Local regression kriging approach for analysing high density data

摘要

This paper demonstrates the development and an application of a local regression-kriging (RK) program. It illustrates the performance of the local RK method for prediction of soil properties using high density crop yield data in time series. We developed a software named RKGuider to carry out a serial steps of RK, especially local RK automatically and using the program, we produce a high resolution soil electrical conductivity map from a coarse survey. The result demonstrated there is spatial correlation between yield data in time series and soil electrical conductivity. High density yield data can be taken as auxiliary variables to predict soil EC using RK.