Presently, the research of soil attribute space simulation for improving simulation accuracy focus on the two facts; one is to get enough sample point as far as possible for improving simulation accuracy, which is by the mean of half variant function and covariance function accuracy. Another is to embody the geographic environmental information (elevation, land use, vegetation type, etc.) to the soil attributes space simulation for improving simulation accuracy. In this article, according to general principles of soil science, which is that the relationship between the soils attributes, the author puts forward the soil attribute space simulation based on the general principles of soil science. And then the author chose the soil organic matter (SOM) as simulation object to simulate the SOM space distribution by using of the relationship between STN and SAK. The methods included Ordinary kriging and Cokriging to predict soil space distribute. The cross-validations results showed that the Cokriging method predicted values and measured values are more consistent (regression line closer to 45° line). At the same time, the regression coefficients between the Cokriging predicted values and the measured values were 0.7342. The regression coefficients between the ordinary kriging predicted values and the measured values was 0.2488. Test data sets validations results showed that the predicted values and measured values were more consistent from the Cokriging method (regression line closer to 45 ° line). At the same time, the regression coefficients by Cokriging predicted value and measured value was 0.2283. The regression coefficients by ordinary kriging predicted values and measured values was 0.2098. Through cross-validation and test data sets validations, the method of Cokriging by the use of soil attribute auxiliary parameter was better for improving simulation accuracy. Therefore, the method of regarding soil sampling point related attributes as auxiliary parameter is a simple and efficient method to improve prediction accuracy, which has extensive application in the spatial simulation of soil attribute.
[1]
Cheng Xu.
Cokriging algorithm applied for Interpolation
,
2009
.
[2]
Victor F. Rodriguez-Galiano,et al.
Downscaling Landsat 7 ETM+ thermal imagery using land surface temperature and NDVI images
,
2012,
Int. J. Appl. Earth Obs. Geoinformation.
[3]
Guo Xudong,et al.
Prediction of the spatial distribution of soil properties based on environmental correlation and geostatistics.
,
2009
.
[4]
Lianfa Li,et al.
Use of generalized additive models and cokriging of spatial residuals to improve land-use regression estimates of nitrogen oxides in Southern California.
,
2012,
Atmospheric environment.
[5]
Yue Tianxiang,et al.
High Accuracy Surface Modeling of Soil Properties Based on Geographic Information
,
2011
.
[6]
M. Nieuwenhuijsen,et al.
Association between GIS-Based Exposure to Urban Air Pollution during Pregnancy and Birth Weight in the INMA Sabadell Cohort
,
2009,
Environmental health perspectives.
[7]
R. Olea.
Geostatistics for Natural Resources Evaluation By Pierre Goovaerts, Oxford University Press, Applied Geostatistics Series, 1997, 483 p., hardcover, $65 (U.S.), ISBN 0-19-511538-4
,
1999
.
[8]
Timothy C. Coburn,et al.
Geostatistics for Natural Resources Evaluation
,
2000,
Technometrics.