hi,
I'm hoping someone might be able to tell me whether I'm on the right track or be able to help. Eric from ESRI has been helping me with many of my questions and has been fantastic but I thought I would add this to the thread so not to bother him with my endless emails.
I am trying to replicate work I have read in Journal papers to determine the optimum density of sites for a network. The method requires altering the number of sample points and examining the change in the variance of the estimation error. Because the location of the point and the variogram model only effect the variance (i.e and not the actual measured value at the point) I should be able to examine the effectiveness of adding or removing fictious sample locations on the variance.
I'm not entirely sure how the steps are performed to do this but this is what I think:
Step one: determine the best variogram model and parameters from my known measured values
Step two: using the model parameters and location of the measured values calculate the average kriging standard deviation or variance for my measured values (as a global indicator of performance)
Step three: Add fictious points to the sample set and recalculate the average kriging standard deviation or variance using the sample model parameters in step one - only x,y, are needed for the fictious points to caclulate the standard deviation.
Step four: repeat this for a number of different site additions - say 10, 20, 50, 100 new sites
Step five: plot the average kriging standard deviation against the number of observation wells assessed- the graph should flattened where the addition of new wells makes little difference to the reduction in standard deviation. At this point I will say this is my optimum number of sites needed for the biggest reduction in error.
Do these steps sound right to anyone?
How do extract the model parameters after variogram analysis
How do I set the model parameters for subsequent analysis and addition of fictious points?
Where do I find the standard deviation of each point added to the network?
Hope this make some sense.
Cheers,
Simon
I'm hoping someone might be able to tell me whether I'm on the right track or be able to help. Eric from ESRI has been helping me with many of my questions and has been fantastic but I thought I would add this to the thread so not to bother him with my endless emails.
I am trying to replicate work I have read in Journal papers to determine the optimum density of sites for a network. The method requires altering the number of sample points and examining the change in the variance of the estimation error. Because the location of the point and the variogram model only effect the variance (i.e and not the actual measured value at the point) I should be able to examine the effectiveness of adding or removing fictious sample locations on the variance.
I'm not entirely sure how the steps are performed to do this but this is what I think:
Step one: determine the best variogram model and parameters from my known measured values
Step two: using the model parameters and location of the measured values calculate the average kriging standard deviation or variance for my measured values (as a global indicator of performance)
Step three: Add fictious points to the sample set and recalculate the average kriging standard deviation or variance using the sample model parameters in step one - only x,y, are needed for the fictious points to caclulate the standard deviation.
Step four: repeat this for a number of different site additions - say 10, 20, 50, 100 new sites
Step five: plot the average kriging standard deviation against the number of observation wells assessed- the graph should flattened where the addition of new wells makes little difference to the reduction in standard deviation. At this point I will say this is my optimum number of sites needed for the biggest reduction in error.
Do these steps sound right to anyone?
How do extract the model parameters after variogram analysis
How do I set the model parameters for subsequent analysis and addition of fictious points?
Where do I find the standard deviation of each point added to the network?
Hope this make some sense.
Cheers,
Simon
Use the Create Geostatistical Layer tool with the xml as the model source. Then specify the new dataset (the one that includes the old locations and the new locations with mock values), and the tool will interpolate on the new dataset with the old model parameters.
Once you have the GA layer, use the GA Layer to Points tool to predict back to the original data (without the new locations added). Don't worry about the field to validate on. Open the attribute table on the point feature class it creates, and the last column will be Standard Error. You can calculate the mean of the Standard Errors by right-clicking the column and choosing "Statistics." You can also use the Summary Statistics tool.
To compare against the old standard errors, do the same process with the GA layer created only using the old locations.