Variograms are used to demonstrate the distance-based variability of spatial data. Understanding and modeling spatial continuity with variograms is important as they are used to estimate point measurements into practical blocks across a wide range of applications such as mining ore grades, oil concentrations, or the environmental contaminants.
Despite open-source options being available to generate variograms, due to their complexity, most users rely on expensive software packages which abstract a lot of the details. This tutorial aims to give a brief introduction to variograms and how the open source Geostatistics Library (GSLib) which can be used independently or with Python to develop variograms.
Here a variogram model is developed on a synthetic mining dataset but the workflow could be used for any kind of spatial data for meteorological applications like temperature, or environmental applications like contaminant tracking.
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
The general idea of variograms are that data points further away from each other are more likely to be more distinct than data points close to one another. The variance of data points further and further apart eventually reaches a point where it is equal to the global variance of the data.
We start with a spatial dataset and can generalize the variogram modeling workflow into a few steps as shown below. First we need to determine adequate search parameters for the variogram. Then identify the major and minor continuity axis. Finally the variograms can be then modeled and subsequently used for estimation or simulation purposes. Each of the steps will be further explained in the following sections…