用改良的面积重叠法自动跟踪中尺度对流系统的试验研究

中尺度对流系统(Mesoscale Convective Systems, MCSs)的活动和演变,是造成致洪暴雨灾害性天气的主要因素之一.多年来,许多学者加大了对中尺度对流系统研究的力度.但是,中尺度对流系统的活动及其移动和传播的研究进展一直受到计算机自动识别和跟踪问题的制约.首先介绍了一种基于面积重叠法的MCSs自动跟踪技术,然后针对它存在的问题进行了改良,并应用改良后的面积重叠法对1998年6月至8月的青藏高原上空MCSs进行了识别与跟踪实验.实验结果表明,改良后的面积重叠法的跟踪准确率可达80%,比改良前提高了10%以上.

A Study of Automatic Tacking Mesoscale Convective Systems with Improved Area-Overlapped Method

The activities and evolvements of mesoscale convective systems (MCSs) is one of the mostly factors which heavy rainstorm arise from Many scholars have increased the interest in the study on MCSs from these years However, the progression on studying activities and evolvements of MCSs is restricted by the technology of automatic identifying and tracking MCSs The thesis first introduces area overlapped method of automatic tracking MCSs Then aiming at its shortage, the area overlapped method is improved Using improved based method, the experiments are performed with data of MCSs over the Tibetan Plateau and its surrounding area in summer of 1998 The experiments show that the precision of tracking MCSs is about 80% with improved based area overlapped method, it is increased more ten percents than before improvement