Global mean surface air temperature (Tglobal) variability on subdecadal timescales can be of substantial magnitude relative to the long‐term global warming signal, and such variability has been associated with considerable environmental and societal impacts. Therefore, probabilistic foreknowledge of short‐term Tglobal evolution may be of value for anticipating and mitigating some course‐resolution climate‐related risks. Here we present a simple, empirically based methodology that utilizes only global spatial patterns of annual mean surface air temperature anomalies to predict subsequent annual Tglobal anomalies via partial least squares regression. The method's skill is primarily achieved via information on the state of long‐term global warming as well as the state and recent evolution of the El Ni?o–Southern Oscillation and the Interdecadal Pacific Oscillation. We test the out‐of‐sample skill of the methodology using cross validation and in a forecast mode where statistical predictions are made precisely as they would have been if the procedure had been operationalized starting in the year 2000. The average forecast errors for lead times of 1 to 4?years are smaller than na?ve benchmarks on average, and they perform favorably relative to most dynamical Global Climate Models retrospectively initialized to the observed state of the climate system. Thus, this method can be used as a computationally efficient benchmark for dynamical model forecast systems. Plain Language Summary Year‐to‐year global temperature variability can be large compared to the long‐term progression of global warming, and such year‐to‐year variability has been shown to have considerable environmental and societal effects. Thus, approximate foreknowledge of yearly global temperature deviations should be of value for anticipating some climate impacts. This study presents a relatively simple, empirical method for predicting year‐to‐year global temperature. We show that information on the global spatial patterns of surface air temperature alone can be used to skillfully predict global average temperature 1 to 4?years ahead of time. We find that the method performs favorably compared to predictions from much more computationally expensive Global Climate Models.
展开▼
