Earlier warning on global warming: a new method for timely tracking and forecasting of global surface temperature change and accelerations

Recent research has raised concerns that the Earth’s global surface temperature (GST) change relative to preindustrial levels (mean temperature 1850-1900), a key indicator closely connected to the planetary boundary for climate change, is in a human-caused phase of acceleration rather than just following a linear trajectory. However, at this point, reliably tracking this acceleration signal in a timely manner and clearly distinguishing it from natural variability remains difficult.

To address this, we propose a new method to regularly forecast the GST change, including both a prediction of the annual-mean of the current year and a projection of the 20-year mean up to 10 years ahead. The forecasts comprise both the global surface air temperature (GSAT) change as primary metric, also used by the IPCC for assessing the degree of compliance with Paris Agreement temperature limits, and for legacy purposes as well the global mean surface temperature (GMST) change, which (mainly) is a blend of surface water temperature over the oceans and surface air temperature over land.

We introduce and demonstrate the method over the 1990 to 2025 timespan. It provides annual-mean results for any current year, and the related 20-year-mean estimates, as early as of July of the year, followed by monthly updates until the current year is observationally complete (within the first quarter of the follow-on year). By combining monthly observational GMST and GSAT data from reliable sources, including reanalysis and seasonal prediction data, a typical GST forecast accuracy within 0.03 °C is achieved as of August of the current year for the annual mean, and a typical 10-year projection accuracy of within 0.05 °C for the 20-year-mean. The latter is a critical metric for early clues on emerging next-decade changes in the Earth system.

We show that the approach enhances the accuracy and timeliness of early-warning estimates of the ongoing GST change, including of GST change acceleration of current-year versus center-year-1990 20-year-mean trend rates and of the related level of exceedance over natural trend-rate variability. As an example, our prediction of September 2025 for the annual-mean GSAT change in 2025 was 1.48 °C, four months ahead of the January 2026 announcement of the EU Copernicus Climate Change Service of a GSAT change of 1.47 °C. By improving in this way our ability to detect and characterize GST change dynamics in a timely and reliable manner, this work provides valuable insights into the warming state of the climate system and its proximity to critical thresholds such as tipping points, along with co-informing on the Earth energy imbalance and potential destabilization tendencies in climate feedback processes. The findings also help inform discussions on the urgency of climate mitigation efforts to avoid exceeding planetary boundaries.