Dear all
I am using spatio-temporal ?regression kriging for estimation of my air temperaturedata. After removing the trend in data with a linear model, I calculate theresiduals. I found empirical semivariogram and fit a sum metric model to it ( Iattach the plot of them). It shows spatio-temporal dependency among theresiduals. But, when I used the variogram for spatio-temporal kriging estimationof the residuals and add them back to the linear model result, it increases theerror of the estimation. Shouldn?t this method decrease the error of estimation?Does anybody know why this is happened?What should I do to resolve this problem?
I really appreciate any helpThanks in advance
Shohreh
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://stat.ethz.ch/pipermail/r-sig-geo/attachments/20151222/40a32a5b/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Sum metric variogram.pdf
Type: application/pdf
Size: 13932 bytes
Desc: not available
URL: <https://stat.ethz.ch/pipermail/r-sig-geo/attachments/20151222/40a32a5b/attachment.pdf>