Importance of Monitoring Frequency for Representation of Dissolved Organic Matter Dynamics in Urban Rivers

Abstract

In-situ dissolved organic matter (DOM) monitoring frequencies have often been chosen for convenience or based on perceived wisdom, without fully assessing their impact on representation of DOM dynamics. To address this gap, we collected 5-min fluorescence data in an urban headwater and resampled it at coarser intervals to investigate the impact of monitoring frequencies on the detectability of DOM dynamics during storms. Expecting hydrometeorological conditions to modify the impact of monitoring frequency, we categorized 85 storm events into groups: Group A (low intensity, short duration), Group B (high intensity, short duration), and Group C (low intensity, long duration). Surprisingly, our analysis indicated that monitoring frequency has minimal influence on commonly used biogeochemical indexes (e.g., maximum, hysteresis and flushing index), which are employed to characterize solute behavior, regardless of storm type. To facilitate a direct comparison between monitoring frequencies, we back-interpolated coarser data into 5-min intervals and calculated mean squared errors by comparing them with original high-resolution data. Our findings indicated that in colder periods with predominately Type A and C storms, a coarser monitoring frequency (>30 min) can capture DOM dynamics. Conversely, in warmer periods when Type B storms dominate, a finer frequency (≤15 min) is necessary to capture key solute chemograph processes (e.g., first flush and dilution). Generally, we suggest a 15-min monitoring frequency as optimal for similar urban headwater systems, and advocate an adaptive approach based on seasonal variations to improve efficiency, especially when power, data transfer, and storage are constraints.