Automatic screening for common issues of microclimate time-series

[manmatej]

Aim: Establishing a Function for Automatic Screening of Microclimatic Time Series for Common Issues

Common issues to address include:

• Unintended data recorded before the field deployment or after the field mission ends.
• Loss of sun shielding, resulting in overheating.
• Changes in logger position due to factors such as animals, weather, or disturbance.

Logger-oriented screening can involve:

• Detecting sudden jumps in the time series. This can be achieved by identifying significant deviations above a specified threshold within a particular period. The function should accept arguments such as threshold (absolute value or percentile) and period. Alternatively, a running mean could be used for comparison.
• Comparing the amplitude of air/soil temperature/moisture within a specific period for loggers with multiple sensors. The function should accept arguments such as period and an amplitude difference threshold.

Loggers group-oriented screening involves:

• Screening for outliers among loggers within a particular study area or loggers group, aiming to identify deviations from the common variability.
• Screening for outliers based on a certain reference, such as a selected logger, a meteorological station, or climatic reanalysis based on coordinates from myClim locality metadata. Climatic station or reanalysis approach requires advanced technical capabilities and introduces additional dependencies. It is primarily suitable for rough screening when the loggers are significantly inconsistent, such as freezing temperatures in temperate summers or warm temperatures in boreal winters.

Additional considerations:
Some functionalities related to anomaly detection have already been implemented in myClim but are conceptually aligned with the topic. These include:

• mc_prep_clean: Identifying missing records (incomplete time series), detecting incorrect time series order, and identifying duplicated records.
• mc_calc_snow: Detecting events when the logger is under the snow, leveraging temperature sensor data.
  Please review the above information and provide feedback or suggestions for the development of the automatic screening function.

---

It would be useful to have the function returning the summary of issue detection. How many of loggers had issues, which logger had issues of which type, how many records were affected by issues.