Examining Variability in North American Temperature and Precipitation Trends Using a Storyline-Based Approach

While mean and extreme temperature and precipitation are expected to change as a result of climate change during the twenty-first century, regional changes in these fields remain difficult to project. In some regions, climate models may not even agree on the sign of change, let alone the regional geographic patterns of change. For a given future emissions scenario, the uncertainty in future projections can be attributed to either model-to-model differences (inter-model variance) or internal variability. As a result of internal variability, models cannot be constrained to a single outcome, but can be used to quantify a range of plausible outcomes and identify the dynamical drivers that would be likely to drive this range of outcomes.

In this study, we examine the range of possible trends in mean and extreme temperature and precipitation fields across CMIP6 models under the SSP3-7.0 emissions scenario across North America. We find that trends in extreme temperature and precipitation are more affected by internal variability than trends in mean temperature and precipitation. We then use a storylines approach - a type of regression-based analysis - to identify the main dynamical drivers that explain the variance in future trends across model simulations and use these dynamical drivers to better contextualize the range of plausible outcomes. The dynamical drivers we consider include trends in global mean surface temperature, the El-Nino Southern Oscillation (ENSO), the Pacific-North America pattern (PNA), the Eastern Pacific dipole (EP), and the North Atlantic Oscillation (NAO). We find that combinations of these drivers can reinforce each other to create high impact storylines. For example, model simulations with large positive global mean surface temperature trends and ENSO trends create a large negative trend in maximum consecutive dry days in the American Southwest.