Simple Kriging Demo
Simple Kriging Demo
Simple Kriging Demo
Michael Pyrcz, Geostatistics at Petroleum and Geosystems Engineering, University of Texas at Austin (mpy
60
50 Data
Unknown
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100
X Coordinate (m)
Description
This sheet provides an illustration of Simple Kriging at a single estimated location.
Step 1: Input the data locations and values, the unknown simulated location. At any point thes
may be changed to observed their influence on the simulation.
Step 2: The distance matrix is automatically calculated, that is the distance between the data a
Step 3: Enter the model of spatial continuity in the form of an isotropic spherical variogram and
should sum to one). This model may be changed at any time to observed sensitivities to spatia
Step 4: Variogram matrix is calculated by applying the distance matrix to the isotropic variogram
Step 5: Covaraince matrix is calculated by subtracing the variogram from the variance (1 for sta
This is applied to improve numerical stability as a diagonally dominant matrix is more readily in
Step 6: The left hand side of the covariance matrix is inverted.
Step 7: The inverted left handside matrix is multiplied by the right hand side matrix to calculate
Step 8: The kriging estimate and kriging variance are calculated with the weights and covarianc
Step 9: With the Gaussian assumption the complete local conditional cumulative distribution fu
niversity of Texas at Austin ([email protected])
4. Variogram Matrix
0.000 0.229 0.357 0.158
0.229 0.000 0.334 0.125
7. Weights 0.357 0.334 0.000 0.248
-0.621 0.341
-0.813 0.462 8. Kriging Results
1.941 0.225 Kriging Estimate 0.131
Sum Weights 1.029 Kriging Variance 0.139
Mean Weight -0.029
p-value f-1(p)
5% -0.48
10% -0.35 9. Local CCDF
15% -0.26 100%
Legend 20% -0.18
90%
Information 25% -0.12
User Input 30% -0.06 80%
Value