In this portion of the lesson you are given a spatial dataset that is made up of the polygons that form the state boundary of Oregon. You also will download a second dataset of world state and province boundaries that we will extract just the Oregon state polygon from.
These are examples of what is typical of datasets that are available for the asking. Neither one is perfect. One has what appears to be a problem with its attribute table. When the datasets are overlaid, a noticeable mismatch is apparent.
You will use the tools from earlier in the lesson, as well as some techniques introduced here, along with something you learned in Geography 484, to decide which of the datasets is most accurate.
A. Download some spatial data — state and province boundaries from Natural Earth.
- First, in the oregon folder that is in your Lesson7 folder, create a folder named NaturalEarth.
Find out more about the Natural Earth dataset.
- Go to Natural Earth Downloads/
- Click the green Cultural button for the Large scale data, 1:10m.
- Scroll down to Admin 1 - States, Provinces, and click the link to Download states and provinces (geoDB). Notice that there are two versions of this data, one comes as a shapefile and the other as a feature class in a file geodatabase. So far in this lesson we have created and used personal geodatabases, but since file geodatabases work similarly and the data is already a feature class, download the geoDB version of the states and provinces boundaries.
- Note: If the geoDB version of the data isn't available, download the shapefile version and when you isolate the Oregon boundary in Section C, steps 9 and 10 below, make sure you then save it as a feature class in your geodatabase (so it has a ShapeArea field in its attribute table).
- Save the .zip file to your hard drive, and then extract the contents of the .zip file to your NaturalEarth folder. The file geodatabase folder should be named ne_10m_admin_1_states_provinces_geodb.gdb (or the shapefile version should be named ne_10m_admin_1_states_provinces.zip
- Note: If you obtained the geoDB version of the data, make sure the folder has the extension ".gdb" after the file name. If it does not, rename the folder so it does have the .gbd extension. Again its full name should be "ne_10m_admin_1_states_provinces_geodb.gdb". The geodatabase will not be visible in ArcCatalog or accessible in ArcMap without that extension.
B. Let's look at these data in ArcCatalog
When accepting geographic data from third party sources it is imperative that you at least check the integrity of the description of the coordinate system, making certain that you will be able to work with it.
Add to what you have already learned about metadada from the Concept Gallery.
- Open ArcCatalog.
- Bring up the Feature Class Properties window for the ne_10m_admin_1_states_provinces_geodb feature class (in the file geodatabase of the same name).
- In the XY Coordinate System tab, look at the information regarding the coordinate system: the datum is described in terms of D_WGS_1984. The Natural Earth data has a world-wide extent, so it is cast in terms of the world referencing system used today, the WGS 1984 reference datum.
Note also the units for the coordinate system are degrees. This simply means that the coordinates are in terms of longitude and latitude. Exit out of the Properties window.
- Let's preview the extent of the data you downloaded from Natural Earth. To Preview the data, right click on the feature class in the geodatabase (still in ArcCatalog). Choose the Item Description option - the Item Description window appears. Click on the Preview tab. Preview both the Geography and the Table options for the feature class via the drop down menu at the bottom of the Preview tab.
In the Geography option, you should see that you have a dataset of states and provinces for the entire world. And, viewing the Table option, if you scroll to the last columns of the (attribute) table you will see the Shape_Area column with values using decimal degree units. Do you know why these units are not usable for comparing different areas? Do you have an idea of what I am going to ask you to do with this data?
- In your oregon data folder is a coverage named or_boundary. Look at its Projection properties. It too is in lon-lat coordinates and based on a datum, but it is based on the NAD83 datum, which is a almost identical to the WGS 84 datum (for North America). We are going to compare these two Oregon state boundaries.
C. Extract the Oregon state polygon from the Natural Earth data
To make it easier for us to overlay and compare these two datasets, we need to export just the Oregon polygon from the Natural Earth dataset.
- In ArcMap, open a new empty map document.
- Save it to your oregon folder as oregon.
- Add the ne_10m_admin_1_states_provinces_geodb feature class. What does this cause the coordinate system of the data frame/map display area to be? (It should be GCS_WGS_1984.)
- Now, add the or_boundary coverage to the map.
You will see the increasingly-familiar warning regarding the fact that its datum is different from that of the data frame.
- Click the Transformations... button, and then look at the options available in the using: menu. For the continental U.S. we will transform the datum in which the or_boundary file is cast in NAD_83 (GCS_North_American_Datum_1983) into WGS84 (GCS_WGS_1984). Examine the settings in the Convert from: and Into: options. This transformation uses the NAD_1983_To_WGS_1984_1 option to transform the NAD 83 data to WGS 84. This transformation NAD_1983_To_WGS_1984_1 can be looked up on page 13 of the PEgt_v9.pdf supplied in your Lesson7data folder. Because we are in the contiguous United States, this transformation is wired into the project-on-the-fly capability of ArcMap as a default datum-level transformation. So in this case, if you did not chose a transformation, the two datasets would correctly overlay on screen.
- Click OK in the Geographic Coordinate System Transformations window.
- Then Close the Geographic Coordinate Systems Warning window.
- Zoom in on Oregon to see the how the two layers overlap and click their visibility off and on to see the differences.
- Using the Select Features tool, select the polygon that makes up the state of Oregon on the ne_10m_admin_1_states_provinces_geodb layer.
- Export the selected feature to a new feature class named oregon_ne (ne = Natural Earth), and for now save it in the ne_10m_admin_1_states_provinces_geodb file geodatabase.
- Click Yes to add the exported data to the map as a layer. At this point, and If you like, you can remove or hide the original Natural Earth layer from ArcMap.
- Save the map document.
D. Time to apply some of what we have been doing
You just got through working with two datasets that you did not create. It is healthy to be skeptical when it comes to free GIS data. If you have not already done so, look at the mismatch between the Oregon boundaries of the two datasets.
I want you to use some of the tools you were introduced to above to investigate this situation.
First, some information:
- Something is corrupt with the attribute table associated with the or_boundary coverage; after you click past an error message, it opens up empty.
- Both the or_boundary coverage and the oregon_ne feature class attribute tables have area fields (even though you cannot see the or_boundary table). The units of measure associated with the coordinate system of both coverages are decimal degrees, so the area values are in units of square degrees. Not something very useful.
- When you create a polygon geodatabase feature class by converting an existing polygon coverage, a new area field (Shape_Area) is created that holds the area values in terms of whatever the projection units are.
So, here is what I want you to do:
- Determine the total area for each of the two Oregon state boundary layers, and
- compare it to a third data source that I supply below.
- calculate the percentage difference between the independent values and the values you calculate from the data layers.
You will need to:
- Create a Personal Geodatabase called oregon.mdb. You will submit the feature classes you create in this geodatabase.
- Convert the or_boundary coverage into a polygon feature class. Again, the feature classes should reside in your oregon.mdb geodatabase.
- Create versions of the two Oregon feature classes (that you created from the Natural Earth dataset and the or_boundary coverage) that are in an appropriate map projection for assessing the area of polygons. (You can convert at the datum level, too, but it may be more analytically robust to retain the data in the datum in which they were originally cast.)
- Place all 4 Oregon feature classes in your geodatabase. Make certain they have names that indicate which dataset they came from and what coordinate system they are in.
- There are multiple polygons in the or_boundary layer. In order to calculate the total area represented you will need to sum the area values.
Here is an independently reported value for the area of the state of Oregon: 248,647 square kilometers. This value does not "include water." The only value I ran across that included water (lakes, etc., I assume) is: 98,386 square miles, rather than square kilometers; you will have to convert to square kilometers. You are also welcome to track down other reported values.
That is it for Part V, and for Lesson 7!
You have just completed Lesson 7. See the Lesson 7 Deliverables for details on what to submit for this lesson. Do not forget...if you have any questions, feel free to post them to the Lesson 7 Discussion Forum.