Using nonlinear dynamics to predict Old Faithful

The current algorithm for predicting eruptions of Old Faithful is observation-based and has an accuracy of plus or minus ten minutes. In this paper, we analyze eruption data from a nonlinear dynamics perspective and develop a mathematical prediction model based on the discrete logistics map. We compare our predictions of the data with those from the observation-based algorithm. We conclude that although our algorithm is not as practical for use in the park setting, it is competitive with the current one.