论文部分内容阅读
Wyss 和 Allmann(1996)(以下简称 WA)的所有结论都是错误的,因为他们的方法不正确。例如,WA 的主要结论说:“估算出所得的预报和地震相关性(23次中有11次)出于偶然的概率,……为96%。”可是当我们严格依照 WA 程序,假设(23次尝试中的)所有23次预报正确,却发现了一个自相矛盾的结果,即概率 P 值大于1。从此例来看,用不着对 WA 的看法作进一步讨论,但我们还是逐点详细答复,以便说明 WA 还犯了几个错误,以及对 VAN 作陈述的真实内容的严重曲解。WA 所作的多种曲解和错误特征性的例子包括:(1)将预报的震级数值直接与M_s(PDE)对比,而 VAN 过去一直清楚地声明预报中提及的震级数值对应于 M_s(ATH),即对应于 M_L+0.5。因为 M_L+0.5大大不同于 M _s(PDE)(平均相差1.0级),他们的这种对比是不允许的;(2)通过增加(或删减)VAN 的关键性措词,从而歪曲 VAN 所表达的真实含义;(3)使用22天预报时间窗,然而,大多数这类预报和单一地震电信号相应(因此对11天预报时间窗);(4)不正确地宣称 Varotsos等(1996a,b)定出ΔM≤1.0为可接受的不确定性,而 VAN 却曾多次发表只有ΔM≤0.7时预报才能认为是成功的;(5)错误地宣称采用 SI-NOA 时,“22次 VAN 的预报中有12次不符合误差范围”,而读者容易核对出在23次中仅有6(或7)次偏离误差范围。此外,WA 大大地高估了应当预报的地震次数。也就是说,VAN 清楚地声明只有预料的震级大于(或等于)5.0级时才发布预报,而 WA 错误地要求 VAN应预报所有 M_s≥4.3或 M_s≥4.0的地震。于是他们将任何没有发布预报的 M_s≥4.3(或4.0级)的地震作为漏报的地震。最后但并非最不重要的,我们记起 Wyss 和 Baer(1981)发表过在希腊会发生预期地震震级达7.75级的长期预报(与这次争辩中讨论的时期相同),其结果完全没有成功。
All the conclusions of Wyss and Allmann (1996) (hereinafter referred to as WA) are wrong because their methods are not correct. For example, the main conclusion of WA states: “The resulting forecasts and seismic correlations are estimated (11 out of 11) by accidental probability, ... 96%.” But when we strictly follow the WA procedure, we assume that (23 All 23 predictions were correct, but found a paradoxical result that the probability P value was greater than one. From this point of view, there is no need for further discussion of the WA’s view, but we provide a detailed point-by-point reply in order to show that WA has made several mistakes and a serious misinterpretation of the true content of VAN’s statement. Examples of the many misinterpretation and error characteristics made by WA include: (1) the magnitude of the predicted magnitude is directly compared to M_s (PDE), which VCP used to state clearly that the magnitudes mentioned in the prediction correspond to M_s (ATH) , Which corresponds to M_L + 0.5. Because M_L + 0.5 is significantly different from M_s (PDE) (averaging 1.0 on average), their comparison is not permissible; (2) distort VAN by adding (or subtracting) the key wording of VAN (3) the use of a 22-day forecast time window, however, most of these forecasts correspond to a single seismic electrical signal (hence to the 11 day forecast time window); (4) incorrectly claim that Varotsos et al. (1996a, b) Set ΔM ≤ 1.0 to be an acceptable uncertainty, whereas VAN has published several times that ΔM ≤ 0.7. Forecasts can be considered as successful. (5) When falsely claiming SI-NOA, “22 VANs Of the 12 forecasts do not meet the error range, ”while readers easily check out only 23 out of 6 (or 7) deviations from the error range. In addition, WA greatly overestimated the number of earthquakes that should be predicted. That is, VAN clearly states that forecasts are only released if the expected magnitude is greater than (or equal to) 5.0, whereas WA incorrectly requires that VAN should predict all earthquakes with M_s ≥ 4.3 or M_s ≥ 4.0. They then used any earthquake with an M_s ≥ 4.3 (or Grade 4.0) that did not release forecasts as an underreported earthquake. And last but not least, we recall that Wyss and Baer (1981) published a long-term forecast of magnitude 7.75 earthquakes in Greece (with the same period discussed in this debate) as expected, with no success at all.