Scholarlink

<p>Mineral deposit formation is a complex process governed by the spatiotemporal interplay of multiple geological factors across various stages, exhibiting spatially non-stationary controls and statistical mineralization heterogeneity. In the field of deep mineral exploration, precisely modeling the ore-controlling factor’s non-stationary and heterogeneous impact on mineralization can not only provide&#xa0;essential constraints for mineral prospectivity mapping but also offer novel insights into deposit genesis. Traditional prediction models are predominantly based on the conditional mean of mineralization, which limits their effectiveness in dealing with heterogeneous mineralization. In this paper, a three-dimensional geographically weighted quantile regression (3D GWQR) model is presented for investigating the spatially heterogeneous influence of ore-controlling factors at five mineralization quantiles (0.1, 0.3, 0.5, 0.7, and 0.9) in the Xiadian gold deposit. The model's effectiveness is conclusively demonstrated by its superior predictive performance and lower uncertainty in comparative evaluation with the quantile regression model and validated through a detailed assessment of its prediction hit-rate and associated spatial pattern accuracy. Based on this, the non-stationary impacts of all controlling factors across mineralization quantiles are demonstrated, as evidenced by the significant spatial non-stationarity in model coefficients (measured by stationarity indices) and the pervasive residual spatial autocorrelation (assessed via Moran’s I). Subsequently, the heterogeneous impacts of geological factors on mineralization are quantitatively investigated through both statistical and geospatial analytical perspectives, demonstrating significant factor-specific contributions at different ore-grade levels. Collectively, our results demonstrated that alteration intensity, with its superior quantile sensitivity and spatial heterogeneity at all mineralization quantiles, is a highly valuable exploration indicator.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Quantifying Spatial and Statistical Heterogeneities in the Relationships Between Mineralization and its Determinants for Quantile-Specific 3D Mineral Prospectivity Mapping

  • 期刊论文

  • 发表日期：

    2026年1月20日

查看全文
  • Jixian Huang,
  • Shijun Wan,
  • Hao Deng,
  • Baoyi Zhang,
  • Xiaoming Huang,
  • Xiancheng Mao

摘要

Mineral deposit formation is a complex process governed by the spatiotemporal interplay of multiple geological factors across various stages, exhibiting spatially non-stationary controls and statistical mineralization heterogeneity. In the field of deep mineral exploration, precisely modeling the ore-controlling factor’s non-stationary and heterogeneous impact on mineralization can not only provide essential constraints for mineral prospectivity mapping but also offer novel insights into deposit genesis. Traditional prediction models are predominantly based on the conditional mean of mineralization, which limits their effectiveness in dealing with heterogeneous mineralization. In this paper, a three-dimensional geographically weighted quantile regression (3D GWQR) model is presented for investigating the spatially heterogeneous influence of ore-controlling factors at five mineralization quantiles (0.1, 0.3, 0.5, 0.7, and 0.9) in the Xiadian gold deposit. The model's effectiveness is conclusively demonstrated by its superior predictive performance and lower uncertainty in comparative evaluation with the quantile regression model and validated through a detailed assessment of its prediction hit-rate and associated spatial pattern accuracy. Based on this, the non-stationary impacts of all controlling factors across mineralization quantiles are demonstrated, as evidenced by the significant spatial non-stationarity in model coefficients (measured by stationarity indices) and the pervasive residual spatial autocorrelation (assessed via Moran’s I). Subsequently, the heterogeneous impacts of geological factors on mineralization are quantitatively investigated through both statistical and geospatial analytical perspectives, demonstrating significant factor-specific contributions at different ore-grade levels. Collectively, our results demonstrated that alteration intensity, with its superior quantile sensitivity and spatial heterogeneity at all mineralization quantiles, is a highly valuable exploration indicator.