Evaluation of rock mass permeability along a borehole by integrating the techniques of geological features and logistic regression: a case study in Taiwan

When the budget for in situ hydraulic tests is constrained, cost-effective approaches determining variations conductivity along a borehole remain enticing helping design and planning of groundwater-related engineering systems. This study proposes practical method with probabilistic outputs fulfilling concerns. First, 474 sets hydrogeological investigation test data fractured rock masses most watersheds mountainous areas Taiwan were collected. Then, seven geological indices [rock quality (RQD), depth index (DI), gouge content (GCD), lithology permeability (LPI), fracture density (FD), width (FW), groundwater velocity (GV)] significantly correlated mass identified. Third, using logistic regression analysis, used as explanatory variables, was utilized an outcome variable (its threshold value 1 × 10−6 m/s) developing prediction models high potential zones borehole. All passed collinearity test, indicating no between indices. To make proposed more flexible applications, total 127 combinations based on combination selection variables explored to develop various models. Through validation Hosmer–Lemeshow Omnibus Wald all developed models, only 77 statistically significant. The evaluated by three measures Nagelkerke R2, SR (success rate), AUC (area under curve) further understand performance each model. results show that accuracy model positively number used. are used, SR, can reach best values, which 0.83, 91.6%, 0.976, respectively. In conclusion, combining analysis provide efficient way understanding conditions site.