Propagation of soil moisture sensing uncertainty into estimation of total soil water, evapotranspiration and irrigation decision-making

• Published in: Agricultural water management Vol. 243; p. 106454
• Main Authors: Sharma, Kiran, Irmak, Suat, Kukal, Meetpal S.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.01.2021
• Subjects: Capacitance; decision making; Electrical resistance; Evapotranspiration; Irrigation; loamy sand soils; rhizosphere; Sensors; silt loam soils; Soil moisture; soil water; Time-domain reflectometry; uncertainty; water management; Irrigation; Soil moisture; Time-domain reflectometry; Evapotranspiration; Electrical resistance; Capacitance; Sensors
• ISSN: 0378-3774; 1873-2283
• DOI: 10.1016/j.agwat.2020.106454

•Accurate reporting of total soil water (TSW) and crop ETc and irrigation triggers (IT) essential.•Evaluated 9 SM sensors in silt loam & loamy sand soils to report TSW, ETc, and IT.•Sensor choice impacts these metrics largely based on soil type and calibration.•Across all sensors and metrics, estimation improved 54 % after site-specific calibration.•Quantified uncertainty in TSW, ETc, and IT associated with all sensors’ use. Soil moisture sensors are subject to uncertainty (inaccuracy) in measuring soil water status, that hinders various applications. User groups (researchers and growers/advisers) rely on these sensors for estimating critical agricultural water management decisions and information such as total soil water in the crop root zone (TSW), crop evapotranspiration (ETc) and predicting irrigation triggers (IT), i.e., when TSW is equal to or lower than readily available water. There is a lack of translation of errors in sensor-reported soil moisture (θv) into TSW, ETc, and IT, which is critical to farm-level decision-making as well as research assessments. Nine soil moisture sensors (based on principles of time-domain reflectometry, capacitance and electrical resistance) were investigated in field conditions for silt loam and loamy sand soils under two installation orientations (vertical and horizontal) during two growing seasons (2017 and 2018). Accurate representation of TSW, ETc, and IT was found to be a function of sensor-type, soil-type as well as calibration-type [factory calibration (F.C.) vs. site-specific calibration (S.S.C.)]. Sensor installation orientation did not affect sensor accuracy. Uncertainties in estimation of TSW, ETc and IT were quantified under each condition of use, and sensors were comparatively ranked for effective selection. It was found that all sensors underestimated ETc in silt loam soil. The deviation of sensor-measured ETc from true ETc ranged from −14 to −31 %, which implies that the choice of sensor under a given soil type impacts the quantification of consumptive use of the soil-vegetation system being monitored. Sensors showed both overstimation and underestimation of ETc in loamy sand soil with deviations of sensor-estimated ETc from true ETc ranging from 14 to −61 %. The S.S.C. resulted in 45 and 17 % improvement in TSW and ETc in silt loam soil, respectively, and 42, 80 and 86 % improvement observed in TSW, IT and ETc in loamy sand soil, respectively. The research findings showed that suitability of soil moisture sensors can differ when different target metrics are used as criteria. These findings emphasize the need for evaluating soil moisture sensors based on practical and application-oriented criteria, in addition to reliance on θv accuracy. To the best of authors’ knowledge, this research is the first to translate traditional θv accuracy assessments into practical and application-oriented criteria and use them to evaluate sensors for these specific applications. Sensor rankings and uncertainty associated with their use presented here will allow diverse users to effectively identify sensors for targeted applications in water management decision-making and research.