Resumo

PAREDES, Franklin  e  GUEVARA, Edilberto. Development and evaluation of a model for predicting meteorological drought in The Plains of Venezuela. Bioagro [online]. 2010, vol.22, n.1, pp.03-10. ISSN 1316-3361.

The Plains of Venezuela are the main units of the country's cereal production, but the drought occurrences cause severe problems in the water supply systems and generate significant economic losses in the cereal market. This research develops and evaluates a model for early prediction of meteorological droughts in Los Llanos region based on oceanic and atmospheric macroclimatic variables (MV), as predicting variables. The precipitation records of 50 rain gauges were used to identify 13 homogeneous regions according to the severity of the annual dry period. The model estimates the SPI3 drought index in each region, and categorizes it in one of three groups: Normal, Dry or Very Dry. Each MV was evaluated, and the one that produced the best differentiation as predictor was selected. The quality of prediction of the MV is evaluated through the probability of detection, false alarm probability, precision, odds ratio, and bias. Five MV were the most suitable for predicting the SPI3. It is recommended to extend this research to other regions of the country.

Palavras-chave : Climatic regionalization; prediction models; seasonal prediction.