Editorial DGVM responses to the latest IPCC future climate scenarios In the spring of 2006, the future looked bright and we were sure our special issue would be coming out before Christmas. It is now spring 2007 and the future has shown us that no one has a crystal ball, not even modelers, to predict the future. A new job and a major family health crisis changed the lives of the two guest editors. Furthermore delays in the original submission of the final versions of the papers after three timely reviews rendered the early 2007 events in the lives of the editors particularly ill-timed. This illustrates however what climate change experts have told us all along: extreme events (such as a life threatening emergency) will shape the response of the earth system to future conditions. This short preamble aims to thank the reviewers of our five manuscripts for their prompt comments. It also aims at thanking the authors for their patience while they waited to see their work finally come to print. Finally it also aims at thanking Elsevier for their patience while waiting for the final package to arrive. We believe this set of papers includes important information about DGVMs and about what the future may hold for world ecosystems. This special issue means to illustrate the power of dynamic global vegetation models and their usefulness in climate change science. Dynamic Global Vegetation Models are generally designed to track simultaneously vegetation changes driven by climatic variability, together with the associated fluxes of water and carbon (possibly also energy exchanges and nutrient flows). In general, DGVMs estimate changes in short and long-term vegetation productivity (both net primary productivity, NPP, and net ecosystem exchange, NEE), competition among lifeforms (plant functional types) for resources, the effects of disturbances, and mortality. External forcings to the DGVMs include both climate and soil characteristics. Cramer et al. (1999) and Scurlock et al. (1999) emphasized the need for improving the validation methods to determine both the sensitivity and the validity of model responses to the forcing climate. Cramer et al. (2001) compared six DGVMs in the first published inter-comparison, comparing simulated vegetation maps to a satellite-derived global map and simulated carbon fluxes with published estimates. Moorcroft (2006) reiterated the need for models to be compared to observations and thus tested for accuracy. Pilkey and Pilkey-Jarvis (2007), following Oreskes (1994), argued in their latest book “Useless Arithmetics” that models cannot be validated and even if they could, their reliability for predicting the future would still remain unproven. In essence, without validation, global vegetation models are just videogames! In this special issue, several authors present validation exercises with different DGVMs. In particular, Price and El Maayar tested the IBIS dynamic global vegetation model at several eddy covariance sites located in Canada and the USA (but withdrew their manuscript). The model was then run into the future with various future climate scenarios. Much of the positive response to future climate could be attributed to projected changes in CO 2 concentration, raising questions about the significance of elevated CO 2 concentrations in large-scale simulations of vegetation change. These results confirmed the importance of an accurate representation of CO 2 effects on plant processes. A similar conclusion could be drawn from Cramer et al. (2001) but at that time, the lack of data concerning CO 2 impacts at the regional scale made it impossible for the modelers to refute either the fertilization impact or the adaptation of plants to rising atmospheric CO 2 concentration. Besides the direct effects of CO 2 on plant processes, indirect effects of CO 2 could also affect simulation results. Govindasamy et al. who did not submit a paper to this special issue but participated at the AGU poster session in 2005 looked at how changes in surface properties due to increased CO 2 would feed back to the atmosphere and affect climate. Their results showed that the direct physical effect of CO 2 fertilization could be a warming over a timescale of a few centuries, mostly due to decreased albedo in the Northern Hemisphere boreal forest regions. This albedo-based warming could partially offset the century-scale cooling effect of additional CO 2 uptake due to CO 2 -fertilization. These results were published elsewhere and are currently the subject of a controversy about the use of planting forests for carbon sequestration at mid- and high-latitudes (temperate and boreal regions), weighing the direct carbon storage benefit against the atmospheric feedback causing additional warming. Bala et al. (2007) showed in an elegant simulation exercise that deforestation (at mid and high latitudes) would likely have a net cooling influence on the earth's climate, mainly because of increased average winter albedo. Hence, an expansion of forested areas in regions subject to seasonal snowcover, could, contrary to popular perception, enhance the greenhouse effect. It should be noted that Bonan et al. (2007) first postulated the enhanced albedo effect of removing boreal forest in a model simulation, and that a modeling experiment by Betts (2000, noted by Bala et al., 2007) first drew attention to the possible negative impact of afforestation. A concern at that time related to the assumption that afforestation would use coniferous evergreen species: Bala et al. did not report on the possible different effects of deciduous and coniferous cover—suggesting an important future experiment for DGVM modelers. Several authors presented results using data assimilation to improve model fit to observations. Baruah et al. first performed a sensitivity analysis to identify a few parameters in the ecosystem model SimCyCLE that could be advantageously replaced by remotely- sensed information. They then compared the simulated NPP with observations. Chen et al. explored the ability of a Kalman Filter to generate distributions and seasonality of model parameter values using observations at multiple forest sites: Howland (Maine, USA), Niwot Ridge Forest (Colorado, USA) and two BOREAS sites (in Saskatchewan and Manitoba, Canada). Their results showed that, Global and Planetary Change 64 (2008) 1–2 0921-8181/$ - see front matter © 2008 Published by Elsevier B.V. doi:10.1016/j.gloplacha.2008.01.005 Contents lists available at ScienceDirect Global and Planetary Change journal homepage: www.elsevier.com/locate/gloplacha