5.2 Effects of direct precipitation on stream temperature in cold regions
Previous studies have concluded the direct precipitation heat flux to be negligible (Evans et al., 1998) and observed stream cooling has been attributed to the resulting flow generation processes (Brown & Hannah, 2007). We began this study with the aim of quantifying these subsurface flows to improve stream temperature models. However, we found that direct precipitation made up a large component of the energy balance during solid precipitation events (Figure 11B) consistent with the findings of Leach and Moore (2017). Furthermore, Hathataga Creek did not undergo any noticeable cooling during rainfall events. This suggested that an increase in advective inputs, from groundwater and hillslope pathways, did not play an important role in altering stream temperatures.
The stream energy balance was dominated by shortwave radiation during the day and longwave radiation at night (Figure 11B). The largest temperature residuals occurred between 14:00 and 15:00 when measured temperatures were up to 1.5 °C warmer than modelled ones. The uncertainties in shortwave radiation came from estimates of shading. This is attributed to shading values that were fixed in the model and therefore did not change throughout the day. The radiation measurements were taken in an open meadow. For most of the day the stream was well shaded as captured by the shading estimates. However, in mid-afternoon the position of the sun was such that it shone along the length of the stream and shading estimates became erroneous. In future modelling studies, shading and view to sky coefficient should be quantified in the field using hemispherical images (Garner, Malcolm, Sadler, & Hannah, 2017) or a densiometer (Gravelle & Link, 2007).
Uncertainties in longwave radiation came from the assumption that canopy surface temperature equals air temperature. In reality, longwave radiation from surrounding vegetation (Lveg, Equation S2-5) is likely greater on sunny days when canopy temperature is greater than air temperature (Pomeroy et al., 2009). The remaining fluxes were relatively small in comparison. Uncertainties in the latent and sensible heat flux came from meteorological measurements made at 3.45 m height rather than 2 m. As a result, wind speed and the latent heat flux were likely lower. Furthermore, measurements were made over the meadow rather than the stream surface. It has been shown that wind speed is higher and more variable at exposed regional weather stations compared to sheltered microclimate sites (Benyahya, Caissie, El-Jabi, & Satish, 2010). In this study the stream surface is located in a topographic low relative to the meadow. Therefore, wind speed was likely lower than measured. During the snowfall event the direct precipitation flux was ten times greater than other energy inputs.
The energy consumption caused by the melt of solid precipitation have important implications for stream temperatures in alpine environments and other cold regions. Equation 2, which describes this process, should be considered in stream energy balance models. Without this flux, parameter estimates from calibration would be less reliable to account for the lower stream temperatures. In other mountain regions, such as the Vernagtbach basin in Austria, there are years when solid precipitation accounted for up to 70 percent of total precipitation during the ablation season. During the same period, there were up to 50 days with snowfall (Escher-Vetter & Siebers, 2007). In the Hathataga catchment, solid precipitation accounted for 23 percent of total precipitation from May to September 2019 with 20 days of snowfall.
This could also have important implications for the field of meteorology where it is a challenge to determine precipitation phase. Greater than expected temperature changes in small streams or of fluid in a precipitation gauge (Geonor, T-200B) could be used as a proxy for solid/liquid phase rather than using an air temperature threshold or expensive laser-based sensors (Campbell Scientific, CS125). The advantage of this method is that cooling associated with the latent heat of fusion only occurs for solid precipitation, irrespective of air temperature. This was especially relevant during summer hailstorms when surface air temperatures were up to 10 °C but stream cooling was observed.