Website for accurate road gradients
 

Hey everyone,

I for the life of me, cannot recall the website that provides more accurate data on ride performance than Strava or Garmin Connect (or a host of other sites). It is especially more accurate in providing road gradients, as it take your computer data and updates any errors from the bike computer.

Can anyone help me?

Probably not the ideal answer but some sites allow for gradient smoothing which makes indoor sessions more realistic as it eliminates the short ramps and gradient spikes in GPX files. I use Rouvy for that feature, you can control max gradient as well as smoothing however total elevation gained remains the same.

Yeah, I was trying to find a website last week to check some local grades I hadn't ridden. I thought I'd bookmarked such a site but darned if I can find it now.

Friday I walked to and from a library for early voting and took a new-to-me route that turned out to have some potential for hill repeats. I couldn't find any info online via my phone during the walk, so I used a clinometer app to physically measure the grade at various points along the way.

It's RidewithGPS.com (RWGPS). There's a function in the More menu in the black box in the upper left. The function says "replace elevation data". I just tried it on a ride and it added over 100' to my total gain, so it probably changed some gradients, too. RWGPS uses geo data for elevation. I don't think there's a better way to do it other than going to the hill with your inclinometer and taking a lot of measurements. As you might guess, RWGPS relies on uploads from a GPS enabled device. Those devices give somewhat inaccurate elevation data because it's barometric data. When we finish a group ride, people frequently say their total gain from their device. There are often spreads of several hundred feet, depending on ride length.

That said, if your GPS location data is off, like off the road, then your elevation data isn't going to be correct either, but if your route is shown on the road, the RWGPS data should be very close.

according to strava the more accurate elevation data from devices come from those with barometric altitude sensors. this barometric altitude data is added to their elevation data base. when i upload via CycleMeter on my samsung phone with no barometric elevation there is a difference between the .fit file i download from strava (original from CycleMeter) and the adjusted .gpx file. see the following...

https://support.strava.com/hc/en-us/...on-Strava-FAQs
https://support.strava.com/hc/en-us/...vation-Basemap

i think if you have a device with a barometric sensor that will be about the best you can achieve. the pressure won't change so much during a ride that it would be off by any noticeable amount.

-scott

Then why, as I mentioned before, after a group ride with everyone using a Garmin Edge, does everyone report different figures for elevation gain, and considerable differences of as much as several hundred feet? Why because these teeny altimeters aren't particularly accurate, nor are they temperature compensated. It could be down to which wall you leaned your bike against during the coffee stop. Who knows. Direct uploads from the Garmin device to websites shows the exact number displayed on the Garmin. However, a recalc on RWGPS using geodata will show the same result for everyone.

FWIW I use a Garmin Edge 200 that uploads the data to Garmin Connect which then sends it to Strava. On almost every ride the elevation data will be different on the unit and the two sites. The result is three different numbers. It's been doing that for the past 4 years that I've been on Strava.

GPS altitude data is OK, but never great. For the curious, do some research on the ASTER Digital Elevation Model. I have quite a few altimeter / barometer apps on my phone, including one that reports ASTER altitude using GPS location lookups. It's also interesting to lock (calibrate) a barometer app to local airport readings (quite accurate), then get differentials computed for iPhone readout (Hunter Research makes one). All kinds of things to experiment with, but at the end of the day, you can get some good comparative / relative numbers for your rides, but nothing you'd use to land a plane. Not at a cyclist-friendly price, anyway.

I cannot answer your first question. i simply do not know. If i were to guess it would be due to differences in model, vintage of model, how well the device is treated by owner, and lack of ability to calibrate. I cannot speak of their accuracy because i don't know the specific type/silicon of sensor inside, i don't have one to open up and find out then get the data sheet on it. so you are right, who knows.

regarding everyone's data being the same after uploading...their data is being compared to the same data set so it is natural that they would all have the same ending results, assuming everyone rode the same exact route.

i still think they are more accurate that GPS only elevation, at least when near canyon walls/sides of mountains.

I don't have an dedicated cycle computer but according to Garmin the model 200 (now obsolete) does not have a barometric altimeter. that being said this could also explain why after a group ride not everyone has the same elevation gain, depending on model of course.

have a garmin explore (no altimeter) and the 1030 with one they are always different on elevation.

Map data doesn't cover every square inch of your travels with an accurate elevation number. A barometric altimeter will cover every square inch of your travels, but it has various limitations too that will always make comparing one to another silly.

If you need to know something related to grade or gain/loss, then the average of many rides will give you the best numbers.

I use the replace elevation function in Ride with GPS that uses the data from the space shuttle mission that mapped the earths elevation.

Or you can go old school: Get the USGS topo map and a pencil and trace your route. This is how we figured out how much gain we had on our hikes before Al Gore invented the Internet.

I agree. and on the strava website they do mention that when the receive uploaded data with real altimeter data they update their digital elevation model (DEM) to improve the accuracy of it. we all benefit from those with real altimeters.

Well, that sure is screwy. I doubt it. You're saying that Strava replaces the geodata set from space mapping with someone's barometric data, which is always inaccurate? Really? If so, that explains why Strava data is always so inaccurate.

"Real alitimeters" might still be found in geo research centers. Maybe. One first has to know what the elevation of the altimeter is and calibrate the altimeter to that known elevation. That immediately raises the question of how does one determine the elevation to be used is calibration? Well, from the geodata set of course. Then one has to have a separate barometer which keeps the altimeter in calibration. There also has to be temperature compensation integral to that altimeter. Your home, auto, or cycling device obviously has none of that. It's always going to be inaccurate.

I think a lot of the confusion in this thread is because there are two ways to determine elevation with a portable device, GPS and barometer. Some of our devices work one way and some the other. The Garmin Edge, with which I am most familiar uses a barometer, which is more accurate than using GPS (though it is still inaccurate, just not as bad). DCrainmaker has a good article about this: https://www.dcrainmaker.com/2010/05/...evice-gps.html

That article speaks directly to the subject of this thread: elevation correction and how to get it.

And that said and noted, Garmin Edge devices will return total elevation numbers, for the exact same course done on a group ride, which may vary by several hundred feet for a long ride, though the variation is usually less than 200'..

Don't let the 10 cm (4 in.) accuracy of LIDAR give you a false impression of how accurate a map or map data can be. That is only vertical accuracy. The smallest resolution horizontally that you can get data from the USGS for is 3 meters. So that's about a 10 x 10 grid or 100 square feet that elevation can be very different, yet reported as one elevation.

And then there is the issue of matching that data to a specific position on the ground that has lots of errors from the angles of observation and many other errors that cannot be overcome completely by any mapping method.

Our GPS's and barometric sensors aren't survey quality instruments. We shouldn't expect as much of them as some are expecting.

i don't see where 10 cm data is available on the USGS site, but it is a big site. i do see LIDAR collected data but that is project specific. most riders probably won't be in that area and it is doubtful that RWGPS, GarminConnect, or Strava are using that data. they are likely using the 1/3 arcsecond seamless DEM:

https://www.usgs.gov/core-science-sy...ducts-services

i found the data sheet for a small sensor. it measures about 6x6mm about 3mm tall. this would fit into a phone easily. the sensor actually does provide temperature and four cal points so temperature compensation can be made but that would be up to the software in communication with it:

https://www.servoflo.com/download-ar...les/607-ms5540

i read the article, its 10 years old but still not really outdated. it mentions Garmin Edges but not all Garmin Edges have a pressure sensors.

https://www.usgs.gov/faqs/what-lidar...ience_products
3rd paragraph.

10 cm is only the vertical accuracy. Horizontally over the earths surface, resolution from well over 140 miles up isn't going to be anywhere near as good.

The 1/3 arcsecond is talking about the horizontal resolution. Think of it as how big the grids are. For each grid there is only one value of elevation.

I admit this stuff is outside anything I have done career wise. But that is how I've understood it from trying to understand how the science of map making works from various other things I've been involved in.

wow, read this from that page...

Is there a difference between 3D Elevation Program (3DEP) DEM data and lidar data?

Yes, lidar data are represented as a cloud of variably spaced point locations (x,y,z) of any terrestrial target reflecting the laser pulse. In addition to bare-earth, the point cloud will typically include buildings, trees, towers, and powerlines. 3DEP DEMs are raster surfaces of bare-earth elevation values sampled at uniform horizontal increments...

so, once you go under a tree or bridge your elevation may suddenly increase.

it is all interesting stuff but what amazes me most though is that it is 2020 and we still really don't have very accurate elevation everywhere that is easily available. last week i was using the USGS API to pull elevation data from a route i routinely ride, boy was it crappy. there were grades of over 30% and the most that hill has is maybe 8%.

Because apparently for everyone but us amateur cyclist, it's been good enough.

yes they can be slow. like I have a block that is 20% grade no way the altimeter can keep up with it.you need maybe 3 blocks.I have a lot of those one or two block hills.

After using Garmin Connect and the Strava route builder / map tools extensively, I think that RWGPS is quite a bit better than either for building routes, and fairly accurate on the elevation gain. Gravel roads are pretty notorious for not being accurate for elevation gain on map tools either, so I don't know how RWGPS improved it.

Can you tell us which Edge models lack this sensor?

according to the garmin site the model 200 and 25 do not have a barometric sensors.

https://buy.garmin.com/en-US/US/p/90675#specs
https://buy.garmin.com/en-US/US/p/508489#specs

this is why in one of my earlier posts i mentioned model and vintage. looks like the latest collection of edges all have the sensor now.

looking over the garmin site the prices of these are not too bad. there might be an edge in my future now...