The Comparative Study Of Different Similarity Criteria In The Similar Forecast Of Surrounding Bohai Area

Similarity prediction has been widely valued by weather reseachers because of its special effect of prediction.It can consider three-dimensional structural characteristics of the weather development process,and take into account the linear and nonlinear variations of atmosphere itself.In addition,a wide range of numerical weather prediction results can be absorbed and the experience of the forecaster is also useful.Similarity prediction method is traditional,but it has vitality in this information era.While in order to measure the degree of similarity between two samples,the quantitative calculation of the degree of similarity of the weather situation field and meteorological field of the two samples is essential.The method(or criterion)to measure the degree of similarity of two samples is called similarity criterion.There are many kinds of similarity criteria,each one of which has its own focus.In this paper,the key research objective is to find a way to select a similarity forecast criterion for the Bohai Sea region.This paper first analyzes the theoretical framework of the calculation methods commonly used in similarity criteria,and then the situation field similarity filter results of different criteria used in Bohai region are compared.Secondly,the stability of similarity criteria and their calculation methods on the filter effect of situation field in Bohai region is proved through the temperature forecast in 2010.Likewise,their effects on the filtering effect of visibility forecast(one of the forecast difficulties)over Bohai area is also proved stable by using the similarity forecast in 2009 and 2010.Through comparison,the results show that tradition way of similarity filter effect is the best only on the 500hPa layer height field in the situation field similarity filter of Bohai region.The combination of three layers(500hPa,700hPa,850hPa)is more effective than single layer in computing the filtering effect of the situation field.At the same time,for those three layers,the simultaneous filtering effect using calculation plan 4 is better than separate effect by one layer at a time,so a more similar sample of situation field can be found.Moreover,as for the similarity filtering effect of visibility forecast in Bohai region,it is found that the effect of nonlinear similarity deviation criterion using the traditional computing solutions is the best compared with other computing methods.In a word,this paper finds out the best similarity criteria for the similarity filtering effect of situation field and element field in Bohai region by comparing the filtering effects of different similarity criteria,which further enriches the thought,examines the methods and develops the content of similarity prediction.At the same time,it is hoped that our explorations can shed an important light on the practical,local,and reasonable selection of similarity criteria in similarity forecast for other researchers and weather forecast institutions.