Future global socio-economic development pathways and their implications for the environment are highly uncertain, as are the technology mixes associated with different global environmental targets.

To develop a range of possible future outcomes, we develop probability distribution estimates for key input parameters of a model of global human activity. Latin Hypercube Sampling is applied to draw 400 samples from the probability distributions for each uncertain input variable, including costs of advanced energy technologies, energy efficiency trends, fossil fuel resource availability, elasticities of substitution, population, and labor and capital productivity. The sampled values are simulated through a multi-sector, multi-region, recursively dynamic model of the world economy.

The results are 400-member ensemble simulations describing future energy and technology mixes as well as GDP and emissions. We find that many patterns of energy and technology development are consistent with various long-term environmental pathways, and that sectoral output for most sectors is little affected through 2050 by the long-term temperature target, but with tight constraints on emissions, emission intensities must fall much more rapidly.

We also combine uncertainty quantification and scenario discovery to investigate scenarios with similar values for one outcome and the range of other outcomes in those scenarios. This analysis illustrates how many combinations of outcomes can be consistent with an outcome of interest. For example, many different technology outcomes can be consistent with high or low economic growth.