ArcMap Interpolation Lab

Due in two weeks

Objective: The objective of this lab is to expose you to some of ArcMap's interpolation capabilities, i.e., to experiment with interpolating a raster surface from a set of measured points. Several ways to do this (IDW, Natrual Neighbors, Kriging, and Spline) are provided two levels down in the ArcToolbox ( i.e., under Spatial Analyst Tools then under Interpolation, and a TIN-based approach is available through 3D Analyst. In this lab you will:

1. import data points (x,y,z) from a text file and convert them to a point shapefile
2. interpolate surfaces from that shapefile using several different interpolation techniques
3. compare the resulting surfaces visually and with the raster calculator.
4. put the results together in a one page layout

Data Needed

Use any of the datasets below. The data are in the usual place, now on Laulima.

• interplab.txt - a text file of 423 randomly sampled data points on the computed surface, z = .25 * sin(6.0 * pi * x * y), with x and y both ranging from -1.0 to 1.0. If you are interested, it can be seen on page 231 of the CRC Handbook of Mathematical Curves and Surfaces.
Or...
• kbayreef_1.txt - a text file of 31,485 SHOALS (lidar) sampled points on the reef in Kaneohe Bay.
Or...
• oahu_xyz.txt - a text file of 3,718 points sampled from contours representing Oahu. These were originally digitized from USGS 1:24,000 topographic quadrangles but I don't remember how dense the sampling from the countours was. Points sampled from contours are not really random and interpolating from them tends to recapture and over emphasize the contours. Try 'em and see what you get.

Preliminaries (the set-up):

1. Start ArcMap and make sure the Spatial Analyst and 3D Analyst extensions are available.
2. And then, left-clicking "Tools" in the main menu, and checking those check boxes, and finding where these extensions have been added to the tools visible on your screen.
3. Set the default database for Arc to use.
4. Set the working directory for ArcMap to use.
5. Import the text data and convert it to a feature set in your geodatabase.

Interpolating A Raster from Sample Data Points

ArcGIS has several geoprocessing tools to interpolate raster data from point data. Try at least IDW, spline, and kriging. Especially with IDW experiment with the interpolation parameter settings. Be sure to interpolate using the 'z' field, not the 'x' or the 'y' fields. (Why can't it figure that out?)

Visually and mathematically (via the raster calculator) compare three surfaces (from those you created). Use Spatial Analyst -> Raster Calculator to compute
[surface1] - [surface2]
i.e., a difference surface, in which to see where, and by how much, the interpolation results differ. (Does changing the exponent or the search radius in IDW interpolation have the effects that you expect from the text and lectures?)

TIN Adventure Ahead

For fullness, also generate a TIN from these data.

In Arc 10.2, there is some interesting background on surfaces, TINs, terrains and more that you might wnat to explore. Maybe start here. You'll need tools from the 3D Analyst toolbox, so check that that extension is enabled. (Customize -> Extensions -> checkbox 3D Analyst)

Then, assuming that you have a set of XYZ points available as a shapefile of points...

Run the "Create TIN" geoprocessing tool found under:
3D Analyst tools :: Data Management :: TIN :: Create TIN
Set the Environment - Workspace Current and Scratch Workspace to a directory (not a geodatabase). Specify a name for the output TIN, and the Input Feature Class. Set the height field (Z? elev? depth?) and SF Type (Mass_Points) --- if Arc did not guess correctly --- by clicking on the fields.

Check-out some of the display options with the TIN. In the table of contents, right click on the TIN name, and go to the Properties, Symbology tab. In the lower left you can turn on/off the hillshading. You can also add/remove several characteristics of the TIN to display (Elevation, Slope, Aspect, Node or Face tags). (They are sll shown under the Fields tab, but) here, you can add them and check on/off which is symbolized.

Does the elevation TIN look like the same surface you would get with IDW by setting the exponent to "1" and the neighbors to "3"?

Map Layout View

I haven't stressed cartographic presentation in this course, but we need to at least touch on Arc's "layout view". The basic idea in the current version of Arc is to have a "Layout View" into which are inserted objects that might be titles, map frames, tables, text blocks, scales, north arrows, and other elements that can be dragged around and modified.

It's all very click-n-drag-n-drop and probably faster to learn through experimentation and the help functions than through a slap-dash step by step handout. Try View -> Layout and then "insert" and position a few elements. (Like a title, some map frames, a text block, etc.) Note that you may need to drag things in the table of contents to newly created 'frames' for them to show-up.

Deliverables

1. Two one-page map layouts. One showing four raster surfaces (one each created by IDW, spline, kriging, and conversion to TIN), with labels so a reader can tell which is which and what paramters were used. The other should be titled "Interpolation Comparison", and should show from three to six maps showing differences between the interpolated surfaces, each labeled as to what it is.
2. A one page write-up briefly describing the differences you note between the surfaces produced by each of these interpolators and the effects of the interpolation parameters with which you experimented.

   Hints and Notes

   Help - is on the main menu, and by now you should be getting comfortable using it.

   Read ESRI's descriptions of these interpolation schemes.