WRF-CMAQ coupled model capable of running in a WRF traditional nesting domain setting and the CMAQ portion of the coupled will be executed in the most inner domain. However, this code is not available to the public.
3. In the WRF namelist settings, are the four-dimensional data assimilation (FDDA) and sea surface temperature (SST) options necessary?
Typical wrf run, fdda is not required. However, in the WRF-CMAQ model, we recommend to have fdda turn on, i.e. provide fdda and sfdda input files. We have seen (other users had similar experience) without them, the simulation results became un-realistic. Even fdda is used, you can see from the namelist, the strength is quite small.
4. When processing emission inventories, the CMAQ-MCIP output requires a GRIDDESC file. How can this file be obtained?
This is a chicken and egg scenario. 🙂 In this instance, I would recommend to use option 0 (I assume you are using the latest coupled model code) to run one day (option 0 means to run WRF only in the coupled model). Then run MCIP with the wrfout just generated to produce GRIDDESC file for your emission process.
On the side, let me emphasize that the WRF-CMAQ coupled model, the meteorology information is from directly from the WRF model at a user defined frequency. The traditional offline CMAQ gets meteorology information at a typical hourly time scale as Dr. Coats mentioned. Interpolation is required to meet the internal time step need (potential source of error).
Cheers, David