Scholarlink

<p>This study brings together hydrogeophysical imaging, hydrochemical evaluation, and machine-learning analysis to better understand how processes at the soil–water interface influence groundwater sanitation within the Akwa Ibom State University environment. The resistivity sections consistently depict a shallow, permeable transition zone marked by low resistivity values (&lt; 100 Ωm), which spatially coincide with elevated electrical conductivity (locally &gt; 1500 µS/cm) and total dissolved solids (&gt; 500 mg/L). Although the groundwater chemistry is still dominated by fresh Ca–Mg–HCO₃ facies, subtle shifts toward Na–Cl signatures suggest the early imprint of anthropogenic inputs. Direct comparison of measured parameters (e.g., TDS, EC, major ions) with WHO and SON guidelines indicates that groundwater quality is generally good (Table 2). In parallel, Bayesian inversion reveals only 1–3% deviation between deterministic and probabilistic resistivity models, attesting to the robustness of the geophysical interpretation. A key novelty of this work lies in the application of machine-learning models to both qualitatively and quantitatively infer water quality and contamination risk directly from integrated geophysical–geochemical inputs. These models consistently identify the shallow soil–water transition zone as the dominant control on vulnerability. Overall, the results show that the groundwater system remains fresh, but is increasingly sensitive to surface-derived sanitation pressures, highlighting the need for improved setback practices and proactive, long-term monitoring.</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

Machine-learning-integrated hydrogeophysical–geochemical assessment of soil–water interface dynamics for groundwater sanitation in Akwa Ibom State University, Nigeria

  • 期刊论文

  • 发表日期：

    2026年2月24日

查看全文
  • Nyakno J. George,
  • Aniekan M. Ekanem,
  • Okechukwu E. Agbasi,
  • Johnson C. Ibuot,
  • Kufre I. Udo,
  • Jude S. Ejepu,
  • Anietie E. Ekot

摘要

This study brings together hydrogeophysical imaging, hydrochemical evaluation, and machine-learning analysis to better understand how processes at the soil–water interface influence groundwater sanitation within the Akwa Ibom State University environment. The resistivity sections consistently depict a shallow, permeable transition zone marked by low resistivity values (< 100 Ωm), which spatially coincide with elevated electrical conductivity (locally > 1500 µS/cm) and total dissolved solids (> 500 mg/L). Although the groundwater chemistry is still dominated by fresh Ca–Mg–HCO₃ facies, subtle shifts toward Na–Cl signatures suggest the early imprint of anthropogenic inputs. Direct comparison of measured parameters (e.g., TDS, EC, major ions) with WHO and SON guidelines indicates that groundwater quality is generally good (Table 2). In parallel, Bayesian inversion reveals only 1–3% deviation between deterministic and probabilistic resistivity models, attesting to the robustness of the geophysical interpretation. A key novelty of this work lies in the application of machine-learning models to both qualitatively and quantitatively infer water quality and contamination risk directly from integrated geophysical–geochemical inputs. These models consistently identify the shallow soil–water transition zone as the dominant control on vulnerability. Overall, the results show that the groundwater system remains fresh, but is increasingly sensitive to surface-derived sanitation pressures, highlighting the need for improved setback practices and proactive, long-term monitoring.