Currently, an inability to accurately measure and model hydraulic properties of soil at large scales plagues the accuracy of climate and crop modeling. Much effort is focused on development of new models and pedotransfer functions which include new variables and soil processes. However, soil is inherently spatially dependent. Therefore, many soil properties and relationships can be analyzed with geospatial statistics. In this thesis, a multivariate geospatial technique called factorial kriging analysis (FKA) was employed to analyze the spatial relationships of soil hydraulic parameters. FKA was also used to develop a novel prediction method to aid in characterizing the spatial distribution of an areas hydraulic properties. Bulk density was found to be the source of short scale variance while paleo-channels were found to be the major source of variance at long scale. The prediction method described within is able to classify hydraulic zones and distinguish between groups of water retention curves.