Exploiting the signal-to-noise ratio in multi-system predictions of boreal summer precipitation and temperature

Abstract

Droughts and heatwaves are among the most impactful climate extremes. Their co-occurrence can have adverse consequences on natural and human systems. Early information on their possible occurrence on seasonal timescales is beneficial for many stakeholders. Seasonal climate forecasts have become openly available to the community, but a wider use is currently hindered by limited skill in certain regions and seasons. Here we show that a simple forecast metric from a multi-system ensemble, the signal-to-noise ratio, can help overcome some limitations. Forecasts of mean daily near-surface air temperature and precipitation in boreal summers with a high signal-to-noise ratio tend to coincide with observed larger deviations from the mean than summers with a low signal-to-noise ratio. The signal-to-noise ratio of the ensemble predictions may serve as a complementary measure of forecast reliability that could benefit users of climate predictions.