Quantifying hydraulic properties of fractured rock masses along a borehole using composite geological indices: A case study in the mid- and upper-Choshuei river basin in central Taiwan

2020 
Abstract Comprehensive information on vertical variations in hydraulic conductivity along a borehole is indispensable to characterize the complexity of hydraulic properties in fractured bedrock aquifers or as input data utilized for groundwater modeling. This study presents a practical-oriented method to satisfy various engineering concerns (e.g., project budget, time for completion, manpower, and direct measurements of hydraulic properties from rock core specimens) for obtaining detailed (continuous) hydraulic conductivity data along a borehole. Based on in-situ hydrogeological data collected from 26 boreholes located in the Choshuei river basin of central Taiwan, eight types of individual and six types of composite geological indices are proposed to examine their correlations with the hydraulic conductivity of fractured rocks through the bivariate analysis. The correlation analysis results conclude that all composite geological indices have higher correlations with the hydraulic conductivity than individual geological indices. Besides, the more individual geological indices are integrated into one composite index, the better the hydraulic conductivity's corrleation tends to be. In light of the correlation results, quantification models of predicting hydraulic conductivity using composite geological indices were developed through regression analysis techniques. The performance of various statistical models was evaluated and compared. The regression analysis results for all six predictive models show that a power-law relationship exists between each composite geological index and hydraulic conductivity, and the coefficient of determination ranges from 0.77 to 0.88. Therefore, the newly developed models can serve as a proxy for vertical hydraulic conductivity data when the on-site hydraulic test is limited by funding.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    73
    References
    1
    Citations
    NaN
    KQI
    []