Gap Filling and Pluviometrics Spacialization Data: Challenges and perspectives

Abstract Precipitation is one of the most important climatic variables for urban and rural planning, to monitor extreme events that may have an impact on society and to assist in urban drainage projects, in order to reduce the risks of floods, or even engineering works, such as dams dimensioning. However, failures in extensive series hamper these studies, and it is necessary to use models to fill them. The present study aims to review the methods of filling in gaps and spatial interpolation of precipitation data. The review of the methods was carried out from the research and reading of bibliographic materials, in order to conceptualize the approaches, identify the advantages and disadvantages of each method and present how recent studies, national and international, have innovated when comparing the performance in different areas of study. Based on this review, the main methods for filling gaps are as follows: i) weighting from Simple or Multiple Linear Regression; ii) mathematical models based on machine learning, such as Artificial Neural Networks; iii) spatial interpolators for filling gaps (Distance Inverse, Natural Neighbor, Krigagem). Finally, there was an evolution in the interpolation and fault filling techniques in the last decades, due to the evolution of computational and technological capacity.