Scholarlink

<p>India’s river systems have extreme variability in streamflow, with over 80% of annual runoff concentrated during the monsoon season, posing challenges for water infrastructure design and flood management. However, few studies have combined multi-basin, multi-temporal, and multi-criteria analyses to systematically assess probability distribution performance at a national scale. This study evaluated the suitability of eight candidate probability distributions (Normal, Log normal, Log-Pearson Type III, Exponential, Gumbel, Generalized Extreme Value (GEV), Weibull, and Generalized Pareto (GP)) for modeling streamflow variability across 13 major river basins in India. Using daily, monthly mean, and annual maximum discharge records from 1965 to 2017, parameters were estimated through the method of moments (MOM), while model performance was evaluated with Kolmogorov–Smirnov and Anderson–Darling tests, root-mean-square error, Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC). Results show that three-parameter distributions, particularly GEV and GP, best represent daily and monthly mean streamflow due to their ability to capture skewness and heavy tails. Conversely, annual maximum flows were more effectively modeled by two-parameter distributions such as Weibull and Log normal, indicating the influence of temporal aggregation. Significant spatial variability was observed, with GP dominating northern and western basins, while southern and eastern basins favored GEV and Log normal fits. Flood quantile estimates up to 500-year return periods revealed substantially higher flood magnitudes in eastern basins compared to drier western basins. These findings highlight the need for basin-specific, scale-dependent frequency analysis to support infrastructure design, flood risk assessment, and climate-resilient water resources planning in India.</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

Assessment of best-fit probability distributions for streamflow across Indian river basins using multi-temporal data

  • 期刊论文

  • 发表日期：

    2026年2月14日

查看全文
  • Naveen Joseph,
  • Adarsh S,
  • Kavyakrishna AS,
  • Malini Swathi MU,
  • Dhanya Ramesh,
  • Divya Ganesh

摘要

India’s river systems have extreme variability in streamflow, with over 80% of annual runoff concentrated during the monsoon season, posing challenges for water infrastructure design and flood management. However, few studies have combined multi-basin, multi-temporal, and multi-criteria analyses to systematically assess probability distribution performance at a national scale. This study evaluated the suitability of eight candidate probability distributions (Normal, Log normal, Log-Pearson Type III, Exponential, Gumbel, Generalized Extreme Value (GEV), Weibull, and Generalized Pareto (GP)) for modeling streamflow variability across 13 major river basins in India. Using daily, monthly mean, and annual maximum discharge records from 1965 to 2017, parameters were estimated through the method of moments (MOM), while model performance was evaluated with Kolmogorov–Smirnov and Anderson–Darling tests, root-mean-square error, Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC). Results show that three-parameter distributions, particularly GEV and GP, best represent daily and monthly mean streamflow due to their ability to capture skewness and heavy tails. Conversely, annual maximum flows were more effectively modeled by two-parameter distributions such as Weibull and Log normal, indicating the influence of temporal aggregation. Significant spatial variability was observed, with GP dominating northern and western basins, while southern and eastern basins favored GEV and Log normal fits. Flood quantile estimates up to 500-year return periods revealed substantially higher flood magnitudes in eastern basins compared to drier western basins. These findings highlight the need for basin-specific, scale-dependent frequency analysis to support infrastructure design, flood risk assessment, and climate-resilient water resources planning in India.