We measured ice cover at Falling Creek Reservoir (FCR) beginning in the winter of 2013-2014 and Beaverdam Reservoir (BVR) beginning in the winter of 2020-2021. We defined ice cover as the presence of ice (any thickness) covering more than 50% of the East-West diameter of the deep hole of FCR or BVR, as could be ascertained by a visual observer standing at the catwalk at the center point of the dam of the reservoir looking in the North direction (FCR) or at the boat launch looking in the North direction (BVR). This ice cover metric should only be applied for the deep hole of the two reservoirs because it may not be always representative of ice cover on the shallow upstream arm of the reservoirs. However, we can confirm from hiking around FCR on four sampling days with complete ice cover at the deep hole during the monitoring period that the ice coverage at the deep hole was the same as upstream and throughout the reservoir as a whole.
Carey received daily ice observations for FCR from Western Virginia Water Authority (WVWA) reservoir water treatment operators on site during most of 2013-2016. After winter 2016, reservoir water treatment operators were not consistently at FCR and thus visual observations from WVWA staff became sporadic. During winter 2017-2021, there were occasional visual observations from members of the Carey Lab at Virginia Tech on FCR or BVR sampling days or WVWA staff. In winters 2021-2022, we added in a new method of visual observation at FCR by taking automated photos of the reservoir with an ice camera (GardePro A3S Trail Game Camera 24MP, Guangdong, China). Two cameras were deployed on the end of FCR’s catwalk, with one facing in the Northwest direction and the other facing in the Southwest direction; both took one photo every hour. We used these data to validate the ice cover dates.
If visual observation data were not available, we used three alternate methods to determine ice cover. First, we used water temperature data from thermistors deployed at multiple depths in FCR from 2018-2022 (Carey et al. 2023b) or thermistors in BVR from 2020-2022 as well as HOBO temperature loggers deployed at 1-m depth resolution at the same BVR deep hole site in 2016-2018 (Carey et al. 2023a) to check for near-zero temperatures at the surface of the water column and inverse thermal stratification, which signified the presence of ice. Following Wetzel (2001), inverse stratification is easily disrupted by a small amount of wind, so if inverse stratification occurred amidst measurable wind at the reservoir (see Carey and Breef-Pilz 2023 for wind data), then we classified that period as having ice cover.
Second, the water temperature data were verified by looking at data from a research-grade meteorological station deployed on FCR's dam. The meteorological station had an upwelling shortwave radiation sensor deployed on the catwalk over the reservoir (see Carey and Breef-Pilz 2023). During periods of ice cover, the upwelling radiation exhibited higher radiation values because of the reflection of light off the ice during daylight hours, which was used to identify days of ice cover at FCR's deep hole. We also referred to the albedo data (upwelling shortwave radiation/downwelling shortwave radiation) as a check on cloudy days (Carey and Breef-Pilz 2023).
Third, the water temperature data were also verified by looking at dissolved oxygen concentrations measured by sensors deployed at the deep hole at 1.6 m, 5 m, and 9 m depths in FCR and 7.5 m, 1.5 m, and 0.5 m above the sediments at BVR (Carey et al. 2023a,b). During periods with more than 1 day of ice cover, dissolved oxygen depletion at depth was noticeable, which indicated inverse stratification.
Years without any data indicate that no ice was detected during those winters after the onset of data collection.
While these methods are not perfect, they represented the most robust way possible to maintain a daily ice cover record over time given the data available. Data flags are conservative and indicate ice periods without visual observations or additional dissolved oxygen data for double verification.
References: Carey, C.C. and A. Breef-Pilz. 2023. Time series of high-frequency meteorological data at Falling Creek Reservoir, Virginia, USA 2015-2022 ver 7. Environmental Data Initiative. https://doi.org/10.6073/pasta/f3f97c7fdd287c29084bf52fc759a801 (Accessed 2023-07-20).
Carey, C.C., A. Breef-Pilz, B.J. Bookout, R.P. McClure, and J.H. Wynne. 2023a. Time series of high-frequency sensor data measuring water temperature, dissolved oxygen, conductivity, specific conductance, total dissolved solids, chlorophyll a, phycocyanin, fluorescent dissolved organic matter, and turbidity at discrete depths in Beaverdam Reservoir, Virginia, USA in 2016-2022 ver 3. Environmental Data Initiative. https://doi.org/10.6073/pasta/4182de376fde52e15d493fdd9f26d0c7 (Accessed 2023-07-20).
Carey, C.C., A. Breef-Pilz, and W.M. Woelmer. 2023b. Time series of high-frequency sensor data measuring water temperature, dissolved oxygen, pressure, conductivity, specific conductance, total dissolved solids, chlorophyll a, phycocyanin, fluorescent dissolved organic matter, and turbidity at discrete depths in Falling Creek Reservoir, Virginia, USA in 2018-2022 ver 7. Environmental Data Initiative. https://doi.org/10.6073/pasta/f6bb4f5f602060dec6652ff8eb555082 (Accessed 2023-07-20).
Wetzel, R.G. 2001. Limnology, 3rd edition. Academic Press, New York.
