Comparing elevation data between UK and France.

[Exit Stage Left]

I wondered if the Pendle 600 would qualify as a Super Randonee. They use Openrunner to assess climbing, that programme puts climbing at 7563 metres,
http://www.openrunner.com/index.php?id=1710310  while Bikehike puts it at 10150m. Openrunner in France uses the IGN database, it's unclear what it uses in the UK, possibly a database with fewer sample points, which would under-record. Obviously there has to be a single standard to judge routes by. But is the underlying elevation data the same in France as in the UK?

[mcshroom]

I don't know, but Bikehike is usually regarded as underestimating the climbing so if it's already 10150m in BH I hate to think what the other estimates would be

[KieronY]

My experience of using Openrunner within France is that it underestimates climbing by anywhere between 20- 30% when comparing results with a GPS fitted with a barometric altimeter.

Although it is possible to overlay IGN maps in Openrunner, I'm not sure they are using IGN data for elevation. I think it is more likely they are using googlemaps api data in conjunction with GPS points with some smoothing applied over longer distances.

I can confirm Openrunner does not use IGN data by plotting a simple route around the block from my house. Openrunner gives the elevation at the start outside my front door as 65m. A few metres down the road there is an IGN verified rèpere de nivellement (French equivalent of a benchmark) that gives the height at 46.72 metres. As the road is pretty flat I can say within half a metre the elevation outside my front door is 47m and so Openrunner is definitely not using IGN data for elevation.

Does BikeHike now use OS data?

[Exit Stage Left]

The principle of the Super Randonnees seems to challenging 600s with an extra 10 hours to complete them. Openrunner has been used for the two that have been done so far.
http://www.openrunner.com/index.php?id=156439
http://www.openrunner.com/index.php?id=767819
So they'd be looking for equally challenging rides worldwide. So it might not be so much 10,000m of climbing as 10,000m of climbing on Openrunner. However the elevation data might work fine on long Alpine climbs, but not on our more varied terrain, so it might not be possible to compare the two different types of ride. I wonder if putting the French rides Gpx into other programmes might reveal anything?

[mattc]

I think there will always be cross-channel inconsistencies, and the choice of a mapping website (as opposed to contour-counting or GPS altimeter track) is fairly arbitrary.

So maybe ask the French if they will accept other approaches? The concept is still quite young.

[Somnolent]

I don't know, but Bikehike is usually regarded as underestimating the climbing so if it's already 10150m in BH I hate to think what the other estimates would be

Really ?
I thought Bikehike slightly over-estimated?   
Bikeroutetoaster is the one which under-estimates IME.

[mattc]

There is no correct answer to the climbing figure for a given route - so it's all relative.

[Exit Stage Left]

I like the idea of 'Super Randonees'. The Pendle 600 took in a lot of very impressive scenery. The principle also allows for a 'touriste' option, at 80km a day. So the formula of 10,000m of climbing, with 50 hours allowed is a good one, otherwise it reserves these rides for a small elite, which is fair enough, but the numbers would be better with 10 more hours.
The main question is whether 10,000m of climbing in the Alps is better modelled by Openrunner than 10,000m of climbing in the UK. If Openrunner underestimates ascent in the varied terrain of England especially, then it's improbable that a 10,000m 600 could be designed based on Openrunner, even though 10,000 metres of ascent is being done, by barometric measures. The Pendle 600 would seem to be the closest we have got so far to 10,000m. It could be measured more accurately to verify it, but ultimately it's up to the ACP to decide if they stick with Openrunner as their standard. It would be nice to see 'International Super Randonnees' take off, but it's a matter of what yardstick to use.

[Exit Stage Left]

It's factors like these that are alluded to on a forum for Memory Maps data that might cause discrepancies.

The elevation data files contain "spot heights" of the Terrain on a grid of 50 metres (for the GB map, it's probably more for the IGN data) and intermediate locations have to be interpolated (a technical word for "guessed" ). So the elevation data has inaccuracies, particularly in mountainous regions. But the most important factor is probably that roads are usually designed to smooth out the terrain (perhaps more so in France than the UK) so the ascent calculated by MM generally exceeds the actual road data (a figure of 2x has been quoted elsewhere). Particularly noatable examples are bridges, tunnels, embankments and cuttings, where you will find the MM elevation profile follows the general land elevation, not the "man-made" route.

http://www.memory-map.co.uk/supportforum/viewtopic.php?f=24&t=1417

Roads in France are generally more 'engineered'.

Then, about roads later in the nineteenth century, "The chief engineer in the Limousin, Pierre Tresaguet, had insisted that a limit sould be placed on [road] gradients. ... The old road east of Morlaix still includes a needless climb of 15 per cent (1 in 7) because the blundering military Governor of Brittany, the Duc d'Aiguillon, preferred straight lines to the more accommodating curve of the older road that runs alongside. Thanks in part to Tresaguet, it is unusual now to find a climb in excess of 8 per cent (1 in 12). This was thought to be the steepest gradient that a fully laden mule could manage. British mountain roads seem to rise in fits and starts like step pyramids. French mountain roads go much higher, but more steadily, and can comfortably be climbed for hours by a fully laden cyclist."

http://iaoreditor.blogspot.co.uk/2010/12/or-and-some-french-roads.html

[Reg.T]

(Slight tangent)

And then there are other oddities like going under bridges. A short route along the Portway in Bristol (alongside the Avon, under the Clifton Suspension Bridge, at around 15-20m) shows a big spike in the elevation profile as you pass under the bridge on both BikeHike (to ~55m) and BikeRouteToaster (to ~48m), but only BikeHike appears to count this in the Ascent/Descent totals.

[simonp]

Well, tracklogs which uses interpolation has the Bryan Chapman at about 12,000m. I really don't think you can trust these tools to give a remotely accurate figure. My gps consistently gives about 7,000m though after-the-event analysis said 8400m which is very close to the official figure.

I have just about zero faith in the climb figures generated by interpolating on a grid.

[Manotea]

Anybody care to find out what openrunner makes of the BCM classic & scenic?

[Exit Stage Left]

Anybody care to find out what openrunner makes of the BCM classic & scenic?

I can do that if there are some Gpx files.

[KieronY]

Using a GPX file I found on Ride with GPS (ZigZag's File) Openrunner gives a positive elevation gain of 5957m for Bryan Chapman classic.

As I said up thread Openrunner routinely underestimates elevation gain and I don't think the country has any bearing.

From routes I've done in France this year:

                                      O/Runner       GPS
BRM200 Argenteuil           1346           1917

BRM400 Flins/Seine          2285           2677

BRM600 Flins/Seine          2341           3209

Just going on what my legs tell me from riding the 200 - the GPS is definitely more accurate.

The GPS reading for the 600 is perhaps the least accurate as day went from searing heat to branch lightning thunderstorms, so wildly varying atmospheric pressure but would still say it is much more accurate than Openrunner.

[Exit Stage Left]

The GPS reading for the 600 is perhaps the least accurate as day went from searing heat to branch lightning thunderstorms, so wildly varying atmospheric pressure but would still say it is much more accurate than Openrunner.

Research suggests that variation to pressure differences affects barometric altimeters by less than 1%. I wouldn't trust them to be accurate without calibration, but it's interesting that there is so much discrepancy against Openrunner.
An interesting exercise would be to map the Blue Ridge Parkway in the USA on Openrunner. There is information on ascent on their website. http://www.nps.gov/blri/planyourvisit/bicyclinginformation.htm
I was looking at likely areas for 'Super Randonees' and the BRP looked interesting. It should be accurately surveyed, so would make an interesting comparison.
The basic question remains one of comparability. If the criterion is 10,000 metres of ascent on Openrunner, then so be it, that's what you'd have to aim for.

[cygnet]

Research suggests that variation to pressure differences affects barometric altimeters by less than 1%. I wouldn't trust them to be accurate without calibration, but it's interesting that there is so much discrepancy against Openrunner.

I rode a short out and back route recently and the max outbound elevation from my etrex dropped approx 20m on the return leg, which is about 12%.
 

Possible mitigating reasons?
a) uncalibrated gps - I reached the max elevation about 25 minutes after setting out?
b) too short a route with not enough climbing? (20m discrepancy = not a lot over the 10000m climbing being discussed)?
c) something else?

I'm new to using a GPS unit so if there's anything I could have done 'wrong' l'd appreciate pointers.

[neilrj]

Research suggests that variation to pressure differences affects barometric altimeters by less than 1%. I wouldn't trust them to be accurate without calibration, but it's interesting that there is so much discrepancy against Openrunner.

I rode a short out and back route recently and the max outbound elevation from my etrex dropped approx 20m on the return leg, which is about 12%.
 

Possible mitigating reasons?
a) uncalibrated gps - I reached the max elevation about 25 minutes after setting out?
b) too short a route with not enough climbing? (20m discrepancy = not a lot over the 10000m climbing being discussed)?
c) something else?

I'm new to using a GPS unit so if there's anything I could have done 'wrong' l'd appreciate pointers.

Which model? Only some have barometric sensors the rest have GPS derived altitude and that's a bit variable (think RNG ) )

[Exit Stage Left]

Air pressure varies with altitude, which is how heights are measured in the absence of detailed surveys. But air pressure varies with the weather, as in high and low pressure. But that's only likely to cause an error of 1%, or 100 metres over 10,000 metres assuming the passage of a single weather front. GPS elevation data depends on calculating the relative angles of a number of satellites orbiting 25,000 miles above the earth, which the GPS unit might see a greater or lesser number of, so the accuracy depends on how many satellites it 'sees'. Beyond that the total amount of climb the GPS unit records depends on how often it records the data, on a 600 km bike ride that might not be often enough to be accurate.
Elevation data seems to be a movable feast, it's as if the length of a 600 km ride could vary between 480 and 720 km depending on how it's recorded. At some point there will be firm data, for the time being it's probably best to go with a single standard if you are seeking to organise rides with an altitude element.

[cygnet]

neilrj, its an etrex30 so I assume it's using a barometric reading. What's RNG?

ESL thanks for the brief explanation - I guess I'll have to do a bit more reading, and maybe check the recoring intervals. Sorry for going a bit OT

[neilrj]

neilrj, its an etrex30 so I assume it's using a barometric reading. What's RNG?

ESL thanks for the brief explanation - I guess I'll have to do a bit more reading, and maybe check the recoring intervals. Sorry for going a bit OT

The Etrex 30 does indeed have a barometric altimeter so doesn't use GPS for a Random Number Generator

[JonBuoy]

But air pressure varies with the weather, as in high and low pressure. But that's only likely to cause an error of 1%, or 100 metres over 10,000 metres assuming the passage of a single weather front.

I guess that it depends on what sort of timescale you are talking about but I believe that it can be significantly more than that.  For example the air pressure at my local weather station has changed by 1.4% over the last 12 hours.

[Exit Stage Left]

But air pressure varies with the weather, as in high and low pressure. But that's only likely to cause an error of 1%, or 100 metres over 10,000 metres assuming the passage of a single weather front.

I guess that it depends on what sort of timescale you are talking about but I believe that it can be significantly more than that.  For example the air pressure at my local weather station has changed by 1.4% over the last 12 hours.

What variation in altitude would that be equivalent to on a barometric altimeter?

[JonBuoy]

But air pressure varies with the weather, as in high and low pressure. But that's only likely to cause an error of 1%, or 100 metres over 10,000 metres assuming the passage of a single weather front.

I guess that it depends on what sort of timescale you are talking about but I believe that it can be significantly more than that.  For example the air pressure at my local weather station has changed by 1.4% over the last 12 hours.

What variation in altitude would that be equivalent to on a barometric altimeter?

About 400'.

[Exit Stage Left]

But air pressure varies with the weather, as in high and low pressure. But that's only likely to cause an error of 1%, or 100 metres over 10,000 metres assuming the passage of a single weather front.

I guess that it depends on what sort of timescale you are talking about but I believe that it can be significantly more than that.  For example the air pressure at my local weather station has changed by 1.4% over the last 12 hours.

What variation in altitude would that be equivalent to on a barometric altimeter?

About 400'.

A 200k Audax equivalent to the sort of 600 we are talking about in terms of effort might have 4,000 metres of climbing, and an average rider might do it in 12 hours, so the error would be in the region of 3%. If a 600k event had 12,000 metres of climbing, which is what is being suggested by the degree of error in Openrunner, there would have to be three such weather events during a 50 hour period to produce a 3% error.

[Reg.T]

Won't the error be dependent on relative altitudes as the front passes?

For example, if you're on a long climb (and I'm maybe thinking more continental than UK here), and the atmospheric pressure is also dropping, a barometric-based calculation would overstate the altitude gain, whereas if the pressure is rising it would understate.

If I've got that right, then the error while riding could vary (either way) from the apparent altitude gain/loss that might get recorded while staying still while a front passes.