Using time domain reflectometry to estimate water content of three soil orders under savanna in Brazil

Abstract Water is a limiting factor in agricultural lands on tropical Savannas. Due to low rainfall and long periods of drought, irrigation becomes a critical factor for increasing productivity. In this scenario, improving irrigation management techniques is crucial for increasing water use efficiency. Soil water content sensors are important tools not only for soil-based irrigation scheduling, but also to calibrate weather-based irrigation scheduling. Time domain reflectometry (TDR) type sensors are an important tool for water content monitoring due to its potential accuracy, automation and real-time measurement capability. The objective of this study was to evaluate the TDR technique for the estimation of water content in five soils of the Brazilian savanna (two Oxisols, two Entisols and one Histosol) using polynomial, square root and logarithmic calibration equations. The influence of soil bulk density, organic matter and clay content, and presence of magnetic materials were also examined. Magnetic permeability was estimated from magnetic hysteresis curves for each soil. The equations were calibrated and validated on an independent data set. Validation was evaluated using the Coefficient of Determination (R2), Root Mean Squared Error (RMSE), Nash-Sutcliffe Efficiency and the RMSE to the standard deviation of observed data ratio. There was no benefit of including bulk density, clay content or organic matter on the calibration equations. The relative magnetic permeability was very close to one in all soils and thus did not influence the calibration. For the soils evaluated in this study a single logarithmic calibration equation calibrated for all samples was the best predictor of volumetric water content.