Having all these web pages based on latitude/longitude, in 2008 I decided to add the TN county high points to my web database. I was astonished to learn that the county high point data was based on narrative descriptions of the USGS 7.5' topo quadrangles -- e.g. "3/4 mile NW of Smith Cemetery" -- and there was no lat/lon data. To my further dismay, many counties had multiple high points. I decided to use the USGS digital elevation data to find the highest point in counties where there were multiple occurrences of the highest contour line.
I have always been fascinated with the accuracy and detail of the USGS 7.5' minute topo maps and have accumulated quite a collection of the maps over the years The elevation accuracy of the USGS 7.5' topo maps is within one-half of the contour interval, e.g. 10' for map with 20' contours. Spot elevations (e.g. the little x on the topo map to the right) are accurate to within one-third of the contour interval. Benchmarks are surveyed and have vertical accuracy of less than an inch. The USGS also provides digitized elevation data (DEM's, digital elevation model) for each topo quadrangle -- in either 10m or 30m grids. The elevation accuracy of this data is typically the same as the topo maps as they are derived from the same photographic data.
I knew that Peter Barr suffered from the county-high-point obsession,
so I told him of my digital elevation work. Peter then wanted to use
the digital elevation data to resolve TN/NC mountain peak ambiguities:
-which of the Chiltoes peaks is higher?
-which is higher Tricorner Knob or Mt Yonaguska?
-where is the high point on Mt Sequoyah?
-where is the high point on Big Cataloochee?
-and many more ...
NC LiDAR data
In March, 2009, Peter discovered that a North Carolina flood-plain project had just completed a LiDAR elevation mapping of the state -- to the tune of $23M. The data provides elevations accurate to less than one foot and at a one to three meter density. Although the focus of the study was the NC coastal areas, the data included the western mountains. A helpful NCGIS staff member provided Peter with LiDAR elevation maps (pdf's) of various peaks of interest. As Peter said, "This is going to revolutionize NC summit rankings and known elevation listings."
Peter forwarded me some of the map pdf's and noted that the Big Cataloochee pdf was interesting in that it showed the high point on a steady rise 1000' north of the spot elevation. (The spot elevation is a light yellow triangle on the pdf.) I fetched the 10m DEM data for the Luftee Knob quadrangle and plotted the grid on a Google map. The 10m DEM data was consistent with the spot elevation and didn't show anything like on the LiDAR map. There are 8 "high points" of 6152'.
The USGS had recently made available a "seamless server" so you could fetch elevation data and such based on a region of your choosing and not have to stitch together 7.5' quads. The seamless server also had the most recent data, so I downloaded elevation data for Big Cataloochee. It matched the NC LiDAR data, and in fact was based on the NC LiDAR. So the USGS data indicated that the high point was north of the spot elevation, and more surprisingly, reported the elevation at the "6155' spot elevation" as 6057'.
I had not bushwhacked to the top of Big Cataloochee, but Peter said he had been three times, going by GPS to the spot elevation. He recalled the forest thick and the terrain gently rolling. He didn't really recall a continuing up slope, but ... ?
The conflicting data was too much to endure -- a field trip was needed to establish "ground truth". Once the park service opened Balsam Mtn Road my wife, another off-trail buddy, and I headed up to Big Cataloochee. We had an altimeter and GPS loaded with coordinates for the spot elevation and the LiDAR "peak". There is a benchmark at Ledge Bald (5184') only 3 miles from the summit, so we could re-calibrate the altimeter. Based on "voting by number of survey ribbons" we reached the apparent high point of Big Cataloochee. There is a small rock outcropping (see photo to right) at this summit (35.67215 -83.17584) which is not far from the spot elevation. The area is undulating and in a thick forest so it is hard to say where the high point is. The altimeter read 6150', and there was no topography that matched the LiDAR map. Guided by the GPS, I did venture off toward the "LiDAR peak", but the terrain rolled and mostly headed slightly down. Clearly, the one-foot accurate LiDAR data was way off -- a phantom peak!
Back at home I downloaded other sources of digital elevation data, SRTM and ASTER. They both showed the high point near the spot elevation. Here's a Google topo map with all the "candidate" high points for Big Cataloochee.
After a few emails, I finally got someone's attention in the NC floodplains mapping group. Peter and I had a few conference calls with the NC folks and sent them our evidence. They discovered they had some "over-scrubbing" issues with several sections of the western NC mountains -- on the edge of their LiDAR data. In June, they provided a corrected elevation map of Big Cataloochee. This pdf is interesting in that it shows the raw bare-earth LiDAR data points. The high point (6152.34' at 35.671907,-83.175883) was now back close to the spot elevation.
Finding the right button to push to get the USGS to fix their data from the corrected NC data was a bit harder. The USGS corrected the 3m, 10m and 30m DEM data on 12/28/09 (high point 6152.2' at 35.67191 -83.17584) -- yippee! The USGS seamless server provides a URL for getting the elevation for a given lat/lon, so here is the 3m DEM elevation at the Big Cataloochee high point -- if it reports 6057', then the data is still in error.