Author(s): Elisabeta-Cristina Timis

Linked Author(s): Elisabeta-Cristina Timis

Keywords: River water quality; Pollution sources; Phosphorus sources and sinks; River Swale; Pollutant sources identification

Abstract: Currently, software tools are crucial in the water quality management field, as they ensure reliable decision support even in the most critical conditions (e.g. in the case of environmental accidents) and also in situations with scarce field data availability. The present work offers a significantly helpful tool especially in those situations when the sinks and sources of pollutant (hereafter termed influences) along a river are difficult to measure or estimate. Such influences are the discharges into the river channel (from point or diffuse sources, controlled or accidental) and also the in-river processes (accumulation or consumption) that are rather complicated or impossible to estimate. These are assessed by the influences module of ADModel. ADModel is a powerful computational tool for the detailed simulation of non-conservative pollutant transport along the river stretches (sub-hourly resolution at under 500 m space steps), that has been validated for nitrogen and phosphorus compounds. The influences module is based on hybrid optimization techniques (genetic algorithms together with traditional ones). The input data includes in-river pollutant concentrations (at the upstream and downstream boundaries of the stretch), as well as information on the already known sources of pollutants. The results are values of concentrations assigned to unknown influences along the stretch. The technique is illustrated for soluble reactive phosphorus (SRP) and organic phosphorus (OP) using field data from a case study on River Swale, a less polluted river stretch of 50.4 km, located in England (UK). The findings are crosschecked against field information. SRP and OP discharges in the main channel are caused mostly by known sources: the three major tributaries, sewage treatment works and commercial or industrial activities associated with the large population centres. The new module of ADModel enabled the estimation of other sources, which are more difficult to assess otherwise, such as farming activities or small touristic activities, together with some of the unidentified in-stream influences, not included in the model during its calibration due to data unavailability. Results reveal that concentrations associated to influences could be correlated to water flow and seasonality for both SRP and OP. At higher flows, correlations are straightforward: SRP sources are prominent above 80 m3/s and OP sinks are prominent above 65 m3/s. At lower flows, processes tend to be more complicated: in-stream SRP sinks (with a maximum value of 0.5 mg/L) are generally larger than SRP sources (with a maximum value of 0.2 mg/L), while for OP the sources are dominant (with a maximum value of 1.05 mg/L compared to 0.5 mg/L attributed to in-stream sinks). Findings are confirmed by the additional field information and could be further linked to the higher residence time at low flows, seasonal activities, temperature variations and extreme weather events (e.g. storms).

DOI: https://doi.org/10.3850/IAHR-39WC25217119202281

Year: 2022