Low-cost sensors for the measurement of atmospheric composition: overview of topic and future applications

Measurement of reactive air pollutants and greenhouse gases underpin a huge variety of applications that span from academic research through to regulatory functions and services for individuals, governments, and businesses. Whilst the vast majority of these observations continue to use established analytical reference methods, miniaturization has led to a growth in the prominence of a generation of devices that are often described generically as “low-cost sensors” (LCSs). LCSs can in practice have other valuable features other than cost that differentiate them from previous technologies including being of smaller size, lower weight and having reduced power consumption. Different technologies falling within this class include passive electrochemical and metal oxide sensors that may have costs of only a few dollars each, through to more complex microelectromechanical devices that use the same analytical principles as reference instruments, but in smaller size and power packages. As a class of device, low-cost sensors encompass a very wide range of technologies and as a consequence they produce a wide range of quality of measurements. When selecting a LCS approach for a particular task, users need to ensure the specific sensor to be used will meet application’s data quality requirements. This report considers sensors that are designed for the measurement of atmospheric composition at ambient concentrations focusing on reactive gaseous air pollutants (CO, NOx, O3, SO2), particulate matter (PM) and greenhouse gases CO2 and CH4. It examines example applications where new scientific and technical insight may potentially be gained from using a network of sensors when compared to more sparsely located observations. Access to low-cost sensors appears to offer exciting new atmospheric applications, can support new services and potentially facilitates the inclusion of a new cohort of users. Based on the scientific literature available up to the end of 2017, it is clear however that some trade-offs arise when LCSs are used in place of existing reference methods. Smaller and/or lower cost devices tend to be less sensitive, less precise and less chemically-specific to the compound or variable of interest. This is balanced by a potential increase in the spatial density of measurements that can be achieved by a network of sensors. The current state of the art in terms of accuracy, reliability and reproducibility of a range of different sensors is described along with the key analytical principles and what has been learned so far about low-cost sensors from both laboratory studies and real-world tests. A summary of concepts is included on how sensors and reference instruments may be used together, as well as with modelling in a complementary way, to improve data quality and generate additional insight into pollution behaviour. The report provides some advice on key considerations when matching a project/study/application with an appropriate sensor monitoring strategy, and the wider application-specific requirements for calibration and data quality. The report contains a number of suggestions on future requirements for low-cost sensors aimed at manufacturers and users and for the broader atmospheric community. The report highlights that low-cost sensors are not currently a direct substitute for reference instruments, especially for mandatory purposes; they are however a complementary source of information on air quality, provided an appropriate sensor is used. It is important for prospective users to identify their specific application needs first, examine examples of studies or deployments that share similar characteristics, identify the likely limitations associated with using LCSs and then evaluate whether their selected LCS approach/technology would sufficiently meet the needs of the measurement objective. Previous studies in both the laboratory and field have shown that data quality from LCSs are highly variable and there is no simple answer to basic questions like “are low-cost sensors reliable?”. Even when the same basic sensor components are used, real-world performance can vary due to different data correction and calibration approaches. This can make the task of understanding data quality very challenging for users, since good or bad performance demonstrated from one device or commercial supplier does not mean that similar devices from others will work the same way. Manufacturers should provide information on their characterizations of sensors and sensor system performance in a manner that is as comprehensive as possible, including results from in-field testing. Reporting of that data should where possible parallel the metrics used for reference instrument specifications, including information on the calibration conditions. Whilst not all users will actively use this information it will support the general development framework for LCS use. Openness in assessment of sensor performance across varying environmental conditions would be very valuable in guiding new user applications and help the field develop more rapidly. Users and operators of low-cost sensors should have a clearly-defined application scope and set of questions they wish to address prior to selection of a sensor approach. This will guide the selection of the most appropriate technology to support a project. Renewed efforts are needed to enhance engagement and sharing of knowledge and skills between the data science community, the atmospheric science community and others to improve LCS data processing and analysis methods. Improved information sharing between manufacturers and user communities should be supported through regular dialogue on emerging issues related to sensor performance, best practice and applications. Adoption of open access and open data policies to further facilitate the development, applications, and use of LCS data is essential. Such practices would facilitate exchange of information among the wide range of interested communities including national/local government, research, policy, industry, and public, and encourage accountability for data quality and any resulting advice derived from LCS data. This assessment was initiated at the request of the WMO Commission for Atmospheric Sciences (CAS) and supported by broader stakeholder atmospheric community including the International Global Atmospheric Chemistry (IGAC) project, Task Force on Measurement and Modelling of the European Monitoring and Evaluation Programme of the LRTAP Convention, UN Environment, World Health Organization, Network of Air Quality Reference Laboratories of the European Commission (AQUILA).