Lab 6: Interpolation and Geostatistical Modeling in ArcGIS

Introduction

This lab teaches you how to do interpolate surfaces and do geostatistical modeling in ArcGIS. Download the sample file (lab6.zip) to your network drive and unzip it there.

Instructions

You will use the following Acrobat book called Using ArcGIS Geostatistical Analyst (Using_Geostatistical_Analyst.pdf) to do the exercise. The pdf file is in the I:\Students\Data\GIS\ArcGIS Documentation\ArcGIS9.1_documentation\ESRI_Library\ArcGIS_Extensions folder. Read Chapter 1 and do the Quickstart Tutorial exercises in Chapter 2. Answer the following questions and produce the following outputs. Labs should be typed, include your name and lab number, be well organized, and be stapled together. All the maps in this lab must include all map elements (legend, scale, title, N arrow, explanatory text if needed, name, source, date).

You should create the maps for this lab as you move through and not all at once at the end. Also, consider the arrangement of layers. Having a raster over a vector layer covers information and is not appropriate.

 Additional Notes:

 There is a version difference between the tutorial (9.1) and the software (10). Please replace the following steps in the Geostatistical Analyst Workbook with the following where noted:

Page 14, getting Geostatistical Analyst going:

*        Click on the Customize Menu > Extensions, Check Geostatistical Analyst to turn it on

*        Click on the Customize Menu and point to Toolbars. Click Geostatistical Analyst

Page 15, Geostatistical Wizard:

*        The wizard is different than shown/described in the PDF. For most things, you should still be able to find what they are describing. Anything that is very different will be noted.

*        For instance, when you start the wizard, select the method in the left section and select the source and data field in the right section.

Page 30, Steps 11 - 13:

*        Change Anisotropy first in the right section under Model #1. Use the drop down menu and change it from False to True - Note the blue eclipse.

*        Also note the automatic direction. This can be changed to manual input for direction by clicking the little calculator button to the right of the direction value.

Page 31, Steps 14 and 15:

*        Just change the direction value. Note how the semivariogram changes

Page 33, Steps 18 and 19:

*        Enter the Y value first under Predicted Value in the right section. Then enter the X value - don't forget the negative sign.

*        Click the Back button to change the Direction angle in the previous window. Click Next to view the results.

Page 34, Steps 20 and 21:

*        Click the Back button to return to where you can change the Direction angle. Click the pencil button to the right of the angle to return it to automatic (a calculator icon should be there) - this will return it to the default

*        Click Next to move on to Kriging step 6 of 6

Page 36, Step 24:

*        You cannot save from here. After you finish the process a Method Report will display and there is a Save button there.

Page 37, Step 28:

*        Right-click on Trend Removed and select Change Output to Prediction Standard Error

Page 39, skip Step 10 and for Step 11: change Anisotropy to True.

Page 41, Step 19:

*        Rename the Indicator Kriging surface you just created to "Indicator Kriging"

Page 42, After Step 4

*        After you check Contours and select another color ramp, click the Classify button.

*        Change the method to Equal Intervals and leave the Classes at 10.

*        Click OK and continue on to Step 5

Page 43, Step 1

*        Change the extent of both of these layers to "the rectangular extent of ca_outline"

 

  1. At the end of Exercise 1, make a map of the result.
  2. Are the mean and median values of the ozone layer similar?  What does this tell you?
  3. Click on the histogram bar with the lowest ozone values.  Where are the lowest ozone levels in California?
  4. What can the QQ Plot tell you when you are looking at one variable?
  5. What does the trend analysis graphic tell you about the East-West and North-South trends in ozone levels in California?
  6. In Exerercise 2, you created a histogram, a QQ Plot, a trend analysis graphic, and a semivariogram.  Each one of the dialogs for these ESDA tools has a button called Add to Layout.  Add all of these graphics to the layout with the interpolated ozone surface, and make a map.
  7. What is the objective of cross-validation?
  8. What is prediction error and how do you measure it?
  9. What is a prediction standard error map?
  10. Print a map of the krigged ozone surface with the trend removed at the end of Exercise 4.
  11. Print the standard error map of the krigged surface.
  12. How do you use the cross-validation comparison dialog to compare surface models?
  13. Print the indicator map at the end of Exercise 5.
  14. Print the final map at the end of Exercise 6. This map should match the one in the tutorial, with the exception of some of the colors. If you choose, you may do this map in color.