Predictive skill of seasonal forecasts of climate indices
Seasonal forecasting models are increasingly being used to forecast application-relevant aspects of upcoming climatic conditions, often summarised by climate indices. Little is known, however, on how the predictive skill of such forecasts of climate indices relates to the predictive skill in forecasting seasonal mean conditions. In a recent study compiled as part of WP22, we have addressed this question.
Using a toy model to generate synthetic forecast-observation pairs, we show that the predictive skill of forecasts of climate indices closely follows the predictive skill in forecasting seasonal mean conditions based on the assumption of predictability in seasonal means only. Skill of forecasts of indices is generally slightly reduced compared to skill in seasonal means. This reduction is strongest for skilful forecasts and thresholds at the tail of the distribution of daily values. Comparing the toy model results with the verification of an operational forecasting system, we find that there is no indication of enhanced (or reduced) skill in forecasts of indices beyond what is expected due to the predictability of the seasonal mean and the systematic effects discussed above. We conclude that forecasts of climate indices can be issued without significant loss in skill as long as the index does not describe a rare or very rare event.
Bhend, J., I. Mahlstein, and M.A. Liniger (2016). Predictive Skill of Climate Indices Compared to Mean Quantities in Seasonal Forecasts. Quarterly Journal of the Royal Meteorological Society, accepted for publication. Doi: 10.1002/qj.2908