GEOSTATS: Logtransform/Logratio transform
- I have two problems which I hope someone may help me with. As a
background, one of the aims of my PhD is to estimate the distribution of
physical and chemical soil attributes across a valley (approx 10K km2).
1. Transforming negatively skewed compositional data
I am analysing soil at six depths for a variety of physical and chemical
attributes. I have three separate compositional data sets. All values are
positive. Within some depths of each set are outliers which skew the data
to the left and some outliers which skew the data to the right. Sampling
and laboratory results have been checked and the values appear to be
correct. What is the most consistent way to treat each set? I understood
that logtransform and logratio transform was not appropriate for negatively
skewed data but am unsure why and what the ramifications are? Is it in the
2. Bimodal distribution
Within one of the compositional data sets (CEC) I have two attributes (Mg%
& Ca%) which at one depth (20-30cm)are skewed bimodally to the left.
Plotting the values against the co-ordinates the peaks fall roughly
east/west. The remaining two attributes are positively skewed. My
question is similar - how to treat the set consistently before kriging?
Thanks in advance
Dept Ag Chem & Soil Science
University of Sydney
*To post a message to the list, send it to ai-geostats@....
*As a general service to list users, please remember to post a summary
of any useful responses to your questions.
*To unsubscribe, send email to majordomo@... with no subject and
"unsubscribe ai-geostats" in the message body.
DO NOT SEND Subscribe/Unsubscribe requests to the list!