A pdf copy of a recent seminar Hybrid continuous-time Box-Jenkins model estimation and forecasting applied to global climate data is available in PUBLICATION DOWNLOADS. The presentation outlines the background theory and motivation of the continuous-time transfer function model identification and estimation routines in CAPTAIN (rivcbj and rivcbjid) and uses these, plus the Dynamic Harmonic Regression (dhr) routine, to analyze and forecast recent climate data. Amongst other things, it shows how there is an underlying quasi-periodic component in the global temperature anomaly data that accounts for the recent ‘leveling off’ in temperature but forecasts that the rise will continue after this, as shown in the plot above. Note that this is a hypothetical plot, in the sense that no model should be used to forecast so far ahead: the forecast interval is selected purely to show the nature of the quasi-cyclical componet.  Indeed, the plot shows that the inherent uncertainty in the data makes future behaviour extremely hard to predict and questions, to some extent, the confidence of climate scientists. Also, note that the analysis accepts the causation suggested by climate scientists and utilizes data produced by them, so it is not surprising that the models and forecasts support the hypothesis of future climate warming. A video of the presentation is available on request (p.young@lancaster.ac.uk).