Interpretation of river water quality data is strongly controlled by measurement time and frequency

Water quality monitoring at high temporal frequency provides a detailed picture of environmental stressors and ecosystem response, which is essential to protect and restore lake and river health. An effective monitoring network requires knowledge on optimal monitoring frequency and data variability. Here, high-frequency hydrochemical datasets (dissolved oxygen, pH, electrical conductivity, turbidity, water temperature, total reactive phosphorus, total phosphorus and nitrate) from six UK catchments were analysed to 1) understand the lowest measurement frequency needed to fully capture the variation in the datasets; and 2) investigate bias caused by sampling at different times of the day. The study found that reducing the measurement frequency increasingly changed the interpretation of the data by altering the calculated median and data range. From 45 individual parameter-catchment combinations (six to eight parameters in six catchments), four-hourly data captured most of the hourly range (>90 %) for 37 combinations, whilst 41 had limited impact on the median (90 % of the range with limited impact on the median in approximately half of the combinations, whereas weekly and monthly data captured this in <6 combinations. Generally, reducing sampling frequency had most impact on the median for parameters showing strong diurnal cycles, whilst parameters showing rapid responses to extreme flow conditions had most impact on the range. Diurnal cycles resulted in year-round intra-daily variation in most of the parameters, apart from nutrient concentrations, where daily variation depended on both seasonal flow patterns and anthropogenic influences. To design an optimised monitoring programme, key catchment characteristics and required data resolution for the monitoring purpose should be considered. Ideally a pilot study with high-frequency monitoring, at least four-hourly, should be used to determine the minimum frequency regime needed to capture temporal behaviours in the intended focus water quality parameters by revealing their biogeochemical response patterns.