Polar Amplification Helps Forecast Northern Temperature Anomalies

Nearly one half of the positive feedback mechanisms that are identified in the literature as potential tipping elements in the climate system and are of serious concern within the next century are located in the northern part of the Northern Hemisphere. Improving forecasts of northern temperatures is therefore critical to improving our understanding and perhaps early detection of tipping points. We propose forecasting northern temperatures using a structural geophysical model of polar amplification, which is defined as the acceleration of warming in regions closer to the poles and the North Pole in particular, that uses anthropogenic forcing and southern temperatures as covariates. We show using pseudo-out-of-sample forecasts over a range of time periods that this geophysical model improves medium-run forecasts over otherwise similar benchmark forecasting models. Using this model, we forecast temperature anomalies in the northern part of the Northern Hemisphere to increase from 1.861C (over the 1961-1990 baseline) in 2023 to 2.214C with a 95% forecast interval of (1.399,3.147) C by 2035.