Discussion: Precision question

Pardon what may be a naive question, but why is 1m LIDAR, which is very expensive) considered crucial for orienteering base mapping and not 10m LIDAR, which is largely free?

At 10m resolution, you have an elevation point for every millimeter of a 1:10000 map. I cannot think of why having an elevation point every 0.1 millimeters provides any practical gain.

Its all a matter of how small the details are that you want to see in your basemap. 10m posting lidar (average of 1 sample every 10m) lets you resolve objects that are 20m in size or larger - Nyquist sampling requires 2 samples per resolution element. So if you'd like to see features smaller than 20m (say, pits, ditches, stream channels, kolbottens, and especially trail treads under canopy) in your basemap, you need higher density sampling.

That's assuming every single sample hits the bare ground. But they don't. In areas with canopy or veg cover, some large fraction of those pulses will never hit the ground, so this further reduces your sampling of the bare earth surface. Most collection specs don't account for the density of the vegetation cover explicitly. I've seen 1.4m posting lidar that was collected leaf-on and there is virtually no fine detail visible on the ground. Pity.

While you can make *reasonable* orienteering basemap contours from 4m data of forested terrain, you don't get much additional info. For example, if there were a bunch of 1m wide ditches in the terrain, you wouldn't see them in the 4m (or 10m) data, but with 1m data they'd be dead easy (note ditches are long features in one dimension, so they don't have to be Nyquist in the short dimension to resolve them).

I don't think I've ever seen lidar with 10m sampling. Thats really super coarse. The lowest I've ever seen is 4m. It wouldn't be worth the flight time to collect lidar at 10m over any terrain from an aircraft. I don't think satellites have reached lidar footprints even close to 10m yet.

In the US, there are 10m DEM datasets produced by the USGS which people sometimes think are lidar, but they are not. These are produced from digitized versions of the 7.5-minute series topos, and contain all the original shortcomings of those products. They are basically useless for orienteering basemap production. You're better of starting from white paper.

This gives you a sense of it:

http://pugetsoundlidar.ess.washington.edu/example2...

Also, take a look at the image at the bottom of this page:

http://www.personal.psu.edu/kls505/Geog_482/Projec...

I should add that 1m posting is not crucial for a good basemap. The rule of thumb I go by is anything 1m or less is fantastic, anything from 1m to 2m is adequate, and anything from 2-4m is fine for the contours but you won't get much else. All of that assuming average mixed forest terrain collected leaf-off. Anything over 4m is probably not worth your time.

Thanks, this clears it up.

But USGS is most definitely claiming lidar coverage:

http://lidar.cr.usgs.gov/LIDAR_Viewer/

Yes, the USGS runs the CLICK lidar clearinghouse, but not all USGS products are lidar. Lidar coverage of the US on CLICK is spotty, but its a good resource, and they are very responsive to questions.

Here's an article about lidar basemaps from the DVOA and SVO newsletters. There is a nice side-by-side comparison of 1m and 2m posting lidar covering the same area, with the corresponding basemap and final fieldchecked orienteering map. The jump from 2m to 1m is pretty impressive, although the 2m data is adequate for making a basemap.