engleski

Total and organic phosphorus as a basis of the regression model for mineral phosphorus prediction

Regression models comparing organic, total phosphorus in the soil and AL extracted phosphorus are created to enable the prediction of mineral phosphorus without the soil analysis, based only on analytical data of one or more different soil properties. Samples were collected in the area of the eastern Croatian, and analysis were carried out on 94 samples and included the pH, organic matter content, organic and total phosphorus. Criteria for selection of samples were pHKCl and organic matter content and all the samples were divided into five categories according to soil acidity and two categories according to organic matter content. The results of total and organic phosphorus in the soil obtained by analysis were used to predict the results for AL method. Models include the basic equation for calculating the concentration of the extracted phosphorus by particular method (AL) based on the amount of organic and total phosphorus regardless of other soil properties. Regression formula were described by equation Y= intercept + OX1 +TX2. Correlation coefficient between AL phosphorus, organic and total phosphorus was r = 0.43* (n = 94), but dividing the samples into five acidity categories, correlation coefficient increased to r = 0.89**. Likewise, the developed regression model was very responsive to the expansion of the input variables and inclusion of information about the organic matter content in the soil what increased model accuracy for 50 %. Although the positive relationship between organic phosphorus and organic matter in the soil has not been established, their interaction contributed to the improvement of the developed model.

total ; organic ; phosphorus ; regression models

nije evidentirano