3.1 Stream discharge and temperature
Stream discharge was manually measured bi-weekly (May-August) or monthly (September-April) using a horizontal-axis current meter (Global Water, FP101) and the velocity-area method (Dingman 2002, p. 609) with an expected uncertainty of up to 9 % (Hood, Roy, & Hayashi, 2006). A stage-discharge rating curve was developed at each of the four stream gauging stations (GS1 to GS4; Figure 1) using at least ten measurements. A root mean squared error below 0.015 m3s-1 was obtained at all four gauging stations with the largest errors occurring during the baseflow period. Water levels were measured by pressure transducers (Solinst, Levelogger). A continuous discharge record was computed by applying the rating curves to the transducer data.
A number of different instruments were used to measure water temperatures to reflect the spatial and temporal scales of interest. Manual measurements were taken with a hand-held thermocouple thermometer (Omega engineering, HH-25TC) during three spring temperature surveys and for verifying continuous data. Temperature sensors (Maxime Integrated, Ibuttons DS1921Z), coated in rubber (Plasti Dip) for waterproofing, were used to measure longitudinal stream temperature profiles (T1 to T11), spring branch temperatures (S1 to S3), and the temperature of the Western tributary (T12). The thin rubber coating allowed the sensors to reach thermal equilibrium in under 5-minutes. Each temperature sensor was tied to a nail that was hammered into the streambed. A flat rock was rested on each nail head to shield the sensor from solar radiation. The pressure transducers (Solinst, Levelogger) also recorded temperatures in the lake and at the stream gauging stations. Manufacturer-specified accuracies are ±1°C and ±0.05°C for the Ibuttons andLeveloggers respectively. The accuracy of the Ibuttons was improved to ±0.25°C following the method outlined in Johnson et al. (2005). Sensors were calibrated in a constant temperature bath (VWR, 1157 Polyscience Refrigerated Circulator) over the range of temperatures observed in the field. Streambed temperature was also monitored by anIbutton , installed in P1 (Piezometer 1), 40 cm below the streambed (Figure 1).
Snapshots of stream surface temperature were captured with a thermal infrared camera (FLIR, E95). The manufacturer-listed accuracy is ± 2 °C and the thermal sensitivity is 0.04 °C. A programmable emissivity of 0.96 was used along with measured relative humidity and air temperature.