Natural catastrophes and point-like processes Data handling and prevision

摘要: A frequent approach when attempting to manage a natural catastrophe is in terms of numerical model, by which we try forecast its occurrence space and time. But, sometimes this difficult or even unrealistic. On more pragmatic grounds can appeal formal analysis the historical time series every concern. Only approximately, however, such be likened point-like process, because "detector-mankind" experienced substantial changes versus Nevertheless, algorithms approximately applied means few suitable assumptions. In ultimate analysis, four basic viewpoints considered: i) either assuming that phenomena are periodic; ii) an event occurs only whenever some energy threshold attained (calorimetric criterion); iii) it system experiences abrupt change boundary conditions; iv), no algorithm viable due scanty observational information, just applying fractal box counting method, other less related and/or equivalent algorithms. The mutual relations, advantages, drawbacks any briefly discussed, with applications. They already lead apparently successful long-range large flood Northern Italy occurred 1994, prevision next explosive eruption Vesuvius. But success application closely determined quality database, physical information fed into rather than mathematics per se have concerned avoiding arbitrary input being added, based on human need for simplicity. present paper gives synthesis several were previously independently simple intuitive basis different case studies, although comparisons discussion their similarities differences.