Contents lists available at ScienceDirect Soil & Tillage Research journal homepage: www.elsevier.com/locate/still Calibration of capacitance sensor for Andosol under field and laboratory conditions in the temperate monsoon climate Kassu Tadesse Kassaye a,b, ⁎ , Julien Boulange c , Hirotaka Saito a , Hirozumi Watanabe a a Tokyo University of Agriculture and Technology, Graduate School of Agriculture, 3-5-8 Saiwaicho, Fuchu, Tokyo, 183-8509, Japan b Kulumsa Agricultural Research Center, Ethiopian Institute of Agricultural Research, P. O. Box 489, Kulumsa, Ethiopia c Center for Global Environmental Research, National Institute for Environmental Science, Onogawa 16-2, Tsukuba, Japan ARTICLE INFO Keywords: Calibration Capacitance sensor Andosol (Kuroboku) Soil water content ABSTRACT Capacitance sensors (CS) were developed to continuously measure soil water content using the dielectric properties of soils. Although CS are calibrated in soils of various types, the default calibrated functions may not work for all conditions as in the case of this soil that has high organic matter content, hence, soil-specific calibrations are sometimes required to achieve reasonable accuracy. This study (1) evaluated the suitability of the default calibrated functions of a Decagon 5 T M CS for Andosol (Kuroboku), and (2) derived soil-specific calibration functions to Andosol. A field monitoring experiment was conducted at Sakaecho (western suburb of Tokyo) from August 2016 to July 2017, where volumetric soil water content (θ v ) was measured in bare field and repacked soils using both gravimetric method with oven drying and CS installed at 5 cm depth. The values of θ v monitored using soil cores sampled from the field under natural condition and CS revealed large errors under both field and laboratory conditions when the default factory supplied calibration function (FSC) was used. The deviations depicted significant (P < 0.001) underestimations of the observed θ v measured in soil cores by 0.117 - 0.199 and -0.004 - 0.131 cm 3 cm -3 for field and laboratory conditions, respectively. Accordingly, soil-specific calibration functions were developed by correlating the dielectric permittivity of the soil with θ v measured from the soil core samples. The θ v recalculated based on the site-specific calibration function under field condition best fitted to the observed θ v . Calibration of CS improved the θ v measurement error from 15% with FSC to ≤ ± 2%. Whereas the improvement with laboratory calibration was from ± 15% with FSC to ± 4% when the function was implemented to the field measured data, hence, it still underestimated the observed field θ v . The deviation between the field and laboratory procedures was attributed to the deformation of the well-aggregated soil structure and its consequent changes in hydraulic properties due to crushing when a 2-mm sieve was used for sample preparation. Quasi-field calibration of the 5 T M CS under natural condition is highly recommended for real-time monitoring of θ v in Andosol. In cases when field calibration is impractical, laboratory calibration further verified with field data could also offer a reliable method for calibration of the 5 T M CS for Andosols. 1. Introduction Soil water in the vadose zone is a key variable in agriculture (Chen et al., 2014) as it regulates the partitioning of rainfall into infiltration, surface runoff, and evaporation (Castillo et al., 2003; Kutilek and Nielsen, 1994; Shein, 2010). Consequently, soil water directly impacts the amount of water available to plants (Stephenson, 1990) but also the Earth’s energy balance by influencing the conversion of incoming solar and atmospheric radiation into radiant, latent, and sensible heat losses (Henderson-Sellers, 1996; Zeng et al., 2009). Hence, accurately de- termining soil water content is highly desirable for various purposes including the optimization of agricultural practices and for calibrating and validating soil water models (Bircher et al., 2016; Geesing et al., 2004; Parvin and Degre, 2016). The agricultural sector alone was highlighted to be responsible for water withdrawal up to 80% of the world’s freshwater resource (Kinzli et al., 2012). In addition, agri- culture not only consumes a large volume of fresh water, but also uti- lizes that water rather inefficiently mainly due to under- or over- watering, which consequently results in reduced crop yields, leaching of nutrients and pesticides from the root zone to deeper soil layers and ultimately pollution of the ground water (Spelman et al., 2013). The precise knowledge of spatiotemporal soil water content using accurate measurements is the first step to efficiently use water in agriculture. Furthermore, hydrological, meteorological, and modeling studies https://doi.org/10.1016/j.still.2018.12.020 Received 7 June 2018; Received in revised form 11 December 2018; Accepted 20 December 2018 ⁎ Corresponding author at: Tokyo University of Agriculture and Technology, Graduate School of Agriculture, 3-5-8 Saiwaicho, Fuchu, Tokyo, 183-8509, Japan. E-mail address: kasstad96@yahoo.com (K.T. Kassaye). Soil & Tillage Research 189 (2019) 52–63 0167-1987/ © 2018 Elsevier B.V. All rights reserved. T