Re: AI-GEOSTATS: moving averages and trend
Hello, Factorial kriging is not very sophisticated, it's just a slight variant of kriging that requires the modification of just a few lines of codes. Anyways, I just posted a program to perform factorial kriging analysis in the download section of my website. I hope your grid is not too big, The program filter.exe (FORTRAN source code filter.f) is a modified version of the Gslib program kt3d.f that allows performing a kriging analysis. Based on the number of nested structures of the variogram model specified in the parameter file filter.par, the program will estimate the values of the noise and noise-filtered signal (1 structure), or the values of the noise, local and regional components (2 structures). Following Goovaerts (1997, page 167), the regional component includes both the long-range component and the trend component in order to attenuate the impact of the search window on the estimation of these long-range spatial components. The zipped folder includes the executable, the source code, as well as a sample parameter file for the Jura dataset. In another paper concerned with noise-filtering of imagery, I run the program for a single pixel to get the kernel weights and then apply the same kernel everywhere (since the data geometry does not change except at the edges of the image). Goovaerts, P., Jacquez, G.M., and W.A. Marcus. 2005. Geostatistical and local cluster analysis of high resolution hyperspectral imagery for detection of anomalies. *Remote Sensing of the Environment*, 95, 351-367.http://home.comcast.net/%7Epgoovaerts/RSE-2005.pdf Cheers, Pierre On Tue, Feb 2, 2010 at 3:39 AM, seba sebastiano.trevis...@libero.it wrote: Hi Pierre I think that for my task factorial kriging is a little bit too much sophisticated (nevertheless, is there any open source or free implementation of it ??? I remember that it is implemented in Isatis.). I have an exhaustive and regularly spaced data set (i.e. a grid) and I need to calculate locally the spatial variability of the residual surface or better I would like to calculate the spatial variability of the high frequency component. Here I'm lucky because I know exactly what I want to see and what I need to filter out. In theory, using (overlapping) moving window averages (but here it seems better to use some more complex kernel) one should be able to filter out the short range variability (characterized by an eventual variogram range within the window size???). Seeing the problem from another perspective, in the case of a perfect sine wave behavior, I should be able to filter out spatial variability components with wave lengths up to the window size. But maybe there is something flawed in my reasoningso feedback is appreciated! Bye Sebastiano At 16.27 01/02/2010, you wrote: well Factorial Kriging Analysis allows you to tailor the filtering weights to the spatial patterns in your data. You can use the same filter size but different kriging weights depending on whether you want to estimate the local or regional scales of variability. Pierre 2010/2/1 seba sebastiano.trevis...@libero.it Hi José Thank you for the interesting references. I'm going to give a look! Bye Sebastiano At 15.46 01/02/2010, José M. Blanco Moreno wrote: Hello again, I am not a mathematician, so I never worried too much on the theoretical reasons. You may be able to find some discussion on this subject in Eubank, R.L. 1999. Nonparametric Regression and Spline Smoothing, 2a ed. M. Dekker, New York. You may be also interested on searching information in and related to (perhaps citing) this work: Altman, N. 1990. Kernel smoothing of data with correlated errors. Journal of the American Statistical Association, 85: 749-759. En/na seba ha escrit: Hi José Thank you for your reply. Effectively I'm trying to figure out the theoretical reasons for their use. Bye Sebas -- Pierre Goovaerts Chief Scientist at BioMedware Inc. 3526 W Liberty, Suite 100 Ann Arbor, MI 48103 Voice: (734) 913-1098 (ext. 202) Fax: (734) 913-2201 Courtesy Associate Professor, University of Florida Associate Editor, Mathematical Geosciences Geostatistician, Computer Sciences Corporation President, PGeostat LLC 710 Ridgemont Lane Ann Arbor, MI 48103 Voice: (734) 668-9900 Fax: (734) 668-7788 http://goovaerts.pierre.googlepages.com/ -- Pierre Goovaerts Chief Scientist at BioMedware Inc. 3526 W Liberty, Suite 100 Ann Arbor, MI 48103 Voice: (734) 913-1098 (ext. 202) Fax: (734) 913-2201 Courtesy Associate Professor, University of Florida Associate Editor, Mathematical Geosciences Geostatistician, Computer Sciences Corporation President, PGeostat LLC 710 Ridgemont Lane Ann Arbor, MI 48103 Voice: (734) 668-9900 Fax: (734) 668-7788 http://goovaerts.pierre.googlepages.com/
Re: AI-GEOSTATS: moving averages and trend
Yes you can. Look at my paper http://home.comcast.net/~pgoovaerts/BIOFERTarticle.pdf Fig. 14, right column. The EC local component is the nuggt effect component and the graph below shows the transect of electrical conductivity values after filtering of that microscale component. You can find other examples in the refereed publication section of my website . Cheers, Pierre On Mon, Feb 1, 2010 at 9:17 PM, M. Nur Heriawan mn_heria...@yahoo.comwrote: Goovaerts, regarding the factorial kriging you mentioned below...is it possible to filter the micro component (nugget effect) from our spatial model? Because the magnitude of nugget effect is related to the magnitude of variance error as well. Thank you. Regards, --- M. Nur Heriawan Earth Resources Exploration Research Group Faculty of Mining and Petroleum Engineering Institut Teknologi Bandung (ITB) Jl. Ganesha 10 Bandung 40132 INDONESIA http://www.mining.itb.ac.id/heriawan -- *From:* Pierre Goovaerts goovae...@terraseer.com *To:* seba sebastiano.trevis...@libero.it *Cc:* José M. Blanco Moreno jmbla...@ub.edu; ai-geostats@jrc.it *Sent:* Tue, February 2, 2010 12:27:09 AM *Subject:* Re: AI-GEOSTATS: moving averages and trend well Factorial Kriging Analysis allows you to tailor the filtering weights to the spatial patterns in your data. You can use the same filter size but different kriging weights depending on whether you want to estimate the local or regional scales of variability. Pierre 2010/2/1 seba sebastiano.trevis...@libero.it Hi José Thank you for the interesting references. I'm going to give a look! Bye Sebastiano At 15.46 01/02/2010, José M. Blanco Moreno wrote: Hello again, I am not a mathematician, so I never worried too much on the theoretical reasons. You may be able to find some discussion on this subject in Eubank, R.L. 1999. Nonparametric Regression and Spline Smoothing, 2a ed. M. Dekker, New York. You may be also interested on searching information in and related to (perhaps citing) this work: Altman, N. 1990. Kernel smoothing of data with correlated errors. Journal of the American Statistical Association, 85: 749-759. En/na seba ha escrit: Hi José Thank you for your reply. Effectively I'm trying to figure out the theoretical reasons for their use. Bye Sebas -- Pierre Goovaerts Chief Scientist at BioMedware Inc. 3526 W Liberty, Suite 100 Ann Arbor, MI 48103 Voice: (734) 913-1098 (ext. 202) Fax: (734) 913-2201 Courtesy Associate Professor, University of Florida Associate Editor, Mathematical Geosciences Geostatistician, Computer Sciences Corporation President, PGeostat LLC 710 Ridgemont Lane Ann Arbor, MI 48103 Voice: (734) 668-9900 Fax: (734) 668-7788 http://goovaerts.pierre.googlepages.com/ -- Pierre Goovaerts Chief Scientist at BioMedware Inc. 3526 W Liberty, Suite 100 Ann Arbor, MI 48103 Voice: (734) 913-1098 (ext. 202) Fax: (734) 913-2201 Courtesy Associate Professor, University of Florida Associate Editor, Mathematical Geosciences Geostatistician, Computer Sciences Corporation President, PGeostat LLC 710 Ridgemont Lane Ann Arbor, MI 48103 Voice: (734) 668-9900 Fax: (734) 668-7788 http://goovaerts.pierre.googlepages.com/
Re: AI-GEOSTATS: (1) Geostatistics in pain
Hello, Your list is far from complete. I would recommened you take a look at the following paper Goovaerts, P. 2009. Geostatistical software. In M.M. Fischer and A. Getis, editors, *Handbook of Applied Spatial Analysis: Software Tools, Methods and Applications.* Springer-Verlag, Berlin, Germany, in press.http://home.comcast.net/%7Epgoovaerts/A7-June.pdf that is available at http://sites.google.com/site/goovaertspierre/pierregoovaertswebsite/publications/book-chapters that provides an overview and comparison of functionalities in a series of geostat software, most of them listed on ai-geostat website. Cheers, Pierre On Wed, Jan 6, 2010 at 12:08 PM, Younes Fadakar yfa.st...@ymail.com wrote: Hi there, This is my first message as a test message checking the usage of the service, working with ai-geostats mailing list. I have many questions to ask you too. To start: The world of Geostatistics seriously needs a tool to present well to novices and professionals. Current availabilities have many of disadvantages. Some are too old, others not user-friendly and the rest more expensive. 1- GsLib seems powerful but too old (DOS-command line in 2010!) 2- WinGsLib is completely confusing despite of logging and automating! no direct input and output!! 3- Variowin is too weak in terms of GUI! 4- mGstat as a Matlab toolbox written too complex not handy program! 5- GeoR as an extention for R makes you to work with R such a command-line environment! what a development rather than GsLib!! 6- Isatis is more expensive; for what?! 7- Gs+ is in pain with weak performance of GUI! 8- Geoeas is something funny just to remember DOS graphics! 9- ... So obviously a serious request remained for more than 20 years without suitable answer! Why? Younes __ See what's on at the movies in your area. Find out now: http://au.movies.yahoo.com/session-times/ + + To post a message to the list, send it to ai-geost...@jrc.ec.europa.eu + To unsubscribe, send email to majordomo@ jrc.ec.europa.eu with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ -- Pierre Goovaerts Chief Scientist at BioMedware Inc. 3526 W Liberty, Suite 100 Ann Arbor, MI 48103 Voice: (734) 913-1098 (ext. 202) Fax: (734) 913-2201 Courtesy Associate Professor, University of Florida Associate Editor, Mathematical Geosciences Geostatistician, Computer Sciences Corporation President, PGeostat LLC 710 Ridgemont Lane Ann Arbor, MI 48103 Voice: (734) 668-9900 Fax: (734) 668-7788 http://goovaerts.pierre.googlepages.com/
Re: AI-GEOSTATS: Unconditional simulation
well Isobel explained how to go from Gaussian to uniform... On Tue, Nov 17, 2009 at 1:59 PM, Anatoly Saveliev s...@ksu.ru wrote: Pierre Goovaerts : For once, I agree with Isobel. sGs is the way to go... sGs - Gaussian by definition; he wants uniform :-) Anatoly Saveliev Pierre On Tue, Nov 17, 2009 at 6:00 AM, seba sebastiano.trevis...@libero.itmailto: sebastiano.trevis...@libero.it wrote: Hi Nick One way is to use simulated annealing (see gslib) putting as objective function your desired variogram and histogram. (but I guess that by means of some data transformation you can do that with a simple sequential gaussian simulation approach) Bye Sebas At 10.06 17/11/2009, Nick Hamm wrote: Dear all I want to simulate a spatially-correlated random field which follows a uniform rather than than Gaussian distribution. Does anybody know a straight-forward way to do this? Nick + + To post a message to the list, send it to ai-geost...@jrc.ec.europa.eu mailto:ai-geost...@jrc.ec.europa.eu + To unsubscribe, send email to majordomo@ jrc.ec.europa.eu http://jrc.ec.europa.eu with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ + + To post a message to the list, send it to ai-geost...@jrc.ec.europa.eu mailto:ai-geost...@jrc.ec.europa.eu + To unsubscribe, send email to majordomo@ jrc.ec.europa.eu http://jrc.ec.europa.eu with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ -- Pierre Goovaerts Chief Scientist at BioMedware Inc. 3526 W Liberty, Suite 100 Ann Arbor, MI 48103 Voice: (734) 913-1098 (ext. 202) Fax: (734) 913-2201 Courtesy Associate Professor, University of Florida Associate Editor, Mathematical Geosciences Geostatistician, Computer Sciences Corporation President, PGeostat LLC 710 Ridgemont Lane Ann Arbor, MI 48103 Voice: (734) 668-9900 Fax: (734) 668-7788 http://goovaerts.pierre.googlepages.com/ -- Pierre Goovaerts Chief Scientist at BioMedware Inc. 3526 W Liberty, Suite 100 Ann Arbor, MI 48103 Voice: (734) 913-1098 (ext. 202) Fax: (734) 913-2201 Courtesy Associate Professor, University of Florida Associate Editor, Mathematical Geosciences Geostatistician, Computer Sciences Corporation President, PGeostat LLC 710 Ridgemont Lane Ann Arbor, MI 48103 Voice: (734) 668-9900 Fax: (734) 668-7788 http://goovaerts.pierre.googlepages.com/
Re: AI-GEOSTATS: Instability in kriging coefficient matrix
Hello, I suspect that the problem resides in the use of 1 for setting up the unbiasedness constraints. In Gslib, the 1's are replaced by a constant that equals the largest value taken by the variogram (i.e. sill or pseudo-sill for unbounded models). While rescaling both the left and righ-hand sides of the unbiasedness constraint won't change the solution, it should make the matrix inversion much more stable. Pierre On Sun, Aug 3, 2008 at 12:10 AM, M.J. Abedini [EMAIL PROTECTED] wrote: Dear All; While conducting ordinary kriging with anisotropic power model, some of the entries in kriging coefficient matrix become unexpectly large creating problem for its inversion. One possible remedy would be to limit OK with moving neighborhood whereby the separation distance is under control. Is there any other solution for this problem? Thanks MJA + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ -- Pierre Goovaerts Chief Scientist at BioMedware Inc. 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 Courtesy Associate Professor, University of Florida Geostatistician, Computer Sciences Corporation President, PGeostat LLC 710 Ridgemont Lane Ann Arbor, MI 48103 Voice: (734) 668-9900 Fax: (734) 668-7788 http://goovaerts.pierre.googlepages.com/
Re: AI-GEOSTATS: Exact interpolant property of kriging
I think that Isobel refers to the implementation of the kriging algorithm in ESRI products where the nugget variability is automatically filtered from the data. Hence, if your variogram has a non-zero nugget effect, the kriged surface won't honor the data at sampled locations. Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 New website: http://goovaerts.pierre.googlepages.com/ On 2/19/08, M.J. Abedini [EMAIL PROTECTED] wrote: Dear Colleagues I thought exact interpolant property of kriging is very applicable to every case regardless of variogram used. But, Isobel's posting implies it is not the case. Furthermore, IDW of any types honor exact interpolant property. It can be proved mathematically. Thanks Abedini -- Forwarded message -- Date: Tue, 19 Feb 2008 11:09:44 + (GMT) From: Isobel Clark [EMAIL PROTECTED] To: Andrea Peruzzi [EMAIL PROTECTED], ai-geostats@jrc.it Subject: Re: AI-GEOSTATS: kriging or IDW in case study of hydrology? Andrea In theory kriging will honour the sample values provided your semi-variogram model takes the value zero at zero distance. Whether the data are honoured or not depends on which computer package you use and what it does with the semi-variogram at zero. You can force this behaviour by replacing any nugget effect with a short range model component. For example a spherical component with a range of influence of 10cm or some such. See our completely free and public domain kriging game, for how the kriging system works. By the way, IDW will only honour your sample values if the algorithms are written with the same criterion. Isobel http://www.kriging.com Andrea Peruzzi [EMAIL PROTECTED] wrote: Dear list, I'm graduate student in hydrogeology, I've to spatialize data of reservoir thickness, and I need to achieve a map having exactly the sampled value in the sampled localization (piezometers). I've little experience in geostatatistics. I had a look at kriging algorithms, but I did understand that kriging does not preserve the sampled value at sampled locations but it tends to smooth results, even if estimates correctly the unsampled space. So I wonder why should I use Kriging instead IDW (which it should preserve my sampled values): kriging respects the spatial variability but do not respect data As I told you before, I've very small knowledge in geostatistics stuff, but I'm interesting in kriging. Could anyone help me? Thanks a lot, Andrea Peruzzi PS: I apologize for writing you again but it's the first time I'm writing you, then I'm not sure how the mailing list works. Thanks :-) + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
RE: AI-GEOSTATS: gslib source code?
All the Gslib source codes + online help menus + additional files for cosimulation/accuracy plot can be downloaded from http://pangea.stanford.edu/ERE/research/scrf/software/ Cheers, Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: [EMAIL PROTECTED] on behalf of Samuel Verstraete Sent: Wed 2/21/2007 3:00 AM To: ai-geostats@jrc.it Subject: AI-GEOSTATS: gslib source code? Hi, Today i was looking for the source code of gslib and it seems to have vanished from the stanford website. AFAIK it used to be available on this url: http://pangea.stanford.edu/ERE/research/scrf/GSLIB/index.html Now all i can find is the link to the statios.com website where you need to register for downloading. AFAIK, gslib is open-source so source should be freely available? gr,S. -- Research group Soil Spatial Inventory Techniques Dept. Soil Management and Soil Care Faculty of Bioscience Engineering Ghent University Department of Soil Management and Soil Care Coupure Links 653 - Block B 9000 Ghent - Belgium Telephone +32(0)9 264.60.42 Fax +32(0)9 264.62.47 E-mail [EMAIL PROTECTED] http://soilman.ugent.be/orbit + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
AI-GEOSTATS: RE: Clarification required
This comment relates to the unbiasedness constraints of the cokriging system. In ordinary cokriging, the sum of the secondary data weights is forced to be equal to zero, hence if only secondary data are used in the estimation (i.e. if no data for the variable of interest or primary variable are found in the search window) the estimator has an expected value of zero and so will lead to a biased estimation... In the case of simple cokriging, if no primary data are found in the search window, a weight of one is assigned to the global mean of the primary variable which is added to the weighted sum of secondary data. Hope it helps Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
RE: AI-GEOSTATS: Rank-order geostats
Hi Mark, This technique is also known as uniform score transform and has been introduced in 1984 by Sullivan as part of probability kriging (e.g. see my book page 301). Like the normal score transform, it amounts at replacing original observations by the corresponding quantiles of a target distribution (e.g. uniform or standard normal distribution). You may want to check the following references that discuss the applications of the rank-order transform and its back-transform. Journel, A.G. and Deutsch, C.V., Rank Order Geostatistics: A Proposal for a Unique Coding and Common Processing of Diverse Data, Geostatistics Wollongong '96, Vol 1, Baafi and Schofield, editors, Kluwer Academic Publishers, September 1996, pp 174-187. Bourgault, G. et al. 1997. Geostatistical Analysis of a soil salinity data set. Adv. Agronomy 58, 241-292. http://www.ars.usda.gov/SP2UserFiles/Place/53102000/pdf_pubs/P1410.pdf http://www.ars.usda.gov/SP2UserFiles/Place/53102000/pdf_pubs/P1410.pdf If you do not want the standardized ranks to be valued 0 or 1, just use the transform rank/n-0.5/n instead of rank/n+1.. For large n, it shouldn't make a lot of differences.. The back-transform is always the critical point. You have the same problem when conducting multiGaussian kriging: the normal score back-transform of the kriging estimate obtained in the normal space will lead to a biased estimation in the original space... You can back-transform any quantile of the distribution though, that is the normal score back-transform of the median will still be the median in the original space. I have 2 comments: 1. The equation [21] given in Juang et al. paper applies only to simple lognormal kriging. For ordinary lognormal kriging, the Lagrange parameter must also be part of the back-transform, see Eq. (6) in Saito, H. and P. Goovaerts. 2000. Geostatistical interpolation of positively skewed and censored data in a dioxin contaminated site. Environmental Science Technology, vol.34, No.19: 4228-4235. 2. As mentioned in Journel Deutsch's paper, the estimates back-transformed according to the middle-point model are only median-unbiased. The idea is that the kriged rank is an estimate of the rank of the unknown original value, hence you simply compute the value with the same rank in the original distribution. Personally I would use the method described for the normal score back-transform in Saito and Goovaerts (2000): you compute the 100 percentiles of the local uniform distribution of probability in the transformed space, then back-transform those 100 percentiles and use their arithmetical average as an estimate of the mean of the local distributions in the original space. Cheers, Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: [EMAIL PROTECTED] on behalf of Mark Dowdall Sent: Mon 2/12/2007 5:13 AM To: ai-geostats@jrc.it Subject: AI-GEOSTATS: Rank-order geostats Hello I have a VERY skewed data set that fails tests for normality and log normality. Variograms are OK for th elower percentiles of the set but as one goes above the median the variograms get quite poor. And that is causing me a bit of a headache for Indicator kriging. I came across a paper by Juang et al (J. Environ. Qual. 2001, 30:894-903) that discussed the use of Rank-order geostats for highly skewed data and had my interest peked. I transformed the data by ranking them, then dividing each rank transformed data point by the total number of data points. And the variogram (omni directional, all data) looked exceptionally well. Enthused, I began reading and searching the archive on ai-geostats but have some questions. 1. Is rank order (as in rank/number of samples) geostatistics known by some other name as there doesnt seem to be too much out there bar a couple of papers? 2. Is n-score geostatistics the same thing? 3. Some people seem to say the rank should be divided by N+1 and others N. Which should it be or have I misunderstood? 4. Juang discusses back transforming the data using a middle point model. I cannot understand how he has acheived this. Has anyone any experience in back transforming the estimates to concentrations? I remember problems I had before with log transformed estimates and whether or not to add half the kriging variance to the back transformation value and would rather not fall into the same kind of problem. If any one has any info on rank order geostatistics and particularly back transforming, I would be very grateful. Thanks in advance M dowdall + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats
RE: AI-GEOSTATS: Categorical Indicator Simulation with Random Results
Hi Kim, You should expect that the simulated categorical maps are less smooth than the maps of probabilities of occurrence generated by indicator kriging. Nevertheless, the simulated maps should honor the data at sampled locations. You might be doing something wrong, but hard to tell until we have a look at the parameter file. Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: [EMAIL PROTECTED] on behalf of Kimberly D. Gordon Sent: Wed 9/13/2006 2:46 PM To: ai-geostats@jrc.it Subject: AI-GEOSTATS: Categorical Indicator Simulation with Random Results Hello to all! I'm having some difficulty with indicator simulation with GSLib's SISIM. I have three categorical variables (based on rock type) that I have kriged to a grid using IK3D. The variograms for the data are surprisingly well behaved considering that this is real data. The kriged results honor the data and show a relatively stratified system (what I expected...). However, when I plug the same input into SISIM, I get seemingly random results and the data is not honored for all of the multiple realizations. I'm not using soft data or Markov chains...What am I doing wrong? Any suggestions are greatly appreciated! Kim Kimberly Gordon INTERA Incorporated 9111 A Research Blvd. Austin, TX 78758 + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
RE: AI-GEOSTATS: Factorial kriging and backtransformation
Hi Gregoire, Spatial components or regionalized factors are mathematical constructs that might help interpreting complex spatial patterns and investigate scale-dependent correlations between physical attributes. I wouldn't try to assign a physical meaning to the value of those components or factors, hence I never bother attempting to backtransform them. Hope it helps, Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: [EMAIL PROTECTED] on behalf of Gregoire Dubois Sent: Wed 8/30/2006 6:02 AM To: ai-geostats@jrc.it Subject: AI-GEOSTATS: Factorial kriging and backtransformation Dear list, I was wondering how senior geostatisticians are dealing with back-transform issues of different factors (or components to be fully correct) obtained by applying factorial kriging to a transformed data set (e.g. log or nscore). Thank you for any hint on this matter. Kind regards, Gregoire __ Gregoire Dubois (Ph.D.) European Commission (EC) Joint Research Centre Directorate (DG JRC) WWW: http://www.ai-geostats.org http://www.ai-geostats.org The views expressed are purely those of the writer and may not in any circumstances be regarded as stating an official position of the European Commission. + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
RE: AI-GEOSTATS: Coregionalization2
Hi Gonzalo, There have been a couple of papers and associated computer codes published in Computers and geosciences: Morisette, J. T., An Example Using SAS to Fit the Model of Coregionalization, Computers and Geoscience, v. 23, n. 3, pp. 317-323, 1997. Pardo-Igúzquiza, E. and Dowd, P. A. 2002. FACOTR2D: a computer program for factorial cokriging. Comput. Geosci. 28, 8 (Oct. 2002), 857-873. DOI= http://dx.doi.org/10.1016/S0098-3004(02)3-1 http://dx.doi.org/10.1016/S0098-3004(02)3-1 Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: [EMAIL PROTECTED] on behalf of [EMAIL PROTECTED] Sent: Thu 8/24/2006 8:36 AM To: ai-geostats@jrc.it Subject: AI-GEOSTATS: Coregionalization2 Hello I would like to enclose more my question, after some of your replies, becasue my question was very broad. I am interested in a program for fitting a linear model of corregionalization for soil physical and chemical properties data. If some of you can give me some advices i woul be very please thank you very much Gonzalo + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
RE: AI-GEOSTATS: validation of Simulation
Hi Stefano, First, I am not sure that stochastic simulation is necessary in your case since you seem to be only interested in what I would call a measure of local (or location-specific) uncertainty. MultiGaussian kriging would in theory give you exactly the same results at less computational cost...of course these days kriging doesn't look as sexy as simulation... To validate the results, I would use the concept of accuracy plot introduced by Clayton Deutch. This concept and others are explained in the paper: Goovaerts, P. 2001. Geostatistical modelling of uncertainty in soil science. Geoderma, 103: 3-26. http://www.terraseer.com/training/geostats/geoder01.pdf that you can download from my webpage. Of course, you might also want to check the reproduction of target statistics, such as histogram and variogram, but again it seems that in your case your focus is on these probabilities of exceeding a particular threshold. I discuss these issues in another publication: Goovaerts, P. 2006. Geostatistical modeling of the spaces of local, spatial and response uncertainty for continuous petrophysical properties. Chapter in book Stochastic Modeling II published by the American Association of Petroleum Geologists... and I could send you a copy if you are interested... An executable to compute the accuracy plot can be downloaded from: http://ekofisk.stanford.edu/SCRFweb/GSLIB/added.html Cheers, Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: [EMAIL PROTECTED] on behalf of Stefano Pegoretti Sent: Tue 7/25/2006 10:38 AM To: ai-geostats@jrc.it Subject: AI-GEOSTATS: validation of Simulation Hallo! I'm a Ph.D. students who works with Indoor Radon Data, and it's the first time I join this list. I've a question for you: after post-processing several Sequential Gaussian Simulation to obtain a probability map of exceeding a given threshold, can someone suggests me a clever way to validate the results? (of course, I've a second dataset of measurements to use in this part of the work ; variography and sgs do not know this data...) Thanks a lot and best regards, stefano + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
RE: AI-GEOSTATS: geostatistical simulation questions
Hi Thomas, I am assuming that you transform your data before conducting your (sequential?) Gaussian simulation. In this case, the backtransform would yield only positive values, assuming of course that like S-GeMS Gstat asks the user to specify the minimum and maximum of the target histogram. Regards, Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: [EMAIL PROTECTED] on behalf of Thomas Adams Sent: Tue 7/25/2006 2:40 PM To: ai-geostats@jrc.it Subject: AI-GEOSTATS: geostatistical simulation questions List: I am interested in doing some geostatistical simulations using GSTAT and have some theoretical questions. I am attempting to model hourly rainfall accumulations over a large region, so there will almost always be zero rainfall somewhere. I can generate random fields of precipitation, using both conditional and unconditional simulations using GSTAT. However, I get negative values as well as (mostly) positive values. The simulated fields otherwise look very reasonable. The data I used (and must use) to estimate my variogram has zero values where no rainfall occurs. What does this suggest to you? I am using gaussian simulation. I have seen some references to more exotic geostatistcal simulation methods using bayesian or some other methods. Is this what I need? From reading the literature, I have seen that some researchers have successfully used indicator kriging and simulation. With GSTAT I can successfully do simulations using 'method : is' rather than 'method : gs', how using GSTAT do I model a continuous (non-binary) variable using the GSTAT syntax? Regards to all, Tom -- Thomas E Adams National Weather Service Ohio River Forecast Center 1901 South State Route 134 Wilmington, OH 45177 EMAIL: [EMAIL PROTECTED] VOICE: 937-383-0528 FAX:937-383-0033 + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
RE: AI-GEOSTATS: KWBP Test Program
Dear Raimon, If the data are not spatially correlated, your variogram will be modeled as a pure nugget effect and all observations will receive the same weights in your block kriging estimation. If you perform a global block kriging (i.e. use of a single search window), your estimate will then be the arithmetical average of your observations and the standard error will be provided by the kriging standard deviation. Cheers, Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: [EMAIL PROTECTED] on behalf of Raimon Tolosana Sent: Sat 7/22/2006 1:30 PM To: ai-geostats@jrc.it Subject: Re: AI-GEOSTATS: KWBP Test Program Dear list, this situation posed by Mr.Merks, in which spatial dependence is not strong enough as to be useful for geostatiscs, might be rather common. I'd like to ask to the list, what kind of estimation of reserves should be done in this case? In the absence of spatial dependence, classical statistics should apply: therefore, shall we estimate the mean value of ore content in the deposit by the arithmetic mean of the samples? And attach an error to it, in the fashion of the standard error of the mean (something like the variance of the sample divided by number of data used)? Or did I grossly misunderstand something in the discussion, with so much bogus-hocus-pocus and 5-line sentences? thanks for the patience Raimon Tolosana En/na JW ha escrit: Hello Readers, More talk and not test. I want to know what the KWBP methodology does with the Bre-X data. Is that too much to ask? Kind regards, Jan W Merks + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
RE: AI-GEOSTATS: Re: generalize kriging variance to average-basedestimators different than
Hi Oriol, It is not clear what you want to do with the kriging variance you obtain... Probably you want to quantify the degree of reliability of the allocation of a particular location to a given facies. This could be measured by the variance or entropy of the distribution of probabilities of occurrence of facies at that location, see my book page 354. This probability distribution is easily computed by indicator kriging or you can use truncated Gaussian simulation if there is any physical ordering of your facies. For your last question, look at Journel and Huijbregts Mining Geostatistics page 451 for the smoothing relations that link the average kriging variance to the variance of observations and the variance of kriging estimates. Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: [EMAIL PROTECTED] on behalf of Oriol Falivene Sent: Sat 7/15/2006 7:28 AM To: [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; ai-geostats@jrc.it Subject: Re: AI-GEOSTATS: Re: generalize kriging variance to average-basedestimators different than Hi all!. Isobel, Jan and Adrian, thank you for your useful suggestions. I have been studying the information you suggested me to read. Please, correct me if I am wrong: It is possible to compute the estimation variance for any weighted average estimator. But in order to do this we need to know the semivariogram of the property. More things regarding estimation variance: Actually, I am working in the interpolation/estimation of categorical properties (i.e. facies). Categorical properties take discrete values which do not need to follow any ordering (i.e. from category A to category C a transitional step with category B may not exist). 1) A possible manner of generating interpolation maps of this type of properties would follow this procedure: 1st) Order categories (this can be difficult to justify in some cases). 2nd) Transform the orderd categorical property into a continuous property by assigning a numerical value to each category. 3rd) Interpolate the numerical values of the continuous property. 4) Truncate the results of the interpolation of the continuous property with a number of thresholds equals to the number of categories minus one (the thresholds should be located between the assigned values), to get the interpolated categorical properties (or facies map). The estimation variance (or kriging variance) refers to the results of the continuous property interpolation (3rd step) and not to the categorical property obtained after truncating the continuous property (4th step). And therefore I do not know if it would be correct to assign this estimation variance to the estimation variance of the categorical property results (probably not), any idea on that? 2) Another option for generating interpolation maps of categorical properties is indicator kriging for categorical variables. The following procedure is used: 1st) The categorical property is transformes into n new properties (one for each category) according to the indicator transform for categorical variables, the value of each new property corresponds to the probability of finding the related category (or facies) at a given position. 2nd) Each new property is interpolated over all the grid and the results correspond to the probabilities for each category to be present in a location. 3rd) In each grid cell the category with the highest probability is chosen to obtain the interpolated categorical properties (or facies map). In this case we have a number of estimation variances related to occurrence probability at each point (one for each category). It is intuitive to chose the one of the selected category, but again I am not sure of how this variances, which are originally related to probability of occurrence, can be transported into variances of the categorical property, any idea on that? And one more question: I would like to know if there is any relationship between kriging estimation variance and the variance of the actual kriging estimates, any idea? Thanks a lot! Oriol Falivene Isobel Clark wrote: Oriol Download for free, my old book Practical Geostatistics. Chapter 4 tells you all about calculating the variance for any weighted average estimator. Follow links from http://www.kriging.com Isobel Oriol Falivene [EMAIL PROTECTED] wrote: Dear Colleagues, I'm a PhD student working on interpolation of categorical variables (like facies). I would like to know if it's possible to generalize the kriging variance to other average-based estimators different than
RE: [Fwd: Re: AI-GEOSTATS: Re: generalize kriging variance toaverage-basedestimators different than]
Hello, If you want to quantify the smoothness of an interpolated map of facies, you should use measure of spatial connectivity. For example, the indicator semivariogram provides information on the probability of transitioning from one facies to another, as a function of the separation distance. Superimposing the variograms computed from different interpolated maps would allow a quick visual comparison of the degree of smoothness of the different maps. Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: [EMAIL PROTECTED] on behalf of Oriol Falivene Sent: Sat 7/15/2006 10:07 AM To: ai-geostats@jrc.it Subject: [Fwd: Re: AI-GEOSTATS: Re: generalize kriging variance toaverage-basedestimators different than] Hi Dr Goovaerts, It is not clear what you want to do with the kriging variance you obtain... Probably you want to quantify the degree of reliability of the allocation of a particular location to a given facies. This could be measured by the variance or entropy of the distribution of probabilities of occurrence of facies at that location, see my book page 354. This probability distribution is easily computed by indicator kriging or you can use truncated Gaussian simulation if there is any physical ordering of your facies. I'm trying to get a measure of the smoothing effect related to a particular algorithm (truncated kriging, truncated inverse square distance, indicator kriging,...) and a particular algorithm set up (searching conditions or number of neighbours used to obtain each facies estimate), applied to interpolate facies distribution in a dense coal mine dataset. A good measure would be the variance of the estimated property, but since I am working with a categorical property (i.e. facies), it is not direct to get this variance (one must assume a certain facies ordering and attribute values to facies, and I'm not sure which would be the effect of this assumptions int the variance of measures). And therefore I was looking for other options like kriging estimation variance. For your last question, look at Journel and Huijbregts Mining Geostatistics page 451 for the smoothing relations that link the average kriging variance to the variance of observations and the variance of kriging estimates. thank you, I will take a look to this. Oriol + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
RE: [Fwd: Re: AI-GEOSTATS: Re: generalize kriging variancetoaverage-basedestimators different than]
To account for the proportion p of the facies, just rescale the semivariogram by the quantity p(1-p).. hence the different semivariograms should be comparable. The fact that the dominant facies tends to be over-represented in the interpolated map is well-known and frequent when using a maximum likelihood classification, see * Goovaerts, P. 1996. Stochastic simulation of categorical variables using a classification algorithm and simulated annealing. Mathematical Geology, 28(7): 909-921. * Goovaerts, P. and A.G. Journel. 1996. Accounting for local probabilities in stochastic modeling of facies data. SPE Journal, 1(1): 21-29. If you want to avoid the smoothing effect and reproduce target proportions for the different facies, you may want to use stochastic simulation as described in the aforementioned papers. Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: Oriol Falivene [mailto:[EMAIL PROTECTED] Sent: Sat 7/15/2006 10:45 AM To: Pierre Goovaerts; ai-geostats@jrc.it Subject: Re: [Fwd: Re: AI-GEOSTATS: Re: generalize kriging variancetoaverage-basedestimators different than] Thank you, I will try with the variograms as you suggested, however as the proportions of each facies are different in the different maps to compare (because of the smoothing), also the sills of the indicator variograms will be different making the comparision non-straightforward. And what about using the proportions of each facies, this seams even more simpler measure of smoothing to me than computing the indicator variograms. For example; as the smoothing increases, the proportions of the dominant facies also increase from that of the original hard data, and the more smoothing the largest the increase. Do you think that computing the proportion of the dominant facies would be of any statistical sense in order to quantify the smoothing effect? Thanks again Oriol Pierre Goovaerts wrote: Hello, If you want to quantify the smoothness of an interpolated map of facies, you should use measure of spatial connectivity. For example, the indicator semivariogram provides information on the probability of transitioning from one facies to another, as a function of the separation distance. Superimposing the variograms computed from different interpolated maps would allow a quick visual comparison of the degree of smoothness of the different maps. Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: [EMAIL PROTECTED] on behalf of Oriol Falivene Sent: Sat 7/15/2006 10:07 AM To: ai-geostats@jrc.it Subject: [Fwd: Re: AI-GEOSTATS: Re: generalize kriging variance toaverage-basedestimators different than] Hi Dr Goovaerts, It is not clear what you want to do with the kriging variance you obtain... Probably you want to quantify the degree of reliability of the allocation of a particular location to a given facies. This could be measured by the variance or entropy of the distribution of probabilities of occurrence of facies at that location, see my book page 354. This probability distribution is easily computed by indicator kriging or you can use truncated Gaussian simulation if there is any physical ordering of your facies. I'm trying to get a measure of the smoothing effect related to a particular algorithm (truncated kriging, truncated inverse square distance, indicator kriging,...) and a particular algorithm set up (searching conditions or number of neighbours used to obtain each facies estimate), applied to interpolate facies distribution in a dense coal mine dataset. A good measure would be the variance of the estimated property, but since I am working with a categorical property (i.e. facies), it is not direct to get this variance (one must assume a certain facies ordering and attribute values to facies, and I'm not sure which would be the effect of this assumptions int the variance of measures). And therefore I was looking for other options like kriging estimation variance. For your last question, look at Journel and Huijbregts Mining Geostatistics page 451 for the smoothing relations that link the average kriging variance to the variance of observations and the variance of kriging estimates. thank you, I will take a look to this. Oriol + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject
RE: AI-GEOSTATS: Re: STK
Hi Rajni, I have done some work on space-time kriging and I will present the results on the occasion of the IAMG 2006 congress. I posted a copy of the proceedings paper on my webpage if you are interested. It is very short (4 pages) but a longer paper is currently in preparation. Cheers, Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: [EMAIL PROTECTED] on behalf of Rajni Gaur Sent: Wed 6/14/2006 9:01 AM To: ai-geostats@jrc.it Subject: AI-GEOSTATS: Re: STK Dear List, I am trying to use space time kriging for my datasets, which i have acquired in time. Could you please give me some references on the space time kriging. Thanks in advance to all. Rajni + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
RE: AI-GEOSTATS: special case of ordinary cokriging
Hello, In general, the primary data will screen the influence of the co-located secondary data, leading to the similarity of the results provided by isotopic cokriging and kriging ignoring this secondary information; see my paper in Math Geol. Goovaerts, P. 1998. Ordinary cokriging revisited. Mathematical Geology, 30(1): 21-42. However, in some situations it is the secondary data that will screen the influence of primary data. It is likely to happen when both variables are strongly correlated and the secondary variable varies more continuously in space than the primary variable, i.e. the secondary variogram has a smaller nugget effect.. See the example in my book pages 219-220. Cheers, Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: [EMAIL PROTECTED] on behalf of Maarten De Boever Sent: Wed 6/7/2006 4:30 AM To: ai-geostats@jrc.it Subject: AI-GEOSTATS: special case of ordinary cokriging Dear all, The potential improvement of cokriging depends on the extend to which the secondary variable has been sampled additionally to the primary. Is there any difference between ordinary kriging and ordinary cokriging in the situation where all observations of the primary and secondary variable are located at the same locations? Will ordinary cokriging have in that situation any advantage over ordinary kriging? Thanks in advantage, De Boever Maarten. -- ir. Maarten De Boever Research Group Soil Spatial Inventory Techniques (ORBIT) Department Soil Management and Soil Care Faculty of Bioscience Engineering Ghent University Coupure 653, 9000 Gent, Belgium Tel. + 32 (0)9 264 6042 Fax + 32 (0)9 264 6247 e-mail : [EMAIL PROTECTED] http://www.soilman.ugent.be/orbit + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
RE: AI-GEOSTATS: multicategory indicator simulation
AI-GEOSTATS Hi Ashton, Sequential Indicator simulation (SIS) is based on the local estimation (i.e. kriging) of the probabilities of occurrence of each of the 7 categories, in your case. Thus, the local mean refers to the local (a priori) probability of occurrence of each of the seven classes based on the calibration of your map. The vectors of local means correspond to the row of your confusion, or error matrix. At the locations of ground-thruthed data, you have non only this vector of local means but also a vector of indicators of occurrence which should include 6 zeros and a one for the category that is observed on the ground. Indicator residuals are computed by subtracting these two vectors. SIS with varying local means is implemented in Gslib program sisim. Cheers, Pierre Pierre Goovaerts Chief Scientist at BioMedware Inc. Courtesy Associate Professor, University of Florida President of PGeostat LLC Office address: 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: [EMAIL PROTECTED] on behalf of Ashton Shortridge Sent: Thu 5/25/2006 10:24 AM To: ai-geostats@jrc.it Subject: AI-GEOSTATS: multicategory indicator simulation AI-GEOSTATS Hello all, I have a land cover dataset with codes 1-7 representing different land cover categories. This data is not too good, but might be better than nothing. Let's call this the map. I have a second dataset - a bunch of point locations at which land cover for the area has been ground-truthed. This is essentially my reference data. I can use these things to construct a confusion, or error matrix, like this: [1,] 0. 0.0222 0. 0.867 0. 0.000 0. [2,] 0.0778 0.1889 0. 0.722 0. 0.000 0.0111 [3,] 0. 0.2417 0.4917 0.250 0.0167 0.000 0. [4,] 0.0333 0.2667 0.0667 0.633 0. 0.000 0. [5,] 0. 0.7500 0. 0.125 0. 0.125 0. [6,] 0. 0. 0.9000 0.000 0.1000 0.000 0. [7,] 0. 0. 0. 0.000 0. 0.000 1. where cell i,j corresponds to the observed probability of observing class j on the ground, where class i was present in the map. For example, a cell with class 3 on the map is actually class 3 about 49% of the time. About 24% of the time it's class 2, and 25% of the time it is class 4. Very rarely (1.7%) it's actually class 5. I would like to employ indicator simulation on this data using simple kriging with locally varying means. I want to generate realizations of reference land cover, using the map landcover data to improve the prediction by serving as the mean estimate. This approach is documented in Goovaerts' book and in a paper by Kyriakidis and Dungan (2001). However, several points are unclear to me. First, simple kriging is employed on residuals from the mean. For multicategorical data of the sort I am investigating here, how would one calculate the mean at a particular location? Second and more practically, I've struggled to discover how to implement this in gstat (R version or standalone), and am wondering if anyone has had success with another software package. Thanks in advance for any assistance you can provide. Ashton -- Ashton Shortridge Assistant Professor [EMAIL PROTECTED] Dept of Geography http://www.msu.edu/~ashton 235 Geography Building ph (517) 432-3561 Michigan State University fx (517) 432-1671 Geography Has moved! Map: http://www.rsgis.msu.edu/images/parking-map.gif + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
RE: [ai-geostats] kriging without a nugget
Hi Niklas, This is a very good question; in fact one of the participants to my last short course asked the same question since he was using ARCview with the option nugget effect excluded and was surprised to see that his observations were not honored by the kriging predictions. This is related to the issue of how to define the nugget effect. On almost all figures in the literature the semivariogram model seems to start on the vertical axis at a value equal to the nugget effect, while in fact the value of the model is set to zero for h=0 in the kriging system. This ensure that kriging is an exact interpolator, which is usually a desirable property. When interpolated nodes correspond to sampled locations, this exactitude property can create spikes in the kriged map; in other words these locations contrast with the general smoothness of the interpolated map produced by kriging and it is one reason why the option to filter the noise, even at the sampled locations was introduced (A general presentation of the filtering properties of kriging can be found in my book p. 172-174). Another application of the filtering method is the use of kriging for finding minimum or maximum in numerical models; see paper Sasena, M.J., Parkinson, M., Goovaerts, P., Papalambros, P.Y. and M. Reed. 2002. http://ode.engin.umich.edu/publications/papers/2002/DETC2002_DAC34091.pdf Adaptive experimental design applied to an ergonomics testing procedure. Proceedings of DETC'02 ASME 2002 Design Engineering Technical Conferences and Computers and Information in Engineering Conference. Montreal, Canada, September 29- October 2, 2002. http://ode.engin.umich.edu/publications/papers/2002/DETC2002_DAC34091.pdf More generally, the question is whether the nugget effect represents measurement errors (variability at the sampled locations) which you might want to filter, or whether it represents small-scale variability in the field. Note that the discontinuities in the map will disappear if you use a simulation method. Hope it helps, Pierre Pierre Goovaerts Chief Scientist at BioMedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: Törneman Niklas [mailto:[EMAIL PROTECTED] Sent: Mon 3/6/2006 3:14 PM To: ai-geostats@unil.ch Subject: [ai-geostats] kriging without a nugget Hi All, I am sort of a beginner within this field and my question might seem a bit simple. Any help would be appreciated however. In commercial all purpose software's such as SURFER there is an option to exclude the nugget effect from the kriging interpolation. The purpose is to ensure that measured values are honoured at their locations. This seems understandable to me since the absence of a nugget ensures that the variance is zero at a distance of zero from the measured point, i.e. the measured value=interpolated value. However, in most projects that I work with (soil pollution problems) there is a significant nugget effect. My question is simply how the interpolation is affected if the nugget effect is excluded when in reality there is a clear nugget present in the data. One reason for this question (apart from a personal interest) is that I am trying to motivate the use of other methods (i.e. SGS) and more specialized software such as SGeMS and GS+. One good easily explainable motivation for this would be if the above mentioned methodology of excluding the nugget is inappropriate, which is suspect that it is. cheers Nicholas http://www.sweco.se/ * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] kriging without a nugget
Hello, The nugget effect is usually interpreted as a combination of small-scale variability and measurement errors. The only way to distinguish between both is to assess the magnitude of measurement errors in the lab, e.g. through replication of the measurement on subsamples. Wherever the interpolated node does not coincide with a sampled location, the nugget variance is filtered. So, I don't see how you could compare the common kriging with factorial kriging in a cross-validation mode. Regards, Pierre Pierre Goovaerts Chief Scientist at BioMedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: M. Nur Heriawan [mailto:[EMAIL PROTECTED] Sent: Mon 3/6/2006 8:39 PM To: ai-geostats@unil.ch Subject: RE: [ai-geostats] kriging without a nugget Dear Pierre, You mentioned that the nugget effect represents measurement errors or small-scale variability in the field. How to differentiate between both of them? Last year I was studying about factorial kriging. Actually on that time I wanted to use factorial kriging to filter the nugget variance. I assumed that my nugget variance only represented small-scale variability, because I was sure (I already checked) that my data set did not have any measurement errors. I also was surprised when I got the estimation result. Beforehand I guessed that the estimation result will be more precise as I excluded the nugget variance, but actually the result showed the contrary. When I plotted the estimated versus real values, the correlation coefficient was much less (compared to include the nugget variance in estimation). Thank you. Nur Heriawan Earth Resources Exploration Research Group Institut Teknologi Bandung Indonesia --- Pierre Goovaerts [EMAIL PROTECTED] wrote: Hi Niklas, This is a very good question; in fact one of the participants to my last short course asked the same question since he was using ARCview with the option nugget effect excluded and was surprised to see that his observations were not honored by the kriging predictions. M. Nur Heriawan http://www.mining.itb.ac.id/heriawan __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] Re: Software for Automatic Semivariogram Estimation
Hi, They are currently writing a book that would be similar to Gslib user manual but tailored to S-GeMS features. In the meantime, you can find some help in the user manual available at http://sgems.sourceforge.net/doc/sgems_manual.pdf Cheers, Pierre Pierre Goovaerts Chief Scientist at BioMedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: Mach Nife [mailto:[EMAIL PROTECTED] Sent: Wed 3/1/2006 3:36 PM To: Pierre Goovaerts; AI Geostats mailing list Subject: RE: [ai-geostats] Re: Software for Automatic Semivariogram Estimation It would be very nice if there would be a tutorial on how to use the variogram modeler. machnife --- Pierre Goovaerts [EMAIL PROTECTED] wrote: Hi Susan, I would recommend the Stanford Geostatistical Modeling Software (S-GeMS) that is public domain and that I use in all my short courses (some of your colleagues have actually be trained by me). The software can be downloaded from http://pangea.stanford.edu/~nremy/GEMS/ Cheers, Pierre Pierre Goovaerts Chief Scientist at BioMedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: Hohner, Susan [mailto:[EMAIL PROTECTED] Sent: Tue 2/28/2006 1:28 PM To: AI Geostats mailing list Subject: RE: [ai-geostats] Re: Software for Automatic Semivariogram Estimation Yikes! I was working through the tutorial for the Geostatistical Analyst Extension when this email discussion popped up. Any recommendations for a traditional geostatistics software package? Thanks, Susan Susan Hohner, Senior Geographer Everglades Division, Mail Stop 4440 South Florida Water Management District 3301 Gun Club Road, West Palm Beach, FL 33406 (561) 682-6801 phone (561) 682-0100 fax [EMAIL PROTECTED] http://www.sfwmd.gov From: Chaosheng Zhang [mailto:[EMAIL PROTECTED] Sent: Tuesday, February 28, 2006 12:25 PM To: AI Geostats mailing list Subject: Re: [ai-geostats] Re: Software for Automatic Semivariogram Estimation Dear all, I have the same concerns with ArcGIS Geostatistical Analyst Extension (v.9.1). I would use a traditional geostatistics software package to fit the variogram models in a very traditional way, and input the parameters to ArcGIS for kriging. It seems that ArcGIS has its own reasons to show variograms in a non-traditional way, but I find it almost impossible to fit the variograms mannually. You can change the parameters, but it is very hard to see how well they fit. By the way, you can change the lag distance or interval in ArcGIS (it is called lag size there). Cheers, Chaosheng -- Dr. Chaosheng Zhang Lecturer in GIS Department of Geography National University of Ireland, Galway IRELAND Tel: +353-91-492375 Fax: +353-91-495505 E-mail: [EMAIL PROTECTED] Web1: www.nuigalway.ie/geography/zhang.html Web2: www.nuigalway.ie/geography/gis * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] Quick question about S-GeMS SGS simulation
Hi Perry, I am assuming you are using the latest version of S-GeMS. In the sgsim function, there is an histogram tab where you can specify the target histogram you want to reproduce (like in Gslib the normal score transform and back-transform will be performed automatically inside the program, and you can choose among power, exponential and hyperbolic models for the lower and upper tail extrapolation). Cheers, Pierre Pierre Goovaerts Chief Scientist at BioMedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: Collier, Perry (TS) [mailto:[EMAIL PROTECTED] Sent: Thu 3/2/2006 6:42 PM To: AI Geostats mailing list Subject: [ai-geostats] Quick question about S-GeMS SGS simulation Quick question about S-GeMS SGS simulation: I had a go at producing an SGS simulation in S-GeMS, but the result was Gaussian (ie: not back transformed as GSLIB SGSIM.exe does) - can anyone advise how to back-transform the simulation in S-GeMS? What about tail extrapolation - how does S-GeMS handle that - there's no place in the user interface to put in any parameters as per GSLIB. I'm light years away from being an expert in this stuff, so it's likely I've missed something obvious... Cheers Perry Collier Senior Geologist Rio Tinto Technical Services Phone: +61 7 3327 7676 Mobile: 0408 015 837 -Original Message- From: Pierre Goovaerts [mailto:[EMAIL PROTECTED] Sent: Friday, 3 March 2006 7:50 AM To: Mach Nife; AI Geostats mailing list Subject: RE: [ai-geostats] Re: Software for Automatic Semivariogram Estimation Hi, They are currently writing a book that would be similar to Gslib user manual but tailored to S-GeMS features. In the meantime, you can find some help in the user manual available at http://sgems.sourceforge.net/doc/sgems_manual.pdf Cheers, Pierre Pierre Goovaerts Chief Scientist at BioMedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: Mach Nife [mailto:[EMAIL PROTECTED] Sent: Wed 3/1/2006 3:36 PM To: Pierre Goovaerts; AI Geostats mailing list Subject: RE: [ai-geostats] Re: Software for Automatic Semivariogram Estimation It would be very nice if there would be a tutorial on how to use the variogram modeler. machnife --- Pierre Goovaerts [EMAIL PROTECTED] wrote: Hi Susan, I would recommend the Stanford Geostatistical Modeling Software (S-GeMS) that is public domain and that I use in all my short courses (some of your colleagues have actually be trained by me). The software can be downloaded from http://pangea.stanford.edu/~nremy/GEMS/ Cheers, Pierre Pierre Goovaerts Chief Scientist at BioMedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: Hohner, Susan [mailto:[EMAIL PROTECTED] Sent: Tue 2/28/2006 1:28 PM To: AI Geostats mailing list Subject: RE: [ai-geostats] Re: Software for Automatic Semivariogram Estimation Yikes! I was working through the tutorial for the Geostatistical Analyst Extension when this email discussion popped up. Any recommendations for a traditional geostatistics software package? Thanks, Susan Susan Hohner, Senior Geographer Everglades Division, Mail Stop 4440 South Florida Water Management District 3301 Gun Club Road, West Palm Beach, FL 33406 (561) 682-6801 phone (561) 682-0100 fax [EMAIL PROTECTED] http://www.sfwmd.gov From: Chaosheng Zhang [mailto:[EMAIL PROTECTED] Sent: Tuesday, February 28, 2006 12:25 PM To: AI Geostats mailing list Subject: Re: [ai-geostats] Re: Software for Automatic Semivariogram Estimation Dear all, I have the same concerns with ArcGIS Geostatistical Analyst Extension (v.9.1). I would use a traditional geostatistics software package to fit the variogram models in a very traditional way, and input the parameters to ArcGIS for kriging. It seems that ArcGIS has its own reasons to show variograms in a non-traditional way, but I find it almost impossible to fit the variograms mannually. You can change the parameters, but it is very hard to see how well they fit. By the way, you can change the lag distance or interval in ArcGIS (it is called lag size there). Cheers, Chaosheng -- Dr. Chaosheng Zhang Lecturer in GIS Department of Geography National University of Ireland, Galway IRELAND Tel: +353-91-492375 Fax: +353-91-495505 E-mail: [EMAIL PROTECTED] Web1: www.nuigalway.ie/geography/zhang.html Web2: www.nuigalway.ie/geography/gis * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following
RE: [ai-geostats] Re: Software for Automatic Semivariogram Estimation
Hi Susan, I would recommend the Stanford Geostatistical Modeling Software (S-GeMS) that is public domain and that I use in all my short courses (some of your colleagues have actually be trained by me). The software can be downloaded from http://pangea.stanford.edu/~nremy/GEMS/ Cheers, Pierre Pierre Goovaerts Chief Scientist at BioMedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: Hohner, Susan [mailto:[EMAIL PROTECTED] Sent: Tue 2/28/2006 1:28 PM To: AI Geostats mailing list Subject: RE: [ai-geostats] Re: Software for Automatic Semivariogram Estimation Yikes! I was working through the tutorial for the Geostatistical Analyst Extension when this email discussion popped up. Any recommendations for a traditional geostatistics software package? Thanks, Susan Susan Hohner, Senior Geographer Everglades Division, Mail Stop 4440 South Florida Water Management District 3301 Gun Club Road, West Palm Beach, FL 33406 (561) 682-6801 phone (561) 682-0100 fax [EMAIL PROTECTED] http://www.sfwmd.gov From: Chaosheng Zhang [mailto:[EMAIL PROTECTED] Sent: Tuesday, February 28, 2006 12:25 PM To: AI Geostats mailing list Subject: Re: [ai-geostats] Re: Software for Automatic Semivariogram Estimation Dear all, I have the same concerns with ArcGIS Geostatistical Analyst Extension (v.9.1). I would use a traditional geostatistics software package to fit the variogram models in a very traditional way, and input the parameters to ArcGIS for kriging. It seems that ArcGIS has its own reasons to show variograms in a non-traditional way, but I find it almost impossible to fit the variograms mannually. You can change the parameters, but it is very hard to see how well they fit. By the way, you can change the lag distance or interval in ArcGIS (it is called lag size there). Cheers, Chaosheng -- Dr. Chaosheng Zhang Lecturer in GIS Department of Geography National University of Ireland, Galway IRELAND Tel: +353-91-492375 Fax: +353-91-495505 E-mail: [EMAIL PROTECTED] Web1: www.nuigalway.ie/geography/zhang.html Web2: www.nuigalway.ie/geography/gis * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] dssim-HR
Hello, I used Compaq Visual Fortran and didn't have any problem compiling the program. Pierre Pierre Goovaerts Chief Scientist at BioMedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ From: samuel verstraete [mailto:[EMAIL PROTECTED] Sent: Fri 2/17/2006 9:51 AM To: Thomas Mejer Hansen; ai-geostats@unil.ch Subject: Re: [ai-geostats] dssim-HR I have tried this intel compiler but i didn't have the same luck as you, would you mind sharing the command you used to compile the module? i used: $ ifort dssim-hr.for which just results in this error list... which is shorter than the gnu f77 compiler error... but still wrong : fortcom: Error: Illegal character in statement label field [M] fortcom: Error: Illegal character in statement label field [O] fortcom: Error: Illegal character in statement label field [D] fortcom: Error: Illegal character in statement label field [U] fortcom: Error: Illegal character in statement label field [L] fortcom: Error: First statement in file must not be continued fortcom: Error: dssim-hr.for, line 9: Illegal character in statement label field [I] IMPLICIT NONE ^ fortcom: Error: dssim-hr.for, line 9: Illegal character in statement label field [M] IMPLICIT NONE -^ fortcom: Error: dssim-hr.for, line 9: Illegal character in statement label field [P] IMPLICIT NONE --^ fortcom: Error: dssim-hr.for, line 9: Illegal character in statement label field [L] IMPLICIT NONE ---^ fortcom: Error: dssim-hr.for, line 9: Illegal character in statement label field [I] IMPLICIT NONE ^ fortcom: Error: dssim-hr.for, line 9: Syntax error, found END-OF-STATEMENT when expecting one of: = = . ( : % IMPLICIT NONE -^ fortcom: Error: dssim-hr.for, line 31: Syntax error, found END-OF-STATEMENT when expecting one of: , ) parameter (MAXNST=4,MAXROT=MAXNST+1,UNEST=-99.0,EPSLON=1.0e-20,VERSION=3.01) ---^ fortcom: Error: dssim-hr.for, line 60: Illegal character in statement label field [E] END MODULE DSSmodule ^ fortcom: Error: dssim-hr.for, line 60: Illegal character in statement label field [N] END MODULE DSSmodule -^ fortcom: Error: dssim-hr.for, line 60: Illegal character in statement label field [D] END MODULE DSSmodule --^ fortcom: Error: dssim-hr.for, line 60: Illegal character in statement label field [M] END MODULE DSSmodule ^ fortcom: Error: dssim-hr.for, line 63: Illegal character in statement label field [p] program dssim ^ fortcom: Error: dssim-hr.for, line 63: Illegal character in statement label field [r] program dssim -^ fortcom: Error: dssim-hr.for, line 63: Illegal character in statement label field [o] program dssim --^ fortcom: Error: dssim-hr.for, line 63: Illegal character in statement label field [g] program dssim ---^ fortcom: Error: dssim-hr.for, line 63: Illegal character in statement label field [r] program dssim ^ fortcom: Error: dssim-hr.for, line 60: Syntax error, found IDENTIFIER 'DULEDSSMODULEMDSSIM' when expecting one of: = = / * , END-OF-STATEMENT ; [ END MODULE DSSmodule --^ fortcom: Error: dssim-hr.for, line 80: Illegal character in statement label field [e] end program dssim ^ fortcom: Error: dssim-hr.for, line 80: Illegal character in statement label field [n] end program dssim -^ fortcom: Error: dssim-hr.for, line 80: Illegal character in statement label field [d] end program dssim --^ fortcom: Error: dssim-hr.for, line 80: Illegal character in statement label field [p] end program dssim ^ fortcom: Error: dssim-hr.for, line 84: Illegal character in statement label field [s] subroutine readparm() --^ fortcom: Error: dssim-hr.for, line 84: Illegal character in statement label field [u] subroutine readparm() ---^ fortcom: Error: dssim-hr.for, line 84: Illegal character in statement label field [b] subroutine readparm() ^ fortcom: Severe: Too many errors, exiting compilation aborted for dssim-hr.for (code 1) On Fri, 17 Feb 2006 15:21:40 +0100 (CET) Thomas Mejer Hansen [EMAIL PROTECTED] wrote: I have succesfully compiled dssim_HR using the Intel Fortran compiler (free for noncommercial use on Linux) : http://www.intel.com/cd/software/products/asmo-na/eng/compilers/219771.htm Have a nice day - Thomas -- Research group Soil Spatial Inventory Techniques Dept. Soil Management and Soil Care Faculty of Bioscience Engineering Ghent University Department of Soil Management and Soil Care Coupure Links 653 - Block B 9000 Ghent - Belgium Telephone +32(0)9 264.60.42 Fax +32(0)9 264.62.47 E-mail [EMAIL PROTECTED] * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] FK and K of spat. comp.
Hi Simone, Your definition is historically the correct one.. Note that the term kriging analysis, sloppy translation for the term analyse krigeante introduced in Matheron's 1992 seminal paper, could also be used. The use of the term factorial kriging for the decomposition of a RF into spatial components based on a nested semivariogram model can be traced back to Gslib user manual. To be consistent, I used the same term in my book and used the expression multivariate factorial kriging when dealing with more than one variable (e.g. regionalized PCA). I think that both expressions are acceptable, albeit confusing. The use of the term factorial does not systematically imply a multivariate analysis and a factor does not need to be a linear combination of variables. Hope it helps, Pierre Pierre Goovaerts Chief Scientist at BioMedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ -Original Message- From: Simone Sammartino [mailto:[EMAIL PROTECTED] Sent: Wed 1/18/2006 9:12 AM To: Geostat newsgroup Cc: Subject:[ai-geostats] FK and K of spat. comp. Dear All a term specification: What is the difference between Factorial Kriging and Kriging of spatial components. I've always believed that the first is the kriging of the factors deriving from the regionalized PCA of multivariate datasets, and the second is the discrimination of the different spatial components deriving from the evaluation of nested variograms. But I can still read in most of scientific articles, about factorial kriging as the estimation procedure related to nested variograms...and it should be not exactly correct! Do I wrong? Thank you Simone - Dr. Simone Sammartino PhD student - Geostatistical analyst - G.I.S. mapping I.A.M.C. - C.N.R. Geomare-Sud section Port of Naples - Naples [EMAIL PROTECTED] - * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] Traditional OCK or Standardize OCK?
Hello, It is indeed correct that as for simple cokriging, the standardized OCK requires knowledge of the population means for both primary and secondary variables, and as I mentioned in my book p. 232 Provided the data are representative of the study area, these means can be estimated from the sample means. Of course, we could also account for the uncertainty attached to those samples means.. but the same can be said regarding the uncertainty attached to the parameters of the semivariogram model... The main reason ordinary kriging is used instead of simple kriging is its ability to accommodate changes in the mean across the study area (what I called global trend in my book) through the use of local search windows. The interesting fact for standardized OCK is that, even if a global mean is used in the standardization, local means are still re-estimated within each search window thanks to the unbiasedness constraint. The main assumption however is that after rescaling by their global means both primary and secondary variables have the same local mean, see Goovaerts (1997, 1998). For me, this might be the main weakness/limitation of the approach. As always, cross-validation is a good way to compare the prediction performances of the different estimators. Pierre Pierre Goovaerts Chief Scientist at BioMedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ -Original Message- From: Heuvelink, Gerard [mailto:[EMAIL PROTECTED] Sent: Thu 1/5/2006 4:31 AM To: Pierre Goovaerts; Adrián Martínez Vargas; Behrang Kushavand; ai-geostats@unil.ch Cc: Subject:RE: [ai-geostats] Traditional OCK or Standardize OCK? The downside of SOCK (often not mentioned) is that as a minimum requirement one must know the difference(s) between the population means (i.e., the means of the random functions) of the primary and secondary variables. In practice, one rarely knows these and uses the differences between the sample means instead, which is incorrect, unless one takes the associated estimation errors into account. However, when the BLUE of the differences between population means is used and the associated estimation errors are taken into account, then I suspect that SOCK boils down to something very close or identical to TOCK. Along similar lines, recall that substituting the BLUE of the population mean in the simple kriging equations yields a predictor that is identical to the ordinary kriging predictor (I think it is in Cressie's book, but in fact it is not that difficult to establish this result). The main (only?) purpose of using ordinary kriging instead of simple kriging is that one often does not know the population mean and cannot simply assume that it is equal to the sample mean or some other combination of the sample data. That is why ordinary kriging is used much more often than simple kriging. It puzzles me why so many geostatisticians so easily replace TOCK by SOCK and ignore the problem above. It is not the right method to avoid large and many negative weights, there are much better ways for that (see discussion of one month ago). Gerard Gerard B.M. Heuvelink Soil Science Centre Wageningen University and Research Centre P.O. Box 47 6700 AA Wageningen The Netherlands tel +31 317 474628 / 482420 email [EMAIL PROTECTED] http://www.sil.wur.nl/UK/ -Original Message- From: Pierre Goovaerts [mailto:[EMAIL PROTECTED] Sent: donderdag 5 januari 2006 0:20 To: Adrián Martínez Vargas; Behrang Kushavand; ai-geostats@unil.ch Subject: RE: [ai-geostats] Traditional OCK or Standardize OCK? Hi, The main difference between SOCK and TOCK is that, in the standardized form, only one unbiasedness constraint is imposed, i.e. the sum of all primary and secondary data weights is one, while in the traditional version a separate constraint is applied for each variable, i.e. sum of primary data weights is one and the sum of secondary data weights is zero for each secondary variable. The traditional constraints lead to larger and more frequent negative weights for the secondary variables. The difference between SOCK and TOCK estimates is expected to increase as differences between the variance of primary and secondary variables increases. The different types of cokriging are described and compared in the following paper: Goovaerts, P. 1998. Ordinary cokriging revisited. Mathematical Geology, 30(1): 21-42. Cheers, Pierre Pierre Goovaerts Chief Scientist at BioMedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ -Original Message- From: Adrián Martínez Vargas [mailto:[EMAIL PROTECTED] Sent: Wed 1/4/2006 12:53 PM To: Behrang Kushavand; ai-geostats@unil.ch Cc: Subject:Re: [ai-geostats] Traditional OCK or Standardize OCK? In the definition of the cross variogram you can see
RE: [ai-geostats] Traditional OCK or Standardize OCK?
Hi, The main difference between SOCK and TOCK is that, in the standardized form, only one unbiasedness constraint is imposed, i.e. the sum of all primary and secondary data weights is one, while in the traditional version a separate constraint is applied for each variable, i.e. sum of primary data weights is one and the sum of secondary data weights is zero for each secondary variable. The traditional constraints lead to larger and more frequent negative weights for the secondary variables. The difference between SOCK and TOCK estimates is expected to increase as differences between the variance of primary and secondary variables increases. The different types of cokriging are described and compared in the following paper: Goovaerts, P. 1998. Ordinary cokriging revisited. Mathematical Geology, 30(1): 21-42. Cheers, Pierre Pierre Goovaerts Chief Scientist at BioMedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 (ext. 8) Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ -Original Message- From: Adrián Martínez Vargas [mailto:[EMAIL PROTECTED] Sent: Wed 1/4/2006 12:53 PM To: Behrang Kushavand; ai-geostats@unil.ch Cc: Subject:Re: [ai-geostats] Traditional OCK or Standardize OCK? In the definition of the cross variogram you can see that it is not adimentional (depend of units Km, %, ppm, etc.), you can avoid this effect using standardize Ordinary Co-Kriging. Adrian -Original Message- From: Behrang Kushavand [EMAIL PROTECTED] To: ai-geostats@unil.ch Date: Wed, 4 Jan 2006 19:55:01 +0330 Subject: [ai-geostats] Traditional OCK or Standardize OCK? Dear All, Is it true that estimation variance of standardize Ordinary Co-Kriging (SOCK) is always equal or smaller than Traditional Ordinary Co-Kriging (TOCK)? What is the advantage of TOCK to SOCK (I think it is about negative weights) and are there any criteria to choice TOCK or SOCK? Thanks Behrang Participe en el V Congreso Internacional de Educación Superior Universidad 2006. La Habana, Cuba, del 13 al 17 de Febrero del 2006 http://www.universidad2006.cu _ Instituto Superior Minero Metalúrgico de Moa Dr. Antonio Núñez Jiménez http://www.ismm.edu.cu * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] KED
Hi Rajni, You can find an application of KED to the mapping of hydraulic conductivity in my latest publication in mathematical Geology: Patriarche, D., Castro, M.C. and P. Goovaerts. 2005. Estimating regional hydraulic conductivity fields - A comparative study of geostatistical methods. Mathematical Geology, 37(6), 587-613. You can download the PDF of the paper from the my webpage: http://home.comcast.net/~goovaerts/publication.html Cheers, Pierre === Pierre Goovaerts Chief Scientist at Biomedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ -Original Message- From: Rajni Gaur [mailto:[EMAIL PROTECTED] Sent: Tue 11/29/2005 9:52 AM To: AI Geostats mailing list Cc: Subject:[ai-geostats] KED Dear List members, Can anyone of the seniors of the geostatistical community, please give me some references regarding the application of kriging with external drift (KED) pertaining to the groundwater systems. I want to apply KED for the estimation of transmissivity values for an aquifer. Looking for the references and response, Thanks to all in advance Regards Rajni * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] Adding kriging variances.
Hi Digby, The variance of a sum of random variables is not equal to the sum of their variances, except if they are independent.. Consequently, the kriging variances cannot be simply combined to compute the variance of the global estimator, see Mining Geostatistics, Page 323 (Journel and Huijbregts, 1978). The most flexible approach is to use stochastic simulation to generate a set of realizations of the block values, aggregate them, and use the empirical distribution of aggregated block values as a model of uncertainty. You can find an example in my publication http://home.comcast.net/~goovaerts/karen.pdf and there is a whole book devoted to the use of stochastic simulation in mining. Journel and Kyriakidis. 2004. Evaluation of Mineral reserves. A simulation approach. Oxford University Press. Cheers, Pierre -Original Message- From: Digby Millikan [mailto:[EMAIL PROTECTED] Sent: Mon 11/7/2005 1:09 AM To: AI Geostats mailing list Cc: Subject:[ai-geostats] Adding kriging variances. Dear list, If you had a set of blocks is it possible to add their kriging variances together, to get a standard error for the mine head grade, is this something that is done in practice or has been done in the past to estimate possible deviations from a kriged head grade? Digby * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] Transformation of zero values for kriging
Hi Christian, It looks like your data are rates and consist of a numerator and denominator. In this case, you cannot apply common geostatistical tools since the reliability of your observations will vary depending on the number of eggs you have sampled for each plant. I am facing similar problems with cancer rates and have been using both Poisson kriging and binomial cokriging. A few papers are available on my webpage and I just submitted another paper that describes Poisson kriging in more details and provides the code to do the analysis. I would gladly share it with you once it is accepted, which should be very soon. Regards, Pierre Pierre Goovaerts Chief Scientist at Biomedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ -Original Message- From: Schlatter Christian [mailto:[EMAIL PROTECTED] Sent: Tue 10/25/2005 4:33 AM To: ai-geostats@unil.ch Cc: Subject:[ai-geostats] Transformation of zero values for kriging Dear list-readers We were studying lepidopteran egg parasitism rates in a grid design (on 6 x 24 plants, totally 144) in the field. The idea was to detect influences in terms of directional dependencies of the parasitism rate (variograms), of base parasitism rate (nugget), and of distance of the effects (range). Now egg parasitism rate was quite low and rare due to different problems (meteorological, area, test site, spraying in field, etc) so we have an average of 10-20 % of parasitized eggs (out of 144 points). For some non-spatial statistical testing this was more or less ok. For the geostatistical work (I was thinking of universal kringing), the number of values is quite little. So I was thinking about transforming them by means of the exponential function (exp^zi), in order to get 144 values and to retransform the result. Can I proceed in this way or am I completely on the wrong way with this approach? Thank you very much for you attention. Best wishes Christian ° Christian Schlatter GIS Pflanzenschutz: Schädlinge - Nützlinge Forschungsinstitut für biologischen Landbau (FiBL) Institut de recherche de l'agriculture biologique Istituto di ricerche dell'agricoltura biologica Research Institute of Organic Agriculture CH-5070 Frick, Switzerland Ackerstrasse CH- 5070 Frick [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] Telefon 0041 62 865 72 75 Fax 0041 62 865 72 73 Unsere Homepage: http://www.fibl.org http://www.fibl.org/ http://www.biogene.org http://www.biogene.org/ http://www.organicxseeds.com http://www.organicxseeds.com http://gis.fibl.ch/WWFMap/viewer.htm http://gis.fibl.ch/WWFMap/viewer.htm ° * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] Indicator semivariogram
Hi Koen, Although I am not a native speaker, you may want to check the few papers I have written on the computation and interpretation of indicator semivariograms, besides my book of course... Most recent papers can be downloaded from my webpage. Your choice should also be guided by whether you apply indicators to continuous or categorical variables. Pierre Pierre Goovaerts Chief Scientist at Biomedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ -Original Message- From: Koen Hufkens [mailto:[EMAIL PROTECTED] Sent: Thu 9/22/2005 8:07 AM To: ai-geostats@unil.ch Cc: Subject:[ai-geostats] Indicator semivariogram Hi list, Does anyone have any references to a clear description of the procedures to calculate and interpretate an indicator semivariogram? I know the principles but my scientific english isn't good enough to formulate it correctly. I got some remarks on a sentence I used in a working paper so I would like to cite from original sources. Thanks in advance, Koen * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] natural neighbor applied to indicator transforms
Hi, In fact, as long as the weights are all positive and sum up to one, your interpolated probability will always be between 0 and 1; so you should be all right.. The approach proposed by Sebastiano is similar to median indicator kriging in the sense that the weights assigned to the observations will be the same across all indicators (here instead of a single indicator semivariogram used to compute the kriging weights, the same weighting set will be applied to all indicators since the data configuration, hence the size of the Thiessen polygons, doesn't change among indicators). Because all the weights are positive and remain the same for the different indicators, this approach should eliminate all order relation deviations (all estimated probabilities will be between 0 and 1, and at each location their sum will be one). Pierre -Original Message- From: Gregoire Dubois [mailto:[EMAIL PROTECTED] Sent: Mon 9/5/2005 7:00 AM To: 'seba'; ai-geostats@unil.ch Cc: Subject: RE: [ai-geostats] natural neighbor applied to indicator transforms Ciao Sebastiano, I realized nobody replied to your question (sorry for have added confusion here). I don't see any objection in applying any interpolator to probability values. However, you should better use exact interpolators to avoid getting probabilities of occurences 1 (or smaller than 0) Cheers Gregoire -Original Message- From: seba [mailto:[EMAIL PROTECTED] Sent: 02 September 2005 10:07 To: ai-geostats@unil.ch Cc: ai-geostats@unil.ch; 'Nicolas Gilardi' Subject: RE: [ai-geostats] natural neighbor applied to indicator transforms I try to reformulate my question. When performing direct (i.e. without crossvariogram) indicator kriging, practically we interpolate probability values by means of ordinary kriging. These probability values could represent the probability of occurrence of some category or the probability to overcome some threshold. My question is: is there anything wrong to interpolate these probability values with other interpolating algorithm like, for example natural neighbor (or triangulation)? In my opinion is all ok . considering also that we have no problem of order relation violations. Again, this technique is applied only for a preliminary data analysis Then a short consideration directed about the importance of boundaries: Quoting Nicolas Gilardi My personnal feeling about the distinction between using a classification algorithm or a regression one is the importance you put on the boundaries.If you look for smooth boundaries, with uncertainty estimations, etc., then a regression algorithm (like indicator kriging) is certainly a good approach. Well, if you use fuzzy classification the boundaries become continuos...fuzzy. Bye S. Trevisani * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] Sum of predicted values
Hi Pete, This is a classical example where stochastic simulation would allow an easy quantification of the uncertainty attached to the aggregated value. Just generate a series of realizations of your process over these 1700 points, sum each set of simulated values, and use the empirical distribution of simulated block values as a model of uncertainty. You can find an example in Goovaerts, P. 2001. Geostatistical modelling of uncertainty in soil science. Geoderma, 103: 3-26. http://www.terraseer.com/training/geostats/geoder01.pdf that you can download from my webpage. Cheers, Pierre -Original Message- From: Pete Gething [mailto:[EMAIL PROTECTED] Sent: Mon 8/1/2005 9:30 AM To: ai-geostats@unil.ch Cc: Subject: [ai-geostats] Sum of predicted values Dear list, I have Kriged predictions of a continuous variable at a set of 1700 points. I want to sum these values and obtain an estimate of the overall prediction variance based on the kriging variances of the individual points (i.e., taking into account the spatial correlation between points). The data are approximately Gaussian. I would expect there to be a standard solution to this problem, but I'm having difficulty finding examples - can anyone help me out, or point me to a reference? Thanks in advance, Pete Peter Gething School of Electronics and Computer Science School of Geography University of Southampton Highfield Southampton SO17 1BJ UK Tel: +44 (0) 23 8059 2013 Email: [EMAIL PROTECTED] * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] SGSIM query
Hi Ellen, I have a version of SGSIM that does that and would gladly share it with you. Note that the public-domain/windows software S-GeMS developed at Stanford (http://ekofisk.stanford.edu/SCRFweb/sgems/) also allows simulation on a grid of points that you load an an object. Cheers, Pierre Pierre Goovaerts Chief Scientist at Biomedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ -Original Message- From: Norm, Ellen, Kathy Emily [mailto:[EMAIL PROTECTED] Sent: Tue 7/26/2005 8:32 AM To: ai-geostats@unil.ch Cc: Subject: [ai-geostats] SGSIM query Hello all, does anybody have (or know where I can obtain) a version of the GSLIB program SGSIM that simulates directly to a points file? I want to simulate at the locations corresponding to my grade control (exhaustive) data locations which are not on a regular grid. Thanking you all in anticipation of a response, Ellen Ellen Bandarian (PhD student) School of Engineering and Mathematics Edith Cowan University 100 Joondalup Drive Joondalup, WA, 6027 * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] modelling trend and kriging type
To add to the excellent comments by Edzer and Gregoire, 1. Universal kriging = kriging with a trend. The second terminology has been proposed by Andre Journel who felt that the term universal was vague and misleadingly ambitious. 2. Kriging with an external drift (KED) is mathematically the same as universal kriging (UK). Secondary variables are simply replacing the spatial coordinates used in UK. 3. Regression kriging denotes all the techniques where the trend is modeled outside the kriging algorithm. There are various methods that can be used to model that trend, ranging from linear regression to neural networks. Kriging is used to interpolate the residuals. In practice these techniques have more flexibility than universal kriging in term of modeling the trend: multiple variables either categorical or continuous can be incorporated easily and many sofwtare are available for this trend modeling. The only limitation is that the trend is modeled globally (i.e. the regression coefficients are constant in space) while in KED the coefficients are reestimated within each search window. Cheers, Pierre Pierre Goovaerts Chief Scientist at Biomedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ -Original Message- From: Recep kantarci [mailto:[EMAIL PROTECTED] Sent: Thu 6/30/2005 9:38 AM To: ai-geostats@unil.ch Cc: Subject: [ai-geostats] modelling trend and kriging type Dear ai-geostats members When the data used has a trend, it is needed to model trend and in this case there exists various types of kriging to apply (universal kriging, kriging with a trend, regression kriging etc). If this is the case, does one should use the same type of kriging or different depending on modeling the trend using coordinates of target variable or using other (namely, secondary or auxillary) variables such as elevation or topography ? That is , are there a dinstinction depending on the type of variables to model the trend while kriging? Best regards Recep _ Yahoo! kullaniyor musunuz? Istenmeyen postadan biktiniz mi? Istenmeyen postadan en iyi korunma Yahoo! Posta’da http://tr.mail.yahoo.com http://tr.mail.yahoo.com/ * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] Interpolation error estimation
Hi, You could try using kriging with non-systematic error where you can directly incorporate information on the reliability of your data in the kriging system. The technique is described in Chiles and Delfiner's textbook. Pierre -Original Message- From: Jose Luis Gomez Dans [mailto:[EMAIL PROTECTED] Sent: Wed 5/11/2005 9:37 AM To: ai-geostats@unil.ch Cc: Subject: [ai-geostats] Interpolation error estimation Hi! I am aware that this question could easily become a whole book, and a rather thick one at that :), but nevermind... I have produced an interpolation using regularised splines. The point data that goes into the interpolation has some measure of error associated to it, so to a first approximation, this can be used as an error indicator for cells of the interpolated grid that contains one such point. The problem arises with grid cells that have no real data in them, as the error is a function of the error of the surrounding points, the distance from the surrounding points (the further you are from a sa, the larger the error) and the error due to the interpolation. I could try and model the error, but there's a scale factor due to the surface presenting different features at different scales, and it all becomes very complicated. So my question is, what is the best way to come up with a grid that has a reasonable error estimation of the interpolated surface, if we know the error of each of the samples that went in? Many thanks for your time and help, Jos ___ Yahoo! Messenger - want a free and easy way to contact your friends online? http://uk.messenger.yahoo.com * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] gslib and simulations with 5 categories
Hi, Just perform the new coding outside Gslib. Excel should do it. Pierre -Original Message- From: gianni [mailto:[EMAIL PROTECTED] Sent: Wed 5/4/2005 6:06 AM To: ai-geostats@unil.ch Cc: Subject: [ai-geostats] gslib and simulations with 5 categories Hi I have 1 categorical variable (transmissivity), with 5 possible outcomes s=1,2,3,4,5 from very low to very high. I have already created simulations with sisim, but i need now to create simulations with 2 outcomes:high and low, but at the same time using the informations i have and then i don't want to give the 1,2 to low and 3,4,5 to high. I would like to know how if it is possible. Thanks Gianni * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] A banal question...
Hi Simone... one way to look at it is to call Z(x) the tail variable and Z(x+h) the head variable. These are two variables describing the relative position in space of values of the same physical attribute. Then, you have your 2 variables to compute the covariance fuction for a given vector h... By pooling together pairs of data from different parts of the study area, you are ignoring the location x and simply using the length (and possibly orientation) of the vector joining these 2 data... which requires the assumption of stationarity of the covariance.. Hope it helps, Pierre -Original Message- From: Simone Sammartino [mailto:[EMAIL PROTECTED] Sent: Mon 5/2/2005 10:14 AM To: Geostat newsgroup Cc: Subject: [ai-geostats] A banal question... Dear all a banal question... I'm not able to understand the stationarity of covariance in second order stationarity theory... On any book or article I can read: covariance between Z(x) e Z(x+h) exist and does not depend on x, but only on h; in fact Cov[Z(x),Z(x+h)]=Cov(h) It is considered so banal that in any text I consulted this part is described with the same sentence...but it is not explicated via mathematical formalism Why should E[Z(x)Z(x+h)]-m^2 be so logically reduced to Cov(h) You'll laugh for my request, but I'm not able to understand why it should be so logical In some text I found also...=Cov(x1-x2)=Cov(h) where distance between x1 and x2 is exactly h, but it does not help me to understand it I can't realize how to calculate Cov(h) that is a variable (it is in reality at least a vector of constant), when usually covariance is calculated between two variables Please have the patience to help me to solve this trick Thanks Simone - Dr. Simone Sammartino PhD student - Geostatistical analyst - G.I.S. mapping I.A.M.C. - C.N.R. Geomare-Sud section Port of Naples - Naples [EMAIL PROTECTED] - 6X velocizzare la tua navigazione a 56k? 6X Web Accelerator di Libero! Scaricalo su INTERNET GRATIS 6X http://www.libero.it * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] Determining background concentrations in soils at industrial sites
Hi Nicolas, Your problem seems to bear some similarities with the problem of target detection in satellite imagery that I presented at the last Geostat congress. The key issue was to detect any outliers/anomalies of a given size that would depart significantly from the background, without any prior knowledge about these background values. It involved geostat filtering to enhance local variations, followed by the application of a spatial statistics (LISA) to compute the probability for filtered values to be local anomalies. This research is summarized in a couple of proceedings paper you can download from my webpage, in addition to a paper that has just been published in Remote Sensing of the Environment and that I would gladly send you in a PDF format. Cheers, Pierre Pierre Goovaerts Chief Scientist at Biomedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ -Original Message- From: Nicolas Jeanne [mailto:[EMAIL PROTECTED] Sent: Thu 4/21/2005 3:46 AM To: ai-geostats@unil.ch Cc: Subject: [ai-geostats] Determining background concentrations in soils at industrial sites Dear list members, I'm currently working on a large industrial site and we are comparing background (natural+anthropogenic) vs. site (point sources) concentration levels for various chemical compounds. Searching through the internet for existing methodologies about this issue, the main documents and references I've found are coming from the U.S.: EPA of course and also the Navy. The methodology described is basically based on: - statistical evaluation of background distribution, - statistical tests to compare background vs. site concentration distributions, or on-site background concentrations vs. off-site concentrations. Geostatistics doesn't seem to be so much applied on the area (I guess this is partly due to the usually important spatial heterogeneity of such background levels), except to investigate for spatial trends in the datasets, such trends being potentially related to hot-spots... Have you heard about or applied or in mind other methodologies to address this issue? Any feedback, reference or papers will be greatly appreciated!! I'll summarize the answers and inform the list. Regards, Nicolas Jeanne. -- http://www.geovariances.com GEOVARIANCES - 49 bis, Av. F. Roosevelt - 77210 Avon - FRANCE Phone: +33-(0)-160.74.91.04-Fax: +33-(0)-164.22.87.28 * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] gslib postik and cpdf
Hello, Implementation of the maximum likelihood allocation (pick up the class with the maximum probability of occurrence) is easy to implement, starting with any Excel spread sheet. Soares algorithms is much less straigthforward to code and I should be able to exhume a Fortran program I wrote 12 years ago to do so. Note however that because Soares' classification does not account for spatial patterns, the less frequent category/class often tends to be allocated to isolated pixels or to the borders of the study are where the probability of occurrence of other classes is smaller than average. Cheers, Pierre -Original Message- From: gianni [mailto:[EMAIL PROTECTED] Sent: Wed 4/20/2005 4:54 PM To: ai-geostats@unil.ch Cc: Subject: [ai-geostats] gslib postik and cpdf Hi I'm italian student from Rome and i would want to give my excuses for my poor english in the first place. I have a problem with the gslib, i would know how to postprocess the cpdf values that the ik3d gives when we use the categorical variables, because the postik seems to work only with continuous variables. I have 1 categorical variable, with 5 possible outcomes s=1,2,3,4,5. I would allocate locations to the category with the largest p under the constraint of reproduction of global proportion. That is the Soares algorithm, but i I have chosen this one because this is the only i found (Goovaerts97) and you could suggest me another one. However I need some fortran libraries about that. Let me do another question This variable is the transmissivity, and i deduced this 5 outcomes from very low to very high. How about using this information like a continuous variable in ik3d? How to fix the first threshold? If i fix the first zk=1 i have in every location the I(u;zk)=1. Thanks for your time Regards * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] Definition of standardize variograms
Hi Gregoire, I agree with you regarding the merits of the standardized semivariogram as implemented in variowin software. In one of my last studies, the rescaling by the lag variance helped correcting the preferential sampling of wells with high arsenic levels, leading to a susbtantial decrease in random fluctuations of the experimental semivariograms. While the general relative semivariogram approximates the lag variance by the square of the lag mean, the standardized semivariogram uses the actual lag variance, hence makes less assumptions. Regarding the terminology, I guess we should used a term like lag-standardized to distinguish the global and lag-specific standardization or rescaling of semivariogram values. Cheers, Pierre -Original Message- From: Gregoire Dubois [mailto:[EMAIL PROTECTED] Sent: Tue 4/5/2005 9:48 AM To: ai-geostats@unil.ch Cc: [EMAIL PROTECTED] Subject: [ai-geostats] Definition of standardize variograms Dear list, While playing around with different software, I encounter different definitions for standardized variograms. Surfer (which is using the terminology of Variowin), uses the term standardized semivariogram for variograms obtained by dividing the semivariance by the lag variance, while GS+ uses the total variance. While the function obtained in GS+ is only a matter of rescaling variograms, allowing so various variograms to be compared, those proposed in Surfer have the same pupose as the local, pairwise and/or general relative variograms (see Isaaks Srivastava, page 163-170), that is to reduce the influence of local means. Interestingly enough, one may note that very few software propose relative variograms while I, very personally, consider these functions as essential for detecting spatial structures of many environmental variables. I have thus here two questions about the use of standardized/relative variogram: 1) What is the correct terminology or definition for standardized variograms? (I personally do not like very much the use of standardized when the standardisation is only applied to each lag...) 2) The general relative variogram (lag divided by the mean of the lag) has properties that are very similar to the standardized variogram (lag divided by the variance of the lag) but both functions differ. How shall one decide what to use and what are the relative properties of these functions? Thank you in advance for any feedback. Gregoire PS: a few points here good be added to Tom Mueller's FAQ on Geostatistical Software Conventions. __ Gregoire Dubois (Ph.D.) JRC - European Commission IES - Emissions and Health Unit Radioactivity Environmental Monitoring group TP 441, Via Fermi 1 21020 Ispra (VA) ITALY Tel. +39 (0)332 78 6360 Fax. +39 (0)332 78 5466 Email: [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] WWW: http://www.ai-geostats.org http://www.ai-geostats.org WWW: http://rem.jrc.cec.eu.int http://rem.jrc.cec.eu.int The views expressed are purely those of the writer and may not in any circumstances be regarded as stating an official position of the European Commission. * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] bi-Gaussian assumption for non-mathematicians
Well, as the author of the green bible I guess I should help out a little bit here... The key idea is that there exists an analytical expression that allows you to compute a priori, for any threshold of a multigaussian random function, the indicator semivariogram models. You only need to know the threshold and the normal score semivariogram of the variable. Then, you just compare the expected or theoretically-derived indicator semivariogram models to the empirical or derived from the data ones. Note that you don't even need to go through the burden of computing the theoretically-derived indicator semivariogram models to know that the underlying assumptions of the multigaussian model are not fulfilled. In many situations, you will notice that your experimental indicator semivariograms are not symmetric with respect to the median; for example the 0.1 decile semivariogram might have a longer range than the 0.9 decile semivariogram. This happens frequently since the low background values tend to be better connected in space than the high values... The next question is what do we do with that?... or in other words How do we know that the differences between expected and empirical indicator semivariograms are significant. You could test it, but I don't think it's worth it in practice... Well, cross-validation has taught me that even if the indicator semivariograms don't look like expected under the multigaussian model, multigaussian kriging might still give you better results than indicator kriging.. so it's hard to come up with cast-in-stone rules regarding the relative merits of parametric and non-parametric approaches.. but I am sure that everyone who has some experience with geostatistics has already realized that.. As I often say during my short-course, geostatistics provides you with a toolbox, and cross-validation and experience will teach you wich tools to use in any particular situation... Cheers, Pierre -Original Message- From: Perry Collier [mailto:[EMAIL PROTECTED] Sent: Tue 3/22/2005 7:55 PM To: ai-geostats@unil.ch Cc: Subject: [ai-geostats] bi-Gaussian assumption for non-mathematicians Hi all from Oz (Australia) First post on this list. I am a mine geo currently doing some post-grad geostats study (Edith Cowan Uni in WA, hi Dr Ute, Prof. Lyn!). Expanding on some very useful feedback from my Uni course director, I would be interested in your learned from the horse's mouth comments (what, why, how, when) regarding the bi-Gaussian assumption for Gaussian simulation and the various means of checking it. I am slightly mathematically challenged, so if anyone could explain the whole thing without too much scary maths, it would be much appreciated. I have Goovaerts' green geostats bible, which is good stuff, but I'm trying to convert some of the maths to English. Any comments from mining practitioners would be interesting... Cheers Perry Collier Senior Mine Geologist Ernest Henry Mine Xstrata Copper Australia Ph:(07) 4769 4527 Fax: (07) 4769 4555 E-mail: [EMAIL PROTECTED] Web: http://www.xstrata.com PO Box 527 Cloncurry QLD 4824 Australia I like rich people. I like the way they live. I like the way I live when I'm with them... From Roger Hammerstein's Sound of Music ** The information contained in this e-mail is confidential and is intended only for the use of the addressee(s). If you receive this e-mail in error, any use, distribution or copying of this e-mail is not permitted. You are requested to forward unwanted e-mail and address any problems to the Xstrata Queensland Support Centre. Support Centre e-mail: [EMAIL PROTECTED] Support Centre phone: Australia 1800 500 646 International +61 2 9034 3710 ** * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] practical range vs range
Hi Els, The key question here is the sampling density and how many data will be included in this search window. If there are many, the screening effect will greatly attenuate the impact of the data further away, hence using a or 3a won't make a big difference. If data are sparser, then usually I set up my search strategy in terms of maximum number of data, not maximum search radius, at least in 2D (in 3D setting the search ellipsoid right is very important). Although simple kriging weights become zero beyond the range, it is not the case for ordinary kriging, which is a reason why you shouldn't systematically discard the observations outside the range of autocorrelation, in particular if the sampling density is low.. Regards, Pierre -Original Message- From: Els Verfaillie [mailto:[EMAIL PROTECTED] Sent: Wed 3/23/2005 5:08 AM To: ai-geostats@unil.ch Cc: Subject: [ai-geostats] practical range vs range Hi list, I want to do ordinary kriging with an anisotropic variogram with GSLIB. My variogram is an exponential model with a practical range of 1800 m in direction 50 and 880 m in direction 320. I'm not sure whether I have to use the practical range (which is 3a) or the value a, which is respectively 733 m and 293 m. Furthermore I wonder which maximum search radius I have to choose: the 3a or the a value? Any suggestions? Cheers, Els ___ Els Verfaillie, PhD student Renard Centre of Marine Geology - Ghent University Krijgslaan 281-S8 B-9000 Gent - Belgium tel: +32-9-2644573 fax: +32-9-2644967 e-mail: [EMAIL PROTECTED] url: http://www.rcmg.ugent.be/ ___ -- No virus found in this outgoing message. Checked by AVG Anti-Virus. Version: 7.0.308 / Virus Database: 266.8.0 - Release Date: 21/03/2005 * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] multivariate geostats - anisotropy
Hi Els, You should expect the directions of anisotropy to be fairly similar if the two attributes are reasonably correlated. Except if you use a linear model of coregionalization, there is no requirement for the directions of anisotropy or the anisotropy ratio to be the same. Pierre Pierre Goovaerts Chief Scientist at Biomedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ -Original Message- From: Els Verfaillie [mailto:[EMAIL PROTECTED] Sent: Thu 2/17/2005 9:22 AM To: ai-geostats@unil.ch Cc: Subject: [ai-geostats] multivariate geostats - anisotropy Hello, I have a dataset of 2 variables: grainsize of the sediment and a digital elevation model of the seafloor. I want to interpolate the grainsize using different multivariate geostatistical techniques like cokriging, kriging with external drift, colocated cokriging, simple kriging with varying local means to compare their results. I have a question about anisotropy: do the directions of anisotropy have to be exactly the same for the two variables? In fact, during my analysis, I have a lot of practical questions like this. Does anyone can advise me some good references about multivariate geostatistics and/or anisotropy? I already have the book of H. Wackernagel and some articles of P. Goovaerts. Especially, I would like to have some practical guidelines. Thanks in advance! Els ___ Els Verfaillie, PhD student Renard Centre of Marine Geology - Ghent University Krijgslaan 281-S8 B-9000 Gent - Belgium tel: +32-9-2644573 fax: +32-9-2644967 e-mail: [EMAIL PROTECTED] url: http://www.rcmg.ugent.be/ ___ * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] kriging with variable measurement error
Hi Gali, You can download the following paper from my webpage: (http://home.comcast.net/~goovaerts/publication.html) Goovaerts, P., AvRuskin, G., Meliker, J., Slotnick, M., Jacquez, G.M. and J. Nriagu. 2004. Modeling uncertainty about pollutant concentration and human exposure using geostatistics and a space-time information system: Application to arsenic in groundwater of Southeast Michigan. In Accuracy 2004: Proceedings of the 6th International Symposium on Spatial Accuracy Assessment in Natural Resources and Environmental Sciences. http://www.biomedware.com/about/pdfs/accuracy_arsenic.pdf and look at the kriging system (5). I have another publication under review for Water Resources Research but the turn over is so slow that it might not get published before the end of this year. Regards, Pierre Goovaerts Chief Scientist at Biomedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ -Original Message- From: Gali Sirkis [mailto:[EMAIL PROTECTED] Sent: Wed 2/16/2005 12:18 PM To: ai-geostats@unil.ch Cc: Subject: [ai-geostats] kriging with variable measurement error Dear list members, Could you please advise regarding readable publications about Kriging with Variance of measurement error? Many thanks, Gali __ Do you Yahoo!? Read only the mail you want - Yahoo! Mail SpamGuard. http://promotions.yahoo.com/new_mail * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] Regression vs. Kriging vs. Simulation vs. IDW
Well... I would say that IDW is still being used by a few consultants that think that kriging is too complicated to apply and that the client will pay them as long as the map looks pretty... and less cynically IDW could give OK results if your data are gridded and the pattern of variability is ostropic. Pierre Pierre Goovaerts Chief Scientist at Biomedware 516 North State Street Ann Arbor, MI 48104 Voice: (734) 913-1098 Fax: (734) 913-2201 http://home.comcast.net/~goovaerts/ -Original Message- From: Darla Munroe [mailto:[EMAIL PROTECTED] Sent: Tue 1/4/2005 3:06 PM To: ai-geostats@unil.ch Cc: Subject: RE: [ai-geostats] Regression vs. Kriging vs. Simulation vs. IDW Just to get the group's opinion on this - When do you use IDW? When is it an advantageous technique, or what purposes does it well serve? Darla Munroe -Original Message- From: Syed Abdul Rahman Shibli [mailto:[EMAIL PROTECTED] Sent: Tuesday, January 04, 2005 2:19 PM To: jyarus; 'Seumas P. Rogan'; ai-geostats@unil.ch Subject: Re: [ai-geostats] Regression vs. Kriging vs. Simulation vs. IDW Perhaps there is some confusion here. Simple kriging, for instance, can be decomposed to the familiar multilinear regression equation since if one assumes all the Z(Xi)s are independent variables, then in the covariance matrix C all of C(Xi,Xj) would be zero except for C(Xi,Xi). So LiC(Xi,Xi)=C(Xi,Xo) The lambdas here being the parameters of the regression equation. The intercept term is the sam, i.e. Lo=E(y)-LiE(xi). Not sure if the previous poster meant this or simply using the location as the independent variable. Cheers Syed On 3/1/05 5:34 PM, jyarus [EMAIL PROTECTED] wrote: Hi Seumas: I thought I would throw my 2 cents in regarding a comparison between kriging and linear regression. While some of the responses have hit a few important differences, like Kriging is a spatial estimator and regression is not, or kriging will honor the original data and regression will not (unless residuals are added back in - not often done). For me, the critical point to be made is between the collocated cokriging application and regression. In collocated cokriging, like simple regression, two variables are being used, one independent and one dependent (of course, this could be expanded to more than one independent variable). The object is to predict a value of the dependent variable from a relationship established between both the independent and dependent observed values. In the ensuing regression equation, there is a slope term. For example, in the equation, Y= c-bX, c is the intercept and b is the slope. As pointed out by one of the contributors, regression by itself is not a spatial estimator, it is a point estimator. As such, the equation contains no information about the surrounding data or about the relationship between the observed data and the unsampled location where a desired estimate of the dependent variable is required. In kriging (or cokriging), the slope term b is replaced by a covariance matrix that informs the system not only about the behavior of the surrounding data points and the unsampled location (similar to distance weighting if omnidirectional), but also about the spatial behavior within the neighborhood - that is, how neighbors are spatially related to other neighbors. Thus, the slope term b is replaced with a sophisticated covariance matrix containing the spatial information. The ramifications of using simple regression instead of true spatial estimator are significant if the results are presented in map form. While this is often difficult to grasp for some, using simple regression as a mapping tool will cause geographic portions of a map to consistently be overestimated and others underestimated! For example, you may find that all the values estimated in the upper left quadrant of the map to be overestimated, and those in the lower right to be underestimated. We would like to believe that a good spatial estimator will be unbiased, and the distribution of the error variances over the area of a map will be uniform - no one part of the map will preferentially over- or underestimated. The bias brought about by the slope term in simple
Re: [ai-geostats] sisim and SK vs. OK
Hi Eric, I am guessing that the sampling density in 3D is small, hence at the beginning of the simulation procedure when the grid is essentially empty the estimate depends mainly on the type of trend model you adopt, i.e. a global mean for SK or a locally re-estimated local mean for OK. It looks like your estimate of the local mean is not very good and this might biase your simulation right from the beginning. Are you using a multiple grid strategy and did you notice a higher proportion of order relation deviations when using OK versus SK? Note that in the recent Stanford software GEMS, sequential indicator simulation is available only with the SK option, which might indicate some problems with the OK option... Hope it helps, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://home.comcast.net/~goovaerts/ On Mon, 3 Jan 2005 [EMAIL PROTECTED] wrote: Hello all. I'm performing unconditional sequential indicator simulation over a 3D domain. As the method requires, I have specified the data cdf at (10) thresholds, and have also defined parameters for an (exponential) variogram at each threshold. When I run the simulation algorithm using simple kriging to estimate the cdf at each threshold for each node, the method works great, i.e., I am able to approximatley reproduce the domain cdf and variograms. However, when I use ordinary kriging, the method falls apart. It seems that reproduction of the domain cdf becomes 'blocky' and looses its smoothness. I have made the search ellipsoid very large and allowed the number of data points that can be used for the OK to be very large, but the results are always poor. Does anyone have any suggestions about what is happening with the OK? I am using GSLib software. Thanks to all! Eric Bhark * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] F and T-test for samples drawn from the same p
Hello, I am currently principal investigator on a major NIH grant that aims to develop software for test of hypothesis using alternate hypothesis specified by the user and that differ from the omnibus spatial independence; we called them spatial neutral models. For example, you can test for clusters of cancer rates above and beyond a regional background in exposure. The p-values are computed using randomization and I applied geostatistical simulation to generate multiple realizations that are then used to derive the empirical distribution of the test statistic. I presented an example during the last GeoEnv conference and I put a PDF copy of the paper, which is in press for the moment, on my website. Cheers, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Sun, 5 Dec 2004, Colin Daly wrote: Hi Sorry to repeat myself - but the samples are not independent. Independance is a fundamental assumption of these types of tests - and you cannot interpret the tests if this assumption is violated. In the situation where spatial correlation exists, the true standard error is nothing like as small as the (s/sqrt(n)) that Chaosheng discusses - because the sqrt(n) depends on independence. Again, as I said before, if the data has any type of trend in it, then it is completely meaningless to try and use these tests - and with no trend but some 'ordinary' correlation, you must find a means of taking the data redundancy into account or risk get hopelessly pessimistic results (in the sense of rejecting the null hypothesis of equal means far too often) Consider a trivial example. A one dimensional random function which takes constant values over intervals of lenght one - so, it takes the value a_0 in the interval [0,1[ then the value a_1 in the interval [1,2[ and so on (let us suppose that each a_n term is drawn at random from a gaussian distribution with the same mean and variance for example). Next suppose you are given samples on the interval [0,2]. You spot that there seems to be a jump between [0,1[ and [1,2[ - so you test for the difference in the means. If you apply an f test you will easily find that the mean differs (and more convincingly the more samples you have drawn!). However by construction of the random function, the mean is not different. We have been lulled into the false conclusion of differing means by assuming that all our data are independent. Regards Colin Daly -Original Message- From: Chaosheng Zhang [mailto:[EMAIL PROTECTED] Sent: Sun 12/5/2004 11:42 AM To: [EMAIL PROTECTED] Cc: Colin Badenhorst; Isobel Clark; Donald E. Myers Subject: Re: [ai-geostats] F and T-test for samples drawn from the same p Dear all, I'm wondering if sample size (number of samples, n) is playing a role here. Since Colin is using Excel to analyse several thousand samples, I have checked the functions of t-tests in Excel. In the Data Analysis Tools help, a function is provided for t-Test: Two-Sample Assuming Unequal Variances analysis. This function is the same as those from many text books (There are other forms of the function). Unfortunately, I cannot find the function for assuming equal variances in Excel, but I assume they are similar, and should be the same as those from some text books. From the function, you can find that when the sample size is large you always get a large t value. When sample size is large enough, even slight differences between the mean values of two data sets (x bar and y bar) can be detected, and this will result in rejection of the null hypothesis. This is in fact quite reasonable. When the sample size is large, you are confident with the mean values (Central Limit Theorem), with a very small stand error (s/(sqrt(n)). Therefore, you are confident to detect the differences between the two data sets. Even though there is only a slight difference, you can still say, yes, they are significantly different. If you still remember some time ago, we had a discussion on large sample size problem for tests for normality. When the sample size is large enough, the result can always be expected (for real data sets), that is, rejection of the null hypothesis. Cheers, Chaosheng -- Dr. Chaosheng Zhang Lecturer in GIS Department of Geography National University of Ireland, Galway IRELAND Tel: +353-91-524411 x 2375 Direct Tel: +353-91-49 2375 Fax: +353-91-525700 E-mail: [EMAIL PROTECTED] Web 1: www.nuigalway.ie/geography/zhang.html Web 2: www.nuigalway.ie/geography/gis/index.htm
Re: [ai-geostats] How can I assign weights based on measurement errors?
Hi Wolfram, You forgot to mention what you want to do with these data. If the objective is to perform kriging, then you can use either kriging with nonsystematic errors or soft indicator kriging to account for the variable level of reliability of your data. Cheers, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Mon, 29 Nov 2004, Wolfram Ruehaak wrote: Dear all, I have difficulties to find information's about the following problem. I have a lot of spatially scattered measurements. These measurements have - resulting from different measurement methods - different measurement errors, which are known. For example some have an total error of 5%, some of 10% and a third group of 20%. I want to give these values a quality-weight in the range from 0.0 to 1.0. (In this case three different weights.) How can I do this? Simple is a weight = 0 which is a value so bad I don't want to use it, and a weight = 1 which could be the value for the group with the best measurements (in this case error = 5%). Is there a statistically firmed way to quantify the weights. Any suggestions will be very welcome. Is there any literature that discusses this matter? Thanks in advance. Wolfram Ruehaak * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
Re: [ai-geostats] regularization
Hi Samuel, I have dealt with similar problems when analyzing the spatial distribution of dioxin and other heavy metals in river sediments. Core lengths can strongly fluctuate from one sampling point to the next. The empirical approach I used was to weigh each sample proportionally to its length both in the computation of semivariograms (use of weighted semivariogram estimators) and in the kriging procedure (rescaling of kriging weights to account for core length). There was no publication on this approach and reports are confidential. These days I would use a less empirical approach and capitalize on the analogy with the treatment of cancer rates, where the reliability of rates is a function of the population size. You could still use weighted semivariogram estimator, but use a kriging with measurement error approach, whereby an error variance term (here inversely proportional to the length of the core) is added to the diagonal elemnts of the kriging matrix. Here is just a suggestion but I am sure that some mining geostaticians will come up with a more elegant solution.. I also think that Jayme Gomez presented a paper on this issue (and the downscaling or disaggregation problem in general) at the last geostat congress in Banff, but since I only caught the last part of his presentation I might be wrong. Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Tue, 26 Oct 2004, samuel verstraete wrote: Hi, I have a 3D data set that has been sampled by a private company. They lacked a complete knowledge of geostatistics so there is no sampling strategy involved. Another thing is that the support of the samples is strongly fluctuating. Horizontally the sampling support is constant and can be considered as a point (about 70cm^2 compared to a few hectares) Vertically the sampling support is not stable and rather huge in comparison with the vertical scale... (sampling can be 0.10 to 1 meter and maximum depth would be 5 to 6 meter or even less) I've read in the literature that there is a possibility to correct for such a things, through regularization. But none of the literature seems to discuss the possibility that the samples themself do not always have the same support, as stated before samples can have a support that is 10 times bigger than the smallest sample. Question is... Is there any other literature that discusses this matter and even more importantly is there any software out there that can take this sampling support into consideration when I'm calculating the variogram or when I start with estimation/simulation of the field. Thanks in advance, -- Samuel Verstraete Ghent University Faculty of Bioscience Engineering Dept. of Soil Management and Soil Care Coupure Links 653, B-9000 Gent, Belgium * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
Re: [ai-geostats] geometric designs for sensor networks
Hi Andrew, A critical piece of information is the objective function you want to optimize. I have worked in the field of optimization of sampling design applied to mechanical engineering and ergonomics, see papers by Sasena et al. that you can download from my webpage. I believe it has great application in the field of design of network of environmental sensors and I am open to collaborations on this topic. Regards, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Thu, 21 Oct 2004, Andrew Baek wrote: Greetings! I am looking for an efficient geometric design for installing bunch of robots in a field. These sensors(or sensor networks) are not static, in the sense that they constantly move and sample data(light, temperature, humidity etc.). Basically, I'd like to find different schemes - not a square/triangle shape from Yfantis et al(1987)- for my problem. Also, this is different from space filling since order matters. I mean, the cost will be quite different whether robot starts from edge or center. My guess is that curve (roulette shape?) will be more favorable in this case, but I don't know... Can someone refer me to some references or suggestions? THX, Andrew * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
Re: [ai-geostats] Simulation maps for anisotropic models
Hi Sanghoon, Your maps look fine and would reflect a strong anisotropy. What is the anisotropy ratio for your variables and did you select a circular or ellptical search window? Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Fri, 10 Sep 2004, Sanghoon Kang wrote: Dear everyone, I'm new to this mailing list and the field of geostatistics, so my question might be too obvious, but I don't have anyone around to get some help. I'm trying to analyze potential factors influencing soil microbes. For those soil characteristics, some of them were anisotropic and some not. I modeled and ran sgsim for stochastic simulation maps to get qualitative comparison among variables. It looked fine for those isotropic variables but there were lines in anistropic variables. Those lines aligned along the major axes, so I guess they might be generated from search radii. However, maps I've seen didn't have those features in them and I'm not comfortable with that. I would like to get some opinions and if possible solution to get rid of them. You can see the maps by clicking the link. http://www.people.virginia.edu/~sk7k/Material/maps.pdf Any general advice or guidelines would be helpful as well. Thanks. - Sanghoon Kang Lab. Microbial Ecology Dept. Environmental Sciences, UVa 434-924-0537 (T) 434-982-2137 (F) http://www.people.virginia.edu/~sk7k http://janicekang.net - My Inbox is protected by SPAMfighter 3995 spam mails have been blocked so far. Download free www.spamfighter.com today! * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
Re: [ai-geostats] FW: spatial relationships
I would agree with Gregoire's assessment. The presence of a global trend does not prohibit the use of geostatistics. As illustrated in the following paper by Journel and Rossi: Journel, A.G. and M.E. Rossi. 1989. When do we need a trend model in kriging? Mathematical Geology, 21(7):715--739. global trends can be easily handled by the use of local search windows in kriging, which allows us to rely on the assumption of quasi-stationarity. Of course if the trend is complex and can be described using process-based models (e.g. urban pollution), it is better to use this physical model for the trend and use geostatistics to interpolate the residuals, provided there is some spatial correlation left. Cheers, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Thu, 2 Sep 2004, Gregoire Dubois wrote: -Original Message- From: Gregoire Dubois [mailto:[EMAIL PROTECTED] Sent: 02 September 2004 09:42 To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: Re: spatial relationships Hi Mark, re-reading Isobel's mail, I thought about a proviso on the proviso. I personally do consider that a semivariogram showing a pure trend is decent. Not in a geostatistical point of view, but it does provide you with some useful information. If you have a trend, the variogram becomes incompatible with the intrinsic hypothesis. but you still have a slope in the experimental correlation functions (semivariograms, correlograms, madogram, etc.). Thus you have a structure, that is you have something there that may provide you with some useful information about your data set that can be used for estimating values of your variable at unsampled locations. If you have a flat correlation function, that is a pure nugget effect, then certainly you are in troubles. Regards, Gregoire Isobel Clark [EMAIL PROTECTED] wrote: Mark I could not agree more with Gregoire (with one proviso, see below). Both geostatistics and any weighted average estimators are based on the same assumptions -- that relationship between values at two locations depends on the distance between them and (possibly) their relative orientation. If you cannot get a decent semi-variogram after trying every type of graph [normal, robust, relative] and every transformation and/or interpretation of your data [logarithm, indicator, rank transforms, Normal scores, mixed populations], you do not have a distance-based relationship. This conclusion also rules out: inverse distance weighting of any kind; Delaunay triangles; Thiessen polygons and so on. My proviso: there are other forms of spatial relationship than pure distance/direction types. The simplest example of this is data with a trend, where the value at a specified point will depend on its absolute position. There may be an added component for the 'residuals' which turns out to be distance/direction based. There are also many examples where, for example, flow characteristics, connectivity and so on play a large part in the structure of your variable. In short: no decent semi-variogram does NOT mean no spatial relationship. It means no simple second-order stationary geostatistical type spatial relationship. Isobel http://geoecosse.bizland.com/whatsnew.htm http://geoecosse.bizland.com/whatsnew.htm ___ALL-NEW Yahoo! Messenger - all new features - even more fun! http://uk.messenger.yahoo.com http://uk.messenger.yahoo.com - Attachment: message-footer.txt MIME Type: text/plain - __ Gregoire Dubois (Ph.D.) JRC - European Commission IES - Emissions and Health Unit Radioactivity Environmental Monitoring group TP 441, Via Fermi 1 21020 Ispra (VA) ITALY Tel. +39 (0)332 78 6360 Fax. +39 (0)332 78 5466 Email: mailto:[EMAIL PROTECTED] [EMAIL PROTECTED] WWW: http://www.ai-geostats.org http://www.ai-geostats.org WWW: http://rem.jrc.cec.eu.int http://rem.jrc.cec.eu.int The views expressed are purely those of the writer and may not in any circumstances be regarded as stating an official position of the European Commission. * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
Re: [ai-geostats] kriging proportions
Hi Marc, You may want to look at the following paper: de Gruijter, J.J., Walvoort, D.J.J., van Gaans, P.F.M., 1997. Continuous soil maps --- a fuzzy set approach to bridge the gap between aggregation levels of process and distribution models. Geoderma 77, 169--195. The authors describe compositional kriging to interpolate class memberships, and they have incorporated additional constraints into the kriging system to ensure that all estimates are positive and add up to a constant (1 in this case). Cheers, Pierre Goovaerts Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Thu, 10 Jun 2004, Marc-Olivier Gasser wrote: Hi everyone, Lets say we have measured three soil particule size values for clay, silt and sand, all adding to one. cl + s i + sa = 1 What would be the best way to take into account each particule size, so the interpolated values still add up to one? Is there any geostatistical process that can handle this? I have tried interpolating parameters caracterising different particle size distribution functions (in the case where there are more than 10 particule sizes) but this adds errors to the modelling and some parameters don't necessarily exhibit spatial correlation. Maximum autocorrelation factor kriging has been suggested such as in: A. J. Desbarats and R. Dimitrakopoulos. Geostatistical Simulation of Regionalized Pore-Size Distributions Using Min/Max Autocorrelation Factors. Mathematical Geology, Vol. 32, No. 8, 2000. but I haven't found many statistical packages implementig this procedure. What other possibilities are there? Best regards, Marc-Olivier * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
Re: AI-GEOSTATS: Automated Variogram modelling
Hello, The issue of automatic versus manual modeling of semivariogram has been the subject of much debate in the past. In my graduate class, I used to ask the students to model their experimental semivariograms first manually (i.e. bye eye), then using non-linear regression. The resulting models were then used in kriging and cross-validation allowed them to assess the prediction performances of both types of models. Most were surprised to find out that manually fitted semivariograms could lead to more accurate predictions than automatically fitted ones. The take-home lesson was that the modeling of the semivariogram is usually a preliminary step towards prediction or simulation, and influence partially their results. Automatic semivariogram modeling is useful to model complex anisotropies as long as the experimental semivariograms are reasonably well defined and also when multiple semivariograms need to be modeled (i.e. indicator kriging). In addition, working now for a software RD company and developing new applications of geostatistics to health science, I have to keep in mind that most users migth not have the necessary background to compute and model semivariograms. The challenge is then to find a procedure to achieve meaningful fits without asking much from the user... The issue of automatic versus manual modeling is particularly important when data are sparse, making the semivariogram erratic... Then the modeling procedure is more than a mere exercice of fitting a curve to experimental values. It aims at creating a model for the spatial variability of the phenomenon under study and it relies greatly on ancillary information (e.g. magnitude of nugget effect, directions of anisotropy) typically provided by expert knowledge. Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Mon, 5 Apr 2004 [EMAIL PROTECTED] wrote: Hi all, I have a couple of questions for the list. I understand that most theoretical variograms are fit by eye, and I was interested in gauging the usefulness of automated (purely data-driven) estimation for theoretical variograms. i.e. Would it be useful to practitioners to be able to fit to be able to fit something like a 'constrained spline' as the theoretical variogram function to give your kriging results? (the spline could be constrained to be positive-semi-definite) 1. Is this something that has been examined in detail in the past? 2. If not - would it be something that geostatisticians would find useful? Any thoughts and references on this matter would be most welcome. Many thanks in advance, Matthew Pawley -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
RE: AI-GEOSTATS: testing sample number
Hi Mark, For each number of samples you wish to eliminate, you really need to repeat the sampling many times in order to account for sampling fluctuations in the assessment of prediction performances. The procedure need to be automated and you won't avoid having to modify a program (i.e. Gslib kt3d whose source code is distributed freely) to implement the hundreds or thousands of run needed for a thorough analysis. Just as an illustration of the kind of fluctuations you can expect if you select randomly 100 subsets of the same size, look at the following paper that can be downloaded from my webpage: Saito, H. and P. Goovaerts. 2000. Geostatistical interpolation of positively skewed and censored data in a dioxin contaminated site. Environmental Science Technology, vol.34, No.19: 4228-4235. Regards, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Wed, 31 Mar 2004, Mark Dowdall wrote: Hello This is a newbie question but I have been all over the Faqs and cannot find an answer. Any help is appreciated and a summary of answers will be posted. I have a set of data that was taken over a particular area. I have kriged and contoured and am happy with the results. But I need to demonstrate, within the bounds of the study/analysis/assumptions, that the number of samples taken was sufficient to describe the area. So my plan was to demonstrate that for a cetain number of samples the variance in the estimates had reached a value that could not be reduced efefctively by increasing sample number. And I thought this could be done by eliminating at random a point (so I have x-1 data points), kriging the remainder with the chosen parameters and checking the relevant parameters. (x is the number of actual samples) Then eliminating two points at random (x-2), repeating and so on. Eventually only 1 point being left. But.if I eliminate the first point at random, does it matter which point is eliminated? Or in the first instance (x-1 points) should I do the process for all possible samples and take an average of the estimation uncertainty? I am using GEOEAS and this could take a long time as all possible permutations of for example x-20 could be quite large. I did download Explostat which has a feature that sems to do this automatically but the manual is not great and the software is a little impenetrable. If anyone can shed light on this I would be most grateful. Thanks in advance M.dowdall Ie. -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: spatial structure
Hi Jose, You can certainly compute semivariograms in the two cases you mentioned in your email, although the interpretation of the graph is not always straigthforward when you compare indicators to continuous variables. Examples of indicator cross semivariograms for categorical and continuous variables can be found in: Goovaerts, P. 1994. Comparison of CoIK, IK and mIK performances for modeling conditional probabilities of categorical variables. In R. Dimitrakopoulos, editor, Geostatistics for the Next Century, pages 18-29. Kluwer, Dordrecht. Goovaerts, P. 1994. Comparative performance of indicator algorithms for modeling conditional probability distribution functions. Mathematical Geology, 26(3):389-411. I have also computed cross variogram between categorical indicators and continuous variables, although this work has never been published.. My plan was to use this semivariograms in cokriging and I looked also at ways to interpret the shape of these graphs. I believe you have should have mentioned the objectives of your analysis in your initial email, I bet you don't compute cross semivariograms for the pleasure... Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Fri, 26 Mar 2004, [ISO-8859-1] José Manuel Blanco Moreno wrote: Hi list, I hope my question is not too basic, but I've been searching in the literature and mail archives from ai-geostats I couldn't find any clear answer. The same way there is the cross-semivariogram to describe the spatial relationship between two continuous variables and the indicator semivariogram (semivariogram of indicator variables): -it is possible to use the cross-semivariogram for two indicator variables? -can be used to describe the spatial distribution of a continuous variable in relation to a indicator variable? -any reference on this subject? Thank you very much. -- José Manuel Blanco Moreno Ph.D Student --- José-Manuel Blanco-Moreno Dept. de Biologia Vegetal (Botànica) Universitat de Barcelona Av. Diagonal 645 08028 Barcelona SPAIN --- phone: (+34)93.402.1471 fax: (+34)93.411.2842 e-mail: [EMAIL PROTECTED] -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: mysterious kriging output
Hello, I agree that in many environmental datasets we could question the assumption of existence of a single population. Although there are ways to split the data into several populations, the key issue is that the study area needs also to be stratified into several populations. In some fields, such as geology, geological maps could provide a stratification of the study area and helps delineating the boundaries between populations. This is far less obvious for environmental data sets. Looking at Noemi's maps, I would agree with Richard's comment that nothing seems to be out of the ordinary. Of course, when dealing with streams the data configuration is far from optimal and screening effects abound. Also, the strong anisotropy ratio means that we deal with a zonal-like anisotopy which might cause sudden changes of covariance for slight difference of angles. In particular, this covariance model could lead to very small correlations off the two main axes of anisotropy, which could explain the larger kriging variance observed along the diagonal directions. Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Tue, 9 Mar 2004, Monica Palaseanu-Lovejoy wrote: Hi, I am working myself with pollution data in soils and i have very high values very close to very low values, and highly skewed distribution. I am more and more concerned with doing kriging on transformed data. This simply means we believe the data came from only one population. But what if it comes from 2 different populations representing 2 different polluting processes? Much more if we do believe there are no gross error measurements. The fact that high values are very close to low values would tell me that the spatial autocorrelation is violated locally. I would try first to see if the outliers (local and global) represent a different population, if these values cluster or not, how significant is the association high- low values, and if the global Moran's I increases if i eliminate the outliers. Maybe the majority of the data which have a higher spatial autocorrelation belong to a better expressed diffusive process, (maybe an older one) while the rest of the data which were identified as outliers before, represent a more patch-y or point source pollution process which didn't have time to diffuse over the entire study area (a younger process, maybe?). Of course if you have proof that the data came from only one population then it is a different story. I will really appreciate to hear other opinions about these thoughts. Thanks, Monica -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: Microgeostatistics
Steven, There have been several papers published about the application of geostatistics to microbiology (e.g. Rossi, Robertson, or Webster), and several of them are listed in the review paper that you can download from my webpage (http://www-personal.engin.umich.edu/~goovaert/publication.html): Goovaerts, P. 1998. Geostatistical tools for characterizing the spatial variability of microbiological and physico-chemical soil properties. Biology and Fertility of Soils, 27(4): 315-334. I also remember a study where geostatistics has been used to study the spatial distribution of bugs on the roots of a plant. Of course, multiple studies have dealt with the study of miscroscopic imagery, including the one I presented at a recent conference on analysis software for microscopy imagery (see pdf slides at http://www-personal.engin.umich.edu/~goovaert/publication.html) Hope it helps, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Sun, 7 Mar 2004, Steven Rogers wrote: Hi list members, I'm working in microbial ecology and am interested in finding information about spatial analysis or interpolation at extremely small scales (millimeters, microns, etc.)., especially as related to survival and fitness of microorganisms. I've coined the term, microgeostatistics to describe this subject. I could be wrong, but it seems like there is comparatively little information on this general area. Any suggestions are greatly appreciated! Steven Rogers Ecostat, Inc. *** Steve Rogers [vhjri.gif] -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: polygon kriging
Dear Lorenz, The easiest way to proceed would be to discretize your polygon using a grid, use stochastic simulation to simulate values within the polygon, and estimate the block value as the arithmetical average of these simulated point values. The advantage is that you can empirically estimate the variance of your block estimate from the distribution of simulated block values. It would be fairly easy to implement this procedure using gslib simulation routines. You just need to make sure that only the grid nodes within the boundaries of the polygon are being simulated (you can not just simulate a rectangular grid and a posteriori eliminate the values simulated outside the limits). This requires a if statement inside sgsim or sisim to skip any grid node falling outside the boundarie (just define a grid of indicator values that would be one if inside and zero if outside the limits of the polygon, and read this grid at the beginning of the simulation). Best regards, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Fri, 20 Feb 2004, Lorenz Dobler wrote: hello list, i would like to assign the principles of block-kriging to user defined (=irregular) polygons. some questions about it: 1. what do you think about this idea ? 2. is there any software that can do this allready ? 3. does anyone of you have practical experiences with this approach ? 4. is there a realistic possibility to change existing gslib libraries to do kriging with irregular polygons instead of regular blocks (problem: finding center [of gravity] for each irregular polygon and defining appropriate search neigbourhood !!) ? hope someone can help kind regards Lenz Dobler Universität Münster Institut für Geoinformatik Robert-Koch-Strasse 26-28 D-48149 Münster Tel.: 0251/83-30089 Fax: 0251/83-39763 email: [EMAIL PROTECTED] -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: Sequential simulation
Hi, Your question is puzzling. First, you should clarify what you mean by zero inflated distribution. Second, a normal score transform should by construction always yield a set of normal scores with zero mean, hence it is unclear why you get an average of -1. Note that if your variable takes only a few values or if there is a large proportion of data below the detection limit or equal to zero, a normal score transform that artificially despikes these similar values might not be the most appropriate approach. Pierre Goovaerts Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Mon, 16 Feb 2004, [ISO-8859-1] José Manuel Blanco Moreno wrote: Hi, list, I'm working on simulation (again) with sgsim (gslib) and I've found something that troubles me... I'm trying to simulate a variable that has a zero inflated distribution (weed counts), and I proceed as follows: -calculate the normal scores variogram for that variable. -fit the model (the total sill sums slightly more than one) -especify the details for simulation ... and run it. And then comes the problem: in the debugging file appears a rather pretty question asking if the variance is near one (when it is clear it is not; my variance is about TWO!) and if the mean is near zero (when, again, it is not near zero, but near minus one). I suppose that this is generated by that zero inflated distribution. Is that true? Another question; Could this affect the validity of simulation? Should I proceed in another way (v.g. Indicator simulation)? Thank you for any light shed on this question. -- --- José Manuel Blanco Moreno Dept. de Biologia Vegetal (Botànica) Universitat de Barcelona Av. Diagonal 645 08028 Barcelona SPAIN --- phone: (+34)93.402.1471 fax: (+34)93.411.2842 e-mail: [EMAIL PROTECTED] -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: working with gslib90
Hello, In order to see error messages when running Gslib programs, you should open a command prompt window and run the program from there (just select the icon of the program and drag it to the prompt window... this window won't close once the program has run). Note that if Gslib program does not find a parameter file in the directory where you are running the program, it will automatically create a default one that you should modify before running the program again. Happy holidays! Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Tue, 30 Dec 2003, [iso-8859-1] snehamoy chatterjee wrote: hi! i want to work with gslib90. i have downloaded the software. it has 13 exe files. when i want to work with any one of those files it wants parameter files. and when i am giving the name of the parameter file the window is disappeared. what is the problem with me? i have not able to proceed furthar. please suggest snehamoy = snehamoy chatterjee dept. of mining engg. iit kharagpur midnapur-721302 Yahoo! India Mobile: Download the latest polyphonic ringtones. Go to http://in.mobile.yahoo.com -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: Detecting spatial autocorrelation in highly non normal data
Hello, I agree that in most situations testing for spatial correlation is not very informative since the null hypothesis of spatial independence is unrealistici and its rejection is trivial. This is why at Biomedware we are working on tests of hypothesis where the null hypothesis is a particular spatial pattern. Stochastic simulation allows one to generate many realizations of this spatial pattern that are then used to derive the distribution of the test statistics. More information can be found in the following publication: http://www-personal.engin.umich.edu/~goovaert/liebisch.pdf Cheers, Pierre Goovaerts Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Thu, 20 Nov 2003, Volker Bahn wrote: The problem is not unique to testing significance of spatial correlation. Any traditional hypothesis test is non-sensical as you describe because we ALWAYS know that the null hypothesis is wrong. The question of interest is how wrong it is and whether the detected effect is of practical consequence. However, a hypothesis test does not test practical relevance of an effect but statistical power to detect it, which depends on sample size, inherent variability and also on effect size. Given a large enough sample size, one will always reject the null hypothesis. Why then do people still hold on to hypothesis tests? Because it gives them a false sense of objectivity. No one wants to admit that the judgement of whether an effect is of practical consequence is a to a certain degree inherently subjective decision (as is the level of alpha etc). Cheers Volker - Original Message - From: Edzer J. Pebesma [EMAIL PROTECTED] To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Sent: Thursday, November 20, 2003 5:11 Subject: Re: AI-GEOSTATS: Detecting spatial autocorrelation in highly non normal data | Trevor, | | I always wonder what the value of testing significance of spatial | correlation is, and never advise to do it. See, if data are spatial, it | is extremely unlikely that they are spatially uncorrelated. Rejecting | the test is usually only a matter of collecting sufficient evidence, | and not at all an interesting finding, because the data were spatial. | | Probably a more real problem is: to what extent does the spatial | correlation present (which may be very weak!) mess up an analysis | that assumes independence of observations. If you choose for | an analysis method that addresses spatial correlation, you're always | on safe ground. | | If your data were collected using some form of random sampling, | analysis based on independent observations is perfectly valid for | estimating areal mean values. This does not imply that data are | spatially uncorrelated, but just that they may be treated independent | because of the sampling scheme. | -- | Edzer | | [EMAIL PROTECTED] wrote: | | Hi Folks, | | I'm hoping someone can help steer me in the right direction. | | I have several sets of data acquired from acoustic surveys conducted on a | small lake trout lake. The data consist of sampling units aligned in | transects. Each sampling unit is 50 m in length. A mean lake trout density | is associated with each sampling unit. | | I'm interested in examining whether any significant spatial autocorrelation | in the exists in the observed distribution of lake trout. The data are | highly non-normal with 75-90% of the observations being zero. Log and Ln | transformations do not normalize the data. | | I've been doing some reading and it seems that most methods of quantifying | spatial autocorrelation require some kind of normality in the data. Any | suggestions on how I might proceed with these data? | | Thanks in advance for all suggestions. | | Apologies for the simplicity of the question, but I'm just beginning my | foray into spatial statistics. | | Trevor Middel | | -- | * To post a message to the list, send it to [EMAIL PROTECTED] | * As a general service to the users, please remember to post a summary of any useful responses to your questions. | * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list | * Support to the list is provided at http://www.ai-geostats.org | | | | | -- | * To post a message to the list, send it to [EMAIL PROTECTED] | * As a general service to the users, please remember to post a summary of any useful responses to your questions. | * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests
Re: AI-GEOSTATS: Information about Kriging 3D and semivariogram in 3D
Hi Lisa, I would recommend that you look at Chapter 16 of Isaaks and Srivastava's book An introduction to Applied Geostatistics which offers a very nice description of how to model semivariograms in different directions and up to 3 dimensions. You can also refer to the following paper: Barabas, N., Goovaerts, P. and P. Adriaens. 2001. Geostatistical assessment and validation of uncertainty for three-dimensional dioxin data from sediments in an estuarine river. Environmental Science Technology, 35(16): 3294-3301. Best regards, Pierre Goovaerts Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Thu, 6 Nov 2003, [iso-8859-1] lisa pizzol wrote: Hello to everybody, I have to build a semivariogram in 3D and I need to know if I can use the same concepts used for the 2D semivariogram or if the theory is different. For example how can I consider the distance? In one article they said that I can first build the semivariogram in the horizontal direction in order to find the values of sill, range, nugget and the equation of the semivariogram. After this I can use these information to find the range in the vertical direction. Do you know any references that explain exactly how to do it? Thank you very much. Lisa - Yahoo! Mail: 6MB di spazio gratuito, 30MB per i tuoi allegati, l'antivirus, il filtro Anti-spam -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: Back transforms and simulations
Hi Chris, The back transform of simulated values is very easy to perform. Just take the exponential of the simulated values since you are not trying to estimate the mean of the local probability distribution in the original space, but only a quantile of this distribution. Note that if you perform SGS using Gslib, there is a built-in normal score transform and back-transform in the program, which is more flexible than the lognormal transform. Cheers, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Fri, 17 Oct 2003, Chris Lloyd wrote: Hello, The subject of logs and back transforms has been discussed a great deal on the list and I've seen much material concerning back transforms following kriging of log transformed data (e.g., the approach outlined by Cressie in his book 'Statistics for Spatial Data' and many other texts). However, I am unsure how to proceed if the objective is simulation. I have applied sequential Gaussian simulation to log (base 10) permeability data and I want to back transform the simulated realisations. I would be grateful for any suggests from list members as to how best to back transform the values in this case. There are too few data to make an indicator approach feasible. I will post a summary of answers. Many thanks in advance. Chris Lloyd -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: negative weight
Hi Laure, One of the main culprits for negative cokriging weights is the ordinary cokriging constraint that the sum of the secondary weights must be zero. I found that the frequency and magnitude of negative cokriging weights greatly decreased when using a single constraint that primary and secondary data weights must sum to one (referred to as standardized cokriging in Gslib software). This issue is discussed in my book and in the following publication: Goovaerts, P. 1998. Ordinary cokriging revisited. Mathematical Geology, 30(1): 21-42. Cheers, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Thu, 4 Sep 2003 [EMAIL PROTECTED] wrote: Hi Mailinglist!!! I have got some problems with negative weight in my cokriging, they induce negative grade. Have you got some advice or publication to help me.I hope you can help me with some of your answers. Thank Laure Laure FONTAINE Services des Réserves COGEMA BUM/DT tél: 33 1 39 26 32 05 Fax: 33 1 39 26 27 31 -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: Variograms fractals
Dear Gregoire, Although I am still living in the Northern hemisphere and today is my wedding anniversary, I will answer your question..:) In fact I have recently reviewed a paper dealing with 2-D fractal analysis and I found the following reference to be of great interest: Butler et al. (2001) Characterization of the structure of river-bed gravels using two-dimensional fractal analysis. Mathematical Geology, 33(3): 301-330. If the variability is isotropic, you can indeed derive the 2D fractal dimension by adding 1 to the dimension estimated from the omnidirectional semivariogram. In presence of anisotropy, the authors present 2 different approaches: 1. Estimate the fractal dimensions from directional semivariograms, and by analogy with the rose diagram of ranges they built rose diagrams of fractal dimensions. 2. Construct variogram maps/surfaces and estimate fractal dimensions from semivariance profiles sampled along specific directions. Again this paper is nicely written and discusses methodological issues related to 2-dimensional fractal analysis. Cheers, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Sun, 27 Jul 2003, Gregoire Dubois wrote: Good day to the Southern hemisphere ! I presume many are on holiday in the northern hemisphere, hence I hope to get more feedback from the South :) In a paper published in Nature (Nature, 1981, Vol. 294, pp. 240-242: Fractal dimensions of landscapes and other environmental data), Peter Burrough investigates the fractal dimension of various environmental data by mean of the slope of the log-log plot of the semivariogram. As a result, Burrough gets for each variable investigated a fractal dimension that is fluctuating between 1 2, as it is the case for 1 dimension. The author suggests the use of D to as guide for further mapping and interpolation. Burrough estimated D assuming that the real data are but a series of regularly spaced samples of the Weierstrass-Mandlebrot function over one-dimensional space or time My questions are the following ones: what are the pratical consequences in making the above cited main assumption, that is using spatial data distributed in 2 dimensions and to consider the variogram as if the data were sampled in one dimension, like in a transect? Can one reasonably extrapolate (that is adding +1 to the fractal dimension obtained above from the log-log plot of the semivariogram) the fractal dimension in a pseudo 1-dimension to a 2-dimensional problem if the investigated phenomenon does not show any anistropy ? Thanks for any feedback. I will summarise useful replies references. Gregoire -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: A question about SASIM
Hello, First you have to see what is the perturbation mechanism used in the program: random swapping or random sampling of target histogram. In the first case, the initial image typically reproduces the target histogram and so it won't be affected by the pertubations. In the second case (which I believe is implemented in your version of sasim), poor reproduction of target histogram might indicate the need for increasing the relative weight assigned to this objective function component. Note that with only 42 observations both target histogram and semivariogram might not be highly reliable and deviations from these targets should be allowed. Regards, Pierre Goovaerts Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Tue, 15 Jul 2003, [big5] ¦¨¥J wrote: Hello all members. I have a question about the subroutine SASIM in Gislib. I have 42 sample data and analysis it's spatial correction structure (variogram) , then i use the information of it's histogram , variogram and indicator variogram into SASIM to generate several realizations. But the histogram of these realizations can't honor original sample data. Is any thing wrong in the procedure and is there any special techincque to adjust the parameter of annealing schedule ? Thank you for reading this message. Best Regards. Fan Cheng Cheng. National Chengkung University in Taiwan. E-mail: [EMAIL PROTECTED] -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: global vs local ordinary kriging
Hi Ulrich, It's not an easy question. First note that the search strategy includes not only the size of the search window but also the maximum number of observations. In may occasions, I set the search radius to a very large distance and use the number of observations as the controling parameter. Using too many observations or too large search windows may lead to oversmoothing, while estimates based on low number of observations (say less than 8 in 2D) might not be very reliable. Of course it depends also on the relative nugget effect. If it is large, even further away observations will receive a significant weight. In practice, global search windows are seldom used because: (1) no reliable semivariogram values are available for so large distances, (2) the size of the kriging system is likely very large, and (3) the stationarity assumption within the search window might become questionable. The best way to proceed would be to do some cross validation using various search strategies and investigate their impact on re-estimation scores. Regards, Pierre Goovaerts Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On 8 Jul 2003, Ulrich Leopold wrote: Dear list, What would you consider the most reliable ordinary kriging estimate? To use a local search neighbourhood (slightly bigger than the effective range) or set to global to include *all* data locations? Ulrich -- __ Ulrich Leopold MSc. Department of Physical Geography Institute for Biodiversity and Ecosystem Dynamics Faculty of Science University of Amsterdam Nieuwe Achtergracht 166 NL-1018WV Amsterdam Phone: +31-(0)20-525-7456 (7451 Secretary) Fax: +31-(0)20-525-7431 Email: [EMAIL PROTECTED] http://www.frw.uva.nl/soil/Welcome.html Check us also out at: Netherlands Centre for Geo-ecological Research http://www.frw.uva.nl/icg -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: Declustering
Hi Oliver, I don't know which algorithm you are using to compute these declustering weights but there is something wrong in your rescaling procedure. These declustering weights are proportional to the size of the polygon of influence of each observation and they can not be negative. Regards, Pierre Goovaerts Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Sat, 5 Jul 2003 [EMAIL PROTECTED] wrote: hello list, to get the global mean of my data set (484 observation wells) and as a prerequisite for Normal Score Transformation with GSLIB, i declustered my data with polygonal declustering. When standardizing the weights to 1, so that the weights sum up to the number of data, I receive some negative weights. this results in negative values for the observations (nitrate concentration). how do i have to interpret this ? does this require a special treatment? skip them? as I mentioned before, i want to normal score transform the data set prior to simple kriging. many many thanx for some help in advance. Regards, Oliver -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: SA using GSLIB
Dear Uwe, The interpretation of the nlag parameter in sasim can be somewhat confusing. It corresponds to the number of lags in all directions to be incorporated in the objective functions, not the class of distance for the semivariogram model. For example, to reproduce the variogram for the 1st lag distance of 0.5 km in all directions, you need to specify nlags=8, which corresponds to the number of cells in the direct neighborhood of the cell being simulated. For reproducing the 2nd class of distance of 1km, nlags should be 24. So the formula to use to reproduce K classes of distances would be nlag=[(2K+1)^2]-1. So if you want to reproduce the variogram up to a distance of 20 kms using a grid size of 0.5km nlag=[(2x40+1)^2]-1=6560! I have recently looked at ways to sample this set of lags since it is clearly impossible to afford such a large number of lags in the objective function. For example, only a random subset of these 6,560 possible lags would be reproduced. This is a limitation of the simulated annealing algorithm (would be even worst in 3D) and any other suggestions would be much appreciated. Regards, Pierre Goovaerts Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Fri, 13 Jun 2003, Uwe Haberlandt wrote: Dear all, I have a question concerning the simulated annealing module within GSLIB (f77v2 and f90). I'm using kilometers as units with grid spacing of 0.5 km and wanted a reproduction of the variogram up to a distance of 20 km. The range is 10 km. However for different nlags set in the parameter file (e.g. 20, 40, 80,120) the debugging file reports variogram values (actual and model) much lower than expected. In the debugging file the model variogram does not reach the sill not even with nlags=120. How is nlags interpreted and what is reported in the debug file? Any help is greatly appreciated Regards Uwe ** Dr. Uwe Haberlandt Ruhr University Bochum Institute for Hydrology, Water Management and Environmental Engineering Universitätsstraße 150 44780 Bochum Germany Tel.: +49 (0)234-32-27619 Fax.: +49 (0)234-32-14153 e-mail: [EMAIL PROTECTED] www: http://www.ruhr-uni-bochum.de/hydrology/ ** -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: Simulation and data
Hi Joe, After the backtransform, the distribution of simulated values at each node is not Gaussian anymore and its variance can be used as a local index of uncertainty, which accounts for both the range of surrounding values and their closeness in terms of data configuration. Regards, Pierre Goovaerts Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Mon, 9 Jun 2003, Joe Geo wrote: Dear ai-geostats A question about simulation I understand in the simulation methods SGS that each node the data a kriging is performed than an value is drawn from a normal distribution with the local mean (from data and previously simulated nodes captured in a search neigbourhood) and variance defined by the kriging variance. My understanding of kriging variance is that this variance is really an index of data configuration independent of the data values. This leads to my question. How a set of multiple simulations capture information about the variability of conditioning data values when the kriging variance is only a data location index. Specifically, it is possible to have the same conditioning data configuration but the variability of the values attached to the data can be quite different. How is this recognised in the simulation process. Thanks Joe _ Hotmail is now available on Australian mobile phones. Go to http://ninemsn.com.au/mobilecentral/signup.asp -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: Raingauge network design
Hello, You may want to visit Jan-Willem van Groenigen webpage agronomy.ucdavis.edu/groenigen/ where you can download Sanos software that implements the constrained optimization of spatial sampling he developed in his Ph.D. Cheers, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Thu, 3 Apr 2003, M.J. Abedini wrote: Dear Colleagues I checked the simulated annealing of GSLIB to see if it is ready to be used for the purpose of raingauge network design. I realized that some further coupling has to be done in this regard. I was wondering if there is any SA software which has already been tailored for this purpose. Your comments with regard to contact person, web site and ... is greatly appreciated. With best wishes MJA -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: factorial cokriging in 2 steps??
Dear Jean-Philippe, Your question is somewhat puzzling. In fact, in all rigor factorial kriging is already a kind of cokriging since you are estimating a variable (i.e. spatial component) from another variable (raw measurements). Are you trying to estimate regionalized factors, that is perform multivariate factorial kriging, or do you want to filter your seimic data and use it as secondary information to estimate another variable? I used a similar approach in: Goovaerts, P. 1999. Accounting for scale-dependent correlation in the spatial prediction of soil properties. In A. Soares, J. Gomez-Hernandez, and R. Froidevaux, editors, geoENV II - Geostatistics for Environmental Applications. Kluwer Academic Publishers, Dordrecht, pages 405-416. Regards, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Mon, 31 Mar 2003, [iso-8859-1] Jean-Philippe Goyen wrote: hello ai-geostatisticians, I am trying to program factorial co-kriging for filtering of seismic amplitudes. But I didn't find anywhere the algorithm or even matrix formulation of the system to solve. As I am more a computer-scientist than a geostatistician, I am not sure of setting the rigth system to solve. So I was just wondering if it would be correct to perform factorial kriging of my two sets of data (as I could obtain decomposed variograms), and then cokrige the results ? (hope my question has a sense). Thank you very much for any advice or reference. Jean-Philippe Goyen - Do You Yahoo!? -- Une adresse @yahoo.fr gratuite et en français ! Testez le nouveau Yahoo! Mail -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: autocorrelation indicators
Hi Christophe, I might have misunderstood your question but since the semivariogram is a measure of dissimilarity while the Moran I is a measure of similarity, I would expect that they vary in opposite ways. The issue is whether you want to test or not whether the difference in correlation is significant, and for this the Moran I would be better suited since expressions exist to compute confidence intervals. Regards, Pierre Goovaerts Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Thu, 27 Mar 2003, Christophe Z Guilmoto wrote: Hello, I was wondering if someone could help me answer a basic question about the spatial autocorrelation (SAC) for a set of socioeconomic variables: is variable X more spatially correlated then variable Y for short distance lags?. I am using for that purpose the values of Moran index (MI) and the semivariogram (SV, expressed as percentage of overall variance) computed for variables X and Y. The sample is large enough (n500), but there is a drift (South-North) that I don't want to correct, as it is part of the phenomenon I want to examine (why this is so is a different matter altogether). My problem is that I sometimes find some kind of discrepancy between both SAC indicators. For example, MI values may be higher for variable X than for variable Y (meaning higher SAC for X), while for the same distance lag, SV is actually lower for Y than for X (meaning higher SAC for Y). I presume that non-stationarity of variables X and Y may be the cause for this. I would prefer to use the SV since its definition (as half the average squared difference between observations) makes it more convenient for comparison. Moreover, Moran's values are also at times greater than 1 for short distance lags (which is a bit embarrassing to explain to non-geostatisticians). However, MI is the most common used SAC indicator in social sciences because of its similarity to ordinary correlation coefficients. So , what is the best way to compare SAC across variables? Which index should I use in case the variable is not stationnary ? Your observations and suggestions are welcome. All further references to papers or other sources examining these measurement problems would also be helpful. Thanks CZG Christophe Z. Guilmoto Demographe, IRD CEIAS-EHESS 54, Boulevard Raspail 75006 Paris France Tél.: 06 67 19 87 10 ou 01 53 72 97 45 -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: border effect?
Dear Gregoire, Extrapolation is always hazardous. If the search window includes some of these positive values, you should expect that the kriging estimate is non-zero since more likely these values are assigned a non-zero kriging weights. Remember that even if observations are located beyond the range of correlation, they still receive non zero kriging weights in ordinary kriging. Hope it helps Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Fri, 21 Mar 2003, Gregoire Dubois wrote: Dear members, analysing a set of data, in which positive values, located in the middle of the investigated area, are surrounded by null values, I obtained non- zero values outside of the investigated area (what I called here a border effect). My variogram is a simple exponential model with no nugget effect (a spherical one generates the same border effect) and I can't find any reasonable explanation of this 'problem'. Any hints ? Thanks for any help, Best regards, Gregoire -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS:GSLIB question
Hi Dave, I have experienced similar problems. You are right that the computation of vertical semivariograms shouldn't be affected by azimuth in the horizontal dimension. If I remember well, you must make sure to specify non-zero bandwiths and angular tolerances. Let me know if it helps. Regards, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Fri, 7 Mar 2003, Dave Rennie wrote: Hi All: Is anyone out there aware of a problem that may exit with GSLIB'S gamv when generating semi-variograms in the vertical direction? I've noticed that the results kicked out by gamv vary dramatically depending on what azimuth you use. Theoretically all variograms oriented vertically should be the same regardless of the azimuth shouldn't they? DR -- === David W. Rennie, P.Eng., Roscoe Postle Associates Inc., Suite 2000, 1066 West Hastings Street, Vancouver, British Columbia CANADA V6E 3X2 Phone:1-604-601-8227 Fax: 1-604-669-3844 === -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: mean of residuals in simple kriging with means
Hi Lorenz, As long as the mean is known and assumed constant across the study area, you can apply simple kriging. There is no need for that mean to be zero. Cheers, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Mon, 24 Feb 2003, Dobler, Lorenz wrote: hello list, in simple kriging with means the mean of residuals is presumed to be zero. if i derive (varying) local means from (correct) regression analysis, residuals usually (should) satisfy the condition of zero mean. but if i use predefined/regional (varying) local means i have some problems because the mean of residual is far away from zero. is it absolutely necessecary in simple kriging with vaying local means that the mean of residuals is really zero ? if yes what can i do because my residuals are far awya from zero? i would be glad if someone could give some advice regards Lenz Lorenz Dobler Bayerisches Geologisches Landesamt (Geological Survey of Bavaria) Heßstrasss 128 D-80797 München Germany Tel.: 0049/(0)89/9214-2756 email: [EMAIL PROTECTED] http://www.geologie.bayern.de/ -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: How reliable are your kriging variances?
Hi Christopher, I believe you forgot a key assumption, homoscedasticity. In most situations this assumption is not realistic and we would like the kriging variance to somehow depend on the local variability of data. Rescaling globally the kriging variance to account for uncertainty about variogram model won't solve this problem. Your map might be globally more accurate but locally it will still fail to indicate where prediction errors might be larger. Regarding statistics to account for reliability of kriging variance, the key question is what do you want to do with that variance. If it's used to derive local probability distributions under the multiGaussian model, you can assess precision and accuracy of uncertainty models using cross-validation. I addressed this issue in the following paper: Goovaerts, P. 2001. Geostatistical modelling of uncertainty in soil science. Geoderma, 103: 3-26. and would be glad to send you a PDF copy of the paper if needed. Regards, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Wed, 19 Feb 2003, Chris Howden wrote: G'day all, I reckon we need to quantify the reliability of the kriging variance map. Because sometimes its going to be an accurate map, and other times its going to be way off the mark. Imagine the situation when there are two maps with similar kriging variances. However when we look at the semivariagram fit one of them closely follow the line of fit while the other has a much larger scatter. This means that one of the maps is actually much more accurate then the other. But as maps are currently presented we would never know!! Could this be a big problem? I think it could. Particularly when the estimation is quite bad, meaning that the variances have been underestimated and should likely be much larger. One solution could be to make the kriging variances proportional to the model fit. Maybe the error between the kriging variance (as estimated using the semivariagram) and the estimation variance (using real data points) could be used to do this? Does anyone know if this has been discussed before? Has it ever been considered. Or am I totally off the trail and should activate my GIS beacon? For those that are interested I'll explain how I got to the above conclusion: Kriging can be summarised by the following: Var(est) = f(weights and semivariance between all points that have a positive weight), and we obtain the Var(krig) by minimising Var(est) with respect to the weights. This is how we get the weights. But in order to do this we need to know what the semivariance between the points is. However if we're estimating a point we don't have then we can't calculate the semi-variance, so we can't find the appropriate weights. However, if we have a model for the semi-variance then we can predict what the semi-variance should be using this model and we can then calculate the appropriate weights. Which is why we require a semivariagram model. So the semivariagram fit is vital in generating not only the estimates, but their reliability also. What this all boils down to is that the most important thing when kriging is the ASSUMPTION that the points used to generate the semi-variagram are capable of representing the semivariance for all points. As well as the ASSUMPTION that the correct model has been fit, and that its a good fit. If either of these assumption fails then the kriging variance is incorrect. More to the point if the model is a poor fit then the kriging variance is less likely to be accurate. This brought me to my question. Should we have some statitistic that quantifies how reliable our kriging variances are? Christopher G Howden Statistical Ecologist Department of Land and Water Conservation (Work) 02 9895 7130 (Fax)02 9895 7867 (Mob) 0410 689 945 -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: weighted cross-validations?
Hi Gregoire, I don't remember of any prior discussion on that topic but I am getting old too..:)) If I understand your problem correctly, you want to weigh differently prediction errors obtained during cross validation. In other words you are more interested in the potential impact of the error than its magnitude, which sounds reasonable to me. The computation of this impact function is largely empirical and similar impact values could be obtained under different scenarios (e.g. large uncertainty at a few sensitive locations balancing small uncertainty at many non-critical locations, or small uncertainty at sensitive locations and large uncertainty at others). Everything will depend on the way you penalize uncertainty at the different locations and you can always map uncertainty values. The interesting question is also: how do you plan to assess uncertainty about the pollutant concentrations? Cheers, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Tue, 18 Feb 2003, Gregoire Dubois wrote: Dear all, cross-validation techniques can be used to evaluate the suitability of the parameters chosen for an estimation problem. Minimising a global error might be too restrictive and might not be suitable to solve some practical problems. In the case of a large-scale pollution problem, one would obviously try to minimise the uncertainty concerning the estimation of the pollutant+IBk-s concentration at sensitive locations (e.g. populated areas or fragile ecosystems). Therefore, information on population densities for example could be used as a conditioning criterion to define the locations where the estimates should be most accurate. Hence, a weighted cross-validation approach might be more suitable under such circumstances. This sounds nice in theory, in pratice it does not make much sense since one may end up with low local uncertainties but with an overall high uncertainty reflecting the inadequacy of the model chosen. Still, there might be a reasonable balance to be found between the global and local errors. I was wondering if some work has already been done on such a topic or is the above complete nonsense ? Thanks for any feedback, Gregoire PS: I think I posted something similar in the past but could not find anything in the archives and don't remember the discussion if there was any... I'm getting old -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS:
Dear Nicole, We discussed this problem in the following paper: Saito, H. and P. Goovaerts. 2000. Geostatistical interpolation of positively skewed and censored data in a dioxin contaminated site. Environmental Science Technology, vol.34, No.19: 4228-4235. We found that a straight back-transform leads to biased estimates and suggested to (1) discretize the Gaussian ccdf (say using 100 percentiles), (2) back-transform each of these percentiles and (3) derive the estimate as the arithmetical average of back-transformed percentiles. I can send you an electronic copy of this paper if you like. Cheers, Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Tue, 11 Feb 2003, Nicole Gerlach wrote: Question about normal score transformation with regard to kriging. I would like interpolate soil-parameters, which have been transformed by the normal score transformation before. What should I consider with regard to the backtransformation? Is a straight back-transform possible or is there any need for a bias correction like in lognormal-kriging? -- Nicole Gerlach Institute for Geoinformatics (IfGI) WWU Muenster -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: Kriging Error vs variance
Hi Russell, I am assuming you refer to kriging error variance. If your semivariogram is bounded and has a sill close to the sample variance, then the simple kriging estimate will automatically be the global mean when the kriging variance is the sample variance (that is when all observations are beyond the range of spatial correlation). Note that you might want to decluster your sample mean before using it as global mean. For ordinary kriging, the kriging variance would actually be greater than the sample variance because of the Lagrangian parameter. I don't think I would adopt a global/sample mean instead of the local mean provided by ordinary kriging even if the variance of the estimator is smaller. However, for kriging with a trend or universal kriging, I wouldn't trust too much the estimate obtained for large kriging variance since the extrapolated trend can be very unrealistic (e.g. negative concentration estimates). Pierre Dr. Pierre Goovaerts President of PGeostat, LLC Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Mon, 27 Jan 2003, Russell Barbour wrote: Dear List members, I am looking for a reference on interpretation of the Kriging error versus the sample variance. Am I correct in assumung that in any kriged interpolation where the Kriging error is greater than the sample varience then the sample mean would be a better estimate at that location? Thanks for your help Russell Barbour Ph.D. Research Associate in Applied Mathematics Vector Ecology Laboratory Yale School of Medicine 60 College St. Rm 600 New Haven CT. 06520 TEL: 203 785 3223 FAX 203 785 3604 email: [EMAIL PROTECTED] -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: Semivariograms
Hi Valquirio, This issue has been discussed in the past and you might find interesting information in the archives. I would suggest the use of indicator kriging to deal with the presence of a large number of zero values. Use zero as the 1st threshold and pick up a few other thresholds, then apply indicator kriging to derive the local distributions of probability the mean of which can be used for estimation. Note that in presence of a large proportion of zeros (say more than 90%) your indicator variogram will more likely still look erratic. Pierre Dr. Pierre Goovaerts Consultant in (Geo)statistics President of PGeostat, LLC and Senior Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Thu, 23 Jan 2003, [iso-8859-1] Valquiria Ferraz Quirino wrote: Dear All, I am working modeling the distribution of tree parameters (Basal area per ha and Number of trees per ha) for an adult strata of a tropical forest. The modeling is being done for (1) all species present in the area, and (2) the seven economically most important species in the area. The final intention is to use the range to improve the sampling technique (if possible less plots for and approximately same precision). As the area has been intensively explored, same species are just present in less than 10 plots (from a total of 357). For the plots were they are not present, I used the number zero to represent a measured 0 m²/ha of basal area (in the first case), or 0 trees/ha (in the second case) on the plot. My questions are: (1) How should I deal with these zeros while modeling the semivariogram? I am asking because I tried using them and the semivariograms look strange (small lags presenting sometimes higher semivariances than large lags). In this case, I also tried to interpolate (using kriging) for values between my plots. Cross validation (Jack knife) shows also an unsatisfatory result (line below the x axis). On my second try, I took the zeros out. The semivariogram looks much better. But the kriging is unsatisfatory estimating very high values for plots were there aren't trees of the studied species at all! Another problem is the number of observations that I used in this case: sometimes just 8. Can anyone give me a help? (2) Can anyone recommend literature that deals with the use of geostatistics to help the planning of number and location of sampling units in forests? Thank you very much!! Valquiria Forst-Ing. Valquiria Ferraz Quirino Kappler Straße 57, Zi. 2121, 79117 Freiburg i. Br. Deutschland Tel.: +49 761 6806-6204 - Busca Yahoo! O serviço de busca mais completo da Internet. O que você pensar o Yahoo! encontra. -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: cokriging and ked
Hello, It is my experience that, for a given data set, the impact of secondary information is usually more pronounced when using KED (kriging with external drift) or SKLM (simple kriging with local means) instead of cokriging. Several factors will control the relative influence of secondary information in cokriging, namely the correlation coefficient, sampling intensity, and relative nugget effect of primary versus secondary variables. As I showed in my book, if the primary variable has a much larger nugget effect than the secondary variable and the two are well correlated, the secondary data may screen the influence of primary data. Try to play with these parameters and see what would be the impact on the final map. Although a detailed map might appear more desirable or better at first glance, beware that the impact of your DEM can be overestimated by some techniques and you might end up getting better re-estimation results for the smooth map. KED could be performed using the program kt3d in Gslib. Hope it helps, Pierre Dr. Pierre Goovaerts Consultant in (Geo)statistics President of PGeostat, LLC and Senior Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Thu, 16 Jan 2003, Alvaro Silva wrote: Hello Thanks to all that give me some help on the cokriging (Tom, Donald, Susan and Tomislav). By reading Goovaerts (hello thanks for your course in Lisbon last November) paper on precipitation estimation including elevation, I notice that cokriging presents smooth surfaces, but I think this isn't always true. I have tested it for the temperature estimation on Madeira Island, and the maps show clearly a relation with altitude although for Portugal mainland, i could not achieve yet this detail. I have tested the Neural Networks with the same purpose and the resulting map for the annual mean air temperature is very good, it captures fine details and presents variability very well, also the r between observed and estimated values with an independent dataset was very good (0.98). When I decided to test also the cokriging to compare the results I was disapointed, because NN presents a much better map. Now i try to understand why the CK doesn't give results as good as I thought it could give. I also would like to test kriging with external drift, does anyone know where can I find a friendly and free software to do so, preferencially with a tutorial. Thanks once again and with my best regards, Álvaro -- José Álvaro Mendes Pimpão Alves Silva Geógrafo - Técnico de SIG Geographer - GIS Technician Departamento de Clima e Ambiente Atmosférico Climate Department Instituto de Meteorologia Portuguese Meteorological Institute Rua C do Aeroporto 1749 - 047 Lisboa Portugal Tel: (+351) 218483961 Fax: (+351) 218402370 Email: [EMAIL PROTECTED] -- -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: Rodograms for large-scale structures ? (Reposting)
Gregoire, I believe I borrowed the term large-scale structure from the 1992 Gslib user's manual, note that the rodogram is not an option in the lastest release of Gslib. The terminology might be misleading since we mean features on the horizontal semivariogram axis (like range) as opposed to relative nugget effect which is inferred from the vertical axis of the semivariogram. This comment is purely empirical and is not backed up by any theory. Also I wouldn't draw any conclusion from a statistics computed on 9 observations. Cheers, Pierre Dr. Pierre Goovaerts Consultant in (Geo)statistics President of PGeostat, LLC and Senior Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Thu, 16 Jan 2003, Gregoire Dubois wrote: Sorry for this re-posting, but a few signs have been improperly converted in my text editor. Here it is again. +++ To my question on the use of rodograms I got the two following references: Paul Harris gave me the following: Journel A. 1988. New distance measures: the route toward truly non-gaussian geostatistics, mathematical geology vol 20 no 4. Pierre Goovaerts mentioned at paper presented at the Geostat congress in Avignon, 1988. Srivastava and Parker. 1989.Robust measures of spatial continuity. in M. Armstrong, editor, Geostatistics, pages 295-308, Kluwer, Dordrecht. I will certainly have a look at these papers. On the basis of what I have found so far, I still have a question about the use of rodograms. In Pierre's book, (page 31), the following is discussed: influence of extreme values can be reduced by using lower values of the order of the variogram (2 = traditional semivariogram, 1= madogram, 1/2= rodogram). On page 86, it is further mentioned that relative rodograms and madograms provide information (range, anisotropy) on large-scale features. This last point is also mentioned in the manual of the GeostatOffice software Rodograms and madograms are useful for investigating large-scale structures, where data are usually rather rough. If p=2 we get the most traditional measure called variogram, it opens finer sides of data correlation and can fail for rough data where rodogram and madogram succeed. (see http://www.ibrae.ac.ru/+AH4-mkanev/eng/gsoffice/HELP/Appendix.html) My original question on rodogram came from a micro scale analysis of 9 points located very closely (measurements of radioactivity made on a grid with nodes separated by 12 cm). The fact that only the rodogram revealed a clear spatial structure (unless you do an audacious regression of the experimental semivariogram) might be an artefact due to the very few samples or a correct approach if it wouldn't contradict the theory claiming that rodograms are more useful for large-scale structures. Is this argumentation based on theory or on experience with environmental data? I would appreciate your feedback about this last point. Gregoire (Has everyone made a new-year resolution on contributing to AI-GEOSTATS ?? The activity of the mailing list has been boosted tremendously since the beginning of this year). -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: Indicato Kriging
Hello, You need to provide more information on your problem. For example, how many threshold values did you use for indicator kriging and did you observe large differences between indicator variogram models (e.g. anisotropy) for different thresholds. I am also assuming you are using the mean of the probability distribution obtained by IK (i.e. E-type estimate). Last how did you interpolate and extrapolate the upper and lower tails of your discrete distributions. Nevertheless I am not surprised by your findings. As long as you are using the same data in these least square interpolation techniques, you will get similar results. Regards, Pierre Goovaerts Dr. Pierre Goovaerts Consultant in (Geo)statistics President of PGeostat, LLC and Senior Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Sun, 12 Jan 2003, [iso-8859-1] Elmidio Estévez Cruz wrote: Dear All: I´m working with a Au gossan deposit. The data come from grade control of 8 benches in the mine. All the histograms show evidences of bimodality .The presence of a complex population is geologically explained by the fact that gold is hosted by different lithological units (limonite, sandstone, clay etc.), which are mixed and it is not possible to separate at present mining scale (5m benches). The variograms of Au show a clear spatial structure with a range of around 30m in all benches. Based on this spatial correlation some people have proposed to used OK to estimate the Au grade in 5x5m block on the other hand the mixture of population indicates the violation of stationarity (homogeneity) indispensable for the application of OK. According to literature the only method that mitigate this problems is IK. I have interpolated the grade using IK but I found a high correlation between OK and IK estimates. I also crossvalidated the methods and found similar errors. What could be the reason for that? Is there any method to demonstrate which interpolation techniques works better in this situation? Best regards, MSc. Emidio Estévez Cruz Dpto de Geología Universidad de Pinar del Río -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: Problem with Variowin 2.2 under Windows XP
Hi Ruben, I am also working on Windows XP and have repeated your experiment. I can read the file created from Excel (option formatted text, space delimited) with prevar2D and then run Vario2D without any problem. Pierre Dr. Pierre Goovaerts Consultant in (Geo)statistics President of PGeostat, LLC and Senior Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Wed, 8 Jan 2003, Ruben Roa wrote: If I understand your question correctly, you are trying to use an EXCEL file as an input for Variowin Prevar. Obviously it won't work for several reasons (a) Variowin uses the same file format as GEOEAS, a plain ASCII file, (b) the data itself is in columns separated by spaces or commas, (c) the variable names are in a header, there is also a line with a count for the number of variables. You could however save the EXCEL as plain ASCII and then use a text editor such as NOTEPAD to generate the required header (Don't use WORDPAD, it will insert various on-numeric items). The missing value indicator for GEOEAS is .1E+32 or 1E+31. In the data columns don't include any non-numeric characters (it is possible to have a last column of non-numeric characters, these must be in single quotes and the program will ignore them. Thanks Donald but it is more complex than that. I'm sorry if i wasn't explicit enough in my question but i did create ASCII files to use with variowin. I'm pretty familiar with that program and GeoEas. First i created my ASCII files with Excel, then with Programmer's File Editor, and then with the Notepad. None of them worked under Windows XP and Millenium. Variowin kept reporting error while reading file. Then i went to GeoEas and tried to make it read its own example.dat file and it couldn't under Windows XP. I have come to the conclusion that due to methods to create ASCII files under the new editions of Windows, after W98, programs that use GeoEas data files are not able to read them any more. Rubén http://webmail.udec.cl -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: curve fitting
Hi Carolina, You can not fit any type of curve to your experimental variograms since the model needs to be permissible, hence the practice to fit only a limited number of models that are known to be permissible. Pierre Dr. Pierre Goovaerts Consultant in (Geo)statistics and Senior Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Mon, 18 Nov 2002, Carolina Garcia Imhof wrote: Dear list members, I have used geoeas until now, which uses a limited number of models. However, I found a program, CurveExpert, which finds the best fitting model, which is usually different from the options in geoeas. For example, for my data, I found that the best fitting models were a 4th level polynomial model and a Hoerl model. Is there any program (downloadable if possible) that would krige with a custom model? Thanks, Carolina Carolina Garcia Imhof Marine Mammal Research Group Marine Science Department 310 Castle Street, PO Box 56 University of Otago Dunedin, New Zealand Fax: 64 3 479 8336 Phone: 64 3 479 5476 e-mail: [EMAIL PROTECTED] [EMAIL PROTECTED] -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: Kriging versus inv. Dist. Weighting
Hi Tomislav, Don't be surprised. It is my experience that cross-validation might sometimes indicate that best interpolation results are obtained using the simplest techniques. If your observations are not too clustered and display no anisotropy, inverse square distance could yield good results. Now, you didn't explain which secondary information was used for cokriging and how many neighboring values were used in the different interpolators. Regards, Pierre Goovaerts Dr. Pierre Goovaerts Consultant in (Geo)statistics and Senior Chief Scientist with Biomedware Inc. 710 Ridgemont Lane Ann Arbor, Michigan, 48103-1535, U.S.A. E-mail: [EMAIL PROTECTED] Phone: (734) 668-9900 Fax: (734) 668-7788 http://alumni.engin.umich.edu/~goovaert/ On Fri, 15 Nov 2002, Tomislav Malvic RGNF wrote: Dear all, This is my first try at geostat mailing list, and maybe my question will not be very professional. I work with data set of porosity in one oil reservoir. Interpolations were done with three interpolation methods: Inverse distance weighting, Kriging (ordinary) and Cokriging (collocated). I done spatial analysis with semivariogram modelling for (co)Kriging. After all, I calculated true error for every included point as difference between real value and estimated value at the same place. I was confused when I saw that Kriging error was higher of Inverse Distance Weighting error! The lowest errors were gained by Cokriging (with the same semivariogram modell as used in Kriging). What could be reason for that? Maybe 14 points is too low set for proper modelling of directional semivariogram analysis (directions=0 and 90 degrees). I tested several lag distances and distance with the highest range was chosen. If chosen distance is too low interpolation map contains mostly areas of bull-eyes. Also, input points are moderately clustered. Thank you and best regards, Tomislav -- * To post a message to the list, send it to [EMAIL PROTECTED] * As a general service to the users, please remember to post a summary of any useful responses to your questions. * To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe ai-geostats followed by end on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list * Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: variogram model of n-scores in SGS
Hi Carlo,
Note that a sill strongly different from 1 would
mean that the assumption of 2nd order stationarity
underlying SGS is not met.
I wouldn't force the sill to be one when
fitting the model and you have to use a unit sill
model in the simulation program for consistency.
If the deviation is not larger than say 0.1,
I would fit the model to the
experimental values and then rescale the
fitted sills so that they sum to 1 for use
in the simulation program.
Pierre
|\/|Pierre Goovaerts
|_\ /_|Assistant professor
__|\/|__Dept of Civil Environmental Engineering
|| The University of Michigan
| M I C H I G A N| EWRE Building, Room 117
|| Ann Arbor, Michigan, 48109-2125, U.S.A
_||_\/_||_
||\ /||E-mail: [EMAIL PROTECTED]
|| \/ ||Phone: (734) 936-0141
Fax: (734) 763-2275
http://www-personal.engin.umich.edu/~goovaert/
On Mon, 8 Jul 2002, Carlo Cardellini wrote:
Hello,
I have a question about the variogram model of normal scores to use as
input in the sequential gaussian simulation: the sill must be 1? and if
this is necessary, if the experimenal variogram of the normal scores
presents an higher sill what is better to do? use model with sill 1 or
use the true model?
Thank you in advance to everyone.
Sincerelly,
Carlo
--
* To post a message to the list, send it to [EMAIL PROTECTED]
* As a general service to the users, please remember to post a summary of any useful
responses to your questions.
* To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and
unsubscribe ai-geostats followed by end on the next line in the message body. DO
NOT SEND Subscribe/Unsubscribe requests to the list
* Support to the list is provided at http://www.ai-geostats.org
--
* To post a message to the list, send it to [EMAIL PROTECTED]
* As a general service to the users, please remember to post a summary of any useful
responses to your questions.
* To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe
ai-geostats followed by end on the next line in the message body. DO NOT SEND
Subscribe/Unsubscribe requests to the list
* Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: Gaussian semivariogram model
Hello,
Problems with the Gaussian semivariogram typically
arise when no nugget effect is specified and
some observations are very close to each other,
leading to covariances matrice with very similar rows.
You can read more about this pathological model in Hans
Wackernagel's book multivariate geostatistics
or the recent book by Chiles and Delfiner.
Pierre
|\/|Pierre Goovaerts
|_\ /_|Assistant professor
__|\/|__Dept of Civil Environmental Engineering
|| The University of Michigan
| M I C H I G A N| EWRE Building, Room 117
|| Ann Arbor, Michigan, 48109-2125, U.S.A
_||_\/_||_
||\ /||E-mail: [EMAIL PROTECTED]
|| \/ ||Phone: (734) 936-0141
Fax: (734) 763-2275
http://www-personal.engin.umich.edu/~goovaert/
On Wed, 1 May 2002, Soeren Nymand Lophaven wrote:
Dear list
I have experienced that the gaussian semivariogram model sometimes leads
to a covariance matrix which is not positive definite. I am aware that the
parabolic behavior of the function near the origin could give these kinds
of problems, but I dont think this is the whole story. Do you about this
phenomenon, and where to read more about it ??
Best regards / Venlig hilsen
Søren Lophaven
**
Master of Science in Engineering| Ph.D. student
Informatics and Mathematical Modelling | Building 321, Room 011
Technical University of Denmark | 2800 kgs. Lyngby, Denmark
E-mail: [EMAIL PROTECTED] | http://www.imm.dtu.dk/~snl
Telephone: +45 45253419 |
**
--
* To post a message to the list, send it to [EMAIL PROTECTED]
* As a general service to the users, please remember to post a summary of any useful
responses to your questions.
* To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and
unsubscribe ai-geostats followed by end on the next line in the message body. DO
NOT SEND Subscribe/Unsubscribe requests to the list
* Support to the list is provided at http://www.ai-geostats.org
--
* To post a message to the list, send it to [EMAIL PROTECTED]
* As a general service to the users, please remember to post a summary of any useful
responses to your questions.
* To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe
ai-geostats followed by end on the next line in the message body. DO NOT SEND
Subscribe/Unsubscribe requests to the list
* Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS: interpretation/testing robustness of variogram
Hi Juliann,
This type of feature is frequently observed when
the variogram is being computed for lags larger
than half the maximum dimension of the study area.
In your case it might also correspond to the
separation distance between the two most densely sampled
subareas.
The most important questions is What do you want
to do with this variogram. If the next step is to apply
kriging and this rise is not spurious (e.g. may
reflect some trend in the data), the issue is whether you
need to model this second part of the variogram.
If the radius of the kriging search window is smaller
than the lag at which the second rise occurs, I wouldn't
bother modeling it.
Pierre
|\/|Pierre Goovaerts
|_\ /_|Assistant professor
__|\/|__Dept of Civil Environmental Engineering
|| The University of Michigan
| M I C H I G A N| EWRE Building, Room 117
|| Ann Arbor, Michigan, 48109-2125, U.S.A
_||_\/_||_
||\ /||E-mail: [EMAIL PROTECTED]
|| \/ ||Phone: (734) 936-0141
Fax: (734) 763-2275
http://www-personal.engin.umich.edu/~goovaert/
On Wed, 28 Nov 2001, Juliann Aukema wrote:
Hi,
I have a question about interpretation and
robustness of a variogram. My variogram rises then
plateaus and then rises again. I interpret this as
meaning that there are two scales at which there is
spatial dependence of the values. However, the number
of pairwise comparisons is quite different for each of
these stages and I am afraid I may just be seeing an
artifact of the sampling. I used variowin and when I
fit a model for just the first rise and plateau, the
sill was at about 1500 meters. The first point has 70
pairs and then from right before the plateau (1000 m)
to the beginning of the second rise, there are 324-396
pairs and finally the points in the second rise have
between 512-534 pairs. I have a total of 66 sample
points, but they are not evenly spaced with two areas
more heavily sampled than intervening areas.
Additional sampling is not feasible. Do I have a
problem? Is there a way to test the robustness of this
variogram (I don't know how to fit a model to a
variogram with two rises, so I couldnt' do cross
validation)?
(Additional information - other data and analyzing the
same data with nested ANOVA, looking at smaller scales
within the data set, support the first rise. Taking
the residuals of a correlated variable - elevation -
removes the second rise but maintains the first rise -
my interpretation is that there are different
processes at the two scales).
I would appreciate any suggestions,
Thanks a lot,
Juliann Aukema
[EMAIL PROTECTED]
__
Do You Yahoo!?
Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month.
http://geocities.yahoo.com/ps/info1
--
* To post a message to the list, send it to [EMAIL PROTECTED]
* As a general service to the users, please remember to post a summary of any useful
responses to your questions.
* To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and
unsubscribe ai-geostats followed by end on the next line in the message body. DO
NOT SEND Subscribe/Unsubscribe requests to the list
* Support to the list is provided at http://www.ai-geostats.org
--
* To post a message to the list, send it to [EMAIL PROTECTED]
* As a general service to the users, please remember to post a summary of any useful
responses to your questions.
* To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe
ai-geostats followed by end on the next line in the message body. DO NOT SEND
Subscribe/Unsubscribe requests to the list
* Support to the list is provided at http://www.ai-geostats.org
Re: FW: AI-GEOSTATS: entering the fray
Hi guys,
I promised myself I would not waste more time
on this futile discussion about covariance and variogram,
but it seems that the discussion has drifted far away from
the initial comment by Isobel or that most people don't
remember what was the initial question.
Isobel's comment originated from my sideline remark
(it was not even part of Celia's initial question)
that the SIMPLE kriging system can not be written in terms of
semivariograms, which Isobel qualified of pure non sense..
It seems that my reference to the excellent
book by Chiles and Delfiner did not convince Isobel.
Let's then use Gslib book by Deutsch and Journel since
it is probably more widely used by members of this discussion
list and Isobel pointed out that anecdote with Andre.
On page 65 of Gslib user manual, 2nd paragraph, I quote:
In the sytem (IV.4) (SIMPLE kriging system!), the covariance
values C(h) cannot be replaced by semivariogram values
g(h)=C(O)-C(h) unless sum_lambda = 1, which is the ordinary
kriging constraint. I guess it's clear enough, and that is
nothing to do with whether we should solve an ordinary
kriging system in terms of covariances or semivariograms
(Everybody knows that you get the same results!), or
whether we should teach students in one way or another...
Given that SIMPLE kriging is rarely used, we might even
argue that all this discussion is pointless...
Again, the reason for that e-mail is to clarify the matter
for students or practitioners who might have been
confused by this exchange of e-mails... I don't have
a book, a software or a consulting company to advertise!
Cheers,
Pierre
|\/|Pierre Goovaerts
|_\ /_|Assistant professor
__|\/|__Dept of Civil Environmental Engineering
|| The University of Michigan
| M I C H I G A N| EWRE Building, Room 117
|| Ann Arbor, Michigan, 48109-2125, U.S.A
_||_\/_||_
||\ /||E-mail: [EMAIL PROTECTED]
|| \/ ||Phone: (734) 936-0141
Fax: (734) 763-2275
http://www-personal.engin.umich.edu/~goovaert/
On Wed, 23 May 2001, Steve Zoraster wrote:
1)What manager in the mining or petroleum industry who has graduated
from college hasn't taken a serious statistics course, including covariances
and correlations?
2)Surely when starting from scratch, educating someone about
geostatistics is more intuitive using covariances? (Just my opinion so far,
speaking as a mathematician who remembers teaching basic college level
statistics to nursing majors, education majors, sociology majors, etc. And
even succeeding occasionally.)
3)I have taken two multi-day courses in geostatistics from well known
industry experts. In each class they included significant material and time
on the first day explaining/justifying variograms by showing their
mathematical relationship to spatial covariance functions. It seems that
those instructors did not trust the variogram to be more intuitive than
spatial covariance functions.
4)The two basic level introductions to geostatistics I have on my
bookshelf replicate the experience at those two classes
Steven Zoraster
-Original Message-
From: Yetta Jager [SMTP:[EMAIL PROTECTED]] mailto:[SMTP:[EMAIL PROTECTED]]
I think part of the difficulty in the semivariogram vs. covariance war is
that modeling is subjective, and the notion of covariance has become more
intuitive for statisticians, while the notion of semivariance has become
more intuitive for geologists.
From: Isobel Clark [[EMAIL PROTECTED]]
I agree that the semi-variogram approach is easier for the non-statistician
to grasp. Difference in value is a simpler concept to grasp than
cross-product, especially when your boss wants to know the likely difference
between what you tell him and what really happens!
--
* To post a message to the list, send it to [EMAIL PROTECTED]
* As a general service to the users, please remember to post a summary of any useful
responses to your questions.
* To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and
unsubscribe ai-geostats followed by end on the next line in the message body. DO
NOT SEND Subscribe/Unsubscribe requests to the list
* Support to the list is provided at http://www.ai-geostats.org
--
* To post a message to the list, send it to [EMAIL PROTECTED]
* As a general service to the users, please remember to post a summary of any useful
responses to your questions.
* To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe
ai-geostats followed by end on the next line in the message body. DO NOT SEND
Subscribe/Unsubscribe requests to the list
* Support to the list is provided at http://www.ai-geostats.org
Re: AI-GEOSTATS:
Hi Isobel,
I never read your book, but in all
textbooks I have consulted so far
the kriging system is derived
in terms of covariances... then,
provided that unbiasedness conditions are
included in the system (that is ORDINARY
AND UNIVERSAL KRIGING cases) the system
is expressed in terms of semivariograms.
Instead of referring to Matheron's book,
let's look at Chiles and Delfiner's 1999
book, which should be close enough!
On page 170, second paragraph
from the bottom, I quote There is
no variogram analog to the simple kriging
system (3.2) because then Z*-Z_0 is not
an allowable linear combination.
Next time, think twice before sending
confusing e-mails that do not help
readers who are trying to learn basic
geostat concepts!
Cheers,
Pierre
PS: If you find Matheron's work with a SIMPLE
kriging system expressed in terms of variograms,
keep it in a safe place.. it must be worth a fortune..
|\/|Pierre Goovaerts
|_\ /_|Assistant professor
__|\/|__Dept of Civil Environmental Engineering
|| The University of Michigan
| M I C H I G A N| EWRE Building, Room 117
|| Ann Arbor, Michigan, 48109-2125, U.S.A
_||_\/_||_
||\ /||E-mail: [EMAIL PROTECTED]
|| \/ ||Phone: (734) 936-0141
Fax: (734) 763-2275
http://www-personal.engin.umich.edu/~goovaert/
On Tue, 22 May 2001, [iso-8859-1] Isobel Clark wrote:
In fact, once the pseudo-sill A cancels out from
the system of linear equations, the system is
expressed in terms of semivariograms
My point exactly and, if you don't do it, you don't
make silly assumptions.
I use to think in terms of covariances since
it's more intuitive, and the simple kriging
system can only be expressed in terms of
covariances anyway...
That is just so much nonsense. Ordinary kriging and
simple kriging are derived on the basis of the
semi-variogram and are simply slight variations on one
another. If you don't trust my book, go back to your
original Matheron, please.
Isobel
Do You Yahoo!?
Get your free @yahoo.co.uk address at http://mail.yahoo.co.uk
or your free @yahoo.ie address at http://mail.yahoo.ie
--
* To post a message to the list, send it to [EMAIL PROTECTED]
* As a general service to the users, please remember to post a summary of any useful
responses to your questions.
* To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and
unsubscribe ai-geostats followed by end on the next line in the message body. DO
NOT SEND Subscribe/Unsubscribe requests to the list
* Support to the list is provided at http://www.ai-geostats.org
--
* To post a message to the list, send it to [EMAIL PROTECTED]
* As a general service to the users, please remember to post a summary of any useful
responses to your questions.
* To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and unsubscribe
ai-geostats followed by end on the next line in the message body. DO NOT SEND
Subscribe/Unsubscribe requests to the list
* Support to the list is provided at http://www.ai-geostats.org
Re: [AI-GEOSTATS: MSE to compare different methods]
Hi Mercedes,
I fully agree with Gregoire's suggestions of performing
a series of jackknifes over a range of sampling densities.
In this way, you account for both the impact of sampling
density and sampling fluctuations in the comparison.
An example of this approach can be found in:
Saito, H. and P. Goovaerts. 2000.
Geostatistical interpolation of positively skewed and censored data
in a dioxin contaminated site.
Environmental Science Technology, vol.34, No.19: 4228-4235.
I can e-mail you a PDF copy of the paper if you like.
Cheers,
Pierre
|\/|Pierre Goovaerts
|_\ /_|Assistant professor
__|\/|__Dept of Civil Environmental Engineering
|| The University of Michigan
| M I C H I G A N| EWRE Building, Room 117
|| Ann Arbor, Michigan, 48109-2125, U.S.A
_||_\/_||_
||\ /||E-mail: [EMAIL PROTECTED]
|| \/ ||Phone: (734) 936-0141
Fax: (734) 763-2275
http://www-personal.engin.umich.edu/~goovaert/
On 7 Jan 2001, Gregoire Dubois wrote:
Dear Mercedes,
doing k fold cross validation (taking out X % of the samples) will not give
you any reliable results unless you repeat the operation several times. Taking
out 15% of the samples one time only will give you an MSE that will depend
strongly on the data you have removed. Has the selection of the 15% been made
randomly? You may get a strong bias if the 15% of the samples have been taken
in one region in particular or if you have taken out extreme values only. At
this stage, I would trust more the results obtained by standard cross
validation (leave one out method).
I didnt check your previous mail but if you have few samples only,
k-fold cross validation wont help you much.
If you have many samples, then you should repeat the procedure at least 10
times to be sure that the way you have extracted the data has not influenced
too much the results.
Also, if you have a phenomenon that fluctuates at different scales, you may
have removed the short scale effect by taking out only few samples (15% is not
much).
My suggestion is the following: it is time consuming but might be worth the
effort. The idea is to take out an increasing number of samples (10, 20, 30,
40, 50, 60, ...,X%) of samples, this 10 times, and see how the average MSE
evolves. You may find out that methods A B work better than C D when only
few samples are removed and that C D give better results than A B when
more than 40% of the samples have been removed. This would mean that C D
describe better the general trend of the phenomenon while A B are more
sensitive to the local structures (since you have more dense data).
If you dont have the time to proceed in such a way, you should use standard
cross validation only and investigate the regions/samples where you have the
highest errors.
Just few thoughts.
Gregoire
"Berterretche, Mercedes" [EMAIL PROTECTED] wrote:
I would like to thank Benjamin Warr for his siggestion about doing
difference images instead of global measures as MSE.
I'm confused because crossvalidation MSE (taking one sample out and
recalculating) and validation MSE (taking 15 percent of the samples out and
recalculating) are giving me opposite results. The validation method would
allows me to compare kriging vs cokriging vs Kriging with an external drift
vs regression , but I don't know if I can trust the results at this point.
Does anybody have any input about this?
Thanks in advance,
Mercedes Berterretche
--
* To post a message to the list, send it to [EMAIL PROTECTED]
* As a general service to the users, please remember to post a summary of
any useful responses to your questions.
* To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and
"unsubscribe ai-geostats" followed by "end" on the next line in the message
body. DO NOT SEND Subscribe/Unsubscribe requests to the list
* Support to the list is provided at http://www.ai-geostats.org
Gregoire Dubois (Ph.D.)
Institute of Mineralogy and Petrography
Dept. of Earth Sciences
University of Lausanne
Switzerland
http://www.ai-geostats.org
Get free email and a permanent address at http://www.netaddress.com/?N=1
--
* To post a message to the list, send it to [EMAIL PROTECTED]
* As a general service to the users, please remember to post a summary of any useful
responses to your questions.
* To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and
"unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO
NOT SEND Subscribe/Unsubscribe requests to the list
