Dynamic Predictive Forecasting with Confidence Intervals

Develop an advanced SQL forecasting procedure that generates multi-period predictions with comprehensive confidence interval calculations. Create a solution that uses time-series decomposition, exponential smoothing, and adaptive trend detection to produce nuanced forecasts. Include capabilities for handling seasonal variations, structural changes, and providing probabilistic prediction ranges.