GIS/MEA582: Geospatial Modeling and Analysis

Spatial interpolation and approximation II

Create the assignment working directory (folder) .\Interpolation2.

Interpolation using splines

Interpolate elevation surfaces from spot point heights using the Spline tool in the Spatial Analyst toolbox.
Use three different weights (0.1, 1, 1000) to control the tension, and thus, the resulting surface character.
For the first interpolation, you'll leave the Weight field blank, accepting the default 0.1 value.
In subsequent interpolations, you'll enter the explicit weights.

ArcMap File->Add Data->Add Data...
Select ncshape.mdb/elev_lid792_randpts
Check out a Spatial Analyst Extension license
Open the ArcToolbox
Select 'Spatial Analyst Tools->Interpolation->Spline'
Set 'Input Point Features' to 'elev_lid792_randpts'
Set 'Z value field' to 'value_'
Set 'Output Raster' to '.\Interpolation2\elev_tps'
Set 'Output cell size' to '2'
Set to Spline Type to 'TENSION'
Leave the other inputs as defaults     
Click 'OK'

Evaluate impact of interpolation function parameters on the resulting surface using hillshade.

Check out a 3D Analyst Extension license.
Set the default output raster cell size environment variable to 2.

Geoprocessing->Environments...->Raster Analysis->Cell Size
Set 'As Specified Below' and enter 2 in the field that follows.
Click 'OK'
Open the ArcToolbox (if it's not already open)
Select '3D Analyst Tools -> Raster Surface -> Hillshade'
Set 'Input Surface' to 'elev_tps'
Set 'Output Raster' to '.\Interpolation2\Hillsh_tps'
Leave the other inputs as defaults
Click 'OK'

Repeat the spline interpolations above, using weights of 1 and then 1000,
and generate the associated hill shade surfaces.
Compare the three hill shade surfaces, and discuss your findings in your report.

Interpolate surface using Radial Basis Functions

Use the Geostatistical Analyst tool's Radial Basis Functions to generate at least two different elevation surfaces by selecting different Kernel functions.
In the first surface you'll use the RBF multi-quadric method.
In the second, you'll choose Regularized Spline.

Check out a Geostatistical Analyst Extension license
Customize->Toolbars->Geostatistical Analysis'
Select 'Geostatistical Analysis->Geostatistical Wizard'
Select the 'Deterministic Methods->Radial Basis Functions' method
Set the 'Source Dataset' to 'elev_lid792_randpts'
Set the Data Field to 'value_'
Click 'Next'
Set the Kernel function to 'Multiquadric'
Click Next/Finish/OK through remainder of dialogs

Rerun the interpolation in the Geostatistical Analyst Wizard using
Kernal Type = Completely Regularized Spline and compare your results.

Optional: compute empirical and model variogram for 'elev_lid792_randpts' and interpolate the surface using kriging with a model variogram that fits the best with the empirical variogram.