Surface energy fluxes and control of evapotranspiration from a Swedish Sphagnum mire

The components of the energy budget for an open Swedish boreal bog were quantified over two growing seasons and an analysis of the influence of factors such as peat wetness, vapour deficit and radiative forcing was made, in order to understand how they can be represented by physically-based models. The presence of a pronounced micro-topography caused a variation of the surface wetness as well as the cover of vascular plants. This was reflected in the spatial variation in soil temperatures and the ground heat flux was almost twice as much in hollows as in ridges. A theoretical function of peat thermal conductance with porosity and water content agreed with calculated damping depths. The surface roughness was low (z(0) = 2-3 cm) and did not change during the season. Bluff-rough effects had to be considered for heat and vapour fluxes. Albedo changed little but significantly with development of vascular plants (from 0.11 to 0.14). The Bowen ratio did not change with temporal fluctuations in peat wetness but depended mainly on net radiation, air temperature and relative humidity. Consequently, there was a trend of falling Bowen ratio both during the day and during the season from May (monthly value 0.9) to September (monthly value 0.6). The bulk surface resistance (r(s)) to evapotranspiration was seldom close to zero and varied little (mean r(s) +/- S.D. = 160 +/- 70 s m(-1)). Its variation depended mainly on vapour pressure deficit and was not directly correlated to peat wetness. At least half the evapotranspiration was estimated to originate from the moss surface. The evapotranspiration was 60-80% of the Penman potential evaporation. The relatively large amount of non-transpiring biomass above ground gave an effect with a relatively high Bowen ratio and a large but stable r(s). It was concluded that there exists a canopy layer, within which the distribution of heat and vapour sources may vary with time. As the sources of evapotranspiration probably vary, the stable surface resistance is not caused by a constant regulation of moss evaporation but by small-scale advection effects. The low coupling to the ambient vapour deficit resulted in the evapotranspiration fitting well with the Priestly-Taylor model, with an alpha value of 0.8. Simple models may then be very effective for a surface type like this. However, the influence of different surface properties on the parameterisation needs to be investigated further. The use of more complicated, but physically sound, models demands more detailed studies of both physical properties of mosses and effects of differences in surface cover and micro-topography. (C) 2001 Elsevier Science B.V. All rights reserved.