Latest Contributions by EricKrause

‎06-05-2015

It was first released in ArcGIS 10.2. It is not in 10 or 10.1.

‎06-05-2015

You'll have to create the first geostatistical layer in the Geostatistical Wizard (or from a geoprocessing tool in the Interpolation toolset). To interpolate a new dataset, open Create Geostatistical Layer and provide the first geostatistical layer as the model source. The tool will analyze the model and decide what kinds of datasets are required. In the simplest case, this would be just a dataset and a field. In a more complex case like cokriging, it might require multiple datasets and multiple fields. You then put the new data into the "Input datasets" parameter of the tool. Give the output geostatistical layer a name and run the tool. It will then create a new geostatistical layer that applies the parameters from the model source to the new dataset.

‎06-05-2015

The key is that you can apply the interpolation parameters to a new dataset. Indeed, it doesn't make much sense to apply it to the same data (this would just make a copy of the layer). Here is a typical scenario that shows how this tool can be useful: You have a monitoring network set up to take daily measurements of pollution levels. Each day, these sensors measure the pollution and store it into a database, and each day you want to create an interpolated map of pollution levels. On the first day, you take a lot of time to create a quality kriging model in the Geostatistical Wizard and output a geostatistical layer. On the next day, you can use this layer as the model source in Create Geostatistical Layer and apply the kriging parameters to the measurements of the new day. By reusing the original geostatistical layer for the data of each new day, you can fully automate this process and not have to manually go into the Geostatistical Wizard every day to create your map.

‎06-04-2015

Generally, yes. Though a standard search neighborhood has more parameters to it. First, it can use sectors, which force particular numbers of neighbors to come from different directions. Second, it has minimum and maximum number of neighbors (per sector) as a parameter. The way the algorithm works is: For each sector, take at least the minimum number of neighbors (even if that requires going beyond the search radius). Keep taking additional neighbors until you hit the maximum number of neighbors or hit the end of the searching radius. So, for a standard search neighborhood, you can get similar behavior by setting the minimum number of neighbors equal to the maximum number of neighbors (no matter what your search radius is). Remember to take the number of sectors into account. For example, if you have four sectors, and you set min=max=8, then each location will use exactly 32 neighbors in the calculation (8 from each of the four sectors). That is, assuming there actually are at least 8 neighbors in each of the sectors.

‎06-04-2015

I should also say that it depends on how fast you need the method to process. In general, it is better to use a search radius that is too big than one that is too small. If it is too small, you are missing relevant information in the neighboring points, and the quality of your predictions will decline. If it is too large, you are pulling in information that is not useful, but these non-informative neighbors tend to get very small weights, so they have little impact on the quality of the interpolation. However, the more neighbors you use, the longer the method takes to calculate. If you aren't concerned with processing speed, you should err on the side of a larger radius rather than a small one. That being said, the default of ~32 neighbors is almost always more than enough, and you generally won't see improvement in predictions by adding more neighbors. But, as always, it depends on your data.

‎06-04-2015

There aren't a lot of criteria of what is a good search radius, but there is plenty about what is a bad search radius. It definitely should not be larger than the range of the semivariogram (as you noted), and it should be large enough that it captures at least 10 points everywhere in the data domain. Other than that, there aren't many recommendations other than comparing validation and crossvalidation statistics for different search radii. In the Geostatistical Wizard, the default search radius for smooth interpolation is calculated such that it attempts to use at least 32 neighbors in each location. It isn't a perfect algorithm, but we have found it to be reliable and robust.

‎05-21-2015

If you choose Include All, all of the coincident points will be used to model the semivariogram, and they will each be given their own weights when making predictions. There is also the issue of number of neighbors in the searching neighborhood: Imagine that you have 10 values all sampled at the same location and use Include All, and you set the maximum number of neighbors to 10 in the searching neighborhood. In this case, when making predictions near the coincident points, the coincident points will fill up the entire searching neighborhood, and this could potentially give strange results. When you choose Mean, the value at the location will be assigned the average of the values of the coincident points. This average will be used to model the semivariogram, and the searching neighborhood will treat it as a single value. The sample size will also be adjusted accordingly. The logic is analogous for Min and Max coincident point options. The Remove All option will treat coincident points as if the value is Null, and none of the values will be used in the semivariogram modeling or prediction stages of kriging. Again, the sample size will be adjusted accordingly. From the geoprocessing tools, coincident points are handled with an environmental setting. You can read more about it here: http://desktop.arcgis.com/en/desktop/latest/tools/environments/coincident-points.htm

‎04-30-2015

StdError is the standard error of prediction. This is equal to the square root of the kriging variance, which is one of the outputs of the kriging equations. Stdd_Error is the standardized error. This is equal to Error divided by StdError. Ideally, these values should follow a standard normal distribution. NormValue is the equivalent quantile on the standard normal distribution for each standardized error. If you make a scatterplot of NormValue versus Stdd_Error, this is called Normal QQPlot, and it is useful for testing whether the standardized errors actually do follow a standard normal distribution. Ideally, the points should create a straight line at a 45 degree angle in the scatterplot.

‎04-20-2015

Hi Krystan, The distance metric that is used is the shortest sequence of straight lines that go around the barrier feature class. This is what is meant by "non-Euclidean." There are no limitations for the geometry of the barrier feature class, but the more complicated the geometry, the longer the method will take to process. If you are experiencing performance issues, you can use the Simplify Line or Simplify Polygon geoprocessing tools to remove vertices from the barrier feature class, and this will dramatically speed up the calculation. You can read the full details of the algorithm in this paper: http://link.springer.com/article/10.1007%2Fs00477-011-0488-2

‎03-31-2015

It depends on what kind of regression kriging you want to do. If the explanatory variables are polynomials of the (x,y) coordinates, we call that Universal Kriging, and it is available in the Geostatistical Wizard. If your covariates are not (x,y) polynomials, you can use the Ordinary Least Squares tool to create a regression equation, and then you can perform kriging on the residuals. To get predictions, add the interpolated residuals to the prediction from the OLS model. Ideally, the kriging parameters and the coefficients from OLS will be calculated simultaneously (rather than doing OLS then doing kriging), but there is currently no way to do this in ArcGIS. Though that will likely change in the not-so-distant future.

‎01-09-2015

Cokriging will actually interpolate the first variable you provide, but it will use information from the secondary dataset. The RMS prediction error is in the units of the first dataset, so if one variable generally has larger values and more variance than the other dataset, you would expect the RMS to be higher (all else being equal). For example, if you have elevation data in feet and you get an RMS of 3, this means that, on average, the predictions will be off by about 3 feet. If your data is in yards, an RMS of 1 would indicate that your predictions will be off by about 1 yard on average. But since one yard is equal to 3 feet, these two models are equally accurate (all else being equal). You should only expect the RMS values to be the same when switching primary and secondary variables if the two datasets are measuring the same thing in the same units and they have similar spatial configurations (ie, roughly the same number of points in the same general locations).

‎01-09-2015

And you're correct that MWK can be implemented as an interpolation method. In our implementation, it allows you to make predictions at a particular set of points, along with getting their estimated range, nugget, and sill. You actually can get it to create a surface if your set of points is on an evenly spaced grid. You can then run the Point to Raster geoprocessing tool to turn the gridded points into a raster surface. However, I don't think there is any way to perform crossvalidation on this surface with ArcGIS tools. That being said, you should probably still use EBK over MWK.

‎01-08-2015

For the kriging step, you should probably use Empirical Bayesian Kriging (EBK). It is our most accurate interpolation method, and it is easy to automate because it is available as a geoprocessing tool (it can also be accessed from the Geostatistical Wizard). You'll want to output a raster and set both the Extent and Mask environments to your boundary shape file. You can also use the Output Coordinate System environment in EBK to define the spatial reference of the output raster. I'm not sure how to do the rest, but hopefully someone else can help with those parts.

‎01-08-2015

You can use this reference to learn about and cite EBK: Krivoruchko K. and Gribov A. (2014) Pragmatic Bayesian kriging for non-stationary and moderately non-Gaussian data. In Mathematics of Planet Earth. Proceedings of the 15th Annual Conference of the International Association for Mathematical Geosciences. Eds: Pardo-Igúzquiza, E.; Guardiola-Albert, C.; Heredia, J.; Moreno-Merino, L.; Durán, J.J.; Vargas-Guzmán, J.A. Springer 2014, pp. 61-64.

‎09-29-2014

If you have a Ordinary Kriging geostatistical layer (made from the Geostatistical Wizard) in your Table of Contents, you can right-click it and choose "Change output to prediction standard error." The layer will then show the kriging standard errors (the square root of the kriging variances). You can then export this layer to a raster with GA Layer to Grid geoprocessing tool.

Online Status	Offline
Date Last Visited	‎07-09-2026 08:16 AM

My Ideas

Latest Contributions by EricKrause

Re: loop through geostatistical layers in a folder

Re: Geostatistical analyst layer

Re: Geostatistical analyst layer

Re: Is there a rule of thumb for the search radius when using empirical bayesian kriging?

Re: Is there a rule of thumb for the search radius when using empirical bayesian kriging?

Re: Is there a rule of thumb for the search radius when using empirical bayesian kriging?

Re: How does the Geostatistical Wizard (Kriging) handle coincidental samples?

Re: Cross Validation/calculate StdError, Stdd_Error, NormValue

Re: Kernel Interpolation with Barriers -- barrier restrictions and algorithm used

Re: Regression kriging

Re: regarding cokriging technique

Re: How to do cross-validation in Moving Window Kriging

Re: Geoprocessing Kriging

Re: How to do cross-validation in Moving Window Kriging

Re: How variance of kriged data is computed within Geostatistics Analyst

Re: Why exporting EBK 3D to voxel layer change org...

Re: Access geostatistical layer created using ArcP...

Re: Why exporting EBK 3D to voxel layer change org...

Re: K-Bessel Semivariogram Equation

Re: Kriging Model Types

R-ArcGIS