4.2 Spring temperature
The clusters of springs are grouped into three spring branches (S1, S2, and S3 in Figure 4). Thermal infrared images of the spring area were taken on August 14, 2019 at 12:15 to better visualise mixing of the three spring branches (Figure 5). Water surface temperatures were between 1.1 and 6.2°C. Distinguishable features include a cold side spring (Figure 5A) and a sharp thermal boundary between water originating from S1 and S2/S3. It can also be seen that GS1 is situated on the warmer stream side which must be considered during interpretation of the temperature data from the gauging station.
To better resolve the temporal variability, temperature sensors were placed in each of the three spring branches from June 11 to October 17, 2019 (Figure 6). Data gaps were the result of sensors being exposed to air (August) or full datalogger memory (July). Initial temperatures in the spring branches were equal at ~1.3°C but these rapidly diverged at the onset of snowmelt (Figure 6). The maximum recorded temperature in each spring branch was not reached simultaneously: S2 reached its maximum on July 15 with temperatures remaining high until late August. Temperatures at S1, on the other hand, were rising during this time and reached a maximum on August 22. By mid-September temperatures in S2 were back down to S3 levels. S1 temperatures had still not been reduced to S3 levels by October 17. Diurnal variations are likely due to heating of spring water following surface discharge. This is especially noticeable in S3 as it is the longest spring branch above the temperature sensor.
The observed spatial and temporal variability in spring temperatures was described above. Below, changes in spring temperature and discharge and changes in lake temperature and water level are explored (Figure 7). Note that stream temperature measured at GS1 was used, which was heavily influenced by high temperature discharge (Figure 5A). Even though this does not represent the spring system as a whole it captures the relationship between the lake and springs satisfying the purpose of this study. Additionally, water levels measured in P5 were used to include days when water levels were below the lake stilling well. True lake water levels were up to 5 cm higher than heads measured in the lake piezometer due to the downward flux into the lakebed. The downward head gradient beneath the lake accounts for the differences in readings between the lake piezometer and stilling well. To capture both spring snowmelt and the lake drying up, across two different seasons, the study period extends from May 1 to October 1, 2018 and 2019.
During the 2018 season the water table under the lake rose to the surface on May 18, after the onset of snowmelt. The maximum water level was 1.8 m on June 24, coinciding with a maximum spring discharge of 0.31 m3 s-1 (Figure 7B). After this, spring and lake temperatures increased rapidly until air temperatures dropped on July 18 (Figure 7C). By August 15, the lake had dried up and spring temperatures began to fall (Figure 7B and 7C). There are some key distinctions between the 2018 and 2019 seasons. Snowmelt started two weeks later in 2019 meaning that water levels reached the surface on May 31. Snowpack, measured at AWS1, prior to snowmelt was 60 cm deeper in 2018, however summer precipitation was greater in 2019. Also, the maximum lake water level was 20 cm higher in 2018. Lastly, the lake dried up later in 2019 and resurfaced following heavy precipitation on September 2.
Prior to snowmelt in late May 2019, temperatures at GS1 remained steady at ~1.3°C (Figure 7F). In 2019, after the surface water appeared in the lake, they continued rising rapidly during the following four days to 1.47 m (Figure 7E). On June 4 stream and lake temperatures reached seasonal lows of 0.9°C and 0.7°C, respectively. Daily average lake temperature proceeded to rise to 4.4°C by June 13 (Figure 7F). During this time, changes in lake water level and spring discharge corresponded to fluctuations in air temperature (Figure 7D). For example, in the first week of June daily average air temperature decreased by 10°C, lake water level decreased by 0.5 m and spring discharge decreased by 0.13 m3 s‑1(Figure 7D, E, F). As air temperatures rose water levels and spring discharge recovered.
Lake water level reached a maximum of 1.62 m on June 18, 2019 and stayed high until mid-July when it began its downward trajectory (Figure 7E). The decline in lake water levels coincided with a rapid increase in both lake and stream temperatures (Figure 7F). By August 25, the lake had dried up. However, the lake briefly appeared again following precipitation on September 2, which was confirmed by a field visit on September 4. This pattern was reflected in stream temperatures as a decrease when lake water level was below the surface and an increase when it resurfaced.