Geostatistical Assessment of Aggregates Stability and Sodium Adsorption Ratio in Salt-Affected Soils AroundUrmia Lake

Salinity and sodicity are the most important land degradation problems particularly in arid and semi-arid regions. Due to the depletion of Urmia Lake located in the northwest of Iran during recent years, the proportion of surrounding saline agricultural lands increased at a past pace. In the salt-affected soils, aggregate stability is weak due to the high contents of sodium. The analysis of spatial variability of mean weight diameter of aggregates (MWD) and sodium adsorption ratio (SAR) is necessary to implement a site-specific soil management especially in the salt-affected soils. The main object of this study was evaluating the effects of different land uses (bare and agriculture) on the spatial variability of MWD and SAR in the salt-affected soils around Urmia Lake. This study was conducted in the agricultural and bare lands of Shend Abad region located at the 15 km of Shabestar city, northwest of Iran (45° 36ʹ 34ʺ E and 38° 6ʹ 37ʺ N). Totally, 100 georeferenced
samples were taken from 0-10 cm soil depth with 100×100 m intervals (80 ha) in agricultural (n=49) and bare (n=51) land uses. Sand, silt, clay, organic carbon (OC), CaCO3, pHe, MWD, SAR and electrical
conductivity (EC), were measured in the collected soil samples. Thewet sieving method was used to determine MWD of wet aggregates. The sieves were: 2, 1, 0.5, 0.25 and 0.106mm. The EC and SAR were measured in 1:2.5 (soil: distilled water) extra. The SAR was calculated from concentrations of Na+ and Ca+ + Mg+. The best fitsemivariogram model (Gaussian, spherical and exponential) was chosen by considering the minimum residual sum of square (RSS) and maximum determination coefficient (R2). Ordinary kriging (OK) and inverse distance weighting (IDW) interpolation methods were used to analyze spatial variability of MWD and SAR. Spatial distribution maps of soil variables were provided by Arc GIS software. The accuracy of OK and IDW methods in estimating MWD and SAR was evaluated by mean error (ME), mean absolute error (MAE), root mean square error (RMSE) and concordance correlation coefficient (CCC) criteria.The CCC indicates the degree to which pairs of the measured and estimated parameter value fall on the 45° line through the origin. According to the results of coefficient of variation (CV) from the study area, the most variable (CV=113.05%) soil indicator was SAR (bare land use), whereas the least variable (CV= 3.52%)
was pHe (agricultural land use). The Pearson correlation coefficients (r value) indicated that there are significant (P <0.01) correlations between MWD with OC (r=0.58), sand (r=0.60), SAR (r=-0.57), EC (r=-0.56), CaCO3 (r=-0.38) and clay (r=-0.48). Also, negative correlations (P <0.01) were observed between SAR with sand (r=- 0.61), OC (r=-0.38) but positive correlations (P <0.01) were found between SAR with EC (r=0.61), pHe (r=0.56), CaCO3 (r=0.40) and Clay (r=0.62). The spatial dependency classes of soil variables were determined according to the ratio of nugget variance to sill expressed in percentages: If the ratio was >25% and <75%, the variable was considered moderately spatially dependent; if the ratio was >75%, variable was considered weakly spatially dependent; and if the ratio was <25%, the variable was considered strongly spatially dependent. The strong and moderate spatial dependences with the effective ranges of 1903 and 1614m were found for SAR and MWD,
respectively. The OC and EC variables had the least (660m) and the highest (1987m) effective range, respectively. The range of influence indicates the limit distance at which a sample point has influence over another points, that is, the maximum distance for correlation between two sampling point. The models of fitted semivariograms were Gaussian for both MWD and SAR. The best prediction according to CCC criterion was obtained by OK for SAR (0.382) and IDW (^2) for MWD (0.325). The spatial maps showed that from agricultural to bare land use by nearing to Urmia Lake, the MWD decreased by decreasing OC and increasing EC and SAR. Results showed that MWD negatively related to the SAR and EC and positively related to the OC and sand in the study area. The SAR correlated positively with EC and Clay and negatively with sand andOC. The spatial dependency was found strong for SAR and the moderate for MWD. The soil OC showed the smallest effective range (660m) among the studied variables. As a suggestion, for subsequent study, soil sampling distance could be taken as 660m in order to save time and minimize cost.