A Study on Soil Property Mapping at National Scale Based on Sparse Grid Soil Samples

Soil property data at national scale are necessary inputs for researches on biogeochemical and hydrological cycles.There are two commonly used methods to obtain accurate soil property data at this scale including soil linkage method and spatial interpolation method.In this paper,soil organic matter(SOM) content in Jilin Province was taken as a case study to evaluate these two methods based on 8～32 km grid soil samples and 1∶1 million soil map,in order to provide suggestions about soil property mapping at national scale based on sparse grid soil samples.Independent validation indicated that the mean error(ME) of the soil linkage method was-0.58,higher than the inverse distance weighted(IDW) interpolation method(0.13),but the mean absolute error(MAE) and root mean square error(RMSE) of the soil linkage method were 1.12 and 1.76,respectively,lower than those of IDW interpolation method(1.43 and 2.12).Although the soil property map through IDW interpolation revealed the general trend of soil property distribution,it had "Bull's eye" pattern and the estimations in areas with few or no soil samples were unreliable.The soil property map obtained by soil linkage method ignored the variations inside the same soil type on the soil map and a break of soil property existed around different soil boundaries,but the soil distribution regularity was well represented in this map and the details could be obtained.According to the validation results of maps through these two methods,the soil linkage method is better than the spatial interpolation method in soil property mapping based on sparse grid soil samples at national scale.