The partners in charge of managing the observation systems of the HYPOX project consortium performed a first evaluation of data quality on the data-set acquired during the monitoring and sensing activity. This resulted in a summary with advices and strategies for in situ marine environment monitoring activities in the project and in similar future water cycle monitoring attempts. Although several critical points were identified in specific observatories, a set of common recommendations was also established.
These are the main conclusions:·
- It is very important to pre-test equipment and to resolve all the issues with measurement equipment before deployment. On the other hand, for complex observatory systems with components of various manufacturers, it is very important to have all the equipments within the system properly tested together before the first deployment.
- When designing an observatory system, a monitoring strategy has to be established for such a, based on the number and type of sensors required, frequency of measurements, position of the observatory, total power consumption and mobility of the system, deployment strategy and capability of support vessels and finally maintenance issues.
- Corrosion should be prevented as much as possible, because observatories are deployed for long periods of time. Including enough sacrificial anodes on instruments and frames, and isolating metal parts are recommended actions.
- It is important to monitor as many output parameters as possible, including those not related to water properties. They can be used to detect and explain malfunction and equipment damage.
- Systems should have enough capacity to cope with established measurement frequency. In this context, using an inadequate monitoring frequency can lead to increased power consumption, higher required storage capacity and more hours for data analysing.
- Collected data should be backed up on several independent levels within the observatory system, even for cabled systems.
- To ensure that the data collection software is capable of re-activating automatically the link between the observatory and the base station if communication is broken, and also retrospectively of collecting data stored during the interruption as well as incorporating it into the database. Preferably, it should also include an alarm system.
- Cabled observatories have several advantages such as unlimited power supply, real-time communication and remote control of instruments. Moreover, measurement intervals can be changed and data can be continuously evaluated. However, adding a cable and a base station is expensive and complex. Nevertheless, acabled observatory could also be assured using Uninterruptible Power Supply (UPS) power systems and internal batteries.
- An Automatic data display can provide a quick view to assess the quality of data and the system performance.
- Sensor accuracy should be verified on a regular basis, by using independent calibrated sensors. In situ calibration should be possible if a 2-way communication is available. However, recovery of instruments for calibrating them is encouraged because in most cases signal drift is coupled to biofouling phenomena. So, if possible, regular visits for cleaning and removing biofouling during warm summer months should be planned for the most sensible components, and for devices deployed in warm oxic flora and fauna rich water.
- Verifying data quality should be done automatically in real time (during the process of importing data from the observatories) and in a delayed mode (run by a data curator). Basic items to be checked during automated reviews should be date/time values, pre-defined thresholds and outliers.
Deliverable D1.3: Report on first data quality checks and recommendations for future observation system.