Automatic near real-time quality control tests for biofouling effect on measurement data

The reliability of water quality measurement is crucial for sustainable use of ocean resources, climate and ecosystem models, and industrial applications. However, measurement stations in remote locations face limitations in terms of power, communication, and maintenance, posing challenges for data quality. Even though some basic (near) real-time automatic tests are proposed in oceanographic measurement guidelines, time- and resource consuming Delayed Mode Quality Control is still required before using measurement data in forecasting, models, or decision- making. To design effective quality control tests for more autonomous sensors with self-validating capabilities in real-time, a good understanding of expected environmental effects or errors on sensor signals is necessary. This paper focuses on the effect of biofouling on the measurement of selected water quality parameters such as conductivity, oxygen, and turbidity. Biofouling remains a major issue despite research on biofouling protection and anti-biofouling sensor design. Biofouling growth on underwater sensors can increase measurement errors and uncertainty, result in shorter operation times, and require costly manual work related to retrieval, cleaning, and re-deployment. For some measurement technologies, biofouling can result in noise, while for others, it may cause systematic drift or delay signal exchange. Here, we propose quality control tests designed to automatically detect and assess the impact of biofouling on sensor signals. These tests are applied to measurement data sets with a known presence of biofouling from Austevoll (Norway). We comment on the challenges of designing tests and setting adequate thresholds. We show that a detailed understanding of biofouling effect on sensors is crucial for designing effective near real-time quality control procedures. Automatic, in-situ tests can save costs related to manual data quality control and increase data quality, thereby enabling well-informed decisions in ocean resource management, climate and ecosystem modeling, and industrial applications.