2.4. Model Development

A vertically-averaged hydrodynamic model (Telemac V7.2; www.opentelemac.org; [32]) was applied to the estuary and surrounding coast. The model was constrained with the most recent available bathymetry data for the region, mapped onto an unstructured mesh that was optimised to adequately resolve estuarine and coastal currents (see Robins et al., 2019 [33] for further details of the model parameterization). Tidal forcing was derived from the Oregon State University TOPEX/Poseidon Global Inverse Solution tidal model (TPXO) global database [34] and the model was spun up to create a steady-state salinity balance under low river flow (Q99 of 1 m³/s) conditions. Comprehensive validation procedures have previously been conducted for hydrodynamics and salinity intrusion [17] which test the suitability of the model in the depth-averaged mode for application to the Conwy.

Based on the findings of the microcosm study, where survival of bla_CTX-M-15 -producing E. coli was enhanced under conditions of high turbidity, we simulated an extreme, but possible, combination event of extreme river flows combined with spring tides lasting two weeks (what could be envisaged as a “worst-case” set of scenarios), during which time the contamination event was modelled as a continuous input. Such scenarios might become increasingly common under climate change predictions, with elevated patterns of high-intensity rainfall that would lead to re-suspension of sediments in water bodies. Under this scenario, realistic high river flows and spring tides occurred in such a way as to produce a possible maximum export seaward of bla_CTX-M-15 -producing E. coli in water of high turbidity.

To simulate the decay of bla_CTX-M-15 -producing E. coli, the die-off rate t₉₀ was calculated to be 2 h and 15 min, based on the microcosm study of high turbidity (interpolated between points; Figure 1). Within the model, we simulated concentrations, C, with a die-off function following Equation (1):

where C0 is the measured concentration at time (t) zero and k=−ln0.1/t90.

Percentage of initial (t₀) bla_CTX-M-15-producing E. coli recovered in the freshwater of different turbidity levels (low, medium or high) over time, following exposure to UV irradiation. Based at the irradiance level of 40 W/m², the equivalent UV dose exposure in winter would be after 14 min, and in summer would be after 4.5 h.

Using gauged river flow measurements for the Conwy (1980 to 2015), we isolated the maximum freshwater inflow to the estuary over a 7-day period, which comprised six high-flow events within six days, with three events of peak magnitudes of 350, 470 and 450 m³/s, which occurred between 30 January–13 February 2004. These flows were used in conjunction with a constant input of bla_CTX-M-15 -producing E. coli from Llanrwst STP to create a 14-day time series of the river concentration boundary forcing at the tidal limit. For this, the input concentrations were multiplied by an estimated flux of water from the WWTP to give an estimated river concentration. At the offshore boundary of the model, the simulation was constrained with the largest annual astronomical tides (tidal range exceeding 6 m), aligning the tidal phase in such a way so that peak river flows coincided with peak ebb tidal flow and maximum concentrations were dispersed offshore.