Air Quality Forecast – App Comparison

I am the kind of person that likes things organized and always planned. During the pandemic, I took up hiking as a new hobby. I spend a lot of time indoors working in front of the computer, so I decided that it was time to find an activity in nature where I will be able to look as far away as possible in order to extend my view field and exercise my eyesight. The location I live allows me to do that hobby reality but air quality is not always optimum.

As you may know, I live in southern Spain and although traffic pollution is an issue, when I hike in the hills there is a different type of pollution that may be present, the Saharan Dust. In 2021 I have noticed that cross-border pollution events are more frequent. There are no constant warnings from local media or local authorities about these kinds of events. Newspapers or TV news will only mention them only when there is nothing else to talk about.

So when I want to go outside and practice that sport, I check not only the weather forecast but the air quality forecast from various apps and web services. In this article, I will compare three air quality services and apps that I mostly use and see if they offer an accurate forecast. The apps are eLichens Air, Plume Labs, and Windy.

The method is simple. The night before hiking, I check the AQ forecast with these three apps, and then the next day I check the actual AQ measurements from my sensors and the apps again.

Friday Night 7-May-2021 | Forecast

Saturday Morning 8-May-2021 | Real-Time

Here is the first comparison. At night of Friday 7th May, I checked the AQ for the next morning at 11am. The forecast data predicted by the Plume Labs and the eLichens Air show an index value and not a mass concentration (μg/m3) like Windy does. In most cases companies use a proprietary AQI and in order to be able to compare them I will need to “decrypt” them. I wasn’t able to find any documentation for the eLichens Air AQI.

Plume Labs gave me an AQI 21 for PM2.5 which according to their documentation corresponds to a value below 25μg/m3. eLichens Air gave me an AQI 16 for PM2.5 which I estimate it corresponds to a value below 12μg/m3. Windy estimated 8μg/m3.

The following day, I checked the measurements at the exact time like the night before plus the real-time values from my outdoor sensors. PM2.5 values were at 5μg/m3 and the eLichens Air, Plume Labs, and Windy gave me almost the same values of 9μg/m3 and 17/18 AQI which are roughly the same.

Saturday Night 8-May-2021 | Forecast

Sunday Morning 9-May-2021 | Real-Time

The second test revealed similar PM2.5 forecast values (20 AQI) between the Plume Labs and the eLichens Air and an estimated value of 14μg/m3 from Windy.

The following morning, the real-time values from the apps were far different from the in-situ values as my outdoor sensor measured 4μg/m3 and the apps gave me a value of 15 AQI and 16μg/m3.


In their defense, there are not lots of available data here because our local AQ station doesn’t share openly PM2.5 data. So they work with models that are not able to give better results. In different parts of the world where spatial coverage is better and official stations share PM2.5 values with the public, they have better results. Some apps rely on satellite imagery, but again satellites pass over Spain once a day and they have a 3.5km by 5km resolution like the sentinel-5 precursor (my favorite so far). I take them into account to make a rough estimation of the following day and of course, I check the data from my local outdoor sensors.


9 thoughts on “Air Quality Forecast – App Comparison

  1. Consider the apps use data from legal AQM mostly. and not low cost ones, which are less reliable.
    better to perform this test on an AQM location.


    • Legal AQM are hard to be found. Low-cost sensors are not as accurate but we shouldn’t be obsessed with accuracy because we will lose the battle. Data from Low-cost sensors when relative humidity is in between acceptable ranges offer more than ok results


      • Maybe. It depends on what level of accuracy you need. Is ±10% of reading good enough? ±5%? 20%?

        As well, what manufacturers claim and reality can be very different. AWAIR v1 gave wildly inaccurate PM2.5 measurements.

        uRAD A3 gives very inaccurate CO2 due to lack of proper temp compensation. We’re seeing readings off by over 200%.

        Foobot gives inaccurate CO2 due to lack of a proper sensor and inaccurate PM due to poor light mgmt. In brighter indoor light it’s off by over 300%.

        These products are all totally useless. Or actually harmful if people think they are relatively accurate and rely on them.

        Liked by 1 person

  2. I found that these apps are not very much reliable in forecasting the air pollution in Athens, Greece as well. Especially Plume Labs. And this is a big city with several “official” AQ monitoring stations from which to get pollution data.

    There is also the Breezometer App that does a forecast, and it is appealing because it does it at a very local level, street by street, and it actually can work well during the warm season. But this too really fails at forcasting PM pollution in winter, because it can’t “see” the local wood smoke or any kind of pollution coming from heating systems. So often in winter it tells you that the air is very clean, when instead it is full of wood smoke! And Plume labs does that too.

    The Windy App works well in my opinion only when there is Saharan dust reaching Greece. My AQ sensors on the balcony tell me the real dust concentrations, and they are usually not so much different from those on the Windy app, so it seems that this app works well to predict this type of Air Pollution. But the app doesn’t predict well local Air Pollution though, be it from traffic or heating systems in winter. Most of the times it is totally far from the real measurements that I can see from my AQ sensors.

    The bottom line for me is that these apps still don’t work well at predicting the kind of pollution that matters most to us, which is the local one. For that we do need our own sensors, or a good web of “official” sensors, which do not exist yet. Hopefully in the future.

    Liked by 1 person

  3. AQ forecasts have not been a big focus for me though we do somewhat pay attention to them. Both EPA Air Now and Air Visual (IQAir) seem relatively accurate for where we live though both miss by quite a bit on occasion. The misses are almost always that actual is worse than forecast though there have been a very few times where the air was cleaner than forecast.

    We are fortunate that the data from our public gov monitors ( is all readily available as is data from a number of others including schools and businesses. Some caution must be used with the latter as monitors are not always very accurate and are sometimes placed in locations that don’t provide accurate info.

    Liked by 1 person

      • We’ve had some record setting poor AQ the past couple of weeks due to wildfires in Canada. My quick take is the Plume Labs is totally useless. I love their interface but their data is often off and their forecasts are not even close. They’ll forecast improving AQ when others forecast declining. Plume is much more likely to be wrong than others.

        BreezeoMeter seems the most accurate both for current conditions and forecast.

        Liked by 1 person

  4. Hello Sotirios,

    What is interesting to me is small particle ( < 2.5µm ) as an indicator of other problems. Being down wind from an oil refinery will result in petrochemical particles. Families near large shipping ports will have diesel smoke. People living near an agricultural area might have small particle readings from substances that are not harmful at all. Etc

    I think it is really important to reference this environmental factor when reporting small particle readings. What is the upwind source ?

    From: See The Air
    Reply-To: See The Air
    Date: Tuesday, July 6, 2021 at 2:32 AM
    To: Dave Clark
    Subject: [New post] Air Quality Forecast – App Comparison

    Sotirios Papathanasiou posted: ” I am the kind of person that likes things organized and always planned. During the pandemic, I took up hiking as a new hobby. I spend a lot of time indoors working in front of the computer, so I decided that it was time to find an activity in nature wher”


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s