Analysis of the predictability of local rain records from different climates

Despite the complexity of the processes relevant for atmospheric convection and precipitation, rainfall has been hypothesized to be a real-world instance of self-organized criticality. Statistical characteristics that support this idea include the universal power-law in the probability density of rain event sizes, and the critical relation found between water vapor and precipitation (tuning and order parameters respectively). The possible consequences of this paradigm for the prediction of atmospheric phenomena still remain unclear. Using 1 minute resolution local rain intensities across different climates, we investigate predictability in both the fast-time dynamics of the activity inside a rain event, and in the series of rain event sizes in the slow time scale. In the first case, the balance between growth and attenuation of activity is analyzed, while in the second the predictability of extreme events is studied by means of a decision variable sensitive to the repulsion between them. A similar behavior was observed in simple SOC models due to finite-size effects. References [1] O. Peters and D. Neelin, Nature Phys. 2, 393-396 (2006) [2] O. Peters, A. Deluca, A. Corral, J. D. Neelin and C. E. Holloway, J. Stat. Mech. P11030 (2010) [3] A. Garber, S. Halleberg and H. Kantz, Phys. Rev. E 80, 026124 (2009) [4] E. Martin, A. Shreim and M. Paczuski, Phys. Rev. E 81, 016109 (2010)