Sequestration is putting a bit of a damper on our travel this year, but we are still planning on attending and presenting at several conferences. Mamata is doing the bulk of the traveling, but our entire team is going to try to make it to State of the Map US in San Francisco in June. Here's a list of the conferences we'll be at this year:
Given the world famous landscapes of Grand Canyon, Yosemite, and other National Parks, including shaded relief on Park Tiles was an easy decision to make. Trickier to decide was just how to design a shaded relief that could do justice to this spectacular terrain while also serving as an unobtrusive background for multi-purpose maps. Another objective was creating a relief that complements the design standards found on the visitor maps produced by Harpers Ferry Center.
On Park Tiles, the shaded relief functions as a ubiquitous background. The relief is kept light to maximize the legibility of park areas, roads, rivers, boundaries, urban areas, and labels that appear on top of it. A relief that prints too dark would lessen overall map legibility. We also wanted a color that could stand on its own as an attractive and pleasing tone — thus, sterile gray was ruled out. After trial and error, we decided on light, neutral beige.
A beige shaded relief makes it possible to present other map information lightly and with clear visual hierarchies. For example, roads appear unobtrusively in a light red. The shaded relief also features reduced contrast for its highlights and shadows. This makes for a less “noisy” background—the shaded relief does not distract from park and thematic information on the map. And because a little tone exists even the brightest highlights, white state and international boundaries show up no matter how rugged the terrain.
We’ve also paid close attention to water. Tones for land and water are complementary yet provide enough figure-ground contrast to distinguish the coast. River and lakes use the same blue tint as the oceans, putting all water features on the same visual plane. Even more important, the land-water boundary appears without any thin coastal lines, making for a cleaner map. A feathered vignette applies darker blue tones along the immediate coast fading to lighter blue further offshore. This graphical effect gives emphasis to coasts without excessively darkening the map.
Creating the shaded relief and ocean vignettes was accomplished with a combination of GIS and graphical software. For example, coloring the shaded relief as beige and adding ocean vignettes occurred in Adobe Photoshop using the Geographic Imager plugin by Avenza, which maintains georeferencing. We used Natural Scene Designer 6.0 to generate the shaded reliefs, saved as GeoTIFFs. To get around the 4GB file size limit for GeoTIFFs, we created smaller, overlapping shaded reliefs for the very large areas covered by zoom levels 9 and higher. We then used GDAL VRT files to reference the multiple GeoTiffs, and then loaded these single pieces into TileMill, giving these relief pieces a seamless appearance on the final online map.
In a previous post, we introduced you to Park Tiles, an online basemap we are working on. In this post, we'll go into more detail about how we're serving Park Tiles and outline how you can use the basemap in your own web maps.
We've all seen the traditional mash-up pattern where data are placed on top of a basemap and, in turn, obscure the underlying (and important!) contextual information. Most “traditional” online basemaps used in web map applications have been designed for navigation first and foremost. Because of this, these basemaps tend to be information rich with a wide variety of highly-saturated colors to denote different road, landuse, and landcover types. The problem with these traditional designs is that information mashed-up with them doesn't always pop off the map and often gets lost. As the use of basemaps as backdrops for spatial information changes, the way these basemaps are designed and served needs to change as well.
Basemaps available through various providers are shifting towards a more minimalist and subdued style where a better figure-ground relationship is established between the basemap and the mashed-up content. This isn't a new idea. It is something print cartographers have done for years. For example, in a National Geographic Atlas, you don't see bird migration data on top of a road map with all of the reference information covered up. You, rather, see these data overlaid on a map background that supports the information instead of distracting from it.
As described in our introductory blog post, the design of Park Tiles is a subtle one. The shaded relief on the basemap is colored to make it sit nicely in the background and the other reference information is equally subdued. The features that are highest in the visual hierarchy are National Park units and their labels. This design ensures that Parks are clearly visible and information that is mashed up on the basemap comes to the foreground (if designed appropriately).
Another pattern that has emerged over the past few years is a change in the way basemaps are served. More and more, we're seeing basemaps that are separated into various components. Some examples include Esri's Terrain with Labels and Light Gray Canvas Map. Both of these maps gives users the ability to put reference information that is typically underneath mashed-up content on top. MapBox's suite of basemaps goes one step further, giving users the ablility to reorder layers, turn groups of features on and off, and recolor groups of features dynamically. With the release of their new basemap designs, Stamen has also created a reference overlay from their Toner style that can be used to provide geographic context with other basemaps. Two of the first online basemaps to demonstrate these new patterns were Esri's Terrain with Labels (released in 2009) and GeoIQ's Acetate basemap (released in 2011).
We have been thinking about how we can take this pattern one step further. We have decided to “explode” Park Tiles into multiple layers. What this means is that each feature on the map is served on its own.
We decided to do this for a few reasons:
We want to support multiple overlay types - including points, lines, and discontinuous/continuous polygons.
We want to get the greatest amount of flexibility out of our basemap so we can mix and match layers to support any number of projects and applications.
We feel like this is new pattern is worth exploring.
Take a Look
You can use our Park Tiles Viewer to explore this pattern in more detail. In the bottom right-hand corner of the map, click the "Toggle Layer View" button.
Under the Overlays layer section, you will find two pre-generated overlay examples that are discussed in more detail below.
Here are some examples of how this pattern works with line and continuous polygon data:
National Trails (Lines)
We want our National Trails data to be shown over Parks and water features, but, as seen in the image below, we don't want them to overrun labels for Parks, cities, and hydro features.
In this case, we can put the tiles for our National Trails underneath the tiles for Park, city, and hydro labels.
Ammonium Deposition in Parks (Continuous Polygons)
The traditional mash-up pattern is to overlay contiguous data on top of a basemap while using transparency to allow the underlying geographic context to be seen.
A more desirable view of the Ammonium Deposition overlay would be a view where the context (roads, labels, Park polygons, etc.) are on top. In this case, we would put the Ammonium Deposition tiles above the shaded relief but below all of the other reference information. The result can be seen in the image below.
The above example still has some issues. The urban areas (symbolized in a pale yellow) can be misinterpreted as data and it is difficult to see the variation in Ammonium Deposition in Rocky Mountain National Park because Park polygons are symbolized with a fill.
The solution is to turn off the urban area and symbolize Park polygons using outlines. This way, there is no other polygon data interfering with the continuous surface of Ammonium Deposition.
By serving Park Tiles as individual tilesets, we're giving ourselves the flexibility to use the individual components of the map for a variety of projects. For example, in some of the Park projects we work on, the focus of the map is only on a subset of Parks managed by the Park Service. In those cases, we would not serve our layer with all of the Parks displayed, but would instead create an alternative Park layer to display only the relevant Parks on the map. The number of use cases addressed by this basemap is huge. Each layer of the map can be mixed and matched into a variety of web maps, and each layer of the basemap can be ordered in a way that best tells the story of the overlaid data.
How to Use and Access Park Tiles
Whether or not you are a Park Service employee or partner, you may find our new basemap and the pattern we've implemented useful in some of your projects. If you'd like to learn more about how to access these tilesets and use them in your own applications, take a look at the Using Park Tiles tutorial in our support documentation.
One of the major projects we have kicked off over the past few months is an online National Park Service-styled basemap we are calling Park Tiles. In this blog post, I'll briefly discuss the design rationale behind Park Tiles and give a brief overview of the data and technology we are using to drive this project.
As more and more Park Service groups and partners move to the web, we've found a need for a customized NPS basemap that can be used as a backdrop to display the wide variety of spatial data that are collected and maintained by the service.
Currently, the NPS uses basemaps from various providers, including Bing, Esri, Google, and MapBox. For certain projects, however, we are looking for more stylistic control over these basemaps. Some of these providers allow us to modify colors and turn some layers on and off, but often when we mash up our content on these basemaps there is a lot of overlap and our information doesn't come to the foreground as much as we would like.
Park Tiles is not meant to be a total replacement for any of the basemaps we currently use. It is, rather, meant to give us an alternative basemap we can use in custom projects where we'd like to see our National Parks highlighted. We'd also like to incorporate more detailed park information (such as campsites, visitor centers, trails, etc.) at larger scales that can serve both NPS employees and visitors.
Most importantly, the National Park Service has a strong cartographic tradition that has been developed by the cartography team at the Harpers Ferry Center in West Virginia. Our primary goal with this online map is to transfer that tradition to the web. We are doing this by using NPS fonts, colors, and symbols. It is also important for us to maintain the graphic identity of the NPS so when we embed web maps in park pages, the maps compliment the existing style.
We have designed Park Tiles with a subdued style and a limited number of colors and information. As you can see below, the shaded relief and reference information on the map are designed using colors that are low in saturation. To contrast, the National Park points, polygons, and labels are symbolized and labeled using colors with higher saturation so they come to the foreground. We have limited the features displayed on the map knowing that if we need to add additional reference information, we have the freedom to do so. All of this has been done with the goal of clearly identifying National Park units and creating a basemap that can support multiple overlay types symbolized using a wide variey of colors.
The data we are using to build this map come from a variety of publicly available sources. The ScaleMaster diagram below shows the data source of each feature by zoom level (or scale). We are using data from the NPS at all scales for park units. Natural Earth data is used for reference information like administrative boundaries, administrative labels, roads, urban areas, and hydrography at small scales. We are using DEMs of varying resolutions for relief from NASA and the USGS.
At larger scales (zoom level 8 and above), we begin to switch data sources. Urban areas switch from Natural Earth to the US Census Bureau, coastlines are drawn from NOAA, and the majority of other features come from OpenStreetMap (OSM). Specifically, we are experimenting with OSM data at large scales to display features like points of interest, trails, and more detailed base data.
Larger scale designs (zoom level 10 and above) are still being prototyped. We prefer to use data from Parks, when available. In cases where individual Parks do not have data, we are currently exploring alternatives, particularly OSM.
We are using a combination of technologies to build and serve this map. To read in data from OSM we are using a couple of different approaches, including Imposm and osm2pgsql. Both of these tools allow us to manage and query OSM data using PostgreSQL and PostGIS. Park units are managed in a SQLite database, and the remainder of the data (for now) are managed in shapefiles.
The primary tool we are using to design Park Tiles is TileMill. TileMill is an open source design studio where multiple data formats can be read, queried, and styled. To create the shaded relief, we're using Natural Scene Designer and Photshop. For icons like highway shields and POI's (at larger scales), we are using Inkscape. We're also using some other tools, including Dymo for labeling and High Road to query and symbolize our road layers at zoom levels 8 and above.
Take a Tour
So far, Park Tiles is available for zoom levels 4-9. We're actively working on incorporating more zoom levels – we'll keep you updated!
You can view Park Tiles here. If you click on the "Switch" button, you'll see our initial large scale prototype of Mt. Rainier National Park that uses all OSM data.
Look for a follow-up blog post where we'll describe, in detail, how we are looking to serve our basemap.
We plan on updating the presentation as we build new web maps and release new versions of our library, so it probably makes sense to consider it a "living document". Enjoy, and please let us know if you have any suggestions to help us improve our communication! You can reach us by email: npmap at nps dot gov.