The Strong Precipitation of the Dry Warm Front Cyclone in Syria and Its Prediction by Data Mining Modeling

Autor: Jianhong Wang, Nour Alakol, Xing Wang, Dongpo He, Kanike Raghavendra Kumar, Chunsheng Miao
Jazyk: angličtina
Rok vydání: 2021
Předmět:
Zdroj: Atmosphere, Vol 12, Iss 12, p 1667 (2021)
Druh dokumentu: article
ISSN: 2073-4433
DOI: 10.3390/atmos12121667
Popis: The Eastern inland of Syria has a Mediterranean climate in the north and a tropical desert climate in the south, which results in a dry south and wet north climate feature, especially in winter. The circulation dynamics analysis of 16 winter strong precipitation events shows that the key system is the dry and warm front cyclone. In most cases (81–100% of the 16 cases), the moisture content in the northern part of the cyclone is higher than that in the southern part (influenced by the Mediterranean climate zone). The humidity in the middle layer is higher than that near the surface (uplifting of the dry warm front), and the thickness of the wet layer and the vertical ascending layer obviously expands upward (as shown by the satellite cloud top reflection). These characteristics lead to the moisture thermodynamic instability in the eastern part of the cyclone (dry and warm air at low level and wet and cold air at upper level). The cyclone flow transports momentum to the local humid layer of the Mediterranean climate belt and then causes unstable conditions and strong rainfall. Considering the limitations of the Syrian ground station network, the NCEP/CFSR global reanalysis data and MODIS aqua-3 cloud parameter data are used to build a multi-source factor index of winter precipitation from 2002 to 2016. A decision tree prediction model is then established and the factors index is constructed into tree shapes by the nodes and branches through calculating rules of information entropy. The suitable tree shape models are adjusted and selected by an automated training and testing process. The forecast model can classify rainfall with a forecast accuracy of more than 90% for strong rainfall over 30 mm.
Databáze: Directory of Open Access Journals