The physiologically based growth model CenW was used to simulate wood-productivity responses of Pinus radiata forests to climate change in New Zealand. The model was tested under current climatic conditions against a comprehensive set of observations from growth plots located throughout the country. Climate change simulations were based on monthly climate change fields of 12 GCMs forced by the SRES B1, A1B and A2 emission scenarios for 2040 and 2090. Simulations used either constant or increasing CO2 concentrations corresponding to the different emission scenarios. With constant CO2, there were only slight growth responses to climate change across the country as a whole. More specifically, there were slight growth reductions in the warmer north but gains in the cooler south, especially at higher altitudes. For sites where P. radiata is currently grown, and across the full suite of GCMs and emission scenarios, changes in wood productivity averaged +3% for both 2040 and 2090. When increasing CO2 concentration was also included, responses of wood productivity were generally positive, with average increases of 19% by 2040 and 37% by 2090. These responses varied regionally, ranging from relatively minor changes in the north of the country to very significant increases in the south, where the beneficial effect of increasing CO2 combined with the beneficial effect of increasing temperatures. These relatively large responses to CO2 depend on maintenance of the current adequate fertility levels in most commercial plantations. Productivity enhancements came at the expense of some soil-carbon losses. Average losses for the country were simulated to average 3.5% under constant CO2 and 1.5% with increasing CO2 concentration. Again, there were regional differences, with larger losses for regions with lesser growth enhancements, and lesser reductions in regions where greater productivity enhancements could partly balance the effect of faster decomposition activity.