Advertisement
If you have a new account but are having problems posting or verifying your account, please email Niamh on [email protected] for help. Thanks :)
New AMA with a US police officer (he's back!). You can ask your questions here

Royal Canal

  • 07-04-2016 4:53pm
    #1
    Closed Accounts Posts: 1,577 ✭✭✭ Indricotherium


    I jogged from Clonsilla to Leixlip on the Royal Canal the other night.

    I was just looking at the data from the GPS. It seems to say that the I descended from an elevation of 62 meters in Clonsilla to 54 meters in Leixlip. All while traveling 'upstream' such that a canal has a flow.

    There are no locks on this route, and even allowing for the towpath being slightly raised in Clonsilla, and a meter or so for the GPS, It is still indicating that the surface of the canal is lower up stream than down?

    Have I got this backwards somehow? or is the gps just wildly inaccurate?

    This is my regular commute and it trends the same each time give or take a meter or so.


Comments

  • Registered Users Posts: 18,889 ✭✭✭✭ murphaph


    I jogged from Clonsilla to Leixlip on the Royal Canal the other night.

    I was just looking at the data from the GPS. It seems to say that the I descended from an elevation of 62 meters in Clonsilla to 54 meters in Leixlip. All while traveling 'upstream' such that a canal has a flow.

    There are no locks on this route, and even allowing for the towpath being slightly raised in Clonsilla, and a meter or so for the GPS, It is still indicating that the surface of the canal is lower up stream than down?

    Have I got this backwards somehow? or is the gps just wildly inaccurate?

    This is my regular commute and it trends the same each time give or take a meter or so.
    With no locks then the water level is the same height above sea level along the entire stretch.


  • Closed Accounts Posts: 1,577 ✭✭✭ Indricotherium


    murphaph wrote: »
    With no locks then the water level is the same height above sea level along the entire stretch.

    That's what I'd have thought. But consistently the GPS is giving a significant fall.


  • Registered Users Posts: 18,889 ✭✭✭✭ murphaph


    That's what I'd have thought. But consistently the GPS is giving a significant fall.
    If the towpath is a constant height above the water level then the GPS is wrong. A stretch of canal between two locks is a giant spirit level...it can't be wrong.


  • Registered Users Posts: 20,858 ✭✭✭✭ Alun


    Elevation is the most inaccurate of any of the measurements that come out of a GPS chipset , and one that is most affected by non-optimal spreads of satellites in the sky above. You could place a GPS receiver in one spot, and over the course of a few days, as the satellites move overhead, the positional error will go up and down like a yoyo. If you add in the effect of nearby buildings masking satellites that are low on the horizon, you can see that you can't really rely on the elevation measurement to be that accurate.

    There's another issue in that the elevation that is generated by the GPS chipset is the elevation above an idealised model of the earth's shape, called a geoid, WGS84. The earth isn't actually that shape, and so local mapping is made in reference to a local geoid that more accurately reflects the shape of the earth for that country. Most GPS receivers will convert the WGS84 elevation to a local one using a very simplified model of the local geoid, and errors creep in here too.

    Hope that helps :)


  • Registered Users Posts: 179 ✭✭ wildgreen


    I was just looking at the data from the GPS. It seems to say that the I descended from an elevation of 62 meters in Clonsilla to 54 meters in Leixlip. All while traveling 'upstream' such that a canal has a flow.
    What make and model GPS watch did you use? The two readings are essentially identical. I saw other internet references that have both locations at around 65 metres.


  • Advertisement
  • Closed Accounts Posts: 1,577 ✭✭✭ Indricotherium


    wildgreen wrote: »
    I was just looking at the data from the GPS. It seems to say that the I descended from an elevation of 62 meters in Clonsilla to 54 meters in Leixlip. All while traveling 'upstream' such that a canal has a flow.
    QUOTE]
    What make and model GPS watch did you use? The two readings are essentially identical. I saw other internet references that have both locations at around 65 metres.

    A polar m400


  • Registered Users Posts: 179 ✭✭ wildgreen


    wildgreen wrote: »
    A polar m400

    Seems to be a good value watch but the altimeter feature is probably better on Garmin and Suunto.


  • Closed Accounts Posts: 27,836 ✭✭✭✭ ThisRegard


    I don't think the M400 has an altimeter so it's using GPS as Alun described above which is not always accurate.

    Are you looking at your altitude on Strava or something similar, as they can also do automatic altitude correction when you import.


  • Registered Users Posts: 6,699 ✭✭✭ plodder


    Alun wrote: »
    Elevation is the most inaccurate of any of the measurements that come out of a GPS chipset , and one that is most affected by non-optimal spreads of satellites in the sky above. You could place a GPS receiver in one spot, and over the course of a few days, as the satellites move overhead, the positional error will go up and down like a yoyo. If you add in the effect of nearby buildings masking satellites that are low on the horizon, you can see that you can't really rely on the elevation measurement to be that accurate.

    There's another issue in that the elevation that is generated by the GPS chipset is the elevation above an idealised model of the earth's shape, called a geoid, WGS84. The earth isn't actually that shape, and so local mapping is made in reference to a local geoid that more accurately reflects the shape of the earth for that country. Most GPS receivers will convert the WGS84 elevation to a local one using a very simplified model of the local geoid, and errors creep in here too.

    Hope that helps :)
    If the same error is seen repeatedly that suggests it's not the random type of error and maybe more likely caused by the altitude model (eg WGS84)?

    Sometimes the websites that host this kind of GPS watch data have an option to apply altitude error corrections based on actual topographical data which would be much more accurate than the simplified models (as posted already above...). That might fix the problem possibly.

    These questions come up on the running forum all the time. Usually, people are surprised to find out that GPS is actually not all that accurate.


  • Registered Users Posts: 20,858 ✭✭✭✭ Alun


    It's also possible that whatever software the OP is using is ignoring the elevation data in the GPS data altogether, and using it's own values generated using a DEM (Digital Elevation Model). It's a common thing, and sometimes the user isn't even aware of it. It's done mainly not only because of the inaccuracy of GPS elevation data, but also because in practice GPS elevation data from handheld devices, especially watches, tends to be very "spiky", with lots of very small jumps. If you add all these up to display an elevation gain figure it ends up way higher than expected, so without applying a smoothing algorithm, is not very useful.

    Since it's free, and OSi want the blood of your first born to pay for their more accurate one, the one used would probably be the SRTM model, generated by a NASA mission back in 2000, and fettled to remove the worst voids and inaccuracies over the years. This information is available as a grid of elevation values at roughly 100m intervals, and mapping applications using it just interpolate between those points to determine the elevation at any particular point. Needless to say this isn't very accurate either, as a lot can happen over a distance of 100m that would mean an interpolated value would quite a way off, such as high ground either side of the point in question, or equally in the case of a canal if it's built on an embankment.

    Out of interest I looked at the Royal Canal, around where you mentioned and the elevation derived from SRTM varied between about 62m and 54m, which would possibly explain what's being seen here.


  • Advertisement
Advertisement