Optimizing Sampling Strategies for Estimating Riverine Nutrient Loads in the Yiluo River Watershed, China

Abstract

Accurately estimating nutrient loads is crucial for effective management and monitoring of aquatic ecosystems. This study evaluated the uncertainty in different sampling frequencies and calculation methods for estimating total nitrogen (TN) and total phosphorus (TP) loads in the Yiluo River watershed, a tributary of the Yellow River in China. Using daily TN and TP concentration data from 2019 to 2020, we conducted a bootstrapping analysis to evaluate the accuracy of nine different load estimation methods at different sampling frequencies. Our results showed that Method 3 (M_3, constant concentration interpolation) and Method 7 (M_7, flow-weighted concentration method), when used with a biweekly sampling frequency, had the lowest Standard Deviation of the Percentage errors (STD) (7.70% and 8.60% for TN, 12.0% and 18.8% for TP, respectively) and Mean Relative Error (MRE) values (0.078% and −1.60% for TN, 0.305% and 2.33% for TP, respectively) on an annual scale. For monthly TN and TP load estimates, M_7 can control the MRE within ±20% at a biweekly sampling frequency. Furthermore, the uncertainty in TN and TP load estimates was generally larger during the summer months (June–September), emphasizing the important role of storm events in nutrient export. Extreme events (<10% of the time) contributed approximately 50% of the annual nutrient loads. The findings of this study provide a scientific basis for optimizing water quality monitoring schemes and management strategies in agricultural watersheds.