Kriging analysis of soil properties

Autor: Gilbert C. Sigua, Wayne H. Hudnall
Rok vydání: 2008
Předmět:
Zdroj: Journal of Soils and Sediments. 8:193-202
ISSN: 1614-7480
1439-0108
Popis: Soil as a landscape body contains wide ranges of physical, chemical, morphological, and mineralogical properties, both laterally and vertically. Soils with similar properties and environments are expected to behave similarly. A statement on land use potential will depend in part on the precision and accuracy of the statements that can be made about the soils. This information has some practical applications in optimizing land management and productivity improvement. The spatial patterns and dependence of some selected physicochemical properties of brackish marsh and surrounding soils were investigated using a 2-D kriging analysis in conjunction with a geostatistical (GS+, Michigan) model. Composite (four subsamples) surface samples (0–20 cm) were obtained across the sampling sites (717 ha). Sampling area was subdivided into north–south transect (NST) and east–west transect (EWT). Subsamples were mixed resulting in 40 composite samples. These samples were stored at 5°C until the different physical (particle size) and chemical analyses were initiated. Soil chemical analyses included electrical conductivity (EC), Cl−1, pH, and water soluble Na, S, Ca, Mg, Fe, and Al by inductively coupled plasma spectroscopy (ICP). Soil extraction was performed via an automatic soil extractor. A 1:2 soil to extracting solution ratio was used for all the extractions. The concentration of Cl−1 in the soil sample was determined using a chloride electrode (Model 94-17B). Some selected physicochemical properties were interpolated by kriging. The procedures of univariate/multivariate analyses and kriging techniques were followed to evaluate the spatial dependence of some selected physicochemical properties of soils in the landscape. The kriging procedures included preliminary data analysis, structural data analysis, log kriging estimations, and image generations of spatial results. Based on the results of the kriging study, a field study evaluating the effects of gypsum and water management on the yield response of selected species of marsh vegetation was conducted in the area where soil samples were taken for the kriging study. The overall yield responses of four species of marsh vegetation to gypsum addition (0 vs. 7 Mg ha−1) and water management (flooded vs. non-flooded) were evaluated for joint grass (Paspalum vaginatum, SW.), marsh hay cordgrass (Spartina patens MuhL.), saltgrass (Distichlis spicata L.), and American three-square (Scirpus americanus Pers.). There was a significant spatial dependence and differences (p ≤ 0.01) for all soil parameters tested. The variability of an individual soil property, as indicated by coefficients of variation (CV), differed widely. Some soil properties in the lower horizons were more uniform than those at the surface because of fewer disturbances and effects of micro-topography. Soils may develop markedly contrasting morphologies and properties that may vary laterally (east–west) and vertically (north–south). There was no plant survival in the non-flooded plots except for the marsh hay cordgrass, which had an average survival rate of 32.8%. Plots receiving 7 Mg ha−1 gypsum had significantly higher dry matter production than the control. Gypsum application increased dry matter yield (Mg ha−1) of joint grass (4.6 to 9.2), marsh hay cordgrass (2.0 to 9.2), salt grass (0.9 to 3.2), and American three-square (1.5 to 3.0) in flooded soils. The successions of high and low micro-relief within landscapes that were identified via kriging analysis may affect the overall growth and biomass production of marsh vegetation. Many of the differences and spatial dependence of soil properties in the study area that vary with topography are because of some combinations of microclimate, soil pedogenesis, geological surficial processes, and the sorting effects of water movement. Plants in the flooded plots had much higher dry matter yield (DMY) than the plants in non-flooded plots. Both gypsum addition and water management should become a major part of any marsh revegetation program. Kriging analysis of the landscape suggested that the majority of the spatial variability of some soil properties within the study area were caused by the interactive effect of micro-topography and hydrologic pattern. Spatial variability will therefore affect soil performance. A uniform application of any soil amendments like fertilizer or gypsum in the area that possessed spatially variable soil would result in an overapplication in some parts of the area and an underapplication in other areas. Plants that were grown in flooded plots with gypsum treatment of 7 Mg ha−1 had a higher dry matter yield. Knowledge of the existing soil variability in the area can be utilized to develop a stochastic model that may describe the potential land use capabilities. The prediction is being made by matching the requirements of specific land use to the characteristics of soils in the landscape. Results of an investigation of this type are of great interest to environmental scientists, water resource planners, regulators, decision makers, engineers, soil scientists, and resource managers. Kriging results were suggesting that both gypsum addition and water management should become a major part of marsh revegetation and restoration programs.
Databáze: OpenAIRE
Externí odkaz: