EXERCISE 6 – BASIC ANALYSIS:
MAP ALGEBRA AND OVERLAYS
Environmental Resources 372:362
Intermediate Environmental Geomatics
We can use GIS to answer many different kinds of spatial questions as we saw in the last lab when we looked at the types of landuse near floodprone areas in the Lawrence Brook Watershed. It turns out that the Lawrence Brook Watershed is a relatively small watershed, about 46 square miles. Today we’ll delve more deeply into the analytical capabilities of ArcGIS using the watershed as a case study. The watershed group there wants to build a new environmental center somewhere within the boundary of the watershed. Based on their specifications, we need to figure out the best place for the building. All of the data we need is in \\ad-rsc\data\teach\intgeo\ClassWork\avdata\lbwpuse\.
Part 1: Finding the Best Site
Condition 1: Distance to Road
1. First, the building should be accessible, so let's keep it within 200 feet of a road. Use the Buffer tool (ArcToolbox | Analysis Tools | Proximity) to buffer the roads \\ad-rsc\data\teach\intgeo\ClassWork\avdata\lbwpuse\lbwstreets) of the watershed by 200 feet. Don’t forget to set the Dissolve Type to All. Call it rdbuff200.
2. However, as an environmental center, it shouldn't be too close to a road. Let's keep it at least 50 feet away from a road. Buffer the roads of the watershed by 50 feet. Call it rdbuff50and set the Dissolve Type to “All” again.
3. Now that we have two separate buffer layers, we need to combine the layers into one layer describing distance to road. Take a peek at the tables for each buffer layer. You'll notice that each has field called Id. When we combine these two layers we want to keep all the attributes in the output. We’ll need a way to tell the 200 foot buffer from the 50 foot buffer after we combine, which means we need to add a unique value to each layer’s Id field.
Add the Editor toolbar to ArcMap (View | Toolbars | Editor). In the toolbar’s dropdown Editor menu, select Start Editing to start an Edit session. This step is necessary because you have to start an edit session to change individual cells in the table. Now open the attribute table for rdbuff50. Type “50” in the ID field of the first record. Now open the attribute table for rdbuff200 and type “200” in the Id field of the first record. Close the attribute tables, save your edits and end the edit session.
4. Now that each layer has a unique ID, let's combine these two different layers to make a single layer. For this, we need the Union tool.
In ArcToolbox | Analysis Tools | Overlay, open the Union tool to combine the two buffer layers. Call the output file rdcomb. Run the tool, then look at rdcomb’s attribute table. We want to find the features that are between 50 and 200 ft of the roads. In the table you’ll see the two ID fields from the original buffered layer: Id and an Id_1.
Id_1 will have a value of 200 in both records, and Id will have a value of 50 in one record and 0 in the other. Areas within the 50 foot buffer of the roads have the value of 50 in their field. Areas within the 200 foot buffer but not within the 50 foot buffer have a value of 0 in this field.
Condition 2: Must be on Available Land
The best site for the building will be on land that is not already developed but can developed. Sounds like we need land use information, so let's combine land use and the road buffer layers. Use the Union tool (ArcToolbox | Analysis Tools | Overlay | Union) to combine landuse and rdcomb into lbwmess. For our purposes, the only land cover type that’s available for development is forest. Of course, we’re simplifying a little—in reality, you’d also need to look at zoning laws, etc.Condition 3: Distance to Streams
Lastly, the site should be within 500' of a stream. After all, it IS a watershed group.
1. Use ArcToolbox to buffer the streams layer of the watershed by 500 feet. Call it streambuff500, and yes, Dissolve Type should be All.
2. Change the value in the streambuff500 Id field to "500". Don't forget to start a new edit session.
3.In ArcToolbox, Union lbwmess and streambuf500 layers into lbwtotal.
4. Look at the attribute table for lbwtotal. Note that there are several fields with names that begin with Id, each one containing a different buffer distance. Which field represents which buffer depends on the order of your unions and the order in which you added the layers when you performed each union. Confused? You should be. That's why we needed to ensure that each Id field had a unique value in it before unioning. Obviously, if we had buffered both the streams and the roads by 200 feet, we would have had to have used a different value for the streambuff Id field. This is one reason why it's important to take notes on your work, especially when you're working on big projects and are creating lots of layers (hint, hint).
The lbwtotal layer is a bit frightful. Let's try restricting the polygons to just those that meet our criteria. We can do this using the Select tool in ArcToolbox | Analysis Tools | Extract. Unlike in ArcMap, when you use the Select command in ArcToolbox you create a new layer with based on your query. Call the output file envcenters. You need to write an SQL Expression to select the best sites (click the SQL button to open the Query Builder). To recap, the best sites will be inside the stream and 200' road buffers (i.e., have a 500 and 200 in their respective Id fields), but outside the 50' road buffer (i.e., have a 0 in their 50' buffer Id field), and be in forested areas (i.e., LUCODE value of 4).
Part 2: Weight and Rate
What if the tiny slivers our analysis left us with were deemed unsuitable? Could we create a larger scoring system that describes the entire watershed? We’ll use a scale from 1 to 10, where 10 is more suitable and 1 is less suitable.
1. Add the following fields (as short integers, precision = 0) to lbwtotal:
2. Open the attribute table and select (using Select by Attributes under the Table Options) all of the polygons inside the 500 foot stream buffer. Then Calculate the StrmScore = 10 for all of those. The calculation will be performed only on the selected records?
Repeat this for each of the roads buffers. Make to sure select those polygons outside the 50 foot buffer and those that are inside the 200 foot buffer. For the landuse, make the forests (LUCODE=4) a 10, the farms (LUCODE=2) a 7 and the rest a 2.
3.Clear the selection, then calculate an additive total (i.e., LandScore+StrmScore+Rd50score+Rd200score) in the TotalScore field. Use the TotalScore field to view the layer with a graduated color scheme. (Hint: It will look better if you eliminate the outlines around the polygons. A shortcut is to right click on the symbols in the Symbology tab of the Properties window and select Properties for All Symbols.)
4. For a different view, we can differentially weight the criteria, allowing us to stress the more important ones and downplay the less important ones. Add a new field called WeighScore (type=short integer). Calculate a weighted total (i.e., 5* LandScore+7*StrmScore+2*Rd50score+3*Rd200score) in the WeighScore field. Take a look at the results in ArcMap using a graduated color scheme.
Assignment 7a: Lawrence Brook
What would happen if you
used a different set of buffers? What if you "scored" areas as more
suitable or less suitable, instead of discarding all areas that are unsuitable?
What if you introduced some address-matched data, like nearby schools? What if
you tried to locate the center near wetlands or lakes? The criteria you could
use for choosing sites are limited only by the available data.
Develop a site-selection system for the environmental center. You must use a weight and rate system. You must also use different buffers than we used AND you must use some address matched data. Use the street.mxc address locator (the one from Lab 5) in \\ad-rsc\data\teach\intgeo\ClassWork\middlese\MiddlesexAddressLocator.mdb. Hand in a map showing your selected site. Your map MUST be grayscale (design it in black/white/grays AND print on the black and white printer (ljup)) and should clearly explain your rationale and scoring on your map.
Assignment 7b: Kenny's Little Kernels
Namtrac Food Products, Inc. has developed a
frozen bag of sugar-coated veggies that kids love called Kenny's Little
Kernels. However, to make it a profitable venture, they figure they need to
produce and sell their product entirely within a single country. They've asked
you to help pick that country. But since they would also consider expanding to
many other countries (without shipping product across borders—in other words, the countries they expand to don’t need to be adjacent to the first country) you should
provide a scoring system that ranks all countries in the order that they might
proceed.
"Well," you thought, "What would improve their chances of
success?" A business book you found suggested that the following factors are key:
- a strong customer base
- good production capability
Copy the demog.dbf file from //ad-rsc/data/teach/intgeo/ClassWork/avdat/world/tables into your directory so that you can modify it. If you so choose, you can copy the whole world directory. Be sure to the use the cntry04 data layer for your geographic data. Your overall approach should be similar to finding a site for the environmental center in the Lawrence Brook Watershed, but in this case you will select the conditions that maximize the chances of success.
ONE solution would work
like this…
Come up with your conditions for success, including which conditions might be more important than others (i.e., relative weights). Remember, the key criterion are strong customer base (factor a) and good production capability (factor b). Some demographic data that might be useful include: Long Life Expectation (b), High Population (a), Stable Growth Rate (a), Lots of Kids (a), and High Convergence of Agricultural and Urban Land (a & b). Note that demog.dbf includes very helpful metadata for deciphering the field names. Records with the value -99 should be considered null values (i.e., not containing data).
Once you have your criteria, add some fields to the demog table and save it under a different name. Calculate a "score" from 1 to 3 in each of the new fields to reflect the degree to which the record satisfies that condition (criterion). Add a field for the total formula score and calculate the formula score with weights included. While you are at it, add a few different fields, and try a few different formulas. Save the edited table.
Join/relate the edited table to the world attribute table and map it.
Come up with your own global solution. You don't need to limit yourself to factors (a) and (b) above, but at least some of your scoring criteria should reflect the importance of these factors. Maps will be graded in part based on their readability and composition. PRINT ONE BLACK AND WHITE MAP. It should be highly communicative. What were your criteria and system? In which country should Kenny’s Kernels be sold?
Assignment due March 23rd. And remember, maps should be black and white.