ABSTRACT
We propose a new framework of proxy based on Cokriging method and apply the framework to build a map for mainland China. This framework utilizes the –topographic slope correlation in the cross‐semivariogram to benefit estimation and has the following benefits: (1) the estimation results are consistent with the measurement data; (2) the estimation uncertainty can be represented by error variance at each unsampled location according to the spatial structure of and topographic slope; (3) the result map does not have artificial boundaries; and (4) the estimation results can reflect the spatial relation between and topographic slope and the spatial relation of local spatial environment of topographic slope. We quantify the performance of this framework and compare it with that of three other topographic slope‐based proxy models, including original models developed from exogenous data and models developed from China local data. The result shows that the framework proposed in this article has the best performance. The framework is applied to 7797 borehole measurements to build a map for mainland China. The map can capture the high values in regions where the geological conditions are dominated by flat‐lying rocks. Moreover, we consider the effect of sample bias that comes from oversampling of borehole profiles in flat terrain regions when applying borehole data in the proposed framework. We utilize the relation of and topographic slope to quantify this bias, and use a distance‐related data spatial declustering method to eliminate it.