Quick-Start Tutorial > Example 2: Estimate Un-sampled Points with Parameters > Understanding Pointset Trend through Experimental Semivariogram

Understanding Pointset Trend through Experimental Semivariogram

In the next step, we can observe the data trend in the semivariogram established by the system defaults. Through the pointset display in the semivariogram, you can examine the spatial correlations between sample points.

Geostatistics hypothesizes that the closer the 2 points are, the higher the correlations is. Therefore, from the semivariogram, you can find that in the “close distance” section (the very end of the right side of X-axis), the semivariance values are smaller (the lower part of the Y-axis). When the distance between a pair of points increases (moving to the right side of X-axis), the semivariance values increase accordingly. However, from the semivariogram, you can find that the fitting model represented by the yellow curve smoothes down becomes parallel to the X-axis after a certain distance. It means that after this distance, all the pairs of points have no more correlations. Note that in the experimental semivariogram, there is a small section that some points in a short distance (small X values) have very high semivariances. We may guess that these points may have some specific regional factors, such as cities having more vehicles and have more serious air pollution, or may be influenced by the whole-region directions.