The quality of forecasts can be measured with a wide variety of indices and formulae. All these approaches rely basically on the relation between the numbers of correct forecasts, wrong forecasts, false alarms and rejected cases. In the case of extreme events damage is the major topic. All extreme events by definition are more or less rare events. In many applications the events frequency of an extreme event is selected to be one event per 100 hundred years. Depending on the application other such event frequencies are in use. The mitigation of damage mainly uses rules for the design structures such as buildings. In principle their proper application would allow damage to occur only if a meteorological event oversteps a certain predefined threshold value. In practice the threshold proves to represent more something like
a soft shoulder and damage is already observed to be caused by events somewhat smaller than the damage threshold value for the extreme weather case. No matter what its exact definition each threshold value is connected to an event frequency. This event frequency is hard to obtain in particular in the vicinity of the threshold of the extreme event case, because it has to be derived from data scarce by definition, however long the observation time series are. Therefore, these threshold values are subject to a certain inaccuracy. In addition, the low frequencies show some variability with time. Recently, climate changes support the idea that also the occurrence frequency of extreme values will change, increase, in the future. Calculating the forecast quality using the basic data leads to two formulations of the forecast quality, both based on the same principles. The fraction formulation correctly is free from any absolute damage height, it is sufficient to find one reference value. When going to the cumulative formulation the role of the effect of the frequency of occurrence can clearified. The two equations allow to compare the effects of long term changes and inaccuracies of the frequency of occurrence of extreme events with the effects of the improvements of the weather prediction.
| Identifer | oai:union.ndltd.org:DRESDEN/oai:qucosa:de:qucosa:15925 |
| Date | 27 July 2017 |
| Creators | Tetzlaff, G. |
| Contributors | Universität Leipzig, Universität Leipzig |
| Source Sets | Hochschulschriftenserver (HSSS) der SLUB Dresden |
| Language | English |
| Detected Language | English |
| Type | info:eu-repo/semantics/publishedVersion, doc-type:article, info:eu-repo/semantics/article, doc-type:Text |
| Rights | info:eu-repo/semantics/openAccess |
| Relation | urn:nbn:de:bsz:15-qucosa-211460, qucosa:14988 |
Page generated in 0.0015 seconds