by Jim Smith, LANDFIRE Project Lead
This is the first in a series of short blogs that will delve into the enigma that is “map accuracy.” Many users of digital spatial data are neither familiar nor comfortable with the concept of thematic map accuracy and I am sure I won’t answer all the questions. In fact, I may raise more than I answer. But my ultimate goal is to enlarge the conversation and encourage deeper and more useful thought about spatial data quality.
How would you describe traveling from New York City to San Francisco on I-80? Does stating that the distance is 2,905 miles really depict that trip? It most definitely does not, because it includes nothing about the beauty of the Appalachians, Great Plains, Rocky Mountains, Great Basin, or Sierras (read Annals of the Former World by John McPhee for that kind of great detail). Describing the drive by noting mileage alone, I contend, is exactly what many are doing when they think they’re describing the “quality” of a thematic map as 50% or 60% or 70% accurate. It isn’t wrong to characterize a map that way, but that it is a woefully incomplete way to indicate the ultimate usefulness of a spatial data set in a specific situation. And after all, isn’t that what users need to know? Can I or should I use this map in my application in my location, and if so why, and if not, why not? Those are the questions that data producers need to answer in order to help users apply data correctly.
Perhaps a wine analogy would help make my point. There are different kinds of wine (red, white, rose, etc.) and there are different kinds of maps (vegetation, land use, etc.). Within each wine group, there are many, many different varieties with different flavors and notes. Neither maps (% accurate) nor wines (% alcohol) can be described by a single number. Will you always prefer a wine with 8% alcohol over a wine with 5% alcohol? Probably not, but similarly I have heard it said that a map that is “60% accurate” is better than a map that is “50% accurate.” That might be the case, but it is not necessarily so.
Effective spatial data sets need the type of information given in a wine description—a burgundy that has an oak essence with undertones of cherry and apricot, for instance. If you think that is silly, remember that maps have many properties besides how well they compare to some reference data set (aka % accurate), properties related to time, geographic extent, thematic content, and spatial resolution. More helpful descriptions of map quality (think usability) could be something like “50% accurate on 15 different types of deciduous forest covering the entire Midwest in 2014 delivered at 30m resolution.” That spatial data set might be more useful to someone than a “60% accurate, 2010 map with two categories of deciduous forest that covers only Ohio at 270m spatial resolution.”
In future blogs I will dive into the “map accuracy pool” -- hopefully not so deep that I drown but deep enough to stimulate your imagination and create a way forward that improves the application of digital spatial data.
Contact Jim
