Technical ArticleApplication of Artificial Neural Networks in Modeling Soil Solution Electrical ConductivityNamdar-Khojasteh, Davood; Shorafa, Mahdi; Omid, Mahmoud; Fazeli-Shaghani, MahmoudAuthor Information Department of Soil Science, College of Agriculture, Tehran University, Tehran, Iran. Mr. Davood Namdar-Khojasteh is corresponding author. E-mail: [email protected] Received November 2, 2009. Accepted for publication July 15, 2010. Soil Science: September 2010 - Volume 175 - Issue 9 - p 432-437 doi: 10.1097/SS.0b013e3181f2a2e9 Buy Metrics Abstract In various applications in soil science and agriculture, there is a need for accurate measurements of soil solution electrical conductivity (EC). Time-domain reflectometry (TDR) has become an important means for measurement of soil water content and bulk EC. The TDR-measured EC (σa) and dielectric constant (Ka) can be used to calculate the soil solution EC (σp). Hilhorst (2000) found that using this linear relationship, measurements of σp can be made in a wide range of soil types without soil-specific calibrations. In the present study, the linear model was evaluated using detailed TDR. We also attempted to model the σp-σa-Ka relationship using artificial neural networks (ANN). To develop ANN models, we used TDR to measure Ka and σa along with five soil physical parameters (sand, silt, clay, organic matter content, and bulk density) in 10 different soil types. In total, 265 Ka and σa measurements were obtained. The ANN estimation of σp was found to have mean square error values between 0.071 and 0.41 dS m−1 for the 10 different soil types, whereas the mean square error of the linear model was 0.315 dS m−1. A sensitivity analysis showed that the ANN model was more sensitive to σa and two soil physical parameters (organic matter and clay content) more than other inputs as they affected the σp-σa-Ka relationship. © 2010 Lippincott Williams & Wilkins, Inc.