Integrated Science for Enhanced Cartography
By Lawrence V. Stanislawski
Figure 1. Raster model estimating relative stream formation potential of the terrain surface determined from runoff, slope, and soil permeability and depth.
The National Geospatial Program is investigating integrated science approaches to enhance cartographic representations of The National Map. For topographic mapping purposes, features represented on a map must be logically integrated within local contexts, which vary spatially with terrain and other geophysical landscape characteristics. For instance, streams should follow valley bottoms and stay within flood plains shown by shaded relief or contour lines of the terrain. Relations among cartographic features, including terrain features, can be used to align map features with the appropriate level of detail at different scales. One example of this concept involves the use of terrain data and other geophysical characteristics to help guide revisions for The National Hydrography Dataset (NHD), which is the database of surface water features for The National Map. In this example, estimates of runoff, terrain slope, soil permeability, and soil depth are combined in a raster model to approximate the potential for stream formation (Stanislawski and others, 2012). Modelled potential values (Figure 1) are then used to weight a process that extracts drainage channels from elevation data. The weights help force extracted channels to follow the physical conditions of the terrain.
Figure 2. Sample of channels derived from an elevation model weighted by the relative stream formation potential of the terrain as estimated by the background raster dataset.
A sample of channels derived by this method at approximately 1:1,000,000-scale (1M) is shown in Figure 2. Local densities of the derived channels vary with the local geographic conditions; densities in the center and northwest parts of Figure 2 are higher than in the northeast section, where relatively lower stream formation potential exists. Without weights, elevation-derived channels form a homogeneous pattern of channels that do not vary in density to reflect the natural conditions of local geography. Patterns of derived channel density that vary with terrain conditions, as previously shown, can be used to revise the content of linear features in the NHD or National Atlas of the United States®. Automated comparisons of derived channels with existing hydrographic features can highlight areas where revisions can be focused.
Figure 3. Top panel shows National Atlas of the United States® 1:1,000,000-scale (1M) hydrographic lines (orange) without canals and intermittent streams overlain by channel lines (light blue) derived from a terrain model weighted by relative stream formation potential. The bottom panel shows derived-channel line density subtracted from the density of the 1M National Atlas of the United States® hydrographic lines. This method highlights features missing between the two datasets and can be used to prioritize revisions.
One technique creates a raster layer of line density for derived channels and subtracts a raster layer of line density formed for existing hydrographic line features. Figure 3 demonstrates this method for channels derived through the weighted elevation process and associated linear features from 1M National Atlas of the United States® hydrography. The bottom density difference panel of Figure 3 maps places where the two sets of lines either deviate positionally or where content is missing for one set of lines. This mapping can be used to guide revisions for the National Atlas of the United States® lines, and a similar process could assist NHD revisions.
Aside from refining the consistency of the content of linear hydrographic features, channels derived by relative stream formation potential provide channel density patterns to guide generalization of the high-resolution NHD. The high-resolution NHD is a multi-scale data layer that must be generalized to a common scale for proper use in map displays and scientific investigations.
Methods described in this article integrate soil and water science concepts and data to enhance and validate cartographic representations. Runoff and soil data were compiled by USGS water scientists, with soil information derived from the State Soil Geographic Database (McCabe and Wolock, 2008; U. S. Department of Agriculture, 1994). Similar integrated science approaches are being evaluated to enhance cartographic representations for other data themes of The National Map.
McCabe, G.J. and Wolock, D.M., 2008, Joint variability of global runoff and global sea surface temperatures: Journal of Hydrometeorology, v. 9, p. 816-824.
Stanislawski, L.V., Doumbouya, A.T., Miller-Corbett, C.D., Buttenfield, B.P., and Arundel-Murin, S.T., 2012, Scaling stream densities for hydrologic generalization. Seventh International Conference on GIScience, September 18-21, 2012, Columbus, Ohio, 6 p., accessed May 1, 2013, at http://www.giscience.org/proceedings/abstracts/giscience2012_paper_124.pdf.
U. S. Department of Agriculture, 1994, State Soil Geographic (STATSGO) Data Base: Data Use Information, National Resources Conservation Service, National Soil Survey Center. Miscellaneous Publication Number 1492. 113 p., accessed May 1, 2013, at http://dbwww.essc.psu.edu/dbtop/doc/statsgo/statsgo_db.pdf.