Climate Dynamics (1997) 13 : 57–74 q Springer-Verlag 1997 Influence of the temperate and boreal forests on the Northern Hemisphere climate in the Météo-France climate model H. Douville, J.-F. Royer Météo-France/CNRM, 42 Avenue Coriolis, 31057 Toulouse Cedex, France Received: 5 September 1995 / Accepted: 12 August 1996 Correspondence to: H. Douville Abstract. A temperate and boreal deforestation ex- periment has been performed at Météo-France using the ARPEGE climate model. A first simulation was performed as a control with a present-day vegetation map, and another one with all forests north of 45 7N replaced by meadows. Prescribed monthly mean clima- tological SSTs were used in both integrations. The AR- PEGE climate model includes a physically based land surface scheme, which has been tested both on snow- free and snow-covered sites, and has a relatively high horizontal resolution. Results of the 4-year integra- tions suggest that forests exert a strong influence on the surface climate of the temperate and boreal re- gions. Deforestation induces a significant cooling which modifies the atmospheric circulation simulated in the high latitudes, and also in the tropics. The most important impact is observed during the melting sea- son which is delayed by the forest removal. This result is consistent with preliminary stand-alone experiments showing that the atmospheric boundary layer can be heated by the forest, even if the ground is covered by snow. The study confirms that vegetation feedbacks should be included when performing future climate studies such as doubled CO 2 experiments, eventhough many uncertainties still remain with regard to other physical aspects of the climate models. 1 Introduction The land surfaces represent a crucial component of the Earth’s climate system. In the past, the Earth has un- dergone large climatic variations which were asso- ciated with important changes in ice and snow cover extent and with strong migrations of the terrestrial eco- systems. A recent numerical study suggests that the vegetation is strongly involved in the initiation of the glaciations (Gallimore and Kutzbach 1996). The results show that the climate feedback from high-latitude for- est retreat might have contributed to the onset of the last ice age. This conclusion is supported by the fact that most GCMs are not able to induce incipient gla- ciation with just orbital and CO 2 forcings. In the future, global warming due to increased con- centrations of atmospheric greenhouse gases might also have a dramatic effect on the cryosphere and the biosphere. The rate of change in surface air tempera- ture might be much faster than the warming following the last ice age and many tree species might not be able to disperse rapidly enough to reach areas of suitable climate. Numerical studies indicate that the warming should be particularly strong in the Northern Hemi- sphere high latitudes, and this could increase the agri- cultural and economical potential of these regions. Conversely, the changes occurring at the Earth’s sur- face will probably feedback into the atmosphere. Many studies based on general circulation models (GCMs) have shown that climate is determined by a dynamic equilibrium in which the atmosphere affects the land surfaces and is affected by them. The pioneer- ing work of Charney et al. (1977) suggested that deser- tification could be explained by an irreversible feed- back between precipitation and surface albedo. The sensitivity of climate to soil moisture conditions has been widely investigated (Shukla and Mintz 1982; Mintz 1984; Delworth and Manabe 1988; Serafini 1990) as well as the important role of snow cover (Yeh et al. 1983; Barnett et al. 1989; Cohen and Rind 1991; Ran- dall et al. 1994). In the 1980s, GCMs were also used to study various climate scenarios. Doubled CO 2 experi- ments have been performed to analyse the possible im- pact of the increase in the greenhouse effect (Manabe and Stouffer 1980; Manabe and Wetherald 1987; Schle- singer and Mitchell 1987) and have shown the impor- tance of different feedbacks involving snow cover (In- gram et al. 1989; Meehl and Washington 1990) as well as vegetation (Henderson-Sellers 1990; Prentice 1990; Graetz 1991). Other experiments have analysed the potential effect of tropical deforestation (Henderson- Sellers and Gornitz 1984; Wilson 1984; Dickinson and