A synergistic collision of technologies is offering a new chance to create air quality networks at an affordable cost. Lower power, lower cost GPS plus GSM networking via mobile phone network are combining with improved low cost gas sensors that sense gas concentrations at parts-per-billion (ppb) levels. The result? Low cost air quality networks that can monitor urban air quality in real time.
Air quality is fast becoming a key concern in many people’s everyday lives. This increased interest in personal exposure and health is now beginning to be matched by the ability to measure at personal scales. The impacts of air quality on health have been recognised with legislation at national and regional levels. This top down governmental activity is being met by a groundswell of interest from individuals, and sensor networks are starting to catch up. We can change the environment, but we first need to understand it.
Air quality in context
Air quality means different things to at-risk people. Compounds classed as pollutants by legislative bodies in the UK 1 include sulfur dioxide, nitrogen oxides, particulate matter, ozone, volatile organic compounds, carbon monoxide and heavy metals, such as lead. This is a broad sweep of species with a range of different sources. Some of these species are mainly associated with industrial sources whilst we are more likely to encounter others on a day to day basis, for example as we travel in the urban environment.
Local air quality has long been of interest to those who live in areas of high obvious pollutant levels such as in mega cities around the world 2. As we begin to be able to collect more information on the environment, so we need more ways of understanding and interpreting this information. We also need to understand how the data relates to the individual.
With this in mind, Cambridge University has been developing low cost, portable sensors for new air quality networks as part of the MESSAGE (Mobile Environmental Sensing System Across Grid Environments) project 3. Software for real-time analysis and dissemination has also been developed as we exploit how this information is part of what makes these sensor networks compelling. Ultimately users should be able to both report and receive information on a given area via personal devices such as 3G or 4G mobile phones, ‘right now’.
Where to start?
Good urban air quality is a noble goal. But what makes good air quality? Our environment is polluted by industry, buildings and transport. To determine what are the “to watch” pollutants, many studies over the years have measured urban air pollutants throughout the world and there is consensus that certain pollutants are the greatest problem to “at risk” members of our society.
The gases of greatest concern are the oxides of nitrogen (nitric oxide and nitrogen dioxide, NO and NO2 respectively, collectively NOx) and carbon monoxide (CO). These gases are of particular interest as they are primarily from traffic sources which are the main daily sources of pollution.
“many studies over the years have measured urban air pollutants throughout the world and there is consensus that certain pollutants cause the greatest problem”
For the main part, NOx and CO are mainly associated with combustion and we are exposed to them as part of our daily lives. Put another way: we are all affected by these gases and most of us are also responsible for them. With an increasingly well informed and urbanised (globally) population, people are beginning to worry about the air they are breathing.
Impacts on health
There are historic correlations between air quality and health. There is data which correlates SO2 (a gas that affects lung function) concentrations in the smog of the early 1950s and weekly mortality. NOx and CO have a range of health effects, especially for those who have pre existing conditions. NOx affects respiratory function and CO affects the pulmonary system.
These effects are thought to be greater with the vulnerable groups of the population, such as the young (where the lungs for example are still developing), the old (where function may be impaired) and those with underlying health problems. NO2 particularly sensitises some people to asthmatic episodes and CO is already well known indoors as a ‘silent killer’. CO is rarely at highly toxic concentrations in the average street but can be dangerously elevated. Ozone (O3) is also a well known urban air problem, whose chemistry is closely related with NOx and can react with many other chemical species and lead to breathing problems.
“with an increasingly well informed and urbanised (globally) population, people are beginning to worry about the air they are breathing”
What is good air quality?
Air quality is regulated by EU air quality directives which have been implemented in the UK via the 2007 Air Quality Standards regulations 4. Concentrations are set over a given time period which should either not be exceeded or not be exceeded more than a given number of times in the year.
Councils must take air quality into account when making planning decisions and declare when areas exceed any of these species. These declarations have implications in terms of mitigating actions and future developments: there are serious cost and planning implications.
Where we are now
Monitoring is undertaken across the UK by a range of networks. The main network is the UK Automatic Air Quality Network 6 (AURN). This network is dispersed across the UK and consists of 126 static stations, monitoring NOx, SO2, CO, O3 and particulate matter. These stations are located in both rural and urban places and usually report the data hourly. This network provides information to both local authorities and the wider public. These traditional networks are based around large static sensor installations which utilise techniques such as chemiluminescence and infra red spectroscopy. Currently there is no data with sufficient temporal and spatial resolution to quantify pollutants in our air as it varies over very short scales of both space and time.
The future: high resolution sensing
We have developed both the sensors and the network to fit into the wider sensing infrastructures to provide a greater spatial and temporal resolution, complementing the long term data collected by Councils and their government bodies.
“atmospheric modelling data needs to be of a sufficient coverage and at suitable intervals for computerised proxies to mimic real world conditions”
In the Cambridgeshire area there are five static sites, one of which is an AURN site. These sites measure NO2, CO, O3 and particulates (but not all measurements are made at all sites) and are taken to be representative of an area of 40.7 km2 and a population of over 110 thousand 7. The data from these sites (in conjunction with diffusion tube monitoring and traffic data) is used as the basis of modelling, which is used to generates maps of air quality gases over the whole area. These maps are then used to estimate levels of pollution, for example at a site of new development before and after construction.
As with all atmospheric modelling, the quality of the outputted products depends on the utility of the inputted data. This real world data needs to be of a sufficient coverage and at suitable intervals for computerised proxies to mimic real world conditions. If data with finer resolution in space is used, then the models can be run more accurately over smaller grids. Similarly with time, if the data is at too coarse a time step, then the model may not accurately follow trends, especially short term events such as school runs.
New networks
As part of the MESSAGE project and in partnership with Alphasense 8 who have incorporated some of their state of the art gas sensors, the University of Cambridge has developed a range of real time mobile air quality monitors suitable for sensing in the urban environment.
The units measured CO, NO and NO2 and are designed for use by the general public. Traditional static sites as used in the AURN network are prohibitively expensive and very large (many kilograms plus powersupplies- certainly not portable).
The University of Cambridge hand-held sensor are relatively cheap (hundreds, not thousands of pounds) and weigh less than half a kilogram. These sensors collate gas concentrations with time and position information via GPS and send these data over existing GPRS phone networks to a server archive in the Mathematics Department for assimilation with other data feeds and further processing. This data can be accessed via the web, allowing,in principle, anyone to access air quality information whether from home, or on the move.
One alternative type of sensor sends data directly to a linked mobile phone over a Bluetooth connection before sending onto the server, allowing the live display of pollution information. These sensors are designed to complement the existing networks by providing data in areas where there would otherwise be no data.
Test run
In the summer of 2009, approximately forty of these sensors were deployed in Cambridge. The sensors were deployed using volunteers on foot, on bikes and in cars. The purpose of this project was to increase the area over which the study was undertaken and to look into the main avenues by which people get around the city of Cambridge.
The information collected is still being sifted (over 120, 000 data points over any 3.5 hour period), but one of the most eye-catching outcomes was a family of maps simply showing concentrations of species during this period. This data highlights a potentially new approach to air quality monitoring. At this stage it is important to note that the data is a ‘snapshot’ of the urban environment over the deployment period. Direct comparison with computer predictions of air quality (which tend to be annual or monthly averages) is not yet appropriate, but computer model verification is under way.
In the plot above, pollutant concentrations are shown by height and we see the NOx data collected from this test study (NO2 in green and the NO in blue). The first thing conclusion is that there are high concentration hotspots. These are broadly in the expected areas of increased traffic numbers or reduced traffic flow. These concentrations vary depending on a range of factors. Knowing the location of these hotspots can be used to feedback, for example, into traffic management or to select low pollution routes at certain times of day.
This plot shows the complexity on very small scales of these urban environments. It also shows that these sensors are relatively easy to deploy. This brings out the intricacy of the system and highlights how changeable concentrations are and how traditional static sites may not be representative beyond their immediate area.
What next
The mobile air quality sensor network developed by Cambridge University and Alphasense are the result of combining components which have become available as a result of the telecommunications boom coupled with advanced sensing technologies. This amalgamation of different technology streams to provide sensitive low cost sensor networks has been very successful. These units are simple to operate and are designed for use by the general public. At this stage they are still only advanced prototypes and a full validation is under way. These mobile pollution monitors have proven to be highly reproducible with multiple species measurement and they allow for high spatial resolution mapping of the hugely complex urban air quality picture.
The urban composition is highly structured, variable and interdependent and is highly dependent on transport modes; more sophisticated analysis methodologies are needed, especially with regard to other species such as particulates. This preliminary study shows that much more information is out there to be harvested.
From this we conclude that making the step from regional air quality sensing to personal monitoring is key and will be coming soon.
References
1 Under EU daughter directives.
2 Those cities with populations greater than 10 million.
3 http://bioinf.ncl.ac.uk/message
4 http://www.airquality.co.uk
5 http://www.airquality.co.uk
6 http://www.bv-aurnsiteinfo.co.uk
7 http://www.cambridge.gov.uk:“Cambridge has a population of 117,700 people (County Research Group mid-2008 estimates)”
8 http://www.alphasense.com
Dr. M. I. Mead, Research Associate. University of Cambridge.
Prof. R. L. Jones, Professor of Atmospheric Chemistry. University of Cambridge.
Dr. J. Saffell, Technical Director, Alphasense.
Rod Jones is Professor of Atmospheric Science at the University of Cambridge, Department of Chemistry with over 25 years of research experience in atmospheric observations, numerical modeling and the developments of novel measurement and sensor techniques.
Dr. Mead is a postdoctoral research worker with over 5 years research experience in instrumentation development atmospheric measurements, and data interpretation. This includes boundary layer measurements and aircraft instrumentation.
Dr John Saffell has been Technical Director of Alphasense Ltd since it was founded in 1997. He has been involved in gas sensing and water quality monitoring for 30 years. He is Chairman of the Council of Gas Detection and Environmental Measurement (CoGDEM) and previous chairman of Sensors for Water Interest Group (SWIG) and works with the Technology Strategy Board, advising on UK sensor stretegies.
Acknowledgements
Cambridge Mobile Sensor Team
Rod Jones, Peter Landshoff, Iq Mead, Mark Calleja, Mark Hayes, Lekan Popoola, Gregor Stewart, Matt McLeod, Tom Hodgson, Jose Baldovi-Jachan, Ray Freshwater, Eiman Kanjo, Michael Simmons
For more information please go to http://www.osedirectory.com/environmental.php
Published: 01st Mar 2010 in AWE International