Comparación de métodos de interpolación para datos de temperatura provenientes de las estaciones meteorológicas de la provincia de Chimborazo

Abstract

The Alternative Energies and Environment Research Group (GEAA) oversees 11 meteorological stations distributed across the province of Chimborazo. Due to the lack of maintenance, there are time series with significant percentages of missing data, preventing detailed studies of the temperature variable in the province. Therefore, the objective of this research was to compare two interpolation methods in order to select the most suitable one to fill in the time series of the temperature variable in question. The chosen methodology for this research includes a documentary design with a mixed approach (qualitative/quantitative) at a descriptive deductive level. This approach was necessary to qualitatively analyze the properties of the methods based on the descriptive deductive method. Simultaneously, a quantitative approach was used as the methods were implemented in the Python programming language to fill in the data. Through this methodology, it was determined that the cubic splines method is more accurate for data imputation due to its higher-order convergence as the spacing h is reduced. Additionally, the practical application revealed that while the methods provide satisfactory data filling, they also have limitations when confronted with a vast interval of missing data. In this context, it is concluded that the cubic splines method offers better theoretical guarantees compared to the Barnes method and achieves more precise data filling for this atmospheric variable.