Short-term environmental controls of heat and water vapour fluxes above a boreal coniferous forest: model computations compared with measurements by eddy correlation

Eddy correlation and stern how measurements were coupled with detailed micr oclimate and soil measurements made in a boreal Scots pine forest in the la te growing season of 1998 to determine sensible and latent heat fluxes from the soil and the canopy separately. A 'resistance/energy' model is constru cted and parametrized in order to reproduce the dynamics of water and heat exchange between the soil, the canopy and the atmosphere as a part of a lar ger forest ecosystem model (FinnFor; Kellomaki and Vaisanen, 1997). Unique features of the present model are that (1) energy flux equations are expres sed in terms of conceptual resistances and their solutions are obtained by closing two surface energy budget equations defined separately for canopy a nd soil surface; (2) the forest canopy is divided into shaded and sunlit fr actions in the radiation transfer submodel and the canopy resistance submod els; (3) a numerical integrating solutions are derived separately for net r adiation absorption in the canopy, bulk canopy resistance and the bulk aero dynamic resistances of the forest; and (4) iterative determinations of cano py water potential based on a classical one-dimensional water how model ena ble the model to represent explicitly the interaction between the above-gro und and the below-ground water dynamics. The model is validated against 19- day flux measurements. In general, the total system sensible heat flux (H), total system latent heat flux (lambda E), canopy latent heat flux (lambda E-c), and soil surface heat flux (G(s)) computed by the model matched well with the measured data. Based on 1/2 h flux measurements, daily lambda E va ried from 0.50-7.38 MW m(-2), H from 0.64-8.3 MW m(-2): and lambda E-c from 0.30-6.93 MW m(-2). The Bowen ratio (H/lambda E) ranged from -4.5 to 9.8, but 82% of the values for the Bowen ratio were within 0.5-2.5. The model co mputations showed that daily lambda E-c and H-c accounted for 21-64% and 43 -66% of the daily total system flux, respectively. Daily soil latent heat ( lambda E,) and soil sensible heat (H-s) fluxes accounted for 0.02-4.5% and 0.05-7.6%, respectively, and the daily energy storage within the canopy (S- c) and G(s) accounted for 0.1-7.2% and 0.8-5.6%, respectively. Plotting of 1/2 h flux data against a single environmental factor indicated that a 68% change in lambda E-c and a 72% change in H-c can be explained by a change i n canopy radiation absorption (R-nc) at the 5% probability level. The high correlation between the canopy fluxes and R-nc could be related to the mode rate weather conditions and high soil water content during the selected day s, whereas lambda E-s, H-s, S-c and G(s) give no significant correlation wi th R-n. As expected. lambda E-c was strongly dependent on canopy resistance (r(cs)), but less impact on aerodynamic resistances during most of the mea suring time. The proportion of energy partitioning in H and lambda E exhibi ted a clear diurnal trend and was mainly controlled by the system total res istance and the vapour pressure deficit, but less related to changes in soi l water content. (C) 1999 Elsevier Science B.V. All rights reserved.