Prediction of soil total nitrogen based on total organic carbon using different models in soils from the Iraqi Kurdistan Region

Document Type : Original Article

Author

University of Sulaimani - College of Agricultural Engineering Sciences, Kurdistan region - Iraq

10.24271/psr.2023.388581.1274

Abstract

Determining the total nitrogen in the soil in a lab takes a long time and requires a lot of different chemicals; this method could be faster and cheaper. Therefore, using pedotransfer to predict total nitrogen (TN) in soil based on soil organic carbon (SOC) is more convenient and economical. Using five equations, including exponential, linear, logarithmic, polynomial, and power models, to predict TN from SOC. The results showed that TN could be predicted as a function of SOC, and the Bland-Altman approach was used to compare the predicted soil TN with the measured TN in the lab. The mean difference between the two approaches at 95% limits of agreement between field and lab measurements of soil TN was 0.000 g kg -1. The calculated values for the soil TN pedotransfer function were - 0.350 and + 0.350 g kg-1. The polynomial model (TN = - 0.667 + 0.203 x OC - 0.004 * OC2) and linear model (TN = - 0.263 + 0.124*OC) equations were considered the best model for predicting soil total nitrogen of studied soils due to their high R2 (0.820, and 0.814) and low RMSE (0.183, and 0.184).

Keywords

Main Subjects