Upper washita river experimental watersheds: data screening procedure for data quality assurance

Area of Science:

• Hydrology
• Environmental Science
• Data Science

Background:

• Long-term hydrologic observation networks can exhibit non-stationarity due to natural or human-induced factors, impacting data reliability.
• Changes in systematic errors within these networks can significantly bias hydrologic modeling and research findings.
• Identifying the drivers of network operation changes is crucial for accurate hydrologic analysis.

Purpose of the Study:

• To apply a data screening procedure to USDA-ARS experimental watersheds data to detect and identify non-stationary conditions and their drivers.
• To assess the impact of changes in systematic errors on hydrologic data.
• To make processed data publicly available for future research.

Main Methods:

• Utilized the SPELLmap application for data processing and statistical analysis of over 1000 time series.
• Applied statistical tests including Anderson test (independency), Spearman test (monotonic trend), Pettitt test (change point), and split record tests.
• Investigated trends in mean and variance for key climate and soil variables.

Main Results:

• Statistically significant monotonic trends and changes in mean and variance were detected in annual maximum air temperature, rainfall, relative humidity, solar radiation, and soil temperature.
• Network operation changes (e.g., calibration, sensor upgrades) and regional weather trends were identified as potential drivers.
• Processed data, including detected spurious data and filled missing data, were made publicly available.

Conclusions:

• A robust data screening procedure is essential for identifying changes in systematic errors and false non-stationarity in hydrologic data.
• Accurate data processing and validation are fundamental prerequisites for reliable hydrologic modeling and research.
• Understanding network operation drivers is key to maintaining data integrity in long-term hydrologic studies.