Identificação e correção da banda brilhante em dados de radar meteorológico

View/ Open
Date
2014Author
Oliveira, Camila de
Metadata
Show full item recordSubject
DissertaçõesTeses
Precipitaçao (Meteorologia)
Radar meteorológico
Neve
Análise numérica
xmlui.dri2xhtml.METS-1.0.item-type
DissertaçãoAbstract
Resumo: A banda brilhante (BB) é uma camada de alta refletividade que ocorre durante o derretimento de neve agregada e cristais de gelo. Essa camada de alta refletividade pode causar erros no cálculo da taxa de precipitação. Por isso, é importante identificar corretamente essa camada nos dados de radar meteorológico para que uma correção possa ser aplicada. Este trabalho utiliza duas técnicas para a identificação desta região. A primeira utiliza os produtos VIL e Zmax para classificar os dados em precipitação convectiva e estratiforme, para então, dentro da região estratiforme, limitar a banda brilhante utilizando um cálculo de sua espessura aparente. A segunda, utiliza a metodologia k-means, que agrupa os dados em três grupos utilizando variáveis obtidas a partir dos dados de refletividade. Estes três grupos representam a precipitação convectiva, a estratiforme e a região de banda brilhante. Após a identificação da região de banda brilhante é aplicada uma técnica para a correção desses dados utilizando o perfil vertical de refletividade. Os valores de refletividade são então convertidos para taxa de precipitação utilizando uma relação Z-R adequada. Em seguida, esses valores são comparados com a precipitação medida por pluviômetros localizados no raio de alcance do radar. As duas técnicas obtiveram melhores resultados quando comparadas com os dados originais. Na comparação entre as técnicas, o k-means mostrou, em geral, os melhores resultados. Palavras-chave: Radar Meteorológico, Banda Brilhante, Perfil Vertical de Refletividade, K-means. abstract:
The bright band (BB) is a layer of enhanced reflectivity due to melting of aggregated
snow and ice crystals. This layer of high reflectivity can cause errors in the calculation
of the rain rate. Therefore, it is important to correctly identify this layer in radar observations
for a correction to be applied. This current study uses two techniques for the
identification of this region. The first uses the products VIL and composite reflectivities
to convective-stratiform precipitation segregation, and then, within the stratiform region,
limit the bright band using a first-guess apparent BB depth. The other method uses kmeans,
wich is used for clustering the data into three clusters, using variables obtained
from radar reflectivity. These three clusters represent the convective and stratiform
precipitation and bright band region. After identify the region of bright band is applied
a technique for the correction of these data using the vertical profile of reflectivity. The
reflectivity values are then converted to rain rate using an appropriate Z-R relationship.
Then these values are compared with rainfall measured by rain gauges located within
radar range. Both techniques present better results when compared with the original
data. However, when comparing the two techniques, k-means showed, in general, the
best results.
Key-words: Radar, Bright Band, Vertical Profile of Reflectivity, K-means.
Collections
- Dissertações [222]