3 replies to this topic

[David Medeiros]

A student at Stanford asked me an interesting question the other day that I didn't have an answer for (that happens a lot working there!).

They have a soils layer with metadata that specifies the data as having a scale of 1:12000. The student wanted to know what that translated into in terms of what is the smallest area (or feature size) that can be considered accurate to display with the data. I guess this is similar to asking, if converting to raster, what minimum cell size would the data need to be converted to avoid downsampling the resolution or over to avoid using more cells than necessary to preserve the data accuracy?

I understand the concept of a data scale but it's hard to reconcile with the concept of a static scale for a map. Is there a way to determine from the data scale what cell size in essence the data would be considered accurate at?

[Strebe]

I understand the concept of a data scale but it's hard to reconcile with the concept of a static scale for a map. Is there a way to determine from the data scale what cell size in essence the data would be considered accurate at?

Not in any rigorous sense because giving a scale to a data set in isolation isn't a rigorous specification. You could get with an order of magnitude but not much better without more information.

Regards,
-- daan Strebe

[DaveB]

Have you looked into minimum mapping unit? I'm not sure how you determine it offhand.

[mfarmer]

I'm not a surveyor, but my brief refresher of National Map Accuracy Standards would lead me to answer: a bit less than 4 square feet can be mapped accurately at that scale.

<Edit: that size should be a bit less than 40 square feet (~34). I wrongly used 1:1200 scale at first.>

Links for NMAS:

Brief explanation:
http://nationalmap.g...dards/nmas.html

A bit fuller:
http://nationalmap.g...ds/nmas647.html

Nice discussion:
http://www.asprs.org...ters/june08.pdf