Resumen

MUNOZ-HERRERA, Wilmer; FERNANDO-BEDOYA, Oscar  y  EDILBERTO-RINCON, Mauricio. Application of neural networks for the reconstruction of time series of precipitation and temperature using satellite information. Rev.EIA.Esc.Ing.Antioq [online]. 2020, vol.17, n.34, pp.73-88.  Epub 28-Ago-2021. ISSN 1794-1237.  https://doi.org/10.24050/reia.v17i34.1292.

Artificial intelligence techniques such as artificial neural networks (ANN) allow solving a wide variety of problems related to different areas of knowledge such as medicine, Bioinformatics and even telecommunications. In many cases, neural networks are used to predict the behavior of a variable based on previous historical data and a set of predictor variables. This article deals with the particular problem of the reconstruction of missing information from meteorological stations using ANNs. The lack of this type of information mainly affects climate studies in which meteorological information is used. These studies can make it possible to avoid significant threats to the sustainable development of our society, natural resources, species and the very life of the human being. This article proposes models based on artificial neural networks and satellite information for the filling of missing data in meteorological stations and spatial reconstruction of the variables of precipitation and temperature for the Valle del Cauca, Colombia. The results obtained reach the correlation coefficients of around 0.9, with more pronounced errors in about 50 mm/month in precipitation and 2 °C in temperature.

Palabras clave : Neural Networks; Data Filling; Remote Sensors; Time Series.

        · resumen en Español     · texto en Español     · Español ( pdf )