Why "valid semivariogram model couldn't be calculated"?

[YukoStender]

Hi,

I have been examining my different semivariogram models to get parameter estimates for my data set using the geostat wizard, ArcGIS 9.3.

A "valid semivariogram model couldn't be calculated" per geostat wizard, when I choose trend as 'constant' using universal kriging. 
However it does give me semivariogram and estimated parameters if I specify '1st order'. 
Similarly, it could not be calculated using ordinary kriging unless I specify to remove 1st order trend.

Under what situation, won't the wizard calculate semivariance and estimate parameters when no or constant trend is specified in the semivariogram modeling?
Does it prevent estimation if there's a trend and 'constant' trend is not adequate for data?

Thanks in advance!

[EricKrause]

Can you send your data to ekrause@esri.com?  I'll take a look and try to figure out what is causing the problem.  Thanks.

[YukoStender]

Hi Eric,

Thanks for your reply.

I found there was no more trouble calculating variogram and parameters after I reloaded my data and made a new point shapefile.  I was able to have the semivariogram modeling section running again.

It was strange, because it the exact same data file (.xls) was imported. 
However somehow the attribute table showed a bunch of rows with zeros in the troubled one while the attribute table of the latter only showed what I actually have in my .xls file. 
Semivariogram modeling run with various model setting with the new shape file.

I still don't understand why the troubled shapefile had rows of zeros and the new one did not, but the immediate operational problem seems to be solved.

Thanks again for your reply.
Yuko

[EricKrause]

Good to hear.  Without knowing the exact problem, I can't say for sure, but this has likely been fixed since 9.3.