Bank Thermal Storage As A Sink Of Temperature Surges In Urbanized Streams

A poorly-studied benefit of bank storage is the ability of the streambed to act as a thermal sink to streams influenced by urban runoff (e.g. bank thermal storage. Headwater streams, with their low thermal inertia, are particularly susceptible to thermal pollution. We utilize numerical modeling to quantify the amount of heat exchanged with the subsurface during temperature surges, which we define as greater than a 1 C stream temperature increase in 15 min. We base our study on Boone Creek, a low-order stream in northwestern North Carolina with stream discharge and temperature data dating to March 2006. The catchment is heavily urbanized, and although the stream is of moderate gradient, it is fed by tributaries that lose up to 200 m/km. The combined effect of urbanization and steep gradient produces a flashy response: stream discharge averages 0.10 m3/s, but may increase up to two orders of magnitude during storm events. These events also affect stream and streambed temperatures. Four summers of monitoring (2006–2008, 2010) indicate that 71 temperature surges occurred with a mean temperature increase of 2.39 C and a maximum increase of 6.36 C. We model generic storm events based on typical Boone Creek storms and streambed hydrogeology with the U.S.G.S. finite-difference groundwater flow and heat transport code VS2DH. The one-dimensional model domain includes a diurnally-oscillating stream temperature and specified head at the upper boundary, a constant streambed temperature and head at the lower boundary, and gaining stream conditions. Reference storm simulations use a temperature increase of 3.66 C and a stream stage increase of 0.66 m. Simulations show that at a depth of 4.5 cm, nearly half of the temperature-surge signal has dissipated and lag times are 30 min. By a depth of 9.5 cm, however, peak temperatures are only one-third of storm levels and lag times are 2 h. At depths beyond 49.5 cm, the perturbation is less than 0.1 C and lags the storm event by more than 17.5 h. Storm influence extends to a depth of 2 m and persists for days. Sensitivity simulations suggest that hydraulic conductivity, sediment heat capacity, and thermal conductivity are the most sensitive model parameters. Calculations show that temperature-surge induced heat storage in the simulated streambed is 72% of the heat storage in the stream