Validation of analytical algorithms for the estimation of soil thermal properties using de Vries model

Abstract

Soil thermal conductivity, and diffusivity together with the damping depth of soil temperature computed using Amplitude decay, Phase shift, Harmonic (amplitude based and phase based), Arctangent, Logarithmic and conduction-convection algorithms were compared with those obtained from de Vries model. The amplitude decay algorithm yielded the most reliable values of the soil thermal properties of all the estimation methods with mean absolute error (MAE), root mean squared error (RMSE) and relative maximum error (RME) of 0.04, 0.05 and 5.63% respectively for soil thermal conductivity. Harmonic algorithm (using the amplitude of the first 4 harmonics) gave values of the soil thermal properties next to the amplitude decay algorithm with MAE, RMSE and RME values 0.41, 0.44 and 47.84% respectively for soil thermal conductivity. Higher error values were associated with the other algorithms. The Arctangent algorithm gave the most deviated values of soil thermal properties with RME of 156.83% for soil thermal conductivity. For soil moisture content between 0.168 and 0.189 (> critical soil moisture content) the values of the soil thermal properties of the loamy sand decreased with increasing soil moisture, while they increased with increasing soil aeration.