Spatial Variability Analysis and Reclamation of Saline-Sodic Soils in the Northern Great Plains

Increased spring rainfall and higher temperatures when combined with changing landuses and extensive tile drainage installation have contributed to the development of sodic and saline/sodic soils in the Northern Great Plains. The objectives of this dissertation were: 1) determine the impact of surface chemical treatments and cover crop on crop yields and soil remediation; 2) determine and describe soil spatial variability and develop a model to identify saline-sodic soils; and 3) evaluate cation impact on dispersion of bentonite clay and selected soils. The research was conducted between 2013 and 2016 at Redfield (Argiustolls, Natrudolls, Calciustolls), White Lake (Argiudolls, Natrudolls), and Pierpont (Hapludolls, Natrudolls), in eastern South Dakota. A randomized complete block design with 4 replications was used. Treatments were cover cropping and surface amendments [gypsum, calcium chloride, elemental sulfur (S), and no amendments]. A mixture of barley (Horedeum vulgare) and sugar beet (Beta vulgaris) was used as the cover crop. At 169 sampling points, yield, soil properties, and reflectance were measured. Spatial class was developed using nugget to sill ratio. The impacts of chemical amendments on reducing soil dispersion were determined. Surface chemical amendment and cover crop treatments did not show significant differences in crop yield and soil properties in most locations. Hence, the amendments did not work in the Northern Great Plain soils with a glacial parent material that has high salt, calcium carbonate, and gypsum levels. Other management strategies that can reduce soil pH and mimic the native prairie grasses (deep-rooted perennial grasses that can use water from deeper in the soil profile) could be useful for future study. The exponential semivariogram model was found to be the optimal model for NDVI and yield with the spatial dependence (nugget/sill ratio) of 14.4 and 0%, respectively. Similarly, the exponential model was the optimum fit for mollic depth, lime depth, pH, EC, and SAR with nugget to sill ratio of 0, 0, 45, 17 and 49 respectively. Local Moran’s I and semivariogram modelling of soil attributes and NDVI data could help locate saline hot spots and quantify spatial heterogeneity respectively in saline-sodic soils. Higher turbidity was recorded in Na salt treated soil and bentonite clay than Ca and Mg salts. Turbidity was useful in measuring clay dispersion and could be used as an indicator of clay dispersion in salt-affected soils