Application of spatially weighted technology for mapping intermediate and felsic igneous rocks in Fujian Province, China

- Trace and incompatible trace elements are used besides rock-forming elements.
- Spatially weighted models can well deal with tendency and non-stationary.
- Spatially weighted logistic regression provides better predict results.

Magmatic activity is of great significance to mineralization not only for heat and fluid it provides, but also for parts of material source it brings. Due to the cover of soil and vegetation and its spatial nonuniformity detected signals from the ground's surface may be weak and of spatial variability, and this brings serious challenges to mineral exploration in these areas. Two models based on spatially weighted technology, i.e., local singularity analysis (LSA) and spatially weighted logistic regression (SWLR) are applied in this study to deal with this challenge. Coverage cannot block the migration of geochemical elements, it is possible that the geochemical features of soil above concealed rocks can be different from surrounding environment, although this kind of differences are weak; coverage may also weaken the surface expression of geophysical fields. LSA is sensitive to weak changes in density or energy, which makes it effective to map the distribution of concealed igneous rock based on geochemical and geophysical properties. Data integration can produce better classification results than any single data analysis, but spatial variability of spatial variables caused by non-stationary coverage can greatly affect the results since sometimes it is hard to establish a global model. In this paper, SWLR is used to integrate all spatial layers extracted from both geochemical and geophysical data, and the iron polymetallic metallogenic belt in south-west of Fujian Province is used as s study case. It is found that LSA technique effectively extracts different sources of geologic anomalies; and the spatial distribution of intermediate and felsic igneous rocks delineated by SWLR shows higher accuracy compared with the result obtained via global logistic regression model.

Because of the tendency and non-stationary of spatial variables, and the existence of some unknown variables, the results are often biased when ordinary regression models are applied for spatial prediction and assessment. Geographically weighted regression has been developed to deal with this problem in Economic Geography. Mineral prospectivity mapping is also belong to spatial prediction and assessment, in which logistic regression is preferred rather than linear regression because the target variable is binary and is often estimated by non-negative probability. The first author developed a spatially weighted logistic regression (SWLR) model. In this study, we mapped the intermediate and felsic igneous rocks in the south-west Fujian Province, China and compared the results from the SWLR method to the results obtained by general logistic regression.Receiver operating characteristic (ROC) curve is one of the most popular models that illustrate the performance of a binary classifier system, and its advantage is that the discrimination threshold is varied. This model is applied in this research to compare general logistic regression and SWLR based on their abilities in predicting intermediate and felsic igneous rocks (Fig. 1).Fig. 1 Area under curve (AUC) values obtained through receiver operating characteristic (ROC) curve method for (a) general logistic regression and (b) spatially weighted logistic regression respectively.AUC (area under curve) is the most popular indicator in ROC method to compare the performance of different predicting results. According to its definition, AUC lies within the range from 0 to 1. Generally, a classifier or diagnosis with the AUC greater than or equal to 0.7 can be considered effective. ROC curves describing the predicting effects obtained by general logistic regression and SWLR respectively were given in Fig. 1, from which it can be viewed that SWLR is more effective than general logistic regression since the former owns a much bigger AUC value.295