hirlam

Abstract This report describes the surface scheme, developed for HIRLAM, and called ”newsnow” duringthe development time. It is physically based on the land surface scheme used at the Rossby Centre(Samuelsson et al., 2006), with some exceptions. The code structure and data assimilation has itsorigin in the earlier HIRLAM surface scheme (Rodriguez et al., 2003).It is a tiled scheme with 7 tiles: Sea, ice, bare soil, low vegetation, forest, snowcovered bare soil andlow vegetation, and finally forest with snow on the forest floor. The basic difference from the earlierscheme is the introduction of the snow scheme, and a completely new forest formulation. The othertiles are in principle the same as the old scheme, but we have replaced the force-restore formulationwithin the soil, with more layers and heat conduction.Thus we have two different snow packs, with separate energy balances. For each tile we compute thelatent and sensible heat fluxes and momentum fluxes, and they are weighted together, according totheir grid fractions, to provide a lower boundary condition for the vertical diffusion. The forest has aseparate canopy temperature, common for the tiles with and without snow on the forest floor, so theyare treated together within the scheme. At present no snow on sea ice is incorporated, but we haveone temperature for the uppermost ice layer and the snow.The water storages treated in the model are, except for soil moisture, intercepted water on low vegetationand on the forest, snow water equivalent on the ground and forest floor. The snow packs alsotreat liquid water within the snow, which can refreeze. We also treat intercepted snow on the forestcanopy, in a simplified way, without a separate temperature.The main differences from the climate model of the Rossby Centre, is that we have three thermallayers in the soil and a relaxation to a deep climatological temperature instead of a no flux lowerboundary condition , and that we have kept the force-restore formulation of soil moisture from theearlier HIRLAM scheme.We have also a simpler estimation of the snow fraction. The deviations from the climate version, hasmainly to do with the adaptation to the surface data assimilation.

Kain, 2004: The Kain-Fritsch convective parametrization. An update. J. Appl. Meteor. 43, 170-181

Noilhan and Planton 1989: A simple parametrization of land surface processes for meteorological models. Mon Wea Rev. 117, 536-549

Zhang et al. 2003: A modified formulation of fractional stratiform condensation rate in the NCAR community atmospheric model. J. Geophys. Res. 108, 1-11

Bott, A., 1989a. A positive definite advection scheme obtained by non-linear renormalization of the advective fluxes. Monthly Weather Review, 117, 1006-1015.

Bott, A., 1989b. Reply. Monthly Weather Review. 117, 2633-2636.