RE: AI-GEOSTATS: Extreme values?

[Myers, Jeff]

Chaosheng Zhang -   I think Marcel Vallee is headed in the right direction on your problem.  There is a good chance that the problem is one of sample and or subsample support.  As mentioned, if you sampled within a foot or tow of a location that displays an extreme or "outlier" value, you may find values an order of magnitude or more below the outlier.  Similarly, you may also have "inliers", where a sample nearby a location with a low concentration may contain a significantly higher value.  Of course, no one gets excited about the inliers that may be unrepresentative, but we get very excited about the outliers!   The possibility of extreme values should be planned for in the initial stage of the sampling program.  Pierre Gy's work has revealed that the physical size, volume, and orientation of a sample and subsample (i.e. the support) are crucial to the concentration estimate obtained.  You are asking a lot to have a 10-g sample represent 400 meters between sample locations in any case.  Unless the support of the original sample and all subsampling stages was sufficient, there is little chance that the samples are highly representative of the true concentration.  Mine areas typically are very heterogeneous and proper sampling support when sampling is essential.  Perhaps you can provide some details.  If the underlying data are not representative due to improper suppoort, you are trying to "contour an illusion", and typically the results are not pleasing.   The way in which the data are used in decision-making is also important.  For instance, if your purpose is to delineate hot spots for risk assessment, extreme values do not pose a problem as they will be addressed.  You may, however, be very interested in getting your best information at an economic cutoff value or risk threshold, since the decision for treatment of values high above or way below the action level is easy.   Jeff Myers Westinghouse Safety Management Solutions 2131 S. Centennial Ave., SE Aiken, SC 29803 803.502.9747 (direct) 803.502.9767 (main) 803.502.2747 (fax) [EMAIL PROTECTED] http://www.gemdqos.com   -----Original Message----- From: Chaosheng Zhang [mailto:[EMAIL PROTECTED]] Sent: Thursday, December 13, 2001 8:02 AM To: [EMAIL PROTECTED] Subject: AI-GEOSTATS: Extreme values? Dear all,   My question is: How to deal with the extreme/outlying values in a data set?   I am dealing with heavy metal concentrations in soils from a mine area. The sample number is 223, and the samples are spatially evenly distributed with the sampling interval of 400 metres. There are several samples with extremely high values, which makes me feel uncomfortable. The percentiles of the dataset are listed as follows (in mg/kg):                  Zn    Cu     Pb     Cd    As         Min     4     1     25    0.0     2          5%    35     6     35    0.1     6         10%    40     7     41    0.2     7         25%    65    13     62    0.3     9         50%   122    18    168    0.6    15         75%   338    27    821    1.5    28         90%   907    56   2799    2.8    58         95%  1986   116   4490    4.2    80         96%  2462   151   4698    4.9    82         97%  3493   178   5413    6.2    91         98%  4697   207   7609    8.3   111         99%  6712   247  11750   12.4   184         Max 11473  1293  16305   48.5  1060 When doing geostatistical and statistical analyses, we need some confidence in dealing with the these very high extreme values which account for less than 2% of the total sample number.    Any suggestions? Cheers,   Chaosheng Zhang =================================== Dr. Chaosheng Zhang Department of Geography National University of Ireland Galway IRELAND Tel: +353-91-524411 ext. 2375 Fax: +353-91-525700 Email: [EMAIL PROTECTED] ===================================