Impact of sampling time on chamber-based measurements of riverine nitrous oxide emissions using relative difference analysis

Abstract

The daily mean riverine N2O flux at a field site is often assumed to be equivalent to the magnitude of occasional once-per-day measurements made using the chamber method. The resulting time series data set therefore has a low temporal resolution. However, significant temporal variations in the N2O fluxes are typically found in streams and rivers, and high-frequency sampling is needed. To minimize the number of observations required without introducing significant bias from the daily averaging of values, we extend the temporal stability method, which has been widely used in soil physics and hydrology. We present the first analysis of the effects of sampling time on the determination of riverine N2O emissions based on high-frequency sampling of N2O fluxes. The results showed that: (1) daytime sampling can overestimate the daily average riverine N2O emissions, and emissions were more spatially stable at nighttime; (2) the spatial stability of N2O was controlled by the patterns of sewage inputs and amounts of dissolved oxygen (DO), NH4+–N, Cl−, and SO42−; and (3) sampling in the early morning (around 05:00h) and evening (around 19:00h) is typically spatially stable and is able to consistently represent the daily average N2O flux rate at a study site. Our methodology potentially provides a simple and rapid sampling protocol that could be used to identify the most suitable sampling times for floating chambers in estimating daily mean N2O flux rate from rivers.