EXERCISE 8 – Spatial Analyst

EXERCISE 8- SPATIAL ANALYST:
MASTER THE RASTER

Environmental Resources 372:362

Intermediate Environmental Geomatics

Spatial Analyst

In ArcMap, turn on the Spatial Analyst extension (Tools/Extensions). You must also add the toolbar for Spatial Analyst (View/Toolbars).

1. Digital Elevation Models

One of the strengths of raster formats (aka grid) is that it more accurately represents things that vary continuously over the landscape such as elevation. We will work with elevation, which is often presented in raster formats by Digital Elevation Models (DEMs). DEMs are a specific file format provided by the USGS at 1:24,000 scale (quad sheet coverage) and are available for most regions of the US .

In ArcMap, add a DEM (\\ad-rsc\databank\usgs2\dem\30m_10m\indiv24k\nj_utm27\hamburg_dem27 - notice there are a lot of other DEM's, the layer 00njquads_u27 provides an index map which tells you what DEM covers which area). The one you have opened is a DEM of a portion of the Highlands Region in north central New Jersey . Changing the colors will help emphasize some of the topographic features. Notice when you look at the symbology, "stretched' is the default. (You can always change this to classified to make transitions more distinct if desired.) Right-click on the "Window" for the Color Ramp and click off Graphic View (black to white will appear in the window). In the pull-down menu, select Elevation #1 color ramp. Where are the high points and where are the low?

Turn on your Spatial Analyst extension and add your Spatial Analyst toolbar if it isn't on already. Let's try hillshading (Spatial Analyst/Surface Analysis/Hillshade) - accept the defaults and put it under the DEM. Make the DEM 60% transparent. How does that look?

Explore what your other options are under Spatial Analyst/Surface Analysis. Contour lines might help us detect where elevations change rapidly - try it with 20 meter contour intervals. (Note you are creating new layers as you go. Where are they being saved? Maybe you should keep an eye on that.)

What type of real analysis can we do with the suite of choices under Surface Analysis? Imagine that you are a developer and are interested in identifying locations in this area that would be suitable for a new residential subdivision. You know you have slope restrictions (maybe 7-10 percent slope is about the cutoff without increasing costs a lot).

Create and reclassify a slope theme into three categories to reflect development potential (developable (0-5); marginally developable (5-8) and non-developable (>8)). (HINT: You create it with Spatial Analyst/Surface Analysis/Slope but you reclassify it with Spatial Analyst/Reclassify...)

You are an environmentally friendly developer and you want your houses to be solar, so they'll need to have Southern exposures. Which areas are available to you? Generate a map of Aspect (using the DEM, not the slope). There are a lot of categories with aspect and it would help if there were fewer - say those with at least some southerly exposure and those without.

Finally, we can combine our reclassified slope and our reclassified aspect using the raster calculator. But that only works if you were careful when you reclassified them. (HINT: Your reclassification needs to turn them into comparable ratings. High values should be Good in all inputs, and low values should be Bad in all inputs.)

KEY MOMENT: Which layers do you want to keep forever (or at least long enough to use them when you next login)? You need to right-click on them and choose to Make Permanent or they will probably disappear as soon as you are done today.

Homework: Produce a well-constructed map that shows the places where the developer should think about looking. You must add an additional criterion other than those we described here. If your additional criterion is based on vector data, you will need to convert it to raster, which is something that we will do in the next section of the lab.

2. Interpolated Surfaces: An Example

You have total annual rainfall data for New Jersey but the data is collected at different weather stations and is summarized by municipality throughout the state. You would like to know what the rainfall is for points between the different municipalities. Spatial Analyst can create a grid by interpolating from the measurements. In the interpolation the cell values between the different municipalities are approximate values of the quantity at that location based on the data from the municipality. Let's see how it works:

Insert a new data frame or open a new map document. Add the layer \\ad-rsc\data\teach\intgeo\ClassWork\spatial\climate.

Change the symbology to show Precip with graduated colors to see how rainfall varies across the state.

With Spatial Analyst, you need set your initial settings correctly if you really want the end results to look right. So, before we do our analysis here, let's get everything else fixed. Go to Spatial Analyst/Options... and set your extent Same as Layer "climate" then set the pixel size as the rather arbitrary 2000 (which means each pixel will be 2000' x 2000'). Hit OK.

Create the interpolated surface (Spatial Analyst/Interpolate to Raster). Use the Inverse Distance Weighting method and be sure you select the correct layer and use Precip as the Z Value Field and go with the other defaults. Press OK.

Display the results with the points on top. Use the 'precipitation' color ramp for the symbology to make some sense of the values. What information do you think was used to create the surfaces beyond New Jersey ? In the interpolation process accuracy progressively drops off farther away from data points. Therefore we cannot put much faith in those surfaces beyond the New Jersey border. We can clean up this problem and make a more accurate and honest map using a mask.

Masks:

Masks allow you to analyze and display only the area you are interested in by removing unwanted areas. A mask is in the simplest sense a grid that defines the area of analysis. It works somewhat like a cookie cutter that causes the analysis to be performed for areas only within the mask.

Create your own mask: Add \\ad-rsc\teach\\intgeo\rowan\avtrn\spatial\climmun. Set your extent equal to climmun. Convert this vector file to a raster (Spatial Analyst/Convert). 'Reclassify' this theme so all the values except "No Data" are equal to 1. Name this NJMask.

To use this as a mask, select Spatial Analyst/Options and set NJMask as the Analysis Extent and as the Analysis Mask under General.

Create another interpolated surface from the point values of precipitation and note the effect of the mask.

Important Note: Don't forget that you have set a mask when you are working with a project. Otherwise you might get 'weird' results when you do other analyses.

Homework: Create a color map that shows the mean temperatures across the state.

Assignment due Monday, March 31st .