Gas flux across the air-water interface was determined by the floating chamber method four times during the year period in 2005 — 2006. The open-bottom floating PE chambers (5L domes with an area of 0.03 m2) were maintained on the water’s surface by a floating body (Styrene) attached to the outside. The chambers (n = 4 — 5) were allowed to float on the water’s surface for a period of 3 hours. Previous measurements confirmed that time to be quite enough to establish linear dependence of concentration change inside the chambers on time for the gas samples collected every 30 min over a 3 hour period. Due to trees on the banks, the chambers at all study sites were continuously in the shade. On each sampling occasion, ambient air samples were collected for determining the initial background concentrations. Samples of headspace gas were collected through the rubber stopper inserted at the chamber’s top, and stored in 100mL PE gas-tight syringes until analysis. Emissions were calculated as the difference between initial background and final concentration in the chamber headspace, and expressed on the 1m2 area of the surface level per day according to the formula:

where F is a gas flux in mg m-2day’1; a is a concentration of particular gas in the chamber headspace in pg L-1; cr is a concentration of particular gas in background air; V is volume of the chamber in L; t is time of incubation in hr; p is an area of chamber expressed in m2 . For each chamber, the fluxes were calculated using linear regression based on the concentration change as a function of time, regardless of the value of the coefficient of determination (cf. Duchemin et al. 1999, Silvenoinen et al. 2008).

In order to assess emissions produced from a total stream area, the stream was divided into five stretches according to the channel width, water velocity and substrate composition. For each stretch we have then chosen one representative sampling site (locality I-V) where samples of both stream and interstitial waters and sediments, respectively, were repeatedly taken. Localities were chosen in respect to their character and availability by car and measuring equipments. For calculation of whole-stream gases emissions into the atmosphere, the total stream area was derived from summing of 14 partial stretches. The area of these stretches was caculated from known lenght and mean channel width (measured by a metal measuring type). Longitudinal distance among the stretches was evaluated by using ArcGIS software and GPS coordinates that have been obtained during the field measurement and from digitalised map of the Sitka stream. The total area of the Sitka stream was estimated to be 181 380 m2 or 0.18 km2. Stretches have differed in their percentual contribution to this total area and also by their total lenght (Table 1).

The total annual methane emissions to the atmosphere from the five segments of the Sitka stream, Ea (kg yr-1) were derived from seasonal average, maximum or minimum emissions measured on every locality and extrapolated to the total area of the particular segment. The total methane emissions produced by the Sitka stream annualy were then calculated according to the following formula:

Ea = (X pi * Fi * 365) / 1 000000 (5)

where Ea is average, maximal or minimal assess of emission of methane from the total stream area in kilograms per year; pi is an area of stretch (in m2) representing given locality; Fi is average, maximal or minimal assess of the methane from a given locality expressed in mg m-2day2.