The sediment grain size data are compositional data and are characterized by non-negative and constant sum. In spatial prediction of grain size compositions, one important aspect of the prediction quality is whether the prediction results meet the requirements of non-negative and constant sum, which is also the essential condition for sediment type recognition and mapping. In this paper, the grain size data obtained from the abandoned Yellow River Estuary are taken as an example to discuss the application of combined logratio transform and ordinary kriging in spatial prediction of grain size compositions and sediment type mapping. Results show that the kriging interpolation results by directly using the grain size data are unreliable owing to their dissatisfaction of non-negative and constant sum, while the prediction results obtained using the logratio-transformed grain size data and kriging interpolation method not only meet those two requirements, but also have a better prediction accuracy of grain size compositions and a relative high sediment type mapping precision. So the new method of combined logratio transform and ordi-nary kriging has reference value for quantitative sediment mapping.%以废黄河三角洲表层粒度分析数据为基础，探讨了对数比转换和kriging插值相结合方法在沉积物粒级组分空间预测和底质类型制图中的应用。结果表明，基于沉积物粒级组分原始数据的kriging预测方法难以保证各组分预测结果的非负和定和要求，因而预测结果的可信度低；而对数比转换 kriging方法不但满足非负和定和要求，而且还有着更优的组分预测结果和较高的底质类型制图精度。新方法对于开展定量化的沉积物粒级组分预测和底质类型制图具有参考价值。
展开▼
