Data Verification Tools for Minimizing Management Costs of Dense Air-Quality Monitoring Networks
详细信息 查看全文
- 作者：Georgia Miskell ; Jennifer Salmond ; Maryam Alavi-Shoshtari ; Mark Bart ; Bruce Ainslie ; Stuart Grange ; Ian G. McKendry ; Geoff S. Henshaw ; David E. Williams
- 刊名：Environmental Science & Technology
- 出版年：2016
- 出版时间：January 19, 2016
- 年：2016
- 卷：50
- 期：2
- 页码：835-846
- 全文大小：640K
- ISSN：1520-5851
文摘
Aiming at minimizing the costs, both of capital expenditure and maintenance, of an extensive air-quality measurement network, we present simple statistical methods that do not require extensive training data sets for automated real-time verification of the reliability of data delivered by a spatially dense hybrid network of both low-cost and reference ozone measurement instruments. Ozone is a pollutant that has a relatively smooth spatial spread over a large scale although there can be significant small-scale variations. We take advantage of these characteristics and demonstrate detection of instrument calibration drift within a few days using a rolling 72 h comparison of hourly averaged data from the test instrument with that from suitably defined proxies. We define the required characteristics of the proxy measurements by working from a definition of the network purpose and specification, in this case reliable determination of the proportion of hourly averaged ozone measurements that are above a threshold in any given day, and detection of calibration drift of greater than ±30% in slope or ±5 parts-per-billion in offset. By analyzing results of a study of an extensive deployment of low-cost instruments in the Lower Fraser Valley, we demonstrate that proxies can be established using land-use criteria and that simple statistical comparisons can identify low-cost instruments that are not stable and therefore need replacing. We propose that a minimal set of compliant reference instruments can be used to verify the reliability of data from a much more extensive network of low-cost devices.