Closed BreadDan closed 6 years ago
avaldebe
commented
7 years ago Hi @BreadDan
It looks like the comment at the end of the 'OM25_SUM' is causing a problem. What compiler did you used? gFrotran is particularly picky with comments inside a namelist.
BreadDan
commented
7 years ago Thank you @avaldebe
Yes, It is gFORTRAN. I would delete the comment next time.
I've just returned to my lab now. It seems that the program used nearly 11 hours to compute a one-month-test (from 2014/1/1 to 2014/1/31) with the default options ( 20 min time step etc), in my lab's PC. Since docker is somehow like a "linux emulator", the performance might be restricted in many ways. But I think it might be acceptable if I'm using it to simulate a smaller space and time scales.
As long as I can see, the base run is designed for the European regions, and I have noticed there is an application of EMEP on China ( http://www.marcopolo-panda.eu/forecast/?id=1&species=pm25&date=20170612&model=7). If I'm going to use EMEP to simulate the air pollution in big cities in China or India, apart from altering the parameters, am I supposed to attach additional data, like emission or methodology ?
avaldebe
commented
7 years ago Hi again @BreadDan
I find your docker setup very interesting. We seldom run the EMEP/MSC-W model on a PCs. Most of our work is done on HPCs in the Norwegian research infrastructure, which are mainly built with of the shelf components (e.g. vilje).
In addition to reduce the model domain and extent of the simulations, you can reduce the model run-time by limiting your output. Reading input/writing output does not really scale with CPUs, as is done almost exclusively by one of the CPUs.
For running the model on a different domain and resolution, you need to provide the meteorology and emissions. I'm not sure if you would also need vegetation parameters, or ones provided with rv4_10 would be adequate for your application. Also, if you plan to run small domains, you will need boundary conditions. For the specific details to run the model on China/India, I would contact @matthieupommier (on a new issue). He is the scientist behind the Marcopolo/Panda simulations.
avaldebe
commented
6 years ago Think the original question is answered, please re-open if not the case.
Hello, I'm a master student from MSVE, ODU (https://www.odu.edu/msve), and I'm trying to make a simulation and visualization on air pollutions. That is why I'm learning to run EMEP in docker on windows. I got my graduate degree in atmospheric science so I think I can understand common parameters and process, but yet I have never operate such software head to toe, So I'm here for asking potentially "silly questions". If my question is already written in the guide or somewhere else, please point it out and I will have a look at it.
Thank you.
===update=== It is , up to now , running in docker in windows 10, running the modrun.sh.
This is the result that I modified the selected line in config_emep.nml.
Before that, It keeps giving error