#Stinkburgh: Predicting Stench in Pittsburgh

Over the last vew years I’ve developed a simple algorithm that I have used to predict days/nights that might be stinky in Pittsburgh. This helps me plan when to exercise outside and when it might be ok to open my windows at night. Turns out that it is actually reasonably easy to make such predictions from mixing height, wind speed, and wind direction data, all of which is forecasted and available from the National Weather Service Forecast Office link (shortened by me): https://bit.ly/pghmixheight .

Here is my algorithm: if the forecast indicates that the mixing height will be low (i.e. 300ft or less) and the wind is slow (i.e. 5mph or less) and coming from the S or SSE (from the Mon Valley to my house – your most concerning wind direction might be different), then there is a fair chance it will be stinky during those conditions (which often happen at night because of the how inversions happen in our area).

Here is a visual sample from the NWS page with a red line (mine) that indicates the level of 300ft. Notice how low the mixing height gets at night!

Mixing Height - prediction - Screen Shot 2020-05-11 at 1.38.39 PM.png

The image below shows what wind speed and wind direction information looks like on the NWS site. The “wind barbs” need a little interpretation and you can learn how to read them here. Basically, the wind is blowing from the barbs towards the anchor point dot.

Screen Shot 2020-05-11 at 1.43.28 PM.png

Based on the two graphics above, I might watch out for stench during Wednesday night, March 13. While the wind is not blowing in any direction, the forecast might change and the stench might find its way to my house. Experiment and develop your own algorithm!

I also like to watch the trends on this PurpleAir map to see when the particles reach a low point, then I go exercise outside at that low point (often around 3-5pm).

Another thing: the ACHD releases their own inversion forecast every day here: https://www.alleghenycounty.us/uploadedFiles/Allegheny_Home/Health_Department/Programs/Air_Quality/data/AirDispPotentialRpt(1).pdf . I haven’t learned yet how to integrate their predictions with my own algorithm above, but it offers some potentially helpful information.

And finally: I heartily endorse the accuracy of the SmellPGH app’s “conditions of stench” predictions. I believe their algorithm is more complicated and reliable than mine — but they don’t usually send the announcement until they’re certain that a stench event is imminent. I like my algorithm because it gives me 24+ hours advanced notice of the *potential* for stinky times — and I regularly adjust my jogging schedule accordingly.

CAVEAT: My algorithm is NOT 100% accurate. It just gives me a heads-up for when I ought to be extra attentive to real-time data so that I can be ready to adjust my personal actions to protect myself and my family.

Ultimately, there is no simple solution. That’s partly why I’m working on a documentary about it all — to learn what people can do to protect themselves AND to build community and political pressure to make fundamental improvements our region’s air quality and as a result, the public’s health. It takes a movement.