Font Size: a A A

Research On Shape Types And Growth Evolution Mechanism Of Ice Crystals In Natural Clouds Over North China

Posted on:2021-04-11Degree:MasterType:Thesis
Country:ChinaCandidate:H X XiaoFull Text:PDF
GTID:2370330647952532Subject:Science of meteorology
Abstract/Summary:PDF Full Text Request
Ice clouds are mostly composed of different ice crystal habits.Different ice crystal habits in natural clouds could significantly impact the singe-scattering properties of ice crystal particles.The single-scattering properties play an important role in the study of cloud remote sensing and the earth's atmospheric radiation budget.Besides,ice crystal habits are related to atmospheric environments such as temperature and water vapor.Therefore,it's of great importance to investigate the habits,distribution,and evolution characteristics of ice crystals in clouds.In this regard,this paper mainly studied ice crystals from the following aspects.1)In this paper,the images of ice crystals observed from airborne Cloud Particle Imager(CPI)in China were used to establish an ice crystal dataset called Ice Crystals Database in China(ICDC),which consists of ten habit categories containing 7,282 images.Besides,we proposed an automatic classification model of ice crystal habits,called TL-Res Net152,which is a deep convolutional neural network based on the newly developed method of transfer learning.The results showed that the TL-Res Net152 model could achieve reliable performance in ice crystal habits classification with an accuracy of 96%,which is far more accurate than traditional classification methods.2)Since different ice crystal habits depend on the factors such as temperature in the cloud,and we studied the distribution of various types of ice crystals at different temperatures and observation height.Accordingly,the TL-Res Net152 ice crystals classification model was used to study and analyze the three ice crystals data observed by aircraft in Hebei,China.To sum up,during three observations,the observed types of ice crystal particles including Ser,Hoc,and Plt are relatively few,while there are lots of the Sir,Cox,and Bud types of ice crystal particles having been observed.3)Additionally,we used the observational data of a stratiform cloud rainfall process observed by airborne microphysics detection instrument in Hebei,China on May 21,2018.The microphysical characteristics of the growth and evolution of ice crystals in clouds are studied in this paper.Then we analyzed the shape,distribution,and growth mechanism of ice crystals in the stratiform cloud system at different periods of precipitation.The results showthat the ice crystal particles gradually evolve into needle-shaped and columnar particles from smaller irregular shapes,and some ice crystal particles are in the aggregation process at the initially of precipitation.Meanwhile,the main growth modes of ice crystal particles may be depositional growth and the Bergeron process.In the middle of precipitation,the shapes of the ice crystal particles include columnar,needle,plate,irregular,and aggregates.Moreover,the main growth modes of ice crystal particles are aggregation growth and may be accompanied by the riming process during the mature of precipitation.
Keywords/Search Tags:Ice crystals, convolutional neural network, transfer learning, images classification, stratiform cloud, aircraft observation
PDF Full Text Request
Related items