For the past week we have had two monitoring stations running concurrently, our normal site and another one about 300 meters away, The idea is to see how representative the main site is of the surrounding area. So far the results look pretty good! The plot below shows the hourly mean concentration of particulates in the two sites (main site is in red, additional site is in blue). There is one very local pollution event that occurred in just one of the sites (late afternoon on the 15/07/15), but most of the variability in the data is found to be similar in both sites.
Figure: PM2.5 concentration (ug m-3) at our semi-permanent site from 04/06/15 to 26/06/15
Bonfires aside, the first set of measurements are pretty encouraging. This is the first time we’ve used these instruments and they are low-cost monitors designed for indoor use, so we weren’t sure if they would give data that was at all robust. But even at first glance we can see that data shows good continuity; concentrations vary gradually over time, with consecutive measurements giving similar readings. The instrument measures continuously and returns data once every minute, so it’s nice to see that this high time resolution data gives relatively little scatter.
The instrument measures the number of particles per litre cubed, but we’ve converted it to a mass-based metric of micro grams per meter cubed as it’s this unit that’s normally used in air quality studies. We need to be cautious with these values, as the conversion between the different units is really uncertain, I’ll go into more details on that in a later post.
The average PM2.5 concentration at the site over this time period (including the bonfire) is 10 ug m-3.
Figure: PM2.5 concentration (ug m-3) at our semi-permanent site from 04/06/15 to 26/06/15 but with a different scale (scale only goes up to 50 ug m-3)
If we plot the same data, but reduce the scale so we can focus more on the large number of readings of lower concentrations, the day to day variability of the data becomes clearer. Some days are more polluted than others; for example the 8th and 9th of June were pretty clean then concentrations increased from the 11th to the 14th. We need a longer time series to understand what’s causing this variability so we’ll have to do a bit more measuring!
When we first devised the experiment we the idea was to hold a few “sampling days” where our volunteers set up monitors in their gardens, and we’d use this data to examine the concentrations of particulates around Saltaire. But after using the instruments outside for just a few days it has become clear that just measuring on a few individual days will give us data that will be very hard to interpret. The problem is that much of the variability in the concentrations of particulates comes from the meteorology and the air mass history. For example, if the air that we are sampling in Saltaire has arrived in our town after travelling northwards over Paris, London and Birmingham then it is likely to be much more polluted than if it had travelled eastwards from North Atlantic.
To get a handle on this variability we need a continuous measurement site that will run over the summer, this will give us an idea of the variability of the concentrations in a single location and will help us analyse the data taken by volunteers on the different sampling days.
It sounds simple, but there were a few issues we’ve had to overcome to set up a semi-permanent site. Firstly, the instruments aren’t waterproof, if we just put them outside they will be ruined within days. To avoid this we’ve semi-waterproofed the instrument by placing it inside two plastic laundry baskets. It’s not pretty, but it does seem to provide enough shelter from the rain while still letting in enough air to measure the particulate concentrations.
Our other problem is that the instruments only have a battery life of about 5 hours. To start off with, I created an almost continuous time series by running three different instruments in rotation. A kind of tag-team approach to atmospheric measurements! But it was a real hassle, and we have a data gap at 2 am because, devoted as I am, I refused to get up in the middle of the night. So to save my sleep we really needed an external power source. The problem is that I’ve no experience with this; I’m lost when it comes to wires, amps and volts. Thankfully we have support from people at Leeds University that were able to wire up a new power source for us. I even got to help with the soldering!
After consultation with both Saltaire residents and scientists at the University of Leeds, we have identified our project aims. These are a balance between what people are most interested in, and what the scientists think will be feasible within our resources and time frames.
- Investigate the geographical distribution of particulate matter air pollution around Saltaire. Specifically, people suggested carrying out a transect study with monitors spaced an increasing distance away from a main road.
- Calculate exposure profiles for a number of volunteer residents of Saltaire.
- Examine whether the reduction in traffic associated with the school holidays has a significant effect on the concentration of particulates.
- Use state of the art computer models to examine the balance between local, regional, national and international sources of air pollution in Saltaire.
We’ll present the data, both on this web site and also at the Saltaire Festival Science Fair (12 the September) so there will be a chance for anyone to come along and hear the results. Thanks for all your comments and ideas.