Zoning of Nutrient’s Elements In Soil and Leaves of Orange Trees Using Gaussian Model (Dezful city as Case Study)

Creating land fertility maps are especially important in terms of determining the areas that need particular nutrients, optimizing the use of agricultural fertilizers, and facilitating the optimal management of soil and plant nutrition. Spatial changes in soil and plant nutrients are common, but knowing these changes is essential for accurate planning and management, particularly regarding agricultural lands. This research aims at zoning the spatial distribution pattern of nutrients, aka nitrogen, phosphorus, potassium, sulfur, calcium and magnesium, in the soil and leaves of Dezful orange orchards trees using the Gaussian model and geographic information system (GIS).

A total number of 130 sampling points are set on the map in the vicinity of orange orchards of Dezful City with an area of 3200 hectares. Factors such as soil, cultivation and irrigation system, slope, elevation, and the manner of orange trees growth are considered to determine sampling locations. Following sampling the soil (0-60 cm depth) and plants, the samples are transferred to the laboratory and the concentration of the most consumed nutrients is measured. After preliminary statistical analyzes on the data, the correlation level of the variables that are measured in the soil and leaves of orange trees, are calculated with the Pearson correlation test. The location of sampling points is simulated via Gaussian model by using the R software. The interpolation is computed using simple kriging and kernel methods. The model sensitivity analysis for the changes applied in the base values for implementing the algorithm, is done based on the replacement of the desired values from the posterior functions as well.

Analysis of dispersion indices show that the highest coefficient of variation is related to phosphorus element in soil and nitrogen element in leaf samples. The results illustrate that the mean square error values for elements of nitrogen, phosphorus, potassium and sulfur are calculated respectively as 0.171, 0.152, 0.132 and 0.153 in simple kriging in soil, and as 0.212, 0.152, 0.229, and 0.166 in kernel method in soil; and respectively as 0.121, 0.188, 0.116 and 0.131 in simple kriging in samples of orange tree leaves, and as 0.184, 0.206, 0.172 and 0.229 in kernel method in the leaves samples as well. The results of the spatial distribution pattern of each of the measured elements in the soil and leaves of orange trees demonstrate that the lowest amount of nitrogen is in the south of the region (0.42 to 1.33 mg/kg) and its distribution pattern is similar to the distribution in the leaves of orange trees (0.9 1 to 1.29 mg/kg). Magnesium has the lowest in the north and part of the south (3.11 to 4.57 mg/kg), and sulfur in most soil of the region is between 21.31 and 26.25 mg/kg.

In examining the effectiveness of the Gaussian statistical model in the distribution of nutrients in the soil and leaves of orange trees in the gardens of Dezful city, the results display that the calculated Pearson linear correlation coefficient  has the highest correlation between calcium and potassium, as well as magnesium and calcium in the soil, but there is no linear correlation between any of the nutrients in the leaves of orange trees. In estimating the best interpolation method, calcium element in soil has the least error in both kernel and simple kriging methods, whereas in plant leaves, magnesium in kernel method and potassium in simple kriging method have less error. The highest error for soil and plant is related to potassium and calcium respectively, in the Cornell method.