Apologies for the long winded explanation, particulalry if you already know this (its obvious you know far more than average).
quote:Originally posted by swebster
I don't even know what maps you are talking about here
I was refering to the 1:50:000 scale maps referred to in the first post of this thread.
I find using the scale of the map to indicate the level of detail to be pretty strange. What difference does it make what scale you plot a map at? It doesn't change the level of detail. I can take a 1:50000 map and enlarge it 2x. That makes it 1:25000 but does not increase the detail of course.
I agree completely. Though, "level of detail" is a pretty ambiguous term in the mapping industry, and one which I have never seen on a contract. Perhaps it refers to how much is shown on the map, as apposed to how accurate the map is (in terms of position). Mapping projects are designed with a level of error
in mind, not a level of detail. Projects are designed to keep the amount of error, not to an absolute minimum, but within limits appropriate for the scale that the maps will be published. For example, a typical, and long running standard for the publication of maps in Canada, is that 90% of objects on the map should be within 1/2 mm of their "true" positions (relative to each other and the coordinate grid on the map), regardless of map scale. You can use the rule of thumb formula m=scale/2000 to calculate what the metre equivalent is for 1/2 mm for any map scale, and hence the error for any scale (or at least the error that 90% of the objects will have less than). For example 10,000/2000 = +/- 5m error for 1:10,000 scale maps. So, map scale is not an indicator of "level of detail" but it is a very strong indicator of the level of error present on the map, provided it was a well made map (this standard only applies to maps made for government publication or engineering purposes; it doesn't apply to commercial or trail-maps, for example).
If we're told to collect data appropriate for making 1:10,000 scale maps, we have enough experience to design a project that will collect data with less than +/- 5m error. Attempts to reduce the error further often lead to going over budget.
Contracts often specify what features are to be included or excluded from the mapping process; which I suppose naturally could be considred the "level of detail", though there is no standard in place to determeine what the "level of detail" is for a particular scale, like there is for level of error. Itwouldn't make sense.
If you're talking about contours, what matters is the spacing of the data points in the elevation data, the contour spacing, the algorithm used to determine the contours (either automatic or by a human I guess) etc.
True, but the spacing of elevation data points is only part what matters when making contour lines. There are many other considerations. Only one of them being the vertical accuracy of those points.
Maybe there is some sort of standard used by the cartographic community that defines the required contour detail for various scales, but it seems like a dumb system.
Again, "detail" is a fairly ambiguous term. The standard for producing contour lines is that, similar to above, 90% (sometimes 95% depending on the contract) of a contour line should be within +/- 1/2 the contour interval. For example, for 10m contour lines, 90% of any contour line should be accurate to within +/- 5 metres. If you have 20m contour lines, a point on the ground that lies along a 100m contour line can be anywhere between 90m and 110m actual elevation, and the contour line will still be considerd to be a good one within standard. Elevation mapping projects are also designed around an error value. If 2 metre contours are required, we have enough experience to collect data that will have less than +/- 1m error.
The interval we choose when making the lines primarly depends on the vertical accuracy and horizontal spacing of the elevation data we collected, but it also depends on the terrain within the map area. If its flat, a tighter contour interval works; if its steep, a larger interval may be more appropriate. As you have stated, this has little to do with map scale. Even if you had elevation data accurate to +/- 1m, you wouldn't publish a 1:20,000 scale map with 2m contours; unless the area was flat you would see nothing else but contour lines.
If you want to PM me for a discussion, feel free. I'm a project manager at an engineering/mapping/surveying company, and an occasional mapping/GIS instructor at BCIT.