Boreal lakes are a known source of greenhouse gases (GHG’s) into the atmosphere especially in the boreal zone, where lakes are numerous. Due to global warming the magnitude of GHG release from lakes is posed to increase. And although this is a well recognized fact, there remains uncertainties in our knowledge on lake GHG fluxes. This is due to several reasons: remote locations of such lakes often make research logistically difficult, the environmental conditions can be rough for research infrastructure and the most commonly used flux chamber method is heavily biased towards daytime flux values. Here, we introduce preliminary results based on a three year long dataset of carbon dioxide and methane fluxes collected via the eddy covariance (EC) method over a large subarctic Lake Pallasjärvi in Finnish Lapland. The site is located in a unique position where on one hand the southern lobe of the EC footprint represents fluxes from a small and shallow bay which receives runoff from the surrounding forest and peatland, and on the other hand, the northern lobe extends to the large and deep main basin of the lake, the data thus representing essentially two different lakes. EC measurements are supplemented by water side measurements of pCO2, pCH4 and temperature stratification as well as flux chamber measurements. This dataset provides a look into the often overlooked aspects of lake gas flux research: diurnal variation of fluxes, fluxes during the spring overturn period and in general observations over a large subarctic lake, on which measurements from past are few and far between.