Introduction

Lakes have been recognized for their important role in the global carbon cycle and are thought to act as a net source of atmospheric CO₂, contributing 1 Pg C yr⁻¹ cumulatively. Lake carbon dynamics are changing under the influence of climate warming — the recent browning of lotic systems and prevalence of CO₂ supersaturated lakes at high latitudes have been attributed to the mobilization of terrestrial organic carbon from surface warming and an enhanced hydrologic cycle. The evasion of CO₂ from the surface waters of lakes is the greatest and most spatiotemporally variable exchange of carbon between lakes and their immediate environment, relative to lateral exchanges in dissolved and solid forms, and enhancement of this source needs to be better understood. Due to the high temporal variability associated with lake-atmosphere gas exchange, flux estimates benefit from continuous, automated monitoring rather than the collection of point samples over space and time. Here we present a study in which the Eosense eosFD CO₂ flux sensor is used for the first time to capture high-resolution temporal variability in lake CO₂ evasion.

An eosFD floating calmly on Lochaber Lake continuously measuring water-atmosphere CO2 fluxes.

Study Area & Equipment

Lochaber Lake is a dimictic, oligotrophic lake located in northeastern Nova Scotia at 45°26’35.88″N 62°00’50.04″W. Lochaber Lake is the only headwater lake that drains into the St. Mary’s River system. The lake is fed by eight streams, three of which are ephemeral, and is drained by one outlet at its southern end. Lochaber is at the southern extent of the Antigonish highlands, and is underlain by sedimentary deposits of the Knoydart Formation, primarily mudstone, siltstone, and shale.

Land uses in the watershed include seasonal and permanent residences, and agriculture. The waters of Lochaber are clear, and macrophytic vegetation thrives along the coastline. The lake has the shape of an elongated oval, with a total shoreline of 18 km, a width of 0.75 km, and a mean depth of 21.8 m. Given the lake shape and bathymetry, the littoral zone dominates Lochaber Lake. Historically, the watershed of Lochaber lake was covered by a dense stand of Acadian forest, with a mix of red spruce, yellow birch, balsam fir, sugar maple, red pine, eastern white pine, eastern hemlock, and American beech. Two centuries ago, much of this forest was cleared for farming. However, farming declined during the mid 19th century and many of these fields are now covered with secondary successional forest species, predominantly softwoods.

In November 2016, two floating platforms were equipped with eosFD CO₂ flux sensors, solar panels, batteries, and a Campbell Scientific radio that was used to download data from shore without disturbing the platform. The platforms were both placed in the littoral zone of the lake and moored to the lake bottom.

Efflux Results and Relationships with Environment

Fluxes of CO₂ from the lake surface were monitored for a period of approximately 2 weeks. Both eosFD sensors tracked each other well, with a whole period average flux of 0.15 μmol CO₂ m⁻² s⁻¹. The fluxes showed considerable temporal variability, both diurnally and over the 2-week period, shown below in Figure 1 (a). To understand the main drivers of variability the temperature and precipitation at a nearby Environment Canada monitoring station were compared with the flux measurements from the eosFD sensors. The clear diurnal pattern in flux is likely associated with either temperature or light availability – but there is no strong correlation between average site temperature and average flux at the site (Figure 2). Similarly, the increase in flux around day 321 doesn’t appear to be directly related to temperature.

Unfortunately during the period between days 324 and 326 we had a power failure in both sensors causing a data gap, but it also gave a clue as to the potential cause of the increased flux. The eosFD temperature data show a distinct drop off in the diurnal temperature variability during that period of time, and the data from the Environment Canada station shows near saturation humidity values (Figure 1b) indicating there was likely thick cloud cover and little solar radiation getting through to the lake surface. This lack of light as well as the impacts of precipitation on the lake nutrients and mixing likely caused this temporary increase in the total CO₂ efflux from the littoral zone.

Figure 1: (a) CO2 fluxes from the two eosFD sensors (green and blue dots with smoothed lines) deployed in the lake with a smoothed fit and daily precipitation (bars). (b) Temperature from the eosFD units (red, orange) and the Relative Humidity at the site (blue).

Figure 2: Correlation between daily average CO2 fluxes and daily average temperature at the site.

Conclusions & Future Work

Using the eosFD CO₂ flux sensor mounted to a moored floating platform we were able to observe the diurnal and synoptic impacts of temperature and precipitation on CO₂ flux in the littoral zone. In other lake studies, researchers have found that lakes tended to be weak sources in the spring, with source intensity growing in late summer and in fall. For example, Vesala et al. [2006] reported May CO₂ fluxes of 0.2 μmol CO₂ m⁻² s⁻¹ and August fluxes of 0.4 μmol CO₂ m⁻² s⁻¹. These results are similar to the fluxes we observed in Lochaber during the study period. During the late fall at this site, the impact of precipitation and light availability are the dominant drivers of variability in CO₂ emissions.

The results presented here are part of a larger study of Lochaber, completed by researchers at the St. Francis Xavier University FluxLab. The results of this study, authored by St. FX student Lynsay Spafford, are under review for publication in the Journal of Geophysical Research.