La persistencia como referencia en la estimación de la habilidad de las predicciones del Tiempo a corto plazo

A persistence forecasting method is presented for the determination of the ability of predictions issued by the Weather Forecast Center of the Institute of Meteorology of Cuba. All this in order to introduce a new approach in the verification of the forecasts before the current simplicity of the prediction’s evaluation. For the construction of the persistent, we proceeded from the procedure of Moya et al . (2013) and was made up of three algorithms for the prediction of: extreme temperatures and wind force in which maximum overlapping intervals were considered, wind direction in hierarchically organized selection criteria were used, as well as cloudiness and precipitation. Also, an adequacy of the method is exposed to the possibility of verifying the forecasts of cloud cover and of the area covered by rain through the satellite images. The validation of the proposal was made with technical predictions valid for twenty four hours in the rainy season of 2016 and observations of surface stations of the Institute of Meteorology of Cuba, yielding four new distributions of maximum overlapping intervals. In this period, the average effectiveness of short range predictions overcame that of persistent prediction. Both types of forecasts show a lower effectiveness in the months of May and October, related to the beginning and end of the rainy season in Cuba, and reaffirm the cloudiness and precipitation as the variables with the lowest historical compliance rates. A measure of the quality of forecast center predictions are given by the positive values of the ability index obtained for all meteorological variables involved in the assessment of predictions, except wind strength reaching negative values in favor of persistent prognosis. The results presented here allow to take into account in the final result of the evaluation the skill of the forecasts, mainly in those days where the variations of the meteorological conditions are remarkable.