Equivalence Assessment of Multiple Datasets by Quantile–Value Diagram (QVD) for Detection of Significant Shifts (Case Study: Geochemical Datasets from Hanza Region, Kerman, Iran)

Abstract

Listen

Translate article

The purpose of equivalence assessement is to recognize a significant shift between values of the same variables (elements in geochemical surveys) in two datasets. To merge two or more independent datasets that are somehow linked together to create a new integrated dataset, the significance of the difference between datasets must be determined. Merging data originated from two or more geochemical surveys can be highly beneficial to establish geochemical maps with an increased resolution and/or a covered area. Different methods of sampling, preparation and analysis are being used in various geochemical surveys which have a major impact on the measurement of the concentration of elements, and there might be a systematic shift in the concentration of the same elements. Therefore, studying the presence of a significant shift between datasets, and leveling them as a pre-processing stage is essential to combine geochemical data and provide an integrated map of a few independent surveys. While combining datasets to provide integrated maps, not assessing the equivalence and leveling the datasets leads to invalid maps. Determining the equivalence assessment between datasets is done by different methods. Quantile–Quantile Dispersion Diagram and Multiple Frequency Histogram methods have already been introduced by researchers. These methods are visual and dependent on expert judgment. In this study the Fisher test, T-student test, and Quantile–Value Diagram (QVD) are introduced to assess the equivalence between two datasets for the first time. Fisher and T-student tests are applied to compare the variance and average of the same elements in two datasets. In this study, we propose an original method for assessing the equivalence. QVD is introduced as a new method of determining the shift between multiple datasets. The case study is the equivalence assessment and shift determination between two independent geochemical surveys in north of Sarduiyeh and south of Rayen sheets (Hanza, southern part of Urmia–Dokhtar metalogenic belt, Iran). The equivalence assessment was performed for twelve elements including Zn, Pb, Ag, Ni, Bi, Cu, As, Sb, Co, W, Mo, and Mn. Based on Quantile-Quantile Dispersion Diagram and Multiple Frequency Histogram methods, some elements of the two databases were considered as equivalent, but the analysis of QVD method shows a significant shift for them. Fisher and T-student tests confirm the results of QVD. QVD is an exact method for the equivalence assessment that is not dependent on the expert judgment. Eventually, Fisher test, T-student test, and QVD are recommended simultaneously to assess the equivalence between datasets.