Magnitude, consequences and correction of temperature-derived errors for absolute pressure transducers under common monitoring scenarios

Document Type

Article

Publication Date

12-19-2021

Department

College of Forest Resources and Environmental Science

Abstract

Continuous water level monitoring using absolute pressure transducers with onboard data logging is common practice in hydrologic studies. While there has been some discussion and study of temperature-derived error (TDE), there has not been a systematic evaluation of the problem. We sought to answer three questions: (1) are current best practices enough to avoid these errors, (2) can laboratory correction be used to correct field data from varying conditions, and (3) what is the scale of the additional uncertainty of the correction procedure? We evaluated the magnitude of such errors under laboratory conditions that mimicked common monitoring scenarios. Using field data, we also demonstrated the impact of TDEs on calculated daily mean water level and diurnal signal decomposition to estimate evapotranspiration (ET). To address instrument and model uncertainty, we fit 1000 possible correction models using a double-bootstrap approach. Correction models fit expected error as a function of water and air temperature and rate of change of air temperature. TDEs were a significant source of error, resulting in recorded data outside of manufacturer-stated instrument uncertainty, with 45% of bootstrap models showing significant but small TDEs under best-practice deployment. Correction equations did introduce additional error, often on a much smaller scale than instrument uncertainty. When tested against a validation data set, correction equations effectively reduced total measurement uncertainty below instrument uncertainty by up to 65%. The effects of TDEs on case-study field data resulted in 56% of daily mean values outside of instrument error bounds (errors: −1.5 to 4.2 cm). Our results suggest that a single laboratory correction equation can be used across monitoring scenarios, though we suggest matching deployment conditions as closely as practical. Identification and correction of TDEs is essential to avoid erroneous conclusions, downstream analyses and water resources management.

Publication Title

Hydrological Processes

Link to Full Text

Share

COinS