My team developed a method to detect PM2.5 hotspots in a city using mobile sensor measurements. Now, we want to simulate a validation dataset to test our method. We want simulated data so that we can set a 'ground truth' for where hotspots are located, and thus compute accuracy metrics.

Does anyone know a free software for this? I imagine it takes pollution source locations, wind speed, etc. as inputs, then outputs a time series of pollution values in a defined region.

Already tried:

  • Polyphemus: looks good, but apparently the software packages have not been updated in 13 years
  • Wikipedia: has a list of atmospheric dispersion models, but it is unclear which are readily available as software
  • 1
    $\begingroup$ Depending on the level of complexity you need, you might try HYSPLIT or AERMOD. If you are looking for multi-source complex models, CAMx and CMAQ are available. Also, there is GEOS-CHEM and WRF-CHEM. $\endgroup$
    – f.thorpe
    Commented Feb 23 at 15:21
  • $\begingroup$ epa.gov/moves $\endgroup$
    – Spencer
    Commented Feb 23 at 17:10

2 Answers 2


In addition to those listed in @f.thorpe's comment (I want to at least point to him to give credit for his choices, which are my first choices too), some other readily available models include:


Thank you for the suggestions, @f.thorpe and @BarocliniCplusplus. I ended up adapting the London Atmopsheric Emissions Inventory dataset for my purpose, since I need data with fine spatial resolution that explicitly models the pollution along road networks. The other suggested softwares did not seem tailored to this use case.

Similar datasets, should it ever interest anyone:


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.