Estimation of Decadal Averages of Annual Riverine Constituent Loads Based on Long-Term Low-Frequency Water Quality Monitoring Data

Abstract

Accurate estimates of annual or seasonal riverine constituent loads, which is the terrestrial mass flux conveyed through rivers, are essential to water pollution control measures and watershed mass studies. An unbiased estimate of river load and its confidence interval can be calculated based on the Horvitz-Thompson (HT) estimator. However, the HT estimator may fail to provide precise load estimates based on low-frequency water quality (WQ) monitoring data such as monthly routine samples employed in WQ monitoring in Japanese public waters. In this study, the authors calculated unbiased estimates and those confidence intervals of decadal averages of annual river loads based on long-term (decadal) low-frequency (monthly) monitoring data at the Yamasaki station in Hyogo prefecture, Japan, located in the upper watershed of Ibo river. The results also showed that a traditional rating curve method tends to provide biased load estimates, whereas a simple average estimate can be used for unbiased load estimation.