An open-source workflow for identifying hydrodynamic water quality events in rivers by continuous water quality monitoring and time-series data processing using R and US EPA CANARY

Abstract

Improving European surface water quality requires urgent action to address diffuse pollution sources particularly from agriculture, with increased frequency and intensity of hydroclimatic events also a key driver of pollutant export to waters and water quality decline worldwide.

However, the need for comprehensive, practical protocols for sensor deployment, sensor maintenance and data management for the adoption of high frequency water quality monitoring has been highlighted, along with the challenges for citizen scientists in analyzing millions of water quality data points and sharing metadata. The practical method presented, with reproducibility built into the workflow, is designed for multiple users and a step-by-step application of the workflow is demonstrated including:

• •
  Deployment arrangement for water quality sondes in two temporary monitoring stations with different site characteristics.
• •
  Data collection and data validation methods.
• •
  Concise, reproducible, open-source workflow detailing the use of R, R markdown and US EPA CANARY software for data import, data cleaning, data visualization, data integrity, along with site-specific CANARY event system configuration for the detection of potential water quality events.

Results for two monitoring stations on different rivers show CANARY successfully identified 100 % (n 47) and 97 % (n 39) of the manually identified turbidity events.