Accurate Elevation Data

[dlovett]

What site if the best for giving the most accurate elevation data?

The route I am working on has the distance of 313km and 4136m on Garmin Basecamp and 312.5km but only 3464m on Ridewithgps.com

Cheers

[psling]

If it's a loop then I'd say the Ridewithgps route has less climbing but the Garmin route would have more downhill so it's a tough call...

Sorry! Being serious, I don't know the answer to your question.

[Richard G]

I read an article on a new public datasource for this sort of stuff recently (some very high accuracy satellite data was made available), but damned if I can remember where I read it.

Some related info on OpenStreetMap: http://wiki.openstreetmap.org/wiki/SRTM (they list other data sources too... and it appears I was thinking of ASTER, which isn't in common use).

[Moder-dye]

Well I wouldn't recommend strava for elevation accuracy when route planning certainly not in my location.

[Richard G]

Interestingly, Strava uses ASTER and SRTM outside of the US.

[Moder-dye]

Don't know what it uses, but in some areas of Shetland I could everest in an afternoons ride. One guy I know it's logging him doing 1600m elevation each way on his 3.3km commute

[Dave Barter]

I read an article on a new public datasource for this sort of stuff recently (some very high accuracy satellite data was made available), but damned if I can remember where I read it.

Some related info on OpenStreetMap: http://wiki.openstreetmap.org/wiki/SRTM (they list other data sources too... and it appears I was thinking of ASTER, which isn't in common use).

It was the EA LIDAR data which is accurate to 1m but does not have full coverage. Mainly covers urban areas and areas of interest.

SRTM and ASTER DEMS (digital elevation models) are both in common use. SRTM2 is more accurate, ASTER has smaller pixel size and more granular.

But which data source you use is meaningless unless you have a decent curve smoothing algorithm as SRTM/ASTER only have height averaged over 30 -90m squares. This makes climbs look like minecraft worlds and introduces all sorts of errors if your GPS point strays.

I am working on an algorithm to consolidate all UK height data including the OS opendata (Terrain 50) to get a decent height profile for all roads and tracks. I'm experimenting with weighted average and curve smoothing approaches to do this. Trust me it is dead hard as anyone who has been bored to death by me about elevation data knows. Sadly the simple answer to the OP question is "who knows?" as most of these sites keep their DEM processing close to their chest. I repeat once more the data source you use is meaningless if your height application approach is pants.

Strava *should* have the best data of everyone as they have many hundreds of GPX measurements they can overlay on top of these DEMS to make them more accurate. If I had to chose, I'd chose Strava above the rest. I use a naughty fake account to test my algorithm.

[robert]

I use every day http://www.openrunner.com/ and I'm very satisfied with the distance and elevation results it gives.

Once I used Strava and I concluded immediatly it was not done for this,

[sean_iow]

I use purchased Ordnance Survey mapping in Tracklogs and this seems to give results that match for the claimed distances/heights when I import a GPX of a route. I assume it must use the Ordnance's height data from the contours? I can try your GPX in on my mapping and let you know the distance and height I get if that would help? This is not just an attempt to get a sneaky look at your draft route

Sean

[ScotRoutes]

Break it down into smaller chunks, overlay it on an OS map and do it the old-fashioned way by counting contour lines. Even if you only do this for a small section it will likely indicate which mapping system is more accurate.

Word of warning though; ascent figures from routes that go directly uphill are likely to be more accurate than those that contour.

[Dave Barter]

I use purchased Ordnance Survey mapping in Tracklogs and this seems to give results that match for the claimed distances/heights when I import a GPX of a route. I assume it must use the Ordnance's height data from the contours? I can try your GPX in on my mapping and let you know the distance and height I get if that would help? This is not just an attempt to get a sneaky look at your draft route

Sean

Nope uses SRTM data.

The 1:50k and 1:25k contours are not available as open or closed data. Only the Terrain 50 product is.

[sean_iow]

Nope uses SRTM data.

The 1:50k and 1:25k contours are not available as open or closed data. Only the Terrain 50 product is.

Never knew that, but it is one if those things you take for granted as just working until you dig a bit deeper to find out how it actually works.

I use MapInfo at work and Southern Water have a contour table which seems to give accurate contours, but I'm only interested in areas where they have manholes as I'm working on drainage schemes and their data man is a bit of a wizard and he may of used the ground level data they have for manhole covers (680,000 of them in their region and most surveyed to within 10mm by site survey) to adjust the contours and increase the accuracy. Not much use for an ITT though as mostly on roads and in developed areas.

[Moder-dye]

I use every day http://www.openrunner.com/ and I'm very satisfied with the distance and elevation results it gives.

Once I used Strava and I concluded immediately it was not done for this,

Ha, result!

Plotted the same route in Open Runner and strava in Shetland in a known (to us) dodgy area. Open runner gives 856m elevation and a sensible profile over an 85km ride, strava gives 12098m elevation and a correspondingly bonkers profile over the same route.

Thanks for that plotting resource Only other benefit of strava for route planing is the heat maps for picking out routes that are acutally used or well used for likelyhood of useability.

[Zippy]

For a limited time, I can get my hands on data from Edina Digimap....

I also have access to stuff off the environment agency geostore - which I think has some detailed lidar datasets - but i think it's also available publically via the data gov thingy anyways

[Dave Barter]

EA Lidar does not have full coverage I use it where I can in my model

[Zippy]

I just looked it up - crums, 11TB of data That's a lot of co-ords!