Geostatistics provides tools to model variables located in space (usually three dimensional) and takes advantage of their spatial structure (continuity) to improve the prediction at locations that have not been sampled and to characterize their spatial texture as a way to assess the uncertainty linked to the limited knowledge provided by the samples.
It is founded in statistical theory and shares many concepts and methods with statistical inference, pattern recognition and other related disciplines. In this set of notes, we review the main concepts and try to provide both intuitive explanations to the different concepts and detailed implementation parameters and examples, to understand the mechanics to operate these techniques.
We will cover concepts related to probabilistic theory, statistical inference, spatial analysis, estimation and simulation. Further to these, we will explain some of the issues linked to constraining the models with geological knowledge, extending these theories to the case of multiple variables, expanding the notions of spatial continuity to pattern statistics, link with classical statistical methods and with machine learning and deep learning techniques.
Author: Julián M. Ortiz
Download the book: https://julianmortiz.com/