Modification of Hilhorst model for saturated extract electrical conductivity estimation of coir using frequency domain reflectometry sensors – A laboratory study

Abstract

Frequency domain reflectometry (FDR) sensors can measure moisture content and electrical conductivity (EC) of growing media, thus providing information for effective water and nutrient management of crops. FDR sensors generally provide the bulk EC (σb) value, which is different from the saturated extract EC (σe) values that most growers and practical studies generally use. To properly estimate σe of media, we modified the previous Hilhorst model that calculates pore water EC (σp) from FDR sensor readings. The 16 coir samples with four moisture and four EC levels were prepared, and their changes in the dielectric constant of bulk medium (εb), σb, and medium temperature from the FDR sensors were recorded and used for analysis. As a result, the εσb=0 (εb when σb is 0) of the coir used in this experiment was 1.64, which was much lower than the εσb=0 suggested for mineral soil (4.1), indicating that the medium-specific εσb=0 should be adjusted. Although σe value estimation through the Hilhorst model with adjusted εσb=0 improved, the values were still unreliable to represent the actual σe values (root mean square error, RMSE = 2.280, R2 = 0.30). To convert the σp derived from the Hilhorst model with adjusted εσb=0 to actual σe, the regression equation was applied as σe = 0.5381 × σp + 1.6493. The adjusted Hilhorst model could provide a more reliable σe estimation (RMSE = 0.929, R2 = 0.62). These results indicate that proper adjustment of εσb=0 for the specific medium is necessary, and adjusting the equation for converting σp into σe would be helpful to properly interpret the σb to practically utilize the FDR sensors for effective water and nutrient management.