An intelligent system to real time rainfall prediction using radar data

• Autor(a) principal: Ulson, Jose Alfredo Covolan [UNESP]
• Data de Publicação: 2001
• Outros Autores: Antonio, MDA, Da Silva, I. N., De Souza, A. N.
• Tipo de documento: Artigo de conferência
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://dl.acm.org/citation.cfm?id=704229; http://hdl.handle.net/11449/36961
• topic: rainfall; radar; Z-R relationships; artificial neural network

This work presents a new approach for rainfall measurements making use of weather radar data for real time application to the radar systems operated by institute of Meteorological Research (IPMET) - UNESP - Bauru - SP-Brazil. Several real time adjustment techniques has been presented being most of them based on surface rain-gauge network. However, some of these methods do not regard the effect of the integration area, time integration and distance rainfall-radar. In this paper, artificial neural networks have been applied for generate a radar reflectivity-rain relationships which regard all effects described above. To evaluate prediction procedure, cross validation was performed using data from IPMET weather Doppler radar and rain-gauge network under the radar umbrella. The preliminary results were acceptable for rainfalls prediction. The small errors observed result from the spatial density and the time resolution of the rain-gauges networks used to calibrate the radar.