Methane ebullition from lake sediment is an important source of atmospheric methane. Previous studies have suggested that temperature variations, water level changes, atmospheric pressure fluctuations and wind-induced current can affect ebullition. However, most of those studies were conducted during open-water season. There is a lack of observations during ice-cover, despite of the abundance of seasonally ice-covered lakes.

In this dataset, we present high-frequency ebullition intensity data, atmospheric pressure data, bottom-water temperature data, and turbidite data from Base Mine Lake (57° 1' N, 111° 37' W in Alberta, Canada) during ice cover. During the study period, the water level in the lake is stable. Also, due to the ice cover, the impact of wind-induced current is negligible. The dataset shows that ebullition during ice cover is regulated by atmospheric pressure variations; the stable bottom-water temperature has on correlation with the ebullition. The dataset also shows that turbidity at depth in the lake increases during ebullition events.

The ebullition intensity data is collected using a single-beam echosounder (Echologger EA400), which is mounted beneath ice. Bursts of 50 pings over 25 seconds (2Hz) are logged once every hour.

The turbidity is collected at 30-minute intervals using a RBRDuo logger with a Seapoint turbidity sensor attached to a mooring chain. The instrument was calibrated by the manufacturer (RBR). To remove spikes, for example, a three-point median can be applied to the raw data .

Water temperature data is collected every 10 seconds using a RBRSolo logger, attached to the same mooring chain as turbidity logger. The instrument was calibrated by the manufacturer (RBR). To smooth the short-term variations, for example, a 10-minute moving average can be applied to the raw data.

Atmospheric pressure data is collected at the center of lake. The pressure data has been compared with the atmospheric pressure at a nearby airport,. The variations in atmospheric pressure at these two stations were nearly identical, except a constant pressure offset.

Notes for echo-sounding data: The echo-sounding data is used to derive the ebullition intensity. The echosounder records the full profiles of the water column beneath it. Every hour, a burst of 50 pings over 25 seconds is set to record the ebullition activities through the water column. Each burst of 50 pings produces an echogram. In the echogram, there are several common elements: rising bubbles, floating reflectors, random noises, a blank after transmit region at the top and the lakebed at the bottom. A filtering  methode discribed below is used to remove the interference while preserving the signal from the rising bubbles.

This filtering method is designed with consideration of the two main features of the echograms: 1) In each echogram, the rising bubbles are diagonal lines and in general have much higher backscattering intensity than the background noise and the average row intensity. 2) the floating reflectors are time invariant and therefore their intensities are similar to the depth specific time average. The filtering method has three steps. Firstly, a certain range of depth (5.2 m to 8.5 m) is chosen to exclude the depths with interference close to the instrument and to exclude the lakebed. Secondly, the influence of the system noise is excluded, i.e. if the intensity of each element is less than the noise it is set to zero. Thirdly, in every row of the echogram, if the value of each element is less than one fifth the time averaged value, it is set to zero. After applying a filtering method, the intensity of the rising bubbles in the individual echogram is averaged and used to represent the ebullition intensity at that hour. 

There are five sheets in the Excel, including ebullition intensity, atmospheric pressure, temperature and turbidity data. The missing data is stored as NaN. There is also an associated README.txt.