Scholarlink

<p>This study aims to evaluate the performance of elliptical and D-vine copula models for predicting discharge at one spring (S5) using observed discharge data from the other four springs (S1-S4) and concurrent rainfall within a single springshed of Central Himalaya, India, in the context that the geological typology of the study springs is similar. The hypothesis that simpler, symmetric dependence models may outperform highly flexible vine structures when hydrological connectivity is strong, and data length is limited, is tested. The results revealed that elliptical copulas substantially outperformed D-vine structures, with the best elliptical model achieving RMSE ≈ 0.11 lpm and R<sup>2</sup> ≈ 0.76 under full conditioning, where both Gaussian and <i>t</i> copulas performed comparably. Among D-vine configurations, simplified hydrologically coherent structures (S1 → S2 → S3 → S5) markedly improved accuracy (R<sup>2</sup> ≈ 0.58<b>–</b>0.59, RMSE ≈ 0.14 lpm) over complex multi-node D-vines, confirming that pairwise correlation strength and physical flow topology are critical determinants. A similar flow variability index and strong inter-spring correlations indicate that groundwater-spring linkages are reliably represented by symmetric dependence structures. This study demonstrates that copula-based dependence modeling offers powerful, data-efficient approaches for discharge prediction in monitoring-limited Himalayan springsheds, where limited hydrological monitoring constrains sustainable water resource management.</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

Unveiling spring behaviour through copula-based inter-spring connectivity in the Central Himalaya

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月22日

查看全文
  • Sandipan Mukherjee,
  • Vaibhav Eknath Gosavi,
  • Shekhar Singh,
  • Mahendra Singh Lodhi

摘要

This study aims to evaluate the performance of elliptical and D-vine copula models for predicting discharge at one spring (S5) using observed discharge data from the other four springs (S1-S4) and concurrent rainfall within a single springshed of Central Himalaya, India, in the context that the geological typology of the study springs is similar. The hypothesis that simpler, symmetric dependence models may outperform highly flexible vine structures when hydrological connectivity is strong, and data length is limited, is tested. The results revealed that elliptical copulas substantially outperformed D-vine structures, with the best elliptical model achieving RMSE ≈ 0.11 lpm and R2 ≈ 0.76 under full conditioning, where both Gaussian and t copulas performed comparably. Among D-vine configurations, simplified hydrologically coherent structures (S1 → S2 → S3 → S5) markedly improved accuracy (R2 ≈ 0.58–0.59, RMSE ≈ 0.14 lpm) over complex multi-node D-vines, confirming that pairwise correlation strength and physical flow topology are critical determinants. A similar flow variability index and strong inter-spring correlations indicate that groundwater-spring linkages are reliably represented by symmetric dependence structures. This study demonstrates that copula-based dependence modeling offers powerful, data-efficient approaches for discharge prediction in monitoring-limited Himalayan springsheds, where limited hydrological monitoring constrains sustainable water resource management.