The standard approach to predicting climate change, assessing its impacts and planning mitigation strategies tends to be compartmentalized, leading to inadequate or incomplete advice for climate policy. Climate models used for future global warming predictions and attribution of past changes generally consider only global climate drivers, ignoring local drivers, such as land use change and urban effects. Impacts studies are generally carried out in isolation from each other and hence ignore interactions between different impacts sectors, such as hydrology, natural ecosystems and agriculture. Feedbacks from impacts to climate change, which often involve land ecosystem-atmosphere interactions, are often neglected. This can result in representations of global changes that are at best inconsistent and at worst completely misleading. Moreover, a number of climate change mitigation strategies, such as carbon sequestration and biofuels involve changes in land ecosystems. In addition to affecting climate through the exchange of carbon with the atmosphere, these land ecosystem changes may affect climate change or its impacts through a variety of additional processes, such as surface albedo change or changes in the surface moisture budget. Failure to account for these may have consequences that are potentially at odds with the aims of climate change mitigation.