As an avid cartophile and hiker, I’ve always enjoyed working with contour maps—vector maps in which lines circumscribe areas of similar or equal value—as they provide just the right amount of context for the vertical aspect of a trail. I routinely use contour maps when evaluating hikes, since being able to see the densely-spaced contour lines at given points is much more helpful than just knowing the total elevation gain of a trail.
So, when I learned that the OpenGeo Suite 3.0 contains a process to generate contour lines from raster data, I jumped at the opportunity to play with it. Neat, I can create my own contour maps! My thoughts turned to Mt. Rainier National Park, where I spent some time this past summer; I was eager to test the contour generation process on the largest eminence in the contiguous United States.
Mount Rainier (Wikipedia)
Mt. Rainier raster data in GeoExplorer
The first step was to acquire topological raster data. Despite a somewhat cumbersome download process I was able to obtain high-quality data via the National Elevation Dataset from the USGS, provided. I eventually acquired a GeoTIFF with one arcsecond detail and loaded the data into GeoServer. While I could have taken advantage of new raster support in PostGIS 2.0, those features primarily support analysis, not visualization.
There are two ways of generating a contour map using the tools in the OpenGeo Suite: statically and dynamically.
I started out with the static case, processing the raster data and generating a vector layer of the contours using the Web Processing Service (WPS), which allows for server-side processing of data in GeoServer. With WPS, you can perform complex calculations and conversions, either by pulling from almost 100 built-in processes or by creating your own. For my case, the
gs:Contourprocess was sufficient for my needs. While quite rudimentary, I used the WPS Request Builder in the GeoServer UI to generate my contour bands, since it’s more straightforward than having to generate the raw XML code for executing the process.
Contour process in WPS Request Builder
The data that I possessed had elevation data in meters, from a low of about 500 meters above sea level to Mt. Rainier’s peak at 4,392 meters. For a nice round number I chose 100 meter increments and then ran this through the WPS process, piping the output back into GeoServer. In this case I saved the output as a shapefile and then imported the layer back into GeoServer, but I could just as easily have chained the output of
gs:Contour to the input of
gs:Import to accomplish the same thing.
With this vector data published, I styled the output by adding rules to label the lines and draw the 500 meter interval lines thicker, all features one would expect from a contour map. I also optimized my cartography for the web by only drawing the 100 meter bands at certain zoom levels to keep the map from becoming too cluttered when zoomed out.
Contour map of Mt. Rainier
Once satisfied that my style worked with the generated vector output, I set about creating a contour map dynamically by utilizing new processing features in OpenGeo Suite 3.0. Using rendering transformations eliminated the intermediate step of generating derivative data by instead applying the
gs:Contour process directly to the raster data using the layer’s style. I only needed to specify the transformation in the SLD and associate it with my raster layer to get exactly the same output in real time—without generating any new vector layers!
Contour map of Mt. Rainier in GeoExplorer
Though I stopped at this point, there are other ways I could have improved this further. For example, instead of using the SLD to selectively display the 100 or 500 meter intervals, I could have tied the interval parameter to the scale denominator of the map. This way, I could zoom the map in all the way and have bands with, say, 10 meter intervals, and then zoom out and render bands with 1000 meter intervals. This would further improve the interactivity of the web map without adding much overhead.
Learn more about creating a dynamic contour map with GeoServer WPS, including the SLD code used in the creation of this map.
Have you tried using GeoServer WPS processes? What have you accomplished with them? Let us know about your experiences by commenting below. And, if you’re ever in Mt. Rainier, I highly recommend the Skyline Trail—just be sure to bring winter clothing, even in summer and at lower elevations there’s likely to be snow!