Dear All,
The attached file showed the prediction and variance map from ordinary kriging interpolation method of one of my yield data sets. The yield data was standardized to mean of 0 and standard deviation of 1 before fitting the variogram model and interpolation. The reason for the standardization is to be able to compare the result of interpolation from one year to the other.
Please could anyone guide me on how to interpret those two maps in the attached file? What do I need to look for? How do I go about reading and interpreting the maps? In some literature, the square root of the kriging variance is obtained to come up with standard error map. Which one is better to present, standard error or kriging variance map?
See attached file please.
Thank you.
Dear Moshood,
if you are using ArcGIS see the following slides for more explanation on how you can validate Kriging results
http://webhelp.esri.com/arcgisdesktop/9.3/tutorials/geostat/Geostat_5_1.htm
Good luck
ArcGIS along with its Geostatistical Analyst Extension directly gives KSE (kriging standard error) map.
You are right. Square root of kriging variance is termed as KSE. So, if you do use ArcGIS, you can directly have the KSE map that can be easily interpreted.
You can see my following article in my profile for KSE map and interpretation.
Optimal design of rain gauge network in the Middle Yarra River catchment, Australia
Dear Sajal,
Where I can find the article " KSE map and interpretation"?
Thank you
Well, when you take the square root of kriging variance you get errors in the same unit of the input variable, thus it makes more sense to present estimation errors to non-statisticians. But, as you standardized the input data, you would need to backtransform your errors to actually have errors in real units. Variance is more sensitive to extreme values, but, with standardized data, variance values result in higher error values: 0.36->0.6; 0.49->0.7; 0.64->0.8... .
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling