**2.3. Data analysis**

(1)

cores [14]. Soil penetration resistance (PR) was measured with a cone penetrometer at depths of 0–10 and 30–40 cm [17], and the soil-crusting index (CI) was calculated by Eq. (1) using soil

> 100 %( ) % % *SOM*

Wheat yield was measured at sampling sites. Field capacity (*θ*0.33MPa) and wilting points (*θ*1.50MPa) were determined with a pressure plate [14], and plant-available water content (PAWC)

> q

**Rapid-very rapid Moderate rapid Moderate Moderate slow Slow Very slow**

0.33*MPa* ) ( 1.50*MPa* ) (2)

*Clay Silt* <sup>=</sup> <sup>+</sup>

*CI*

*PAW* = - (q

Concave Na N N N N N <1 N N N L M H 1–5 N VL L M H VH 5–10 VL L M H VH VH 10–20 VL L M H VH VH >20 L M H VH VH VH

, negligible; VL, very low; L, low; M, medium; H, high; VH, very high.

**Table 1.** Indices determined for surface runoff classes in the study area.

organic matter (SOM), clay, and, silt contents [1]:

**Figure 2.** Locations of soil sampling points in the basin.

96 Land Degradation and Desertification - a Global Crisis

was calculated by Eq. (2):

**Slope, % Permeability classes**

Na

Descriptive statistics of mean, standard deviation (SD), coefficient of variation (CV), kurtosis, skewness, maximum, and minimum were calculated for the variables of particle-size components, coarse material, wheat yield, soil loss, available water content (AWC), runoff, crusting index, penetration resistance, saturated hydraulic conductivity (Ks), soil organic matter, electrical conductivity, pH, cation exchange capacity, available K, available P, and available micronutrients (Fe, Cu, Mn, Cd, Zn, and B).

Factor analysis was conducted separately on topsoil and subsoil using Statistical Package for the Social Sciences (SPSS; Chicago, IL) to summarize correlations among variables [20]. First, correlation matrices, eigen values, and eigen vectors were calculated. Second, main factors were determined by the maximum likelihood method [21] and scree analysis [20]. Factors with a loading of >0.5 were retained. Finally, principal components were determined [22]. Relations among the variables were explained using the factor loadings. The principal components derived from the prepared correlation matrices were subjected to an orthogonal rotation of axes (varimax rotation) when multiple loadings occurred. Nine factors (Factor 1: "erodibility factor"; Factor 2: "soil fertility factor"; Factor 3: "soil chemistry factor"; Factor 4: "soil-crusting factor"; Factor 5: "soil erosion factor"; Factor 6: "soil conductivity factor"; Factor 7: "plantavailable water content factor"; Factor 8: "macroelement factor"; Factor 9: "crop yield factor") for topsoil and seven factors (Factor 1: "microelements factor"; Factor 2: "soil physics factor"; Factor 3: "soil fertility factor"; Factor 4: "soil chemistry factor"; Factor 5: "yield factor"; Factor 6: "soil potassium factor"; and Factor 7: "soil cadmium factor") for subsoil were retained. The loading (or eigenvectors) of a variable in a factor is similar to the correlation between the variable and the factor.
