[ai-geostats] questions of variogram

[Chunhua Zhang]

Hello list,

I calculate variogram with my field data (21 
quadrats in each plot). And I find that experimental variogram fluctuates 

around a constant value and has very high nugget. 
It may be due to the reason that the sample 
interval is too big. 
Are there any other reasons? Actually I have 
sampling data for 30 plots.Is it 
reasonable to put data for all plots
to calculate variogram? I 
think absolutely the variogram will be less 
fluctuate. Are the results meaningful?
And during sampling design,normally sampling points dirtributed 
randomly orregularly distributed in the study area?
How aboutthe sample size?

Thanks a lot!

Chunhua

Re: [ai-geostats] problem regarding variogram

[Isobel Clark]

snehamoy 

when you say positively skewed do you mean skewed to
the left or to the right. If you calculate the
skewness coefficient, postive is skewed to the left
with a long tail to the right. If your data is skewed
to the right, it is negatively skewed.

For negative skewness, we have had success with
subtracting limestone values from the maximum
chemically possible Ca content. Then using a lognormal
approach. That is take 58(?)-Ca and then take logs. If
you have a variance of 140, you didn't take logarithms
yet.

Your variance should be the same whether you do the
linear transform or not, only the logarithmic or other
non-linear transformation would change the variance.


Can you clarify?
Isobel
http://geoecosse.bizland.com/course_brochure.htm






Yahoo! Messenger - Communicate instantly...Ping 
your friends today! Download Messenger Now 
http://uk.messenger.yahoo.com/download/index.html