| Popis: |
Merging several already existing geochemical datasets can be highly profitable for obtaining geochemical maps with increased resolution and/or coverage area. However, this practice can lead to potential problems due to varying quality of the datasets and biases between datasets. In this paper, we present a general method which addresses these issues thereby making it possible to pool together multiple pre-existing datasets to produce one reliable dataset for geochemical mapping. The method first consists in performing four data quality checks that evaluate the relevance of existing data for the intended geochemical map. The checks pertain to georeferencing, temporal variability, spatial structure and presence of censored values. Next, a procedure is applied to detect biases between datasets and, where needed, to level the data using linear transformation. We then applied this method to a case study in Wallonia (Southern Belgium) involving the mapping of nickel in agricultural soils and create a new dataset based on multiple pre-existing geochemical datasets. We conclude by quantifying the new dataset's improved reliability for geochemical mapping thereby demonstrating the potential benefits of applying this proposed method to real-life situations. |
Full Text from ScienceDirect