Intro to Atlas GIS 2.0
Mackenzie, Tanjuakio and Sparco
Intro to Atlas GIS 2.0
The main part of this section introduces some functions at the very heart of GIS analysis: splitting one geographic layer by another, apportioning attribute values among portions of split features, creating buffer zones, etc. There are countless possible uses for these functions; we cover a few illustrative applications here.
AtlasGIS's Edit-Geographic-Move-Position function lets you move geographic features around on your map. Strategic Mapping's documentation suggests you might want to shift a copy of a feature slightly and shade in the original so that the feature looks elevated. You might use this to re-position a grid on your map. Or you might just want to see what Delaware would look like if it were located in the middle of the Gulf of Mexico.
To move a selected feature a precise direction or distance relative to its original location, use the F6 key to specify the azimuth and distance. If you move a cluster of selected features at once, specify a target location for the center of the cluster relative to the original center of the cluster. To move a selected feature or cluster of selected features to an absolute location (specified in lat-long), use the F5 key.
Edit-Geographic-Turn lets you rotate a selected feature about its centroid, or a cluster of selected features about the central point of the cluster. Edit-Geographic-Resize lets you rescale vertical and/or horizontal dimensions of selected line or region features. Resizing multiple selected point features moves them closer to each other.
The Edit-Geographic-Vertex functions let you add, delete or move individual vertices of features. You have to select a feature singly (Select-One) before you can edit it. Edit-Geographic-Vertex then highlights the component vertices of the feature; you pick one at a time with the mouse. If you want to make vertex changes permanent, Save these changes before you return to the main menu. Be aware that when you change a vertex of a geographic feature, the boundaries of all features which share that vertex will be changed. This includes features in other geographic layers!
It may be easier to shift complex boundaries by simply creating (with the mouse or digitizer) the new boundary as a new feature with Edit-Geographic-Add) and then Union or Split the feature to correct it. Another option is to use the EDITVERT script, which lets you edit multiple vertices quickly. This script is accessed via File-Run (see Part 9).
You can browse and edit records in the active geographic and attribute files in tabular form with Edit-Attribute-Browse. Any white cell in a browse worksheet can be edited by simply typing over it. Be careful! Any changes you make will be saved with the geographic or attribute file.
The /-View-Filter tool restricts the Edit-Attribute-Browse table display to a desired subset of selected records. /-View-Order sequences records in any desired sort order. File-Attribute-Index-New lets you create an index file (*.NDX) for quick re-sorts of your combined geographic/attribute file. First name the index file, then enter the dBase-style expression on which the attribute file should be indexed. To browse the attribute file according to this index, use Edit-Attribute-Browse, with the /-View-Order tool, and choose the *.NDX file you created from the pick list.
You can select geographic features and associated attributes directly from the Edit-Attribute-Browse spreadsheet by toggling records with the Spacebar, F3 or F4 key in the "Sel" column (the left-most column in the spreadsheet). Selected records are indicated by little white squares. A full set of logical select tools is also available from within the browse worksheet.
You can add a new field in an attribute file with File-Attribute-Tools-Structure. Move down to the last (blank) row of the Structure spreadsheet and edit the characteristics of the new field (NAME, TYPE, SIZE, etc.) as desired. By default, this new field will appear as the rightmost (empty) column of the Edit-Attribute-Browse spreadsheet. (You can restructure the order of the field if you wish.)
To put data in this newly-created field, use Edit-Attribute-Replace-Expression, specify the new field in the pick-list, and enter a database expression. Or you can edit the Edit-Attribute-Browse spreadsheet directly, using the /-Edit-Replace-Expression tool.
The /-Edit-Replace-Distance tool lets you fill an attribute field with linear distances of geographic features from a reference feature.
To add a new geographic feature to an existing or new layer, use Edit-Geographic-Add. Specify the target layer or name the new layer, then specify the new feature's _NAME and _NAME2 in the entry form, and then place the new feature's vertices on the screen with the mouse or digitizer tablet.
The / key calls up the Tools menu. You can Grab existing vertices or Common boundary segments for inclusion in your new feature. (This helps prevent accidental creation of slivers due to re-ditizing shared boundary lines.) You can also zoom or pan the map. These tools are discussed in more detail in the digitizing section (Part 6).
To delete geographic features, select them and use Edit-Geographic-Delete-[Partial/Delete]. Delete-Partial deletes the geographic feature but not its associated attribute record. Delete-Delete deletes both. Remember: deleted geographic features remain in your file, merely tagged as "deleted," until you fully erase them with File-Geographic-Tools-Compress.
Edit-Geographic-Copy-Layer copies geographic features from one layer to another; Edit-Geographic-Copy-Position copies features from one position to another within the same layer.
Edit-Geographic-Union merges selected multiple features into a single feature, dissolving common boundaries of adjacent features and combining non-adjacent features as islands. . Specify the new feature's _NAME and _NAME2 in the entry form, then specify how you want each attribute field aggregated: Sum, Average, Weighted Average (specify the weight field), minimum, maximum, count, blank or first.
Edit-Geographic-Split-Overlay splits one geographic feature (line or region) by another;
Edit-Geographic-Split-Islands splits multi-island features into separate features.
Edit-Layer-Add lets you create new region, line or point layers in the active geographic file. You can put features in this new layer with Edit-Geographic-Add using the mouse or tablet and the associated Tools (/) commands.
Or you can copy selected features from other layers into this new layer using Edit-Geographic-Copy-Layer. This is useful for creating aggregates (Edit-Geographic-Union) or splits (Edit-Geographic-Split) of features without losing the original features.
When you copy point features to a line or region layer, you create circular polyline or region features. AtlasGIS asks you to specify the radius (in miles) and number of vertices (minimum 3, forming triangles, maximum 4096) for these circles.
When you copy line features to points, you obtain the center points of the lines. (Note that the centroid of a line in AtlasGIS is located on the line, not inside the feature.) When you copy region features to points, you obtain (interior) region centroids.
Be careful when you copy line features to a region layer: AtlasGIS performs a linear closure of each unclosed arc. This may not be what you want. For example, if you convert an S-shaped linear feature to a region, you will obtain a feature comprised of two islands touching at a common point.
When you copy region features to a line layer, you obtain closed polyline perimeters.
Suppose you wanted to define a new region bounded by multiple existing line features. (Say you are trying to import polygons from some other GIS, but can only get them as boundary line segment features which you now have to piece together.) If you Copy multiple line features directly into a region layer, AtlasGIS will automatically close each line feature on itself, and you will get multiple odd-looking region features around the perimeter. You could Union the line features forming a polygon first, and get a correct region, but any line features which are shared boundary segments would then be unavailable for incorporating into adjacent regions.
The solution is to create (Edit-Layer-Add) a new temporary line layer (call it "Junk" or whatever) into which you can copy and then union sets of line segments as closed polylines. You would select sets of line features from the Highways layer which circumscribe individual regions you want to create, and copy these into the temporary line layer with Edit-Geographic-Copy-Layer. Note that the line features in the Highways layer are de-selected and the copies in the temporary layer are now selected. Use Edit-Geographic-Union to combine these into a single line feature. Then copy the single closed polyline feature into the new region layer. If you are creating multiple region features in this manner, you must remember to deselect the newly-created region feature before selecting source line features to create another region feature.
Note: This technique is particularly useful when you digitize region features as individual line boundary segments and then want to piece the segments together to form perfectly contiguous region features (see Section 5).
The Edit-Geographic and Operate menus support similar operations: overlaying layers, aggregating or splitting one layer by another, etc. You can merge multiple geographic features into composite features with Edit-Geographic-Union or Operate-Union, or split geographic features in one layer by features in another layer with Edit-Geographic-Split or Operate-Split. In general, the Edit-Geographic functions are meant to be used on individual features, while the Operate functions are intended for wholesale use on sets of selected features or even entire geographic layers. Operate includes additional functions as well: you can create buffer rings or zones around geographic features, overlay regions on lines and transfer region attributes to left and right sides of lines, and generalize (remove unnecessary vertices from) geographic features.
Operate-Split splits an "input" layer by "overlay" layer to create a "target" intersection layer of new features with associated new attributes, and apportions attribute values from the input and/or overlay features to the new features. You specify how you want each attribute field to be apportioned. The new features can take on attributes values of either source layer; numeric attributes can be weighted by the area of the new region relative to the area of the source region.
Note: Operate-Split only works when the overlay layer is regions; the input layer may be either lines or regions. To split regions by lines, you can extend the lines (perhaps copied to a temporary line layer) to create encircling polylines, then copy these to a temporary region layer which you would use as the overlay layer.
Operate-Union combines geographic features from an input layer via a grouping expression into composite features in a target layer. You can specify how you want each attribute field aggregated: summed, averaged, weighted-averaged, minimum value or maximum value, or count. Union aggregates sets of points into one point at their average location. It aggregates touching line features into single lines; non-touching lines may be aggregated as islands or remain separate features. It aggregates touching regions as single regions; non-touching regions may be aggregated as islands or left as separate features.
You can Union regions into regions or points; lines into lines or points, or points into points. If you Union regions into target point layers, AtlasGIS locates each target point at the area-weighted average of the centroids of the source regions. If you Union lines into points, AtlasGIS averages the centroids of the source lines.
Operate-Split, discussed above, splits geographic features and splits (apportions) attribute values associated with those geographic features as well. Operate-Union, discussed above, aggregates both geographic features and their associated attribute data. The other Operate functions discussed below split, combine or pass attribute data between different geographic layers, but do not necessarily split or combine geographic features themselves.
Operate-Total-Inside aggregates attribute data from multiple features (point, line or region) in an "Input" layer into attribute data for features in a "Target" layer. On the command screen you specify: the input layer which will be aggregated; the target layer which will have the new aggregate feature; how you want attribute data from the input layer split (apportioned by area, or 100% to the target feature containing the input feature's centroid); what input layer attribute fields you want aggregated; what target layer attribute fields you want created; and how you want each attribute field in the input layer aggregated (Sum, Average, Weighted Average--you have to specify which field provides the weights, Minimum, Maximum or Count).
For example, given a geographic file containing a Census block region layer, with associated population attribute records, and a watershed region layer lacking data in the population attribute field, you could calculate the population in each watershed by using Operate-Total-Inside to determine population totals for each watershed.
Operate-Total-Matching aggregates attribute data from an input layer to a target layer via logical matching expressions. Again, you specify how you want each attribute field in the input layer aggregated.
Operate-Inside assigns attribute values from region features in an "overlay" layer to the features in an "input" layer whose centroids lie inside those regions. You can overlay a
region layer on an input layer of smaller regions, lines or points. Operate-Inside can also assign attribute data to a point input layer from the nearest points or lines in an overlay layer. These functions are typically used to fill in missing attribute data with regional averages or whatever.
Operate-Left/right overlays a region layer on one or more input line layers, and assigns region attributes to left- and right-side attributes of the line features. For example, some roads delineate Census tracts, and this function can create new road attributes identifying the Census tract on either side of road features. (This is already done in TIGER/Line data.) The AtlasGIS documenation notes that a line feature may either (1) be inside a region; (2) bound a region on the "right" (its sequence of vertices matches the clockwise sequence of the region); (3) bound a region on the "left" (its sequence of vertices is the reverse of the sequence of vertices in the region); (4) be outside the region; or (5) be indeterminate vis-a-vis the region (some vertices of the line feature are inside, bounding or outside the region while others are not). Note that you can split (Operate-Split) any indeterminate line features as necessary so that all input line features are determinate.
Operate-Buffer creates buffer regions (to define zones of influence or whatever) around geographic features. Operate-Buffer-Ring creates buffer rings around point features; Operate-Buffer-Zone creates buffers around line and/or region features. Operate-Buffer creates one buffer region per selected feature. You can merge overlapping buffer regions as necessary with Edit-Geographic-Union.
You can specify both inner and outer buffer radii; this lets you create multiple concentric buffer regions around features. You can create variable buffer widths by using an expression to define radii.
Note: you can also create buffer regions around points using Edit-Geographic-Copy to copy points to a region layer.
Operate-Generalize thins (removes vertices providing unnecessary detail from) region and line layers, writing thinned features to a new layer. The thinning algorithm calculates how much each vertex deviates perpendicularly from an imaginary line drawn between its two neighboring vertices. If the perpendicular distance is less than the user-specified value, the vertex is eliminated, and the line or region boundary is slightly smoothed. This utility can substantially reduce file size, conserving disk space and speeding up processing.
We briefly discussd the principal tools for controlling map and page displays in Chapter 2, and one of the Exercises in Chapter 2 introduces many of these tools. To recap, Display-Layer-Settings controls overall displays of map features and labels. The Display-Page commands control the displays, sizes, positions, texts, fonts, etc. of page elements. This section notes some additional tools for controlling displays of map labels, and for adding freehand annotations to the page.
The Display-Layer-Settings control sheet lets you choose how to display feature labels: Visible (always), Fits (visible if they fit), Hidden or Off. You can refine your labeling with the Display-laBel commands which control the display and positioning of individual feature labels or sets of labels.
A frequent problem with road features from TIGER line files is that every segment of road has a label, and if you display them all, your map is virtually blacked out with over-written labels, while few labels will actually fit between segments if you use the "Fit" option. Unique displays only one instance of each type of label.
AtlasGIS offers a broad range of freehand page annotation tools (Display-Freehand) for adding and editing text, symbol and freehand graphic objects anywhere on the page. These are easy to figure out when you need them. Experiment with them when you have time.
The objective of this exercise is to analyze the spatial distribution of hardwood, softwood and mixed forest biomass by zip code in Delaware, and relate this to the incidence of residential use of firewood as a primary heating fuel.
1. Merge the FOREST geographic file containing forest polygons in Delaware with the DE_ZIPS geographic file containing Delaware zip code point and region features. The forest polygon file _NAME2 field specifies whether the polygon is deciduous (code 41), coniferous (code 42) or mixed (code 43). Create a thematic map showing the forest polygons in well-differentiated shades of gray or green within labeled zip code regions. Add appropriate titles and other notations, and print this map.
2. Import the STF3B Census file for Delaware, by zip code, including the fields H0300001 through H0300009 (occupied housing units using various types of heating fuel). The field H0300006 represents the number of households using wood as a primary heating fuel in each zip code region. Match the CENSUSID field in the Census file to the _NAME field in the geographic file. Name the new attribute file whatever you want. Create a 7-level thematic map showing percents of households using firewood as a primary residential heating fuel, by labeled zip code. Add appropriate titles and other notations, and print this map.
3. Calculate the land area in each type of forest (deciduous, coniferous or mixed) within each zip code region. Calculate the total forested land area in each zip code region. Calculate the percents of each zip code area in each type of forest, and the percent in any type of forest. (You will have to restructure the attribute file, adding new fields as necessary to receive these values.)
4. Export the necessary data to a spreadsheet or statistical analysis program where you can analyze the correlation between biomass availability and fuelwood use by zip code. Which is more closely related to the incidence of household reliance on fuelwood--hardwood density or overall forest density? Returning to AtlasGIS, what percent of the standing hardwood biomass would you estimate is used as residential fuelwood annually in each zip code region?
The objective of this exercise is to forecast population growth by Census block between 1990 and 2000 in Southern New Castle County, using information on land development plans filed with the New Castle County Planning Department. Since many of these developments are only conceptual at this point, you will be mapping and analyzing them as circles.
1. Merge the SNCC_BLK and NEW_DEVS files. SNCC_BLK contains Census block regions for New Castle County south of the C&D Canal. NEW_DEVS contains 71 points representing current or pending developments. Load the block-level Census data file MOTSTF1B, which contains STF1B (block-level) data on population and housing units. Create a 7-level thematic map showing housing units per acre in 1990 (assume Census blocks with missing data don't contain any housing units). Add appropriate titles and other notations, and print this map. Extract summary statistics on 1990 population and housing units for this area, and calculate the average number of occupants per housing unit.
2. Create two new fields in the attribute file to store acres to be developed and numbers of new housing units to be added from the new developments. Enter the acres and housing units planned in these fields for each development point feature (from the table in the County Planning Department's draft report). Create circular buffer regions representing each development, with each circle sized to encompass the acreage to be developed. Apportion the new housing units in these development circles to the Census block regions by relative area. Calculate how many new housing units will be added to each Census block when these developments are completed as planned (about the year 2000). Create another 7-level thematic map showing future housing densities per acre (1990 units plus new units, divided by area), using the same thematic settings. Superimpose the new development circles.
3. Extract summary statistics on projected (ca. year 2000) housing units. Calculate the percent increase in housing units for this portion of the county. Note that the official growth projections for the area currently call for 12,470 new housing units and 33,400 new residents in this area by the year 2020 (Delaware Population Consortium). How realistic are these projections according to your analysis?
The objective of this exercise is to quantify gross and net changes in acreages in each land use category in Sussex County, Delaware, between 1974 and 1984.
1. Load the mapfile file SXLULC which contains the geographic file SXLULC. The mapfile contains pre-defined settings for a thematic map of the LANDUSE layer, which contains 1984 Land-Use/Land Cover (LULC) regions in Sussex County with classification codes _NAME field. This map layer was developed from Landsat MSS satellite imagery.
The SXTIGER file contains TIGER line feature layers for the county. You can extract some of these layers to a separate file, and merge them with the SXLULC are layers are also included in the geographic file for visual referencing. (NOTE: Since you'll need a clean copy of the same geographic file for Exercise 4.4, you may want to close it immediately and File-Geographic-use_As a copy of it for this exercise.)
Merge in the SXLU74CH file, which contains a layer of regions which underwent a land-use change between 1974 and 1984. This layer's _NAME field contains LULC classification codes for 1974. This map layer was developed by comparing aerial photos from 1974 against the 1984 LULC map.
2. Create the following thematic maps with road, rail and water
features overlaid for visual orientation:
(a) 1984 LULC (basically the SXLULC mapfile with any cosmetic
changes you want to make)
(b) Change regions showing their ex ante (1974) land uses
(basically the SXLU74CH map using the same map layout and thematic
settings).
(c) The same change regions showing their ex post (1984)
land uses. (Use Operate-Inside to extract these regions'
1984 classifications from the 1984 LULC regions in which they
are located, inserting these into the _NAME2 field.)
(d) 1974 LULC (essentially "backing out" the change
regions from the 1984 LULC map).
3. Construct an 8x8 land use transition matrix, with columns representing Level-1 1984 LULC categories (the first digit of the classification code is the Level-1 category), rows representing Level-1 1974 LULC categories, and each cell in the table reporting the total number of acres transferred from each 1974 LULC category to each 1984 LULC category. When you have extracted all these values from AtlasGIS, enter them in a spreadsheet table.
4. Calculate --
(a) total acres in each LULC category in 1984;
(b) gross acres lost from each LULC category between 1974 and
1984 (row sums from your transition matrix);
(c) gross acres gained by each LULC category between 1974 and
1984 (column sums from your transition matrix--note that your
total of row sums should equal your total of column sums);
(d) net acres gained (lost) by each LULC category (gross acres
gained minus gross acres lost); and percent change from 1974 in
acreage in each LULC category.
(e) total acres in each LULC category in 1974.
The State of Delaware is evaluating 3 candidate parcels in Sussex County as potential state wildlife management areas, and needs an evaluation of each parcel's suitability as habitat for Delmarva fox squirrel. Remnant populations of this endangered species are only found in large, unbroken tracts of mixed forest.
1. Load the mapfile file SXLULC which contains (a clean
copy of) the geographic file SXLULC. The mapfile contains pre-defined
settings for a thematic map of the LANDUSE layer, which contains
1984 Land-Use/Land Cover (LULC) regions in Sussex County with
classification codes _NAME field. The HIGHWAYS layer is included
for visual referencing; the PARCELS layer contains the three candidate
parcels to be evaluated. Select all the LANDUSE regions with
_NAME="43" (mixed forest regions).
(a) How many of these are larger than 2 square miles?
(b) How many mixed forest regions have concentration ratios
(defined as _AREA/_LENGTH) greater than 0.12?
(c) Do any of the 3 candidate parcels contain any of these large
or high-concentration mixed forest regions?
2. Create 1/4-mile buffers around each parcel. Select the 3 PARCEL regions, the 3 buffers, and all LANDUSE regions touching any parcel of parcel buffer. Write these to a separate geographic file. Close the SUSXLULC geographic file, load this new and merge in the SXTIGER line file containing complete roads, rail lines and water features. To speed up subsequent processing, you may want to select just the TIGER line features touching a parcel or parcel buffer; then write the parcels, buffers, LULC regions, and all touching line features to a separate file: this smaller file will be much quicker to analyze. Restore the original thematic map settings.
3. Determine the acreage and perimeter of each parcel (1 square mile = 640 acres).
Split the LANDUSE layer by the PARCEL layer, putting the split
regions into a new layer.
(a) How many acres of each LULC category are located inside each
of the 3 parcels?
(b) Calculate the concentration ratios of the mixed forest regions
in each parcel. In your judgment, which site has the best overall
concentration of mixed forest?
4. Create buffers around the neccesary features to determine how
many acres in each parcel are located--
(a) within 1/10 mile of any highway
(b) within 1/10 mile of any rail line
(c) within 1/20 mile of any secondary road
(d) within 1/10 mile of any stream.
5. Create a map of LULC regions in each parcel, showing: parcel region boundaries and parcel buffers; the highway, rail line, secondary road and stream segments located in each parcel or parcel buffer; and the buffers regions around these line features. Knowing that the Delmarva fox squirrel prefers mixed forest habitats near streams and away from roads and rail lines, highlight the ideal habitat areas in each parcel map. Print the 3 parcel maps and a county map showing the location of each parcel.
Rank the 3 parcels' suitabilities as habitat for this species.