Innovations and limits in methods of forecasting global environmental change

Environmental science has developed a diverse set of theories, analytical tools and models to understand and predict ecological responses to human impacts. We review recent innovations in the family of methods used to forecast global environmental change, and offer constructive critiques of five common approaches: phenomenological projections, storyline scenarios, integrated assessment models, decomposition-identity approaches, and global climate simulations. Overall, there is a lack of coherent, empirically based validation for many methods and their assumptions, and only partial incorporation of underlying uncertainties in both parameter estimates and interrelationships of model components. The greatest improvements in global environmental forecasting will likely come from a more systemic approach to quantifying the aggregate socio-economic drivers of the agents of change, along with better integration of multi-disciplinary approaches.