Tower fluxes measured at the Boreal Ecosystem-Atmosphere Study southern study area old aspen (SSA-OA), southern study area old jack pine (SSA-OJP), and northern study area old jack pine (NSA-OJP) sites during the three 1994 intensive field campaigns (IFCs) (May 24 to June 16 (IFC-1), July 19 to August 10 (IFC-2), and August 30 to September 19 (IFC-3)) were compared to fluxes simulated by a land surface model for the same period. Comparisons were limited to the average diurnal cycle for these periods to mitigate large day-to-day variability in the observations and problems with missing data. For consistency with the global implementation of the model, vegetation and soil parameters were not set to site-specific values but rather were the generic needleleaf evergreen and broadleaf deciduous vegetation and the generic sandy and loamy soil used in the global model. Despite the use of generic vegetation and soil, the model reasonably simulated the diurnal cycle of sensible heat, latent heat, net radiation, and CO2 fluxes for the SSA-OJP and SSA-OA sites. The main errors were that the model did not reproduce the midday reduction in latent heat seen at the SSA-OJP site during IFC-1 and IFC-2 and had less photosynthetic CO2 uptake than observed at the SSA-OA site. Differences in vegetation structure and physiology between the two sites were important to accurately simulate the fluxes. The needleleaf evergreen vegetation resulted in higher net radiation and a higher Bowen ratio than the broadleaf deciduous vegetation. Soil differences were less important. The NSA-OJP site was not so well simulated: midday latent heat flux was overestimated, and photosynthetic CO2 uptake was underestimated during each IFC. The only difference in the simulated southern and northern jack pine sites was in their atmospheric forcings; vegetation structure and soil types were the same. These results suggest the model is able to reproduce variability between vegetation types but not within vegetation types. ¿ 1997 American Geophysical Union