Poisson renormalized entropy as a possible precursor to large earthquakes

The concept of Poisson renormalized entropy is presented as an observable of background seismicity with precursory possibilities. The usual way of estimating entropies evaluating probabilities directly from the normalized number of incidences is shown to be lacking in sensitivity and seriously subject to saturation. As an example, the concept of Poisson renormalized entropy and the associated method are applied to the seismicity previous to the 2011 Tohoku M W = 9.1 megaearthquake, and shown to be able to identify precursory changes in the seismicity. Results are compared with those from fractal dimension and Gutenberg-Richter b -value studies.