Evaluation and Determination of the Most Effective Index on Soil Sensitivity to the Surface Sealing Using Ridge Regression

The surface seal formation can have destructive agricultural, hydrological and environmental effects. Seal formation is a complex mechanism that controlled by a wide range of factors such as soil properties, rainfall characteristics, and flow conditions. The present research was conduct in order to identify an efficient index to determine the sensitivity of soil to the seal formation in the soil of Kowar plain region of Fars province.

In the present study, according to the area of the region, 160 sampling points selected, then according to the characteristics of the studied area and type of land use, to achieve the goals of the research, first, the land use, geology and topography maps of the area is prepared and by doing several times of field work in different seasons of the year, the topographies and shapes in the maps and pictures were adapted to the environmental conditions. To carry out this research, 80 composite soil samples (0-20 cm) were prepared from the Kovar plain of Fars province. Then the parameters of soil particle size distribution, mass moisture, organic matter, electrical conductivity and pH of saturated mud extract measured. Then the parameters of soil particles size distribution, volume wetness, organic matter, electrical conductivity and pH of saturated mud extract, measured. Also, mean weight diameter(MWD), geometric mean diameter(GMD), calcium carbonate equivalent, fractal dimensions, saturation percentage, sodium, calcium, magnesium, sodium absorption ratio (SAR), saturated hydraulic conductivity (Ks) and bulk density were determined. Assessing the sensitivity of the soil to the surface seal formation by comparing the regression equations of seven different indices, including the soil structural stability index (SSI), crusting index (CI), water aggregate stability (WAS), crusting susceptibility index (CSI), consistency index (C5 - C10), penetration resistance (PR) and relative sealing index (RSI) carried out in data preprocessing, Descriptive statistics of variables such as mean, maximum, minimum and variance, data distribution diagram and data distribution obtained using Minitab- 19 software. In addition, in case of remote data or mistakes in entering the data, appropriate measures taken and a table of descriptive information was prepared. The normality of the frequency distribution of these features evaluated using the significance test of skewness. To explain the ability of indicators and statistics in seal formation, First, linear correlation between variables determined by SPSS-22 software and by using Pearson's correlation coefficient (r), the correlation relationship between the indices and early characteristics of the soil was obtained and for data analysis, Pearson's correlation method and ridge multiple linear regression were used by stepwise backward method and using Statistica, SPSS-26 and Minitab-19 software.

In terms of the soil structural stability index, the studied area was in the danger range of aggregate destruction, which can considered relevant to the high percentage of silt in the region. The sodium surface absorption ratio of the majority of soil samples in the study area, had a non-sodium rating. 95% of the samples with agricultural land use in terms of organic matter were located in the weak to medium with structure and structure stability group. The majority of samples had more than 20% clay. Most of the measured variables had a low coefficient of variation and in terms of data distribution and dispersion, had a relatively favorable situation. Clay was the common variable in five regression equations of the indices, which refers to the importance and dual role of clay in the aggregates stability and reducing the sensitivity to compaction. The relatively low value of the standard deviation in the soil structural stability index indicated the greater accuracy of this model in estimating the coefficients. In the present study, the best-fitted regression model, to describe the sensitivity of the soil to the surface seal formation, compared to other models, belong to the soil structural stability and crusting indices.

Based on the results obtained, the soil aggregate stability index with the modified coefficient of explanation R2=0.92 had a high capability in predicting the sensitivity of the soil to the seal formation and is the most effective index in expressing the changes in soil aggregate stability and the surface sealing. The results of the data analysis expressing the extreme limitation class of the mean weight diameter was in the soil of the region.