Bushwalk Australia

I am contemplating a Western Arthurs trip - full traverse and wondering about the pros and cons of returning to Scotts Peak or walking out to Tahune in terms of time and mud. Obviously if the car is at Scotts Peak then a no brainer. The distance to Tahune seems a bit longer but not that much. I've only walked into Blakes Opening from Tahune but that was in 1971 doing a full E & W Arthurs traverse so have not "enjoyed" the YoYo section or the apparent mud bath of the Arthurs Plains.

Summertime the plains aren't that bad - the only invariably muddy bit is around Junction Creek, although how muddy any particular section is depends on the amount of rain, of course. Roseanne to Scotts Peak Rd is an easy one and a half to two days.

I had the same choice this past January. As I recall the estimates in Chapman were essentially the same. I decided on the Huon track because I thought it would be easier to hitchhike at/near Tahune than Scotts Peak Dam.

I left Lake Rosanne at 11:30am and without any undue effort arrived at the junction with the main road leading to Tahune at 2pm the next day. The first car came by about five minutes later and the guy drove me all the way to Hobart. He even offered to let me stay at his house if I couldn't find a hotel!

Times can vary a lot, but Chapman's distances are a long way from being accurate. It's about 30km from the Scotts trackhead to Cracroft Crossing. Not sure of the distance between there and Roseanne.
If I didn't have a car parked at one end, I'd take the other route out just for the variety.

north-north-west wrote:Times can vary a lot, but Chapman's distances are a long way from being accurate. It's about 30km from the Scotts trackhead to Cracroft Crossing.

Chapman's book gives a distance of 27.4km. I'm not sure how he measured it. How did you?

From Cracroft to the Huon car park is 25.4km according Chapman. Then it's another km or so to Tahune. So pretty similar distances, if you believe those numbers.

Even though time estimates are much, much fuzzier I think they are often the more useful metric. Chapman's estimate for Cracroft-Huon car park is 6.5-9.5 hours (took me about 8hrs) versus 7.0-10.5 hours for Cracroft-Scotts. Again, very similar.

north-north-west wrote:If I didn't have a car parked at one end, I'd take the other route out just for the variety.

That was part of my calculus as well. Unfortunately the Huon track is closed in and rather dull, or at least I found it to be so.

Orion wrote:

north-north-west wrote:Times can vary a lot, but Chapman's distances are a long way from being accurate. It's about 30km from the Scotts trackhead to Cracroft Crossing.

Chapman's book gives a distance of 27.4km. I'm not sure how he measured it. How did you?

GPS tracking whilst walking. It's quite accurate. 30.2km from the carpark to the track junction. Most of Chapman's distances are straight line map distances, with some minor allowance for jinks in the track. He seldom actually measures the on ground distance walked. The OT is the only walk I've ever done where my GPS comes close to agreeing with his measurements.

From Cracroft to the Huon car park is 25.4km according Chapman. Then it's another km or so to Tahune. So pretty similar distances, if you believe those numbers.
Even though time estimates are much, much fuzzier I think they are often the more useful metric. Chapman's estimate for Cracroft-Huon car park is 6.5-9.5 hours (took me about 8hrs) versus 7.0-10.5 hours for Cracroft-Scotts. Again, very similar.

There's not a lot of difference in distance and time. The Yoyo doesn't have the big soggy bogs near Junction, which is one good reason for using it.

I've wanted to stash a packraft on the Cracroft. Walk from Scotts Peak and float into Huonville

north-north-west wrote:GPS tracking whilst walking. It's quite accurate. 30.2km from the carpark to the track junction

Thanks for that. Do you know *how* accurate it is?

north-north-west wrote:Most of Chapman's distances are straight line map distances, with some minor allowance for jinks in the track.

Interesting. It's surprising to me that his updated books don't include GPS based measurements.

So his distances tend to be underestimates. Well it's good to know that I've been walking faster than I thought!

Orion wrote:

north-north-west wrote:GPS tracking whilst walking. It's quite accurate. 30.2km from the carpark to the track junction

Thanks for that. Do you know *how* accurate it is?

More so than the car's odometer.

north-north-west wrote:More so than the car's odometer.

That's not a very accurate answer. Car odometers can be off by more than you claim Chapman is.

I use my GPS to measure distances of walks sometimes but I have no valid way to determine its accuracy.
How did you determine the accuracy of your GPS?

By having the odometer checked and then comparing the two over a pre-determined route.

Aren't you quibbling a bit here? It's easy to see the inaccuracy in some of Chapman's distances by simply measuring the track on a large-scale map or even on GoogleEarth.

I had 26.2km from Cracroft to Scotts Peak using my Suunto watch - accuracy plus or minus 5m, with a ping every 30 seconds. Huon Track was 28.3km, about 3km longer than Chapman's guide.

north-north-west wrote:By having the odometer checked and then comparing the two over a pre-determined route.

Aren't you quibbling a bit here? It's easy to see the inaccuracy in some of Chapman's distances by simply measuring the track on a large-scale map or even on GoogleEarth.

Quibbing? Well, yeah. The whole discussion about accuracy of track length in this thread is a quib as it has little bearing on the OT question. That doesn't mean it isn't interesting.

I know I could compare my GPS on a road. But it would be a best case comparison. Walking on a track where one would weave, switchback, and be under tree cover or in canyons at times would worsen the comparison. A winding single track route that was open to both walkers and bicyclists would be better I would imagine, although determining the accuracy of your bicycle odometer in a variety of situations might also be problematic.

I measured a particular day walk twice and the calculated distance differed by 2%. I expect that the spread of values would be something larger than that if I repeated it a number times. And that's with a single GPS in reasonably good conditions for reception. If I compared multiple devices in more challenging terrain there would be a larger margin of error. By saying the track is 30.2km you are implying an accuracy on the order of 0.1km, which is better than 0.5%, and unquestionably better than my GPS is capable of.

Orion wrote:I measured a particular day walk twice and the calculated distance differed by 2%. I expect that the spread of values would be something larger than that if I repeated it a number times.

Possibly the GPS sampling is a source of this difference. By that I mean the GPS calculate that distance n one trip may have correctly plotted a corner peak and therefore been able to a fair degree of truth. The next trip and it may have missed that corner peak and so it has rounded out that corner and you have less distance and there for less truth. Upping the sampling rate on the GPS can be done on some GPSes, at the expense of battery life.

The bicycle wheel lacks this problem and so may be more truthful in this regard.
The same can be said for pedometers.
These two methods do have other problems though so they too can have accuracy problems but from other sources.

In any case a simple single reading of distance would not give confidence in the measurement unless the system and operator has lots of experience and have calculated out their uncertainties (accuracy in 'common peoples speak'). Without that experience to obtain some mean full incite into that uncertainty I would think 6 reading should be taken and then some calculations performed to determine the mean and standard deviations and distribution.

Warin wrote:Possibly the GPS sampling is a source of this difference.

Looking at the tracks superimposed it's clear that the GPS location error is part of the difference. Simply put, the tracks are slightly different, even on straight sections where sampling rate alone doesn't play a role. The software does its best to clean up the jitter but can't invent signal that is lost in the noise.

Here's a closeup trace from a return walk that I did on two different days. This section of the trail is wide open to the sky.




Another thing I don't know is if the error bar is centered around the true distance or is biased long or short. The accuracy could be a lot worse than 2%. How can I know? I'm never going to push a carefully calibrated wheel in front of me for 18km just to figure this out.