Effect of sampling frequency and streamflow on nutrient source apportionment in subtropical rivers

Accurate estimation of nutrient contributions is essential for effective pollution control, yet remains challenging due to substantial uncertainties arising from limited sample sizes and dynamic hydrological regimes. This study employs a process-based load apportionment model (LAM), integrating daily flow records and high-resolution water quality data from 41 monitoring stations across the Pearl River Basin (PRB), to quantitatively distinguish point-source versus non-point-source contributions to total nitrogen (TN) and total phosphorus (TP) loads. Statistical T-tests were systematically applied to evaluate the sensitivity of source apportionment results to monitoring frequency and streamflow variability. The results indicate that: (1) Non-point sources dominate nutrient fluxes, contributing 85.95 and 92.13% of annual TN and TP loads respectively, acting as the largest sources averaging 83.41% (TN) and 90.88% (TP) of the period (average <i>R</i>² > 0.70); (2) Regional heterogeneity exists, with the Beijiang sub-basin exhibiting significantly lower non-point-source TN contributions (66.15%) compared to other sub-basins; (3) Monitoring frequency exerts greater influence on TN source partitioning (<i>P</i> < 0.05 at 65.85% stations) than TP (46.34% stations). These findings highlight the necessity of region-specific management strategies and underscore the value of high-frequency monitoring coupled with multi-source data fusion to enhance the robustness of pollution source identification.