Discussion on MOZBC program

[Indranil Nandi]

 Respected members,

I am willing to understand the need for using MOZBC program in a year-long WRF_Chem simulation with more than 1.5 month spinup time. 

When I am trying to simulate the model without any initial & boundary condition for the chemical species using MOZBC program with WACCM Global outputs the model is running smoothly.  But after implementing both IC and BC using the MOZBC, the model unable to start the simulation and stops with an error message  "*** Error in `./wrf.exe': munmap_chunk(): invalid pointer: 0x00000000205da230 ***" with few other lines which primarily seem to be an error related to memory and computation. Although I have tried to solve the issue with increasing no. of processors I end up with the same.

To all the WRF_Chem experts, is it strictly necessary to run WRF_Chem with IC and BC  from MOZBC? If we avoid the MOZBC step and run the model for a year-long simulation with 2 months spinup time how much there will be an error in my results for Trace gas and Aerosol species?

I have attached my namelist file. I am in a very confusing situation. Please help me.

--

Indranil Nandi

Ph.D. Research Scholar
Centre for Atmospheric Sciences
IIT Delhi, Hauz Khas, New Delhi 110016, India

Contact No. 8902005948 

[Gabriele Pfister]

Hi Indranil

yes, you will absolutely need to have chemical boundary conditions for such a long run. Do you use a single wrfbc file for the entire time period? 

I did notice a typo in your namelist: 

 have_bcs_chem                       = .ture.,   .true.,   

should be 

 have_bcs_chem                       = .true.,   .true.,   

Gabi

--

================================
Gabriele Pfister, Acting Director

Atmospheric Chemistry Observations & Modeling

National Center for Atmospheric Research

Email: pfi...@ucar.edu

Phone: +1 303 497 2915

Web: https://staff.ucar.edu/users/pfister

[Indranil Nandi]

Respected Gabriele Ma'am,

First of all, thank you for your kind response and for pointing out the typo in my namelist. This means in the MOZBC namelist "do_bc = .true."  is needed for me to get correct simulations in the long run. What about the "do_ic" option in the same? Is it  also needed to be "do_ic = .true." or  "do_ic = .false." is sufficient if I allow a good spin-up time in the simulation?

No, due to the computational resource limit I generally split the run into small parts. Each part contains a total 3-month simulation (1-month spin-up & last 2-month result for analysis). For each part, I use a single wrfbdy file.     

[Gabriele Pfister]

Hi Indranil, 

you will just need to set 

have_bcs_chem = .true. 

Gabi

[Alice Burton]

Hi Indranil,

May I ask you please to share your mozbc.inp script. Cause once I used my own mozbc script the simulated PM10 would be strongly underestimated. I believe I should map the PM10 in my script but I do not know how can I do this.

Thanks