We developed a dynamic regional vegetation model to address problems of responses of regional vegetation to elevated ambient CO2 and climatic change. The model takes into consideration both local ecosystem processes within a patch or grid cell, such as plant growth and death, and mass and energy flow, such as plant migration, across adjacent grid cells. The model is able to couple vegetation structure dynamics and primary production processes. The normalized differential vegetation index from meteorological satellite AVHRR was used to parameterize the model. Plant migration rates were derived based on effective seedling distribution around parent plants. The model was applied to Northeast China Transect at a spatial resolution of 10 min latitude by 10 min longitude per grid cell and a temporal resolution of 1 month. The results indicated that with doubled CO2 concentration, a 20% increase in precipitation and a 4 ¿C increase in temperature, the model predicted that net primary productivity (NPP) of Larix forests, conifer-broadleaf mixed forests, Aneurolepidium chinense steppes, Stipa grandis steppes, and wetland and salty meadows would decrease by 15% to 20%. However, NPP of deciduous broadleaf forests, woodland and shrubs, Stipa baicalensis meadow steppes, and desert grasslands would increase by 20% to 115%, as predicted by the model for the same climatic scenario. The average NPP of natural vegetation over the whole transect would decrease slightly, largely because of the compensation between the positive effects of increased CO2 and precipitation and the negative effect of increased evapotranspiration induced by increased temperature. ¿ 1998 American Geophysical Union