Evaluation and Development of Pedotransfer Functions and Artificial Neural Networks to Saturation Moisture Content Estimation

Modeling of irrigation and agricultural drainage requires knowledge of the soil hydraulic properties. However, uncertainty in the direct measurement of the saturation moisture content (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msub><mi>θ</mi><mi mathvariant="normal">s</mi></msub></semantics></math></inline-formula>) has been generated in several methodologies for its estimation, such as Pedotransfer Functions (PTFs) and Artificial Neuronal Networks (ANNs). In this work, eight different PTFs were developed for the (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msub><mi>θ</mi><mi mathvariant="normal">s</mi></msub></semantics></math></inline-formula>) estimation, which relate to the proportion of sand and clay, bulk density (BD) as well as the saturated hydraulic conductivity (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msub><mi>K</mi><mi mathvariant="normal">s</mi></msub></semantics></math></inline-formula>). In addition, ANNs were developed with different combinations of input and hidden layers for the estimation of <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msub><mi>θ</mi><mi mathvariant="normal">s</mi></msub></semantics></math></inline-formula>. The results showed R<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mrow></mrow><mn>2</mn></msup></semantics></math></inline-formula> values from <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.9046</mn><mo>≤</mo><msup><mi mathvariant="normal">R</mi><mn>2</mn></msup><mo>≤</mo><mn>0.9877</mn></mrow></semantics></math></inline-formula> for the eight different PTFs, while with the ANNs, values of R<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><msup><mrow></mrow><mn>2</mn></msup><mo>></mo><mn>0.9891</mn></mrow></semantics></math></inline-formula> were obtained. Finally, the root-mean-square error (RMSE) was obtained for each ANN configuration, with results ranging from <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.0245</mn><mo>≤</mo><mi>RMSE</mi><mo>≤</mo><mn>0.0262</mn></mrow></semantics></math></inline-formula>. It was found that with particular soil characteristic parameters (% Clay, % Silt, % Sand, BD and <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msub><mi>K</mi><mi mathvariant="normal">s</mi></msub></semantics></math></inline-formula>), accurate estimate of <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msub><mi>θ</mi><mi mathvariant="normal">s</mi></msub></semantics></math></inline-formula> is obtained. With the development of these models (PTFs and ANNs), high R<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mrow></mrow><mn>2</mn></msup></semantics></math></inline-formula> values were obtained for 10 of the 12 textural classes.