Route Around Removed Pemi Wld. Bridge?

[DougPaul]

I think WGS84 (or NAD83 which I think you'd agree is virtually identical for all practical purposes) is to be taken unless someone states another datum. I took Hillwalker's reference to be WGS84. All the points I've mentioned are WGS84.

The difference between WGS84 and NAD83 is less than 2 meters in the CONUS--which is too small to be resolvable by a consumer GPS. I believe NAD83 is being used on some aeronautical and nautical charts.

Years ago I used the USGS quad topos so heavily that I used NAD27 CONUS with all my software settings. When I converted I think I remember noting that the typical difference in the Whites was around 100 feet between the 27 and 84.

I made a comparison of a point in the Crawford Notch and got a difference of ~230m between WGS84 and NAD27. Enough to be a serious issue for backcountry navigation.

Older USGS topos are NAD27, some of the newer ones are WGS84. Local maps are often in local datums (eg a state datum).

I still, for my own private purposes plot WGS84 on the topos when all I need is a general idea. The situation is complicated in that a lot of commercial digital distributions of the USGS topos have ALREADY been converted. Add all this to GPS inherent margin of error (some GPS screens note it, but I've seen errors beyond even that) and all this information from amateurs like me should be taken only as approximations.

WGS84 is used internally in GPSes and is the default for screen I/O. (You can set the screen I/O datum to be any of a variety of datums, the standard for track and waypoint files is WGS84*.) I personally use WGS84 everywhere.

* In the past, some GPS models have violated this standard by using the screen datum for the files. Hopefully all such GPSes are thoroughly obsolete by now.

The "error" or "accuracy" indicated on the GPS is not reliable or even very meaningful. It only takes a subset of the factors into account, so the real error is greater than or equal to the indicated error. The nominal accuracy for a consumer GPS is a 95% probability of being within 10 meters of the correct coordinates under good signal conditions (ie with a good skyview, such as from a bare mountain top, in the air, or at sea). The accuracy is generally worse when in trees, valleys, poor antenna placement, etc.


So while your guess of Hillwalker's datum is likely to be correct, I would still like to get the info directly from him.

Plea to all waypoint posters: please post the datum with the coordinates. Your waypoint could be dangerously confusing without it.

Doug

 

[Will]

I made a comparison of a point in the Crawford Notch and got a difference of ~230m between WGS84 and NAD27. Enough to be a serious issue for backcountry navigation.

It would be a _hugely_ serious issue. [EDITED TO ADD: but it isn't, in fact] My experience is you've slipped a decimal point there or something, but I'm sure you're correct [EDITED TO ADD: not correct] and my observations have been wrong [EDITED TO ADD: accurate, as it turns out]. I have not claimed to be an expert.

I just looked at some old conversions and, again, I'm getting _much_ smaller differentials, tiny compared to your 230m. I must be seriously misunderstanding the WGS84/NAD27 issue. If I am wrong I'm sorry if I've misled anyone. [EDITED TO ADD: I was not wrong and I did not mislead]

 

[Hillwalker]

Doug's question

Hi, Doug, In answer to your question:WGS84. However when I made those hikes GPS was just a dream. Therefore the Lat/Lon I gave was just an approximation of my location then.

It's really interesting seeing railroad steel lying right where it had been placed during J.E. Henry's time. Right where the tracks end, there is a two or three foot long of rail lying between the tracks that looks like it was sawed from a rail. Knowing just hard hard RR rails are, I have often wondered about that piece. If it was flame cut, the person who cut it was really good.

 

[Stan]

Any speculation on the best crossing with xc skis? By this I mean an area where the snow might drift deep enough to form a secure bridge, the water generally shallow enough not to drop you in to your hips if it isn't secure and the banks low enough to scamper up without removing the skis.

 

[DougPaul]

It would be a _hugely_ serious issue. My experience is that there's a slipped decimal point there or something, but I'm sure you're correct and my observations have been wrong. I have not claimed to be an expert.

I don't claim to be an expert either--just a knowledgeable layman and hardly infallible.

I just looked at some old conversions and, again, I'm getting _much_ smaller differentials, tiny compared to your 230m. I must be seriously misunderstanding the WGS84/NAD27 issue. If I am wrong I'm sorry if I've misled anyone.

I just redid my estimation of the distance and got 226 meters from NG TOPO! and 217 meters from Garmin Mapsource. (I originally made this estimate in 2005 using procedure 2 on NG TOPO!--got the same exact number.)

Procedure 1 (direct):
* Place a waypoint (in this case, on Crawford House in Crawford Notch)
* Read out its UTM coordinates in WGS84
* Place a second waypoint using the above numerical coordinates under NAD27
* Measure the distance between the waypoints

Procedure 2 (indirect):
* Place a waypoint (in this case, on Crawford House in Crawford Notch)
* Read out its UTM coordinates in both WGS84 and NAD27
* Compute the distance between the two sets of coordinates

Note: UTM reads out in meters, thus the distance is in meters.

Procedure 1 is direct and procedure 2 is indirect but both should give the same answer. My results:
* 227m, NG TOPO!, procedure 1
* 226m, NG TOPO!, procedure 2
* 217m, Garmin Mapsource, procedure 2

Obviously these procedures are testing the datum mappings in the respective programs--hopefully the mappings are reasonably accurate. I presume that one could also perform the same estimation procedures in a GPS, again hoping that the mappings are reasonably accurate.

Doug

 

[DougPaul]

Hi, Doug, In answer to your question:WGS84. However when I made those hikes GPS was just a dream. Therefore the Lat/Lon I gave was just an approximation of my location then.

Thanks. The estimate of the accuracy is also useful.

It's really interesting seeing railroad steel lying right where it had been placed during J.E. Henry's time. Right where the tracks end, there is a two or three foot long of rail lying between the tracks that looks like it was sawed from a rail. Knowing just hard hard RR rails are, I have often wondered about that piece. If it was flame cut, the person who cut it was really good.

When I was over in that region, I was on skis and never saw the ground or anything lying on it...

D.

 

[DougPaul]

Any speculation on the best crossing with xc skis? By this I mean an area where the snow might drift deep enough to form a secure bridge, the water generally shallow enough not to drop you in to your hips if it isn't secure and the banks low enough to scamper up without removing the skis.

In 2002, I crossed the N Branch on skis at the obvious spot on the TFT described in my earlier post to a spot near Jumping Brook. There was an obvious rail corridor paralleling the river on the N side. I only followed it for a short distance.

Doug

 

[MichaelJ]

We all agree that NAD83 and WGS84 are close enough that to a consumer GPS the difference doesn't matter.

This chart shows the difference of NAD27 to NAD83 in New Hampshire should be in the 35-40m range.

Confirming that, according to this NOAA utility, using a sample point mentioned earlier in this thread of N44 07.522 W71 30.484, the datum shift is 38.9m.

Still a big distance, over 100', and thus could make recovering that waypoint difficult out in the wild; however, not nearly as bad as the estimate of over 200m.

 

[DougPaul]

We all agree that NAD83 and WGS84 are close enough that to a consumer GPS the difference doesn't matter.

This chart shows the difference of NAD27 to NAD83 in New Hampshire should be in the 35-40m range.

Obviously inconsistent with the conversions that I am getting from NG TOPO! and Garmin MapSource which are in agreement with each other. (I also tested my Garmin 60CSx--it agrees with MapSource.)

Confirming that, according to this NOAA utility, using a sample point mentioned earlier in this thread of N44 07.522 W71 30.484, the datum shift is 38.9m.

If you read the notes on the figure, it is derived from NADCON which is the source of this utility. So it is not an independent confirmation of the figure.

Still a big distance, over 100', and thus could make recovering that waypoint difficult out in the wild; however, not nearly as bad as the estimate of over 200m.

Agreed 130 ft (40m) is still big enough to cause serious problems.

I don't know why there is the factor of 5 difference between the TOPO!/MapSource and NADCON WGS84--NAD27 distances. (Somehow I would expect the conversions made by the products of one of the premier commercial GPS producers to be fairly accurate...) I don't know how independent the TOPO! and MapSource conversions are--it would be nice to find an independent third source. (TOPO! may already be an independent third source...)


A note:
There are multiple NAD27 datums (MapSource lists Alaska, Bahamas, Canada, Canal Zone, Caribbean, Central, CONUS, Cuba, Greenland, Mexico, and San Salvadr) and many sources, including the NADCON facility, list only a single version named "NAD27" without qualification. (I'm guessing that this refers to NAD27 CONUS.) I compared Central with CONUS (using MapSource)--it was only 13 meters change.

Doug

 

[DougPaul]

I think I have figured out what is happening... (Both numbers are correct... )

If I assign waypoints to the same set of Lat-Lon coordinates in WGS84 and NAD27, they are 38 meters apart.

If I assign waypoints to the same set of UTM coordinates in WGS84 and NAD27, they are 227 meters apart.

(These numbers are from NG TOPO! using locations close to the Crawford House site in Crawford Notch.)

So there is something about the UTM mappings which depends upon the datums in a way that causes a much larger position shift as a function of the datum.

Doug

 

[MichaelJ]

Because UTM is a Mercator Projection, perhaps?

Note that if you look on the bottom-left corner of the USGS Crawford Notch quad, it states:

Horizontal Datum ……………… 1927 North American Datum
To place on the predicted North American Datum of 1983, move the project lines as shown by dashed corner ticks (3 meters south and 39 meters west).

So:
a) It appears the 30-40m is the correct one for on-the-ground distance.

2) Ahh … THAT's what those little tick marks on the map are!

 

[DougPaul]

Because UTM is a Mercator Projection, perhaps?

Don't know--I'm no expert on projections from a spheroidal earth to flat pieces of paper.

Lat and Lon are well matched to locating positions on a spheroidal earth so I regard them as the "best"* means of specifying locations.

In contrast, UTM is a collection of 60 local projections of zones from the spheroid to flat maps. Perhaps the difference is due to errors in these projections. Or perhaps the projection from lat-lon to UTM is sensitive to the datum.

* While either method can specify a location to any accuracy, lat-lon allows one to calculate the relationships (eg distance and direction) between two points accurately whereas any projection to a flat representation contains distortions which will introduce errors in the relationships.

Horizontal Datum ……………… 1927 North American Datum
To place on the predicted North American Datum of 1983, move the project lines as shown by dashed corner ticks (3 meters south and 39 meters west).

So:
a) It appears the 30-40m is the correct one for on-the-ground distance.

Agreed for a lat-lon representation, but my evidence suggests that it is not true for a UTM representation.

<comment>
My initial assumption was that the change due to the datums should be the same for either position representation (my 226m figure was computed in UTM based upon this assumption), however the evidence is suggesting that it is not the same.
</comment>


I think it is worthwhile to be aware that position errors due to using the wrong datum are much greater (at least in the Crawford Notch area) using UTM than when using lat-lon. However, I've spent enough time on this issue and will be happy to wait for someone adequately knowledgeable about map projections and datums to explain why.

My basic point remains: one should specify the datum with the coordinates.

Time to get back to the Pemi Wilderness?

Doug

 

[forestgnome]

Any speculation on the best crossing with xc skis? By this I mean an area where the snow might drift deep enough to form a secure bridge, the water generally shallow enough not to drop you in to your hips if it isn't secure and the banks low enough to scamper up without removing the skis.

I've been thinking about this as well. I've been wanting to try xc skiing and this was the route everyone described as the best in the Whites for easy skiing Too bad I never got to enjoy it before it was taken from the public. I'm hoping that at some point a natural bridge will form at some point if you go far enough upstream. I would be nervous to be over water with planks attached to me feet. Anyone?

 

[Becca M]

I've been thinking about this as well. I've been wanting to try xc skiing and this was the route everyone described as the best in the Whites for easy skiing Too bad I never got to enjoy it before it was taken from the public. I'm hoping that at some point a natural bridge will form at some point if you go far enough upstream. I would be nervous to be over water with planks attached to me feet. Anyone?

End of Feb last year, it was common to see ski tracks crossing the Pemi River where the old suspension bridge was. Snow bridging looked good, but, I would be scared to cross the Pemi River there because once ya sink a ski there, you might as well have a boat anchor around your neck :0 Near the Thoreau Falls Trail Bridge, the water was moving too fast and deep to create a snow bridge, so, hopefully *that* bridge will be there for a while!!!!!

 

[DougPaul]

End of Feb last year, it was common to see ski tracks crossing the Pemi River where the old suspension bridge was. Snow bridging looked good, but, I would be scared to cross the Pemi River there because once ya sink a ski there, you might as well have a boat anchor around your neck :0

The crossing didn't look too bad... (there are ski tracks visible over on the right)

There were also ski tracks on portions of the N Branch. (FWIW, Goodman suggests skiing on the river in his guidebook.) We probably could have crossed without difficulty at the location of the old crossing that I mentioned earlier in the thread--I did it out and back in 2002.

Near the Thoreau Falls Trail Bridge, the water was moving too fast and deep to create a snow bridge, so, hopefully *that* bridge will be there for a while!!!!!

Keeping my fingers (and ski tips) crossed...

Courtesy of my pics (their memory is better than mine...) it looked crossable above the bridge but not below. Pics at http://mysite.verizon.net/dbpwebjunk/pemi-lollipop-2011-02-22/index.html

Doug

 

[Becca M]

another question

Another question:

Do people who have been to the mouth of the North Fork into the Pemi River think a river crossing of the North Fork would be easier to find coming from the north side of the Pemi (from the Lincoln Woods Trail) or from the East side of the North Fork (just after the Thoreau Falls Trail Bridge)?

 

[Becca M]

Today I roughly followed the North Fork crossing location - at the mouth of the North Fork into the Pemi. The crossing right at the mouth definitely looked the most favorable compared to upstream. It was pretty wide and semi-shallow, but, a LOT of rock hopping as a result of the width.

Although I continued north (rather than west along the pemi to the Black Brook/molly monument area), the woods in general seemed reasonably open.

 

[Will]

The crossing right at the mouth definitely looked the most favorable compared to upstream. It was pretty wide and semi-shallow, but, a LOT of rock hopping as a result of the width.

We found that spot wide and slow, but with several yards pretty deep compared to upstream a little bit. One described it as requiring "swimming" through that short stretch of the crossing, but I'll certainly take another look.

Although I continued north (rather than west along the pemi to the Black Brook/molly monument area), the woods in general seemed reasonably open.

Yes, that's a major issue here. You go high up and it seems it may be easy *, or keep low to the East branch and it's easy, with some interruptions. It's that mid-level, 20-150 yards or so up the East Branch bank that has some horrors.

* and there's that old rail bed up high as you get closer to the monuments. My attraction to staying low to the East Branch bank is that it's such a pretty walk (not that open woods high up isn't nice too).

 

[Will]

I just did the Molly Route from the foot of the Bondcliff Trail to the Thoreau Falls Trail for the first time in years, with a newer GPS to check old data, and things are pretty much as I reported and the data points accurate. Amazingly, cheap consumer GPS gadgets are apparently a lot more accurate than a lot of maps! I was again walking at times in the middle of the River or even on the other side, according to the maps.

The abandoned section of the Wilderness Trail is clearer than it was: the NFS seems to have given up hiding it from us. Most of the saplings are gone and the debris kicked to the side. It’s a trail. The path from the top of the removed bridge down to the river is less clear, that’s puzzling, but it’s still there. Lots of footprints in the sand along the river moving towards the Thoreau Falls Trail.

I stayed on the River’s edge for all but maybe 5-6 very short sections where I had to go only a few yards in, twice about 20 yards in but never deeper into the woods. There lie monsters. The first detour was right at the beginning, but very quickly back out to the “easy ledges” I had described. I’m sure there are good paths way further up the bank but I enjoy walking along the River as long as the flow level allows it. Sometimes you temporarily rock step out into the river to keep going, counter-intuitive, just look ahead for your easiest steps. Better stepping on low flatish rocks out into the river than scrambling over boulders right at the shore.

When I got to the island I talked about, I decided to just walk the island itself. Very nice. End to end. Easy walking and some nice resting spots.

Crossing the N branch of the E Pemi was much easier than it was as I remembered. I did it much lower ( just a short way N of the fork) and dry booted across. My convoluted route from there to N of the the needlessly-destroyed Thoreau Falls Bridge and the Trail is still there. I’m sure there is easier, but what I reported works. The two campsites I mentioned show disuse, also puzzling; I guess the destruction of the three bridges in this area has drastically reduced visitation, I guess that was the Plan. The Thoreau Falls Trail itself seems much less travelled, I really felt alone.

A thoroughly enjoyable trip, Been away from it too long.