Landscape-scale CO2, H2O vapour and energy flux of moist-wet coastal tundra ecosystems over two growing seasons

1 Eddy covariance was used to measure landscape-scale fluxes of CO2, H2O vapour and energy in moist-wet sedge tundra ecosystems of coastal arctic Alaska. 2 Net radiation (R-n) was partitioned equally into sensible (H) and latent heat (L-e), although a larger proportion of R-n was partitioned into L-e during and after the early season snow-melt period, while proportionally more R-n was partitioned into H as the season progressed. Ground heat flux (G) comprised an average of 15% of R-n throughout the growing season. 3 Interannual variation in evapotranspiration (ET) was negligible despite a 30% increase in total precipitation in 1995. On average, seasonal estimates of ET were 120 mm (1.3 mm day(-1)), which are similar to estimates found in the literature. 4 Net CO2 influx generally peaked during the mid-day period when photosynthetic photon flux density (PPFD) was at a diurnal maximum, while peak net efflux occurred between 23.00 and 02.00 hours when PPFD was at a diurnal minimum. Most of the variation in diurnal net CO2 exchange was explained by diurnal variations in PPFD, with the functional response best described as a rectangular hyperbola. 5 The coastal tundra landscapes were net sources of 0.02-0.05 mol m(-2) day(-1) during the early season snow-melt period. During the mid-season peak in productivity, which occurred as early as 10 July in 1994 and as late as 10 August in 1995, these landscapes were net sinks of on average 0.09 mol m(-2) day(-1). Sink strength decreased considerably after the mid-season peak in productivity, presumably because of reductions in PPFD, an increase in water table depth, relatively high temperatures, and a decline in leaf area. 6 Over both growing seasons, the coastal tundra landscapes were sinks of 1.7 mol m(-2). This is comparable with the 'long-term' estimate of net CO2 exchange calculated from the literature.