A Smart Sensor Network for near Real Time Data Quality Flagging and Archiving of Environmental Data

Abstract:

Large interdisciplinary teams of environmental scientists, especially those conducting field research, generate diverse datasets using a range of monitoring instruments often in remote regions. While it is of key scientific importance to generate high-resolution spatial and temporal data for a range of environmental measurements, it is equally crucial to make sure that the data being generated are of a high quality and free from errors due to human operation, environmental conditions, and other accidental occurrences. Here we discuss our vision for developing a smart sensor network which could be used in monitoring field data for near real-time flagging of the data according to quality indicators. A smart sensor network would add assurance metrics to data gathered from sensors in harsh environments (e.g. Arctic, boreal, and tropics) and promote intelligent archiving. Field operations come with a high cost, and improved processes could significantly improve data quality and reduce overall operational outlays. We discuss plans to monitor the range of data collected from different sensors in the field and generate quality metadata in near real-time to reduce the cost of field operations and minimize uncertainties in error propagation due to poor quality data, field operation, etc. Our work will have implications for field research programs and other data-intensive monitoring systems.