Algorithms for the spatial interpolation of environmental data

Spatial interpolation is a technique used widely in the environmental sciences to estimate values between measurements obtained from remote sensors. Deterministic algorithms such as Inverse-distance Weighting and Radial Basis Functions and statistical methods like Kriging have been the most preferred methods for this kind of problem in the past. More recently, machine learning algorithms have begun to adapt to this problem. This works attempts to make a survey of the various commonly used and novel methods that can be used to perform spatial interpolation. We make an empirical study where various techniques are used to estimate significant wave height measurements using data obtained from the National Data Buoy Center (NDBC) of the United States’ Oceanographic and Atmospheric Administration (NOAA). We show that Machine Learning methods can be reliable and more accurate alternatives to other commonly used methods.Ciencia e Ingeniería de dato