Estimation Methods for Annual River Loads Based on Water Quality Monitoring Data from Public Water Bodies

― A case study in the upper Ibo watershed ―

Abstract

This study evaluated the validity of the unbiased Horvitz-Thompson (HT) estimator and the precise bias-corrected regression estimator (BCRE) in estimating the annual river loads based on long-term monthly routine water quality monitoring data in public water bodies in Japan. The annual loads of total nitrogen and phosphorus in the upper Ibo River, Japan, were evaluated in this study. The daily concentration data in the upper Ibo River were synthesized using monthly data in the upper Ibo River and daily data in nine watersheds in the U.S. As a result, BCRE is recommended for both the annual river load estimate based on annual monitoring data and the average annual river load estimate based on 10-year monitoring data because the probabilities of the confidence intervals bracketing the true load value of the HT estimator tend to have significantly low values compared to the expected confidence level.