Solar cycle prediction: Combining precursor methods with McNish and Lincoln technique

Medium-term and long-term prediction of the magnitude of the maximum of smoothed sunspot numbers, and thus of the solar cycle time profile, is a basic input for many space environment predictions. The widely used statistical technique of McNish and Lincoln is systematically compared to predictions based on precursors, either related to the cycle time profile characteristics or to geomagnetic indices. It is shown that when cycles 13 - 23 are considered, all prediction methods give, at least for one of the cycles, an error much larger than 20%, an inadequate result. None of the methods is fully reliable. Thus, it is proposed to combine the predictions based on precursors and to improve McNish and Lincoln results with them in order to limit such rare but large errors and to improve significantly the reliability of the predictions performed in the course of the solar cycle ascending phase.