Improvement of the Canadian Land Surface Scheme (CLASS) under Stable Conditions

CLASS (The Canadian Land Surface Scheme), a land surface model currently used in the Canadian GCM, has been tested extensively against various field flux measurements in past five years. Comparisons between modelled and measured energy and water vapour fluxes under stable conditions were difficult because of the uncertainty in the flux measurement. Since canopy temperature can be measured using infrared thermometers, it is very useful to evaluate land surface models by comparing modelled and measured canopy temperatures under stable conditions. In this study, we found that the canopy temperature and sensible heat flux simulated by CLASS were often significantly underestimated during the nighttime, i.e. under stable conditions. Several corrections were applied: (1) canopy air temperature instead of canopy temperature was used in the calculation of the bulk Richardson number, (2) a free convection term was added to turbulent transfer coefficients to enhance heat transfer between the radiatively cooled canopy leaf surface and the air below under stable conditions and (3) the original soil evaporation formulation was replaced by an improved formulation. Our results showed that each correction contributed positively to the reducing of errors in the simulations of canopy temperature and sensible heat flux under stable conditions. When all the corrections were included in CLASS, the least errors in the simulations were obtained.