Latest Contributions by EricKrause

If you just want a rough measure of relative pollution exposure, consider using the Kernel Density tool in Spatial Analyst.  Sum up the pollution released from each location, and use it as the Population field. You won't be able to actually estimate the total exposure at a given location, but you'll be able to say that certain locations are more exposed than others.  As for estimating the actual amount of exposure, I don't know how you would do that.  You would probably need to talk to a physicist. ... View more

You need to ask yourself what an interpolation would mean for your data.  If you used your data to predict a value at an unmeasured location, how would you interpret the value of the prediction?  If your data is about the amount of pollution released, then the interpretation of the prediction would be "the amount of pollution released at this location over the year."  But what if there is no factory at the new location?  In the case of pollution release, it's only being released from particular locations.  It doesn't really make sense to interpolate a variable that only occurs at discrete points on the map. If you want to make a map of pollution levels, your data needs to be random samples of pollution levels, not measurements of pollution release from discrete locations.  There may be physics models that can predict pollution levels from data about pollution release, but ordinary kriging is not the way to do this. ... View more

If you're using ArcGIS 10, you can right-click the parameters in the Wizard and Copy All.  This is shown in the attached graphic.  You can then paste them directly into Excel, and it will format nicely.  Let me know if you're using a version earlier than 10. ... View more

Probably the most structured and safe way to do this is to save your geostatistical layers to .lyr files using the "Save to Layer File" geoprocessing tool.  When you pull the .lyr file into ArcMap, right-click the layer and choose Properties.  On the Method Report tab, you'll see all the parameters in the same structure as the Method Report window in the Geostatistical Wizard. Do you really need an xml file separate from the geostatistical layer?  If you really need it, let me know. ... View more

This is a complicated question, but the short answer is that it is not valid to krig on residuals of kriging in ArcGIS.  To get an idea why it is not valid, you need to know a few things: 1) Kriging is a exact interpolator when you do not use measurment error, so without measurement error your residuals will all be zero.  There would be nothing to krig. 2) ArcGIS only supports homogeneous measurement error where the percent of the nugget explained by measurement error is constant and uncorrelated.  3) Because we assume the measurement error is uncorrelated, it should have no spatial autocorrelation, so kriging would not be valid.  If your residuals do show spatial autocorrelation, it means that your original kriging model was not fitted correctly. ... View more

If you're interested in looking at residuals (the difference between the predicted value and the actual value), then use GA Layer to Points.  If you're interested in seeing if your kriging model actually fits the data, use cross-validation (or validation by splitting your data in half). Getting good at geostatistical modeling requires study, practice, and often a bit of luck.  The best place to start is our help documentation.  Here's a quick list of topics to review:  Gaussian distributions, transformations, semivariograms, histograms, QQ plots, voronoi diagrams, stationarity, trend removal, spatial autocorrelation, searching neighborhoods, cross-validation, and output types (prediction, standard error, probability, quantile). Needless to say, that is too much to cover in a blog or forum post. If you read the help and decide you want more information, a colleague published a book last year on performing spatial statistics (and geostatistics) in ArcGIS: http://esripress.esri.com/display/index.cfm?fuseaction=display&websiteID=194 ... View more

The results are different because these two methods are doing slightly different things. On the cross-validation page of the Geostatistical Wizard, you're seeing statistics calculated in the following way: 1.  Remove the first point in the dataset, then use the remaining (n-1) points to predict the value at the location of the point you removed. 2.  Repeat step 1 for all n points in the dataset, and calculate the statistics. The idea is that if your model is good, you should be able to use (n-1) points to closely predict the value of the nth point. When you do GA Layer to Points with a field to validate on, you will not be removing the points before the prediction.  In other words, when you predict at an input point location, you will use the measured value at that location to make the prediction.  Clearly, including the measured value will provide more accurate predictions than not using it (in fact, if you turn measurement error off, it will do a perfect prediction every time). Validation with GA Layer to Points is designed to let you use an entirely different dataset for your validation (it doesn't really make sense to validate on the same dataset that you used to build the model).  It's a common practice to split your data in half and build the model on one half and validate (via GA Layer to Points) on the other half.  The idea here is that if your model is good, it should be able to accurately predict the values of the dataset that you didn't use to build the model. Was that explanation clear?  I'm happy to clarify anything. ... View more

The global vs local trend removal slide bar in ArcMap 9.3.1 is used to determine how locally the trend will be removed.  100% global will remove trend using global polynomial interpolation, and anything else will use local polynomial interpolation.  Read about these two methods in our help, and feel free to ask questions about anything you have trouble understanding.  Note that in ArcMap 10, you can actually see the trend removal parameters, so you have more control over how the trend is removed.  As for how you decide how locally to remove the trend, there are a couple of things to watch.  First, if you remove the trend too locally, you may find that the semivariogram on the next page is near flat.  That means that you removed too much trend.  You want to remove just enough in order to still fit a good semivariogram.  Second, you can go by crossvalidation statistics, which I'll talk about below.  The biggest things to look for in crossvalidation is that the average standard error should be similar to the root-mean-square, and the root-mean-square standardized should be close to one.  If you find several models that have these properties, choose the one with the lowest root-mean-square.  You can also see if the mean standardized is close to 0, but if the other statistics are good, this one is usually good too. ... View more

The formula you posted is for simple kriging, not ordinary kriging.  The formula for ordinary kriging looks similar, but it uses a Lagrange multiplier, which makes it a lot harder to understand. For simple kriging, if you want to get the expected value and variance in the log-space (before the back transformation), then you should do the transformation manually.  Create a new field in your feature class, and calculate the log of the data with the Field Calculator.  Then perform simple kriging (with no transformation) on the logged data. The prediction surface will correspond to the estimated value, and the prediction standard error surface corresponds to the square root of the estimated variance. ... View more

I'm not sure what you mean by "units".  Can you clarify your question with more information and maybe an example?  Thanks. ... View more

This page has a list of our seven ESDA tools.  I'm still not sure which one you're talking about. http://help.arcgis.com/en/arcgisdesktop/10.0/help/index.html#//00310000000m000000.htm ... View more

The default extent of a geostatistical layer is the rectangular extent of the input points.  If you want to change the extent, right-click the layer in ArcMap's table of contents, and choose "Properties".  Go to the Extent tab, and you can define a different extent (like the extent of California). For more information about changing the extent and clipping a geostatistical layer, see the link below.  Skip down to steps 12-20 in the section titled "Create a surface using the default options." http://help.arcgis.com/en/arcgisdesktop/10.0/help/index.html#/Exercise_1_Creating_a_surface_using_default_parameters/0031000000p0000000/ If you have any further questions about geostatistics, you'll get the quickest response by using the Geostatistical Analyst Forum: http://forums.arcgis.com/forums/100-Geostatistical-Analyst Good luck, and I hope this helped. ... View more

The clip layer in polygon declustering must be a polygon feature layer. Polygon declustering works by creating Thiessen polygons from the input points, and the bigger the Thiessen polygon, the more weight that point gets.  The problem is that points on the boundaries will often have Thiessen polygons that are way too big.  I've attached two graphics that show the Thiessen polygons of our tutorial data, one with clipping and one without.  It's clear that without clipping, the polygons near the border of California are given too much weight.  By clipping to the border of California, the weights will be much more reasonable. ... View more

Any idea why it distorts the kriging when I convert to a raster? This blog should help: http://blogs.esri.com/esri/arcgis/2008/09/29/understanding-geostatistical-analyst-layers/ The main idea is that a geostatistical layer is really a mathematical model that takes an (x,y) location and returns a value (as opposed to a raster that has values stored in each cell).  A geostatistical layer doesn't "know" its own values until you ask it to calculate them.  The contours that you see in a geostatistical layer are made from a quick contouring algorithm: the model makes the fewest calculations that it can in order to draw roughly accurate contours. When you export a geostatistical layer to a raster, all it is doing is calling the mathematical function at the center of every raster cell.  This obviously requires many more calculations than drawing quick contours, so the surfaces may appear a bit different (sometimes very different). ... View more