Abrupt climatic changes have occurred several times in the past, leading to large-scale modifications of vegetation patterns with important consequences for the global carbon cycle. Dynamic global vegetation models (DGVM) constitute an advanced tool for reconstructing past or predicting future shifts in vegetation distributions in response to climatic change on a global scale. The Lund-Potsdam-Jena (LPJ) model is a DGVM that also includes a complete description of terrestrial-vegetation carbon cycling. Here, it is used for a long-time integration simulating terrestrial ecosystem responses to an abrupt climatic change event. Climate data from an 850-year-long coupled ocean-atmosphere model (ECHAM3/LSG) experiment representing a highly idealized Younger Dryas (c. 12 ka BP) like event are used to study the reactions of the vegetation distribution and changes in terrestrial carbon storage. The main feature of the Younger Dryas simulation experiment is the suppression of the Atlantic thermohaline circulation leading to a significant cooling of the Northern Hemisphere accompanied by a large-scale precipitation decrease. The simulation exhibits a significant shift of the vegetation distribution in the Northern Hemisphere during the cold period in conjunction with a change in global total terrestrial carbon stocks of 180 X 10¹² kg C as a consequence of the climatic change event. The response time of the terrestrial biosphere lags the climatic changes by about 250 years for vegetation and 400 years for soil-carbon pools.