[lil byke]

Are the elevation gain numbers always so fricken wrong for so many trails? excepting the manually entered ride data in trailforks.

For example, Buffalo Creek on the front range, almost every trail has numbers that are off, sometimes by thousands of feet, checking ride data and comments left by riders. I use trailforks, strava and mtbproject and they all suck. I also track my own mileage and elevation gain on a garmin as well.

Is there anything we can do as a community? It sucks when you go out to ride a trail and get something completely different then the elevation described. Also ride descriptions in every app are so bad and vague, it seems like we could crowd source this information alot better than we do and it would benefit everyone.

Maybe we need yet another better app?

 

[upstateSC-rider]

Not a big deal to me, I start 'down here' and want to ride 'up there', whether it's a 800' or 1500'.

 

[natas1321]

Never really paid attention to it, I just ride the trail and enjoy what it has to offer.

Sent from my moto g(7) supra using Tapatalk

 

[Sparticus]

I rode mountain bikes about 20 years before I got my first Garmin.
Trailforks didn't come along until much later yet.
Imagine riding a bike without any data at all.
Walked to school uphill both ways, too.
Whippersnappers!
=sParty

 

[Jayem]

Have strava corrected it based on the TOPO, that works pretty well if your other methods aren't working.

 

[mtnbkrmike]

I just want to get to the summit, regardless of what the elevation gain is. I know the rides I’m doing. I don’t need metrics to understand what I am in for.

I guess it might matter if I were riding in foreign terrain on trails the first time, but that seems like a distant memory. A dream from years gone by. And even then, I have a very good idea of what I am getting into, without studying data.

 

[John Kuhl]

The only thing I trust now is my Garmin. I've seen way too many things that were off.

 

[Le Duke]

So, this will be long winded. Sorry in advance.

These apps use the “trail”, as recorded by users, and compare it to what is basically a grid of squares making up the entire surface of the earth, as measured by the NASA SRTM project. SRTM = Shuttle Radar Topography Mission; this data was collected by one of the actual Space Shuttle birds, using what is called Synthetic Aperture Radar, to measure the elevation of the entire surface of the Earth. These squares are commonly referred to as pixels, and they make up an image called a digital elevation model (DEM).

Those squares, nominally 30m x 30m, are an “average” of that entire 30m x 30m area. So, one portion of the 30m x 30m square could be at oh, 8500ft, another portion at 8480ft and another portion at 8450ft. But the average elevation of that 30x30m square will be “8459.65ft” (that’s a made up, but plausible number).

So, then the trail, which might wind around a bit within that 30x30m square, is basically recorded as being at that “average” elevation , despite the fact that it might go from 8450 to 8500ft in that area…but recorded as 8459.65ft.

(The best way to imagine this is to think of a 30m x 30m square drawn on a ski slope. If its perfectly flat, the average elevation is the average of the highest and lowest points and in the middle of the square. But with any imperfections or minor changes in slope (like a flat service road running up through it) that average elevation is a) not located in the middle and b) not JUST the average of the high and low points.)

So, why do they do this? Because data. To get a more accurate measurement, you’d have to go down to 10x10m, 3x3m, 1x1m, 30x30cm, etc. But, 10m pixels require 9x as much processing and storage, 3m 100x, 1m 900x, 30cm 10,000x.

Then, let’s compare this to your GPS. It records a new elevation every second or so. It has its own problems, though. Elevation is the most inaccurate measurement it takes, for starters. Also, the steeper the surrounding terrain the greater chance of positional inaccuracy due to a phenomenon called multipathing.

You’re using a free or very cheap app that has to process elevation data for trails all over the world. If you don’t like it, download and store high res elevation data (it’s free, provided by the US government), plot your trails in GIS software and get better elevation profiles.


Sent from my iPhone using Tapatalk

 

[plummet]

Choose the one that gives you the biggest elevation...... then throw the others in the bin so you have no other reference. Then come on here and boast at the extensive elevation you have ridden for that day......

 

[lil byke]

So, this will be long winded. Sorry in advance.

These apps use the “trail”, as recorded by users, and compare it to what is basically a grid of squares making up the entire surface of the earth, as measured by the NASA SRTM project. These are commonly referred to as pixels, and they make up an image called a digital elevation model (DEM).

Those squares, nominally 30m x 30m, are an “average” of that entire area. So, one portion of the 30m x 30m square could be at oh, 8500ft, another portion at 8480ft and another portion at 8450ft. But the average elevation of that 30x30m square will be “8459.65ft” (that’s a made up, but plausible number).

So, then the trail, which might wind around a bit within that 30x30m square, is basically recorded as being at that “average” elevation , despite the fact that it might go from 8450 to 8500ft in that area…but recorded as 8459.65ft.

(The best way to imagine this is to think of a 30m x 30m square drawn on a ski slope. If its perfectly flat, the average elevation is the average of the highest and lowest points and in the middle of the square. But with any imperfections or minor changes in slope (like a flat service road running up through it) that average elevation is a) not located in the middle and b) not JUST the average of the high and low points.)

So, why do they do this? Because data. To get a more accurate measurement, you’d have to go down to 10x10m, 3x3m, 1x1m, 30x30cm, etc. But, 10m pixels require 9x as much processing and storage, 3m 100x, 1m 900x, 30cm 10,000x.

Then, let’s compare this to your GPS. It records a new elevation every second or so. It has its own problems, though. Elevation is the most inaccurate measurement it takes, for starters. Also, the steeper the surrounding terrain the greater chance of positional inaccuracy due to a phenomenon called multipathing.

You’re using a free or very cheap app that has to process elevation data for trails all over the world. If you don’t like it, download and store high res elevation data (it’s free, provided by the US government), plot your trails in GIS software and get better elevation profiles.


Sent from my iPhone using Tapatalk

This is interesting, thanks for elaborating. Where do I find this data and what software is good for plotting?

 

[Le Duke]

This is interesting, thanks for elaborating. Where do I find this data and what software is good for plotting?

This is a bit of a rabbit hole to go down, but you can get a personal use license for ArcMap or ArcGIS Desktop (same thing) from a company called ESRI. Unless it's changed in the last year, it's $100/year. That may seem steep, but a professional license costs a lot, lot more.

ArcGIS Desktop | Desktop GIS Software Suite

ArcGIS Desktop is a complete desktop GIS software suite that allows you to create maps, perform spatial analysis and manage data. Learn more about capabilities.

www.esri.com

Choose the "for individuals" option.

You can find quite a few tutorials on YouTube on how to make trail profiles, etc.

You can get Digital Elevation Model data here. 100% free, hosted by the USDA, collected by NASA and processed by the United States Geological Survey.

USDA:NRCS:Geospatial Data Gateway:Home

datagateway.nrcs.usda.gov

 

[kampgnar]

This is a bit of a rabbit hole to go down, but you can get a personal use license for ArcMap or ArcGIS Desktop (same thing) from a company called ESRI. Unless it's changed in the last year, it's $100/year. That may seem steep, but a professional license costs a lot, lot more.

Highly recommend QGIS for situations like this. Free and open source.

Welcome to the QGIS project!

www.qgis.org

 

[baker]

Never really paid attention to it, I just ride the trail and enjoy what it has to offer.

When I see comments like this, I always wonder where the people ride.

In Colorado, the elevation gain is often more important to me than the distance. The elevation gain number is like a pain index.

I also enjoy seeing how much climbing I can do. So far this year, I'm at 234,000 feet of climbing.

And Le Duke, thanks for the technical explanation.

 

[slimat99]

It's because those apps are about gathering your data, not providing data to you.

 

[kapusta]

Not sure if this has been mentioned, but even if you have perfectly recorded, 100% accurate data, how you quantify "elevation gain" over a ride is still not so straight forward.

Do you count every little dip and rise in the trail? Of course not. So you need to decide what amount of gain counts. When I have played around with this parameter on some tracks I have recorded, the resulting gain numbers can change dramatically on trails with a lot of short ups and downs.

 

[Sidewalk]

I semi care about elevation because a 40 mile ride, with 4000' of gain is likely going to be FAR more manageable than 40 miles a 8000'. THe first ride is something I can do after work, the second something that will likely take all day. So if I am traveling somewhere to go ride, I want to get a good idea what I am in for so I will know if I need to figure out water, clothing, bail out options, etc.

Personally, I look for riders on Strava who have done the trails I am looking at, and find one that recorded with a device that has a barometer.

 

[Le Duke]

Not sure if this has been mentioned, but even if you have perfectly recorded, 100% accurate data, how you quantify "elevation gain" over a ride is still not so straight forward.

Do you count every little dip and rise in the trail? Of course not. So you need to decide what amount of gain counts. When I have played around with this parameter on some tracks I have recorded, the resulting gain numbers can change dramatically on trails with a lot of short ups and downs.

Yep. And the speed at which you are riding/walking/running will dramatically change the elevations you record as well.

Think about it this way: if you are on a trail that is 100m long, traveling at 5m/s, , and recording at once per second, you will theoretically record 20 data points on that trail. If you are traveling at 10m/s, you will only record 10 data points. Your ride will appear more "segmented", as it is basically playing connect the dots and creating straight lines on a trail that is likely full of turns and very few perfectly straight lines. The exact same thing would apply to elevation as well.

 

[frdfandc]

I just turn on my Garmin and sync my Strava to it. Whatever the elevation gain is at the end of the ride is what I rode. I don't really worry about elevation unless I want to get some serious climbs in.

 

[Harold]

This is interesting, thanks for elaborating.

LeDuke has only gotten started. There are so many extra directions to take this. Looking at the absolute basics, ask yourself the question, HOW do we measure elevation? What's "zero"? Zero is considered to be "mean sea level" but ocean levels are not identical between the Atlantic and Pacific. Plus the oceans "slosh" because of the pull of the moon's gravity. And wind causes water movement, as well.

Further, the globe is not a perfect shape of any sort. It sort of follows a general shape, but it's not exactly anything in particular. Elevations and positions are partly based on the geodetic model of the globe, which is a mathematical generalization of the shape of the globe.

So, how to you measure elevation? There is NO DIRECT WAY to measure elevation (you can't put a measuring tape to the mountain)! Everything we've got is indirect and even stuff like barometric altimeters are not just indirect, but a sort of vague estimate.

Everything LeDuke said is ALSO absolutely true. But there's a caveat with using DEM data to assign elevations. That method is dependent on the accuracy of the positions you are recording. If your DEM data is based on 30m pixels, then your positional data has some wiggle room before it's going to affect the elevation. But if you move to the super high res LiDAR-derived DEM data that has centimeter level accuracy (usually tens of cm, with somewhere between 1ft and 1m pixels), then EACH position on your path had better be dead on perfect otherwise you're going to be adding elevation change where elevation change didn't occur. And if your path is on steep terrain, that could be a SUBSTANTIAL amount of elevation change over the course of a ride.

All this to say that our consumer grade handheld/fitness equipment and our social activity sharing websites are never going to be anything more than a vague estimate that WILL vary hundreds to thousands of feet of climbing over the course of a ride for the foreseeable future.

Where do I find this data and what software is good for plotting?

I bought a license for ESRI's ArcGIS several years ago. Cost me about $1600, and this was before they had an annual "personal" license level. Nice that they have that now. Still, I bought the software to make money from it, so it wouldn't have been relevant to me, anyway.

I like QGIS, too. It's important to note that there's going to be a pretty big learning curve to using software like this. This stuff is no joke. Just learning about file structures and data conventions for geospatial data will take some time.

 

[Le Duke]

LeDuke has only gotten started. There are so many extra directions to take this. Looking at the absolute basics, ask yourself the question, HOW do we measure elevation? What's "zero"? Zero is considered to be "mean sea level" but ocean levels are not identical between the Atlantic and Pacific. Plus the oceans "slosh" because of the pull of the moon's gravity. And wind causes water movement, as well.

Further, the globe is not a perfect shape of any sort. It sort of follows a general shape, but it's not exactly anything in particular. Elevations and positions are partly based on the geodetic model of the globe, which is a mathematical generalization of the shape of the globe.

So, how to you measure elevation? There is NO DIRECT WAY to measure elevation (you can't put a measuring tape to the mountain)! Everything we've got is indirect and even stuff like barometric altimeters are not just indirect, but a sort of vague estimate.

Everything LeDuke said is ALSO absolutely true. But there's a caveat with using DEM data to assign elevations. That method is dependent on the accuracy of the positions you are recording. If your DEM data is based on 30m pixels, then your positional data has some wiggle room before it's going to affect the elevation. But if you move to the super high res LiDAR-derived DEM data that has centimeter level accuracy (usually tens of cm, with somewhere between 1ft and 1m pixels), then EACH position on your path had better be dead on perfect otherwise you're going to be adding elevation change where elevation change didn't occur. And if your path is on steep terrain, that could be a SUBSTANTIAL amount of elevation change over the course of a ride.

All this to say that our consumer grade handheld/fitness equipment and our social activity sharing websites are never going to be anything more than a vague estimate that WILL vary hundreds to thousands of feet of climbing over the course of a ride for the foreseeable future.



I bought a license for ESRI's ArcGIS several years ago. Cost me about $1600, and this was before they had an annual "personal" license level. Nice that they have that now. Still, I bought the software to make money from it, so it wouldn't have been relevant to me, anyway.

I like QGIS, too. It's important to note that there's going to be a pretty big learning curve to using software like this. This stuff is no joke. Just learning about file structures and data conventions for geospatial data will take some time.

Oh. Yeah.

I should have specified that I have a MS in this stuff (GIS), and did it for a living. If Harold, kampgnar, myself or the other GIS guys here make it sound "simple", that's because for us, something like this is just that. For the average lay person, I must apologize, because what would take me 15-20 minutes (after downloading the data) would likely take you days or weeks. Unzipping a DEM file, organizing files, creating a mosaic of multiple DEM files, clipping a DEM raster file, importing a .gpx, extracting elevations from a DEM along that segment, etc. All of that is stuff that, unless you had someone looking over your shoulder or you miraculously found a series of YouTube videos with exactly this process in mind, would take you a long, long time to figure out on your own.