Discussion: GPS vs LiDAR - Position Disagreement

Hello -

A very careful selection and GPS (Juniper Geode, <1m, usually <0.3m, by its app's estimation) capturing of 29 features on a prospective Sprint map (here in Colorado) and subsequent analysis of the differences revealed a median 1.39m total error (a large 0.38 sd) with median component errors of -1.2mE and 0.46mN. Easting error ranged from -0.3m to -1.8m and Northing error ranged somewhat troublingly from -0.9m to 1.2m.

It's clearly systemic and not random, as least for the Easting part, For whatever combination of reasons, even reprojecting to a WGS-based projection using OCAD or R (PROJ v7.2.1) result in a map file that my GPS points agree with. I've been through the HTDP calculation process using local benchmarks and am convinced that my GPS receiver is not systematically off.

I can easily process the LiDAR in R such that all Xs and/or Ys are shifted by amounts I select, thus bringing my median total error down from 1.38m to about 0.5m, based on my 29-point sample set. I can then geo-adjust the imagery to align with the LiDAR (or my GPS points), and thus have three sources (GPS, LiDAR, and Imagery) in quite good alignment.

Here's my question:
Are there any practical or philosophical reasons why I should not take this approach (assuming I disclose what I've done in my mapper's notes)?

Thank you.

Many of us without a sub-1m GPS have probably got this, or some other systematic discrepancy in our sources of information.

I'm amused by thinking what the mapper's notes would say. In terms of oldskool technology, how do these discrepancies compare with the thickness of an 0.5mm pencil line on 1:7500 field notes? Am I reading this correctly that you're talking about distances that are roughly arm's length? That are around the range of SI-Air? What is it that you think people need to know?

Thanks both, of you.

@gruver Yes, I'm sure with a non <1m receiver one would have the same WGS84 vs LiDAR datum-generated median discrepancy, just with an even larger standard deviation.

@jjcote
That basically answers my question. Somewhere along the line I gathered that <2m accuracy was expected of top-grade ISOM maps and <1m accuracy was expected of top-grade ISSProM maps ('tho I'm pretty no such accuracy requirement is in the actual specifcations for obvious reasons). If this is not the case then the only point of adjusting the LiDAR would be to relieve the systemic annoyance.

To answer yours, my notes would state the amount of adjustment applied to the LiDAR. This way a subsequent mapper seeking to expand the map would have a fighting chance at getting abutting LiDAR tiles' contours, etc., to align properly. Rubbersheeting is only so much fun.

I'm just wheeling in the latest imagery for the urban mapping round here. Its great (.075m resolution) but many of the buildings have moved. Should I report this to the Earthquake Commission? Or just put it in the mappers notes?

File a complaint against the surveyor for having directed the buildings to be put in the wrong place. :-P

even reprojecting to a WGS-based projection using OCAD or R (PROJ v7.2.1) result in a map file that my GPS points agree with.

Not sure that sentence makes sense - are you you saying that reprojecting the lidar means the lidar and GPS match? If that is the case then the problem is solved surely.

JJ is right that under 2 m error is insignificant if you are working on a 'forest' map (10000/15000) but it might be noticeable on a 'sprint ' map (4000) . But still it is not a lot and any problematic discrepancies could easily be fixed by a mapper working in the traditional way - drawing what they see in real time in the terrain (either on Mylar or using a tablet). I imagine it would only be a real noticeable problem if you are trying to just record gps tracks with waypoints and do all the drawing/interpretation back at the office.

Actually I find with a good base map of sprint terrain (parks/urban) generated from lidar and high quality orthophotos, GPS is mostly unnecessary (in fact it is often more of a distraction). The base map is going to be more than accurate enough and the real skill the mapper needs will be in knowing how and when when to distort reality a little to make things understandable and readable ie: how to work with symbols that have footprints larger than the feature on the ground and when to displace objects to maintain minimum gaps, how to exaggerate contour features (not just use the lidar contours exactly as they are), etc, etc. Stuff like that is best done in the field and GPS isn't much help.

And I wouldn't bother explaining all that projection stuff in the event info - it's irrelevant to the runner and confusing. TMI

but many of the buildings have moved.

what do you mean?

If I remember correctly Australia moves north about 7cm per year, position was fixed last time 90's or so, so Australia is now about 2 meters off compared to GPS. Unless they have updated it.

uh-oh - we'll soon be part of China.

Isn't that sort of continental drift addressed by the national coordinate system - in Australia it is GDA. There is a 2020 version and a 1994 version - surely GDA 2020 would be pretty accurate.

All continents move - not just Australia. Japan moved 2.4m in one day in March 2011. https://earthobservatory.nasa.gov/images/148036/te.... Which might explain why the coordinate system there was updated in 2011 even though it had been updated quite recently (2000)

Australia might become part of China soon enough but it wont have anything to do with continental drift.

That discussion reminds me a bit of footnote 1 in a report I once wrote on a cold wave in the Australian tropics.

Around here we move just 2.5 cm/year and currently are 70 cm off. We also rise about 1 cm / year, still recovering from the continental glacier we had quite recently.

Anyway, controls move during a race so slowly we don't have to aim off.

I now understand the original poster to be asking about info to be passed on to the next mapper, rather than to the competitors, and wondering whether the map should be true to the lidar georeferencing or the GPS readings. A reasonable question, though one that I personally wouldn't worry about given the magnitude of the numbers.

I take a very pragmatic approach here: The government mapping authority base material (LiDAR and vector topo data) is assumed to be exactly according to WGS84, any GPS offset is just that, local (in time) GPS offset. Using survey grade (sub-2cm) GPS gear does tend to correspond very closely to the existing mapping data, but I never move any of the base data unless very clearly wrong, and in that case I report it back to my friends in Kartverket. :-)

@robplow Dec 20, 2021 12:16 AM
My apologies, a bum sentence on my part.

What I meant to say is that I expected the locations of features in a base map created in a WGS84-based projection to agree with the locations of those features as returned by my GPS.

What I found is that the unexpected WGS84 <-> GPS disagreement is about the same magnitude and direction as that for NAD83 <-> GPS.

I.e., reprojecting to a WGS84-based projection using OCAD or R (PROJ v7.2.1)does not solve the problem.

Helpful comments to the effect that GPS doesn't really add a lot on a sprint map and that accuracy, after a point, doesn't actually make an O-map any better.

Thanks to all.

Bear in mind that many of us cut our teeth on maps that were certainly not georeferenced (that concept had no meaning), and where magnetic north and scale might be suspect. Also worth considering is that although these days there's a post-race step where GPS comes into play, when actually out on the course, variations in humidity can result in scale differences that swamp the discrepancies that you're seeing, if the map is printed on ordinary paper. These differences are large enough that it's considered inadvisable to fieldcheck with a paper basemap, due to the potential problems in getting the fieldnotes registered properly when drafting. These days it's possible to adjust the scale after scanning, but in the pen-and-ink days, this could be a serious annoyance.

Still, meaningful or not, for academic reasons, or just curiosity, if someone with genuine GIS knowledge is able to explain properly the offset in question it would be cool to know.

Could it have something to do with continental drift? I mean WGS84 and NAD83 both sound like they are nearly 40 years old. How do they account for moving continents? I would imagine the US west coast would be fairly mobile given the seismic activity. Maybe not so much in Colorado,but still . . .

@robplow Dec 20, 2021 12:16 AM
My apologies, a bum sentence on my part.

Yeah, I guessed you had mis-written - but wasn't sure.

JJ, pre lidar/gps O maps were georeferenced in a way - photogrammetrists used existing topo maps to set up their models using aerial images which were otherwise not georeferenced. They just didn't use that word. But certainly you are right that the concept of georeferencing (whether or not the O map conformed to a coordinate system) was totally irrelevant to the end user and even to the mapper.

Actually good control of scale/angle was an important part of the work. At the start of a project I always drew a 200 or 250m grid aligned to magnetic north (my own coordinate system if you will) which made it easier to control for distortion due to photocopying/scanning errors or paper expansion/ contraction, etc (although I did try to avoid using paper for the reasons you mentioned - always tried to use mylar copies of base maps) . Also I always had a mylar copy of the original base map as the lowest layer on the light table so that I could line up the fieldwork to that base map every time - and adjust slightly from time to time if there was any distortion.

Failure to do stuff like this is why a lot of old O maps are horribly distorted and we need rubbersheeting when trying to georeference them. The average orienteer would never have noticed the distortion but I hate to think how many navigational errors were caused by such distortion but the orienteer blamed them on their own poor technique.

It is true that the basemap was originally set up to real-world coordinates, but that was usually lost in the process, since nearly nobody had any way of utilizing it anyway. I think I do have a very few printed maps from back in the day that had some marks in the margins with the coordinates, but they were really just a curiosity. Almost nobody ever paid attention to the numbers written on the edges of the basemap.

Magnetic north had to be added by whoever did the fieldwork, and I received at least one set of fieldnotes to draft where it was definitely wrong. (I caught that one before I started drafting.) And more than one time, I got the fieldwork in little pieces, one by one, without a master basemap to align them to. The stuff of nightmares.

Your tales of horror certainly sound very familiar.

Yep, magnetic north was calculated by taking bearings along straight objects (fences etc) identifiable on the base map. Ideally you would do that for several objects spread across the area to be mapped and averaging the result. But it wasn't always easy to find enough straight objects.

I had one base map where the north was different in different sections (2-3 deg different) - the inexperienced photogrammetrist had not been careful enough setting up the models of different photos. After wasting a day or two wondering why nothing would line up properly and eventually figuring out the problem I used two different norths for the fieldwork and averaged them for the final map.

On that same base map there was a contour that just stopped right in the middle of the map. I had to add in the contour for half the map - and adjust the nearby contours up and down to make room.

I guess I need to explain my "moving" buildings. But if you have to explain a joke then its clearly not a very good one:-((

The buildings which I referred to were drawn according to some previous geo-referenced photos, positioned by OCAD according to their world files. The latest geo-referenced photos, positioned in the same way, show a very small mis-match to the building outlines. Although we are geologically active, any ground movement is likely to be millimetres. What could well be a metre or two is the quality of the photos and the ortho-rectification, and the effects of non-vertical sight lines in some places. I'm not an expert in how all that is done, but I imagine the building envelope as seen from the sky can well seem to have moved. Perhaps easier to follow is that the effects of shadows, and eaves, require interpretation by the mapper, and can be different with a new set of photos.

What is happening is similar (but a lot smaller than) the way a feature "moves" in Google Earth when you use historic imagery. I feel no need to move my buildings in general, but I expect to discover some misinterpretations from the earlier drawing.

Did the Wellington area move appreciably in the Kaikoura earthquake?

No simple answer. Parts of the South Island have moved 5m closer to the North Island, and there were similar vertical displacements. But the lower North Island is itself moving slowly as the pacific techtonic plate slides under the Australian plate, in slow-slip events that do not result in shaking. As a result of googling your question I couldnt find a straight answer but now think the ground movement between my two sets of photos may be centimetres, not millimetres. I still prefer the theory that interpreting the photos is the likely cause of what I have observed.

We had one long, skinny map (in Colorado) where we were guessing for a while that multiple models had been set up that were not consistent, such that the north, middle, and south sections were slightly rotated with respect to each other, and the photogrammetrist had fudged things together where they joined. This also provided an explanation for why there was a place where it seemed like you had to run too far compared to what you expected from the map. (Although eventually somebody went out and did some additional work on that map and it might be that he was able to fix it by just correcting magnetic north, I'm not sure.) In any case, it was immediately adjacent to another basemap that never got fieldchecked, one of the reasons being that there were many, many places where the contours crossed each other.

An unrelated magnetic north issue is on Surebridge Mountain, site of the WOC93 Classic race. It's an area that had iron mines, and there are effects on compasses. I stood in one place and lowered a compass from waist level to the ground, and watched the needle move 90 degrees. There's another large hillside (hundreds of meters long) where the compass indicates north to be 45 degrees off from where it is everywhere else. One of the mappers suggested in all seriousness that the magnetic north lines should be drawn with kind of an S shape to show that. What we actually did was to keep the WOC courses away from the problematic spots.

Interpreting orthophotos for base map purposes is tricky. Like you say it can be hard to distinguish between shadows and objects. Also you need to beware of the potential parallax errors of tall objects like trees, buildings, power poles, etc. If there is an apparent discrepancy in the location of objects between lidar derived images and the orthophotos I always go with the lidar.

I never trust Google imagery - I only ever use it as a last resort. Their algorithms seem to do weird stuff to the images - accurate, consistent georeferencing is not their highest priority.

One of the mappers suggested in all seriousness that the magnetic north lines should be drawn with kind of an S shape to show that

Mr Key?

No, not Steve, it was one of the first-pass mappers, János Sőtér. At least, I'm almost positive that's what he was saying. His English is a lot better than my Hungarian, but that's not saying much. Sometimes we communicated in German, and I don't speak German.

Georeferenced an old (very well used) map and discovered that it was 1:14500 in one axis and 1:16000 in the other. Despite the map being used for over 50 events, this scale information was news to everyone. I think this demonstrates the potential for unnecessary precision with the latest in GPS technology. The solution was remapping with a lidar base.

I hope that wasn't one of my maps?!

unnecessary precision Are you suggesting that just because no one noticed the incorrect scale of the old map it was acceptable? That it is "unnecessary" to get the scale right? Just because no one noticed doesn't mean they weren't negatively effected.

Expecting the scale to be accurate (at least a lot closer than 14500 or 16000) is not "unnecessary precision" - it is just a reasonable expectation of minimum quality.

I remember running a WOC team selection race back in the 80's. The printer, (unbeknownst to the organizers) had reduced the size of the map to fit on a standard paper size. The printed map was around 1:17000 I seem to recall (over 10% out). I remember having a terrible race - always stopping short of controls, making mistakes, etc and being thoroughly demoralized by the end of a very long race in tough terrain. I wasn't the only one. During the race we didn't 'notice' that the scale was wrong but we were most certainly affected by it.

The organizers figured out the problem on the day (clearly neither course setter nor controller had picked it up which is a worry!) but decided not to inform the runners because they were afraid people might protest and disrupt the team selection process. I only found out about the incorrect scale six months later. I left that race feeling quite demoralized at my poor technique that day - but it turns out it wasn't my fault.

Also if the scale is wrong that means the course lengths will be wrong and the runners' km rates will be wrong. After that event it looked like we were running slower than we were (a further source of demoralization).

I also recall some discussion once about the fastest ever km rate run in Australia: the claim being it was set in the 70's on on the Eppalock South map. Except it turned out later the map scale was wrong (this time around 12000 -13000 I think) and people were not running as fast as they thought they were.

How did the sport ever survive some of the mistakes we mappers made back so many years ago. I shake my head at people worrying about a difference of meters or fractions of a degree.
I think I may have told some of this story before but settle in- or skip over as you wish. Back around 1973 Orienteering Quebec ordered several base maps from a Scandinavian company of flawless reputation. The maps of areas in the Montreal (Lachute) and Ottawa area were to be used for major events from 1975 to 1978 including provincial, national and North American championships as well as the O'Ring Quebec 5-Day.
Despite the difference being some 13 degrees none of the field workers - self included - noticed that the meridian lines were drawn to grid north, not magnetic north. In Scandinavia that difference would haver been minimal but in Quebec 13 degrees is 13 degrees.
In my defence I'll say that the rest of the base map was so good that I was rarely shooting in a point object from more than 100 meters or so away and cliffs could be placed on the contour bends without checking by compass. All the marshes were already there, the base maps having been made from spring photos.
So we got through the competitions without protests. There were some questioning why they were most often off their compass headings to the right but I think it wasn't until a second round of field work was due that we noticed our compass bearings did not align exactly with the roads we were checking.
One of the maps was quietly corrected. The others unfortunately have not been seen since those heydays of the 1970s.
So worry about your fractions of a meter or a degree if you wish but I'll keep in mind what a good friend of mine used to say, "it's not going to be noticed by someone riding by on a flying horse."

Great story!

At the risk of repeating myself unnecessarily - just because "it's not going to be noticed by someone riding by on a flying horse." doesn't mean people weren't affected by it. And clearly the races on those maps were not 'fair' - clearly people were making major compass errors which makes the results rather random - which is the antithesis of what a serious competitive sport is about: rewarding skill/ability not chance.

Sure we should keep things in perspective - I think we have all agreed that the 1-2m lidar/gps discrepancies that inspired this thread are not significant. But I would argue that a 15000 map actually being printed at 16000 IS significant. Maybe a 1-2 deg north error is insignificant but certainly a 13 degree north error IS.

I don't mean to cast shade on anyone who produced these maps - I have been guilty of plenty of my own mapping sins - but it seems to me the response when you realise these errors have been made should be 'how can we make sure this doesn't happen again' rather than 'no one noticed so it doesn't matter if it happens again',

I find this 'if no one notices then it is OK' attitude somewhat worrying. It almost guarantees the mistakes will be repeated.

Also, think about how many people have negative experiences getting lost again and again on maps that have such systematic errors. How many people give up orienteering because they felt they were not good enough - when in fact the problem was the map not them. That is the real take away from my story about the race on the map with the wrong scale - I didn't know the scale was wrong but it certainly affected me and that negative experience put a serious (albeit short lived) dent in my confidence. Stuff like that DOES matter.

I agree, but I think the point was that if a 10% scale discrepancy or a 13 degree angle error, both of which are unquestionably horrible, managed to go unnoticed, then a shift of <2 meters (from an invisible reference) isn't something to lose sleep over. I don't think anyone is say that either of the first two were actually okay.

Rob

First, rest easy. It wasn't one of your maps.

Second, I wasn't arguing that a scale difference of that magnitude didn't matter. We did a complete remap using lidar to remove the problem. But if no-one noticed that error in 50+ events, then the 1-2 metre differences that started this thread don't have much significance out in the forest.

10% scale discrepancy or a 13 degree angle error, both of which are unquestionably horrible, managed to go unnoticed

The wrong North direction at Saylor Park definitely was immediately noticed, on the leg to control #1 in fact.

Regarding 10% scale difference, on the old maps (some made without any base map at all -- based on s'emochnoe obosnovanie ) one of the essential skills was sensing that *locally* the scale is smaller or larger than nominal, due to in part the mapper spreading a cluster of visually related objects to overcome the minimal size limitation. Rob would be demoralized.

Ok let me put it like this:

I would far prefer to work with someone like CompassCoyote who notices a 1-2m discrepancy and investigates and asks around about it, than someone who just automatically thinks: 1-2m nothing to worry about, no one will notice. Even if it turns out 1-2m is insignificant the CompassCoyote type (who pays attention to the detail) is far less likely to make those big mistakes that are significant - like getting the scale and north badly wrong.

Yurets: the old Soviet mappers were famous for their ability to make excellent maps with no base map at all. To do that you need to be very good at bearings and measuring distance, which means pace counting. With those skills you are unlikely to miss a north alignment that is more than 1 or 2 deg out or a 10% scale error. Major event maps have been made like that even until quite recently - I remember reading about the making of the WOC2007 maps around Kiev with no base maps - using a home made clinometer to map the complex sand dunes contours. With skilled mappers like that around it is a bit demoralising when others make a mess of the basics like scale and north.

That's a tenuous connection Rob. I've seen the outline of an area covered in 1m paving slabs drawn as a zigzag. And the same mapper forget about the grid-magnetic north angle entirely. Around here its 23 degrees. (Not getting at you at all CC. Your skills may even include the recognition of feeble humour:-))

Let me rephrase that again: I think someone who is concerned about a 1-2m discrepancy between lidar and gps is far less likely to make a big georeferencing mistake like getting the scale or mag north wrong. I certainly don't think having an awareness about that sort of discrepancy is nit picking.

What level of detail to map (eg the edge of that paving) and what is or isn't overmapping is a whole different kettle of fish.

Yurets is absolutely correct that the Saylor Park magnetic north error was questioned immediately by the competitors, the first time the map was used. This is also true of other maps I know of with incorrect magnetic north. The "unnoticed 13 degrees" I was referring to was the one that Gord mentioned, that reportedly went unnoticed by the mappers, but not by the competitors. I know of at least one other case like that, where the mappers didn't notice. This is more likely to happen if the basemap is very good, I think, because it means that there's less need for the compass while doing fieldwork.

Mistakes in mapping will happen and they should be avoided. When discovered they should be corrected. If our 1970s meridian mistakes had been discovered before the events for sure we would have reprinted. It was not as someone seems to suggest that anything was covered up. We just trusted the photogrammetry so much that we did not notice.
On an otherwise good orienteering map in the hands of good to better orienteer there is going to be more than enough information to help the orienteer make compensation for the unknown errors.
For instance on the map I worked on which had the misdrawn meridian lines all the marshes were in the right place compared to the contours which were also in the right places thus the nearby cliffs and boulders were close enough to being correct.
Those who ran at COC at Lac Phillippe in '76: do you still have your map? Is your route marked on. Go back and see if you see a pattern.

I certainly wasn't suggesting the Canadian meridian error was covered up.

The case of scale error I mentioned in Australia certainly was covered up. Interestingly in that case the mapper (Rob Vincent) also ran the course that day (he was already preselected for the WOC team) and it was him who noticed the error. I guess since he mapped the area it gave him extra insight but I also I think because he was running the course more as a training he had the time to consider why things didn't feel right and work out exactly what was wrong - a luxury you don't have when you are in full competition mode. In a case like that it is actually pretty to easy to work it out - you just have to measure the distance between the northlines and do some very simple math.

Personally I map with the compass attached to the mapping board (something I originally learnt from Steve Key) and always keep the compass lined up with north I never use other features to orient the map - always orient the map with compass then sight other objects and draw. That way there is a double check built into everything I do - if a previously mapped feature is mapped in the wrong position it will quickly become apparent because something wont line up properly when the map is oriented using the compass. Mapping like that I am confident that I would always pick up any north error greater than about 1 deg. Something that has actually happened a few times. The other advantage of doing it like that is I never have to stuff around taking a bearing, rotating the housing then putting the compass on the map in the right place, etc etc. I have never understood why people map with a loose compass.

I do it the way you do it, but for people who like using a sighting compass, that works better if it's not on the map board. (But it's always seemed to me to be a very slow operation.)

To do that you need to be very good at bearings and measuring distance, which means pace counting

Pace counting will not be good enough. In practice a long rope of a known length was used. A crew of 2 people went around the perimeter of the area, using the rope for distance, and taking measurement of both horizontal and vertical angles between two consecutive nodes.
In the end (after going around the closed countour) the accumulating plane and vertical errors were spread over, and the resulting data was the starting point for mapping the interior.

JJ people who use sighting compass can still put a fixed compass on the board as well (unless they are within a few cm of each other there is no interference). I used to do both - used the sighting compass for longer bearings - until I realised I was just as accurate (and much quicker) with fixed compass as I was with sighting compass.

Yurets, Fascinating. Very innovative. But I assume once the initial work with the rope was done the mappers were just pace counting. How did they measure the vertical angles?

Reminds me a bit of the Great Trigonometrical Survey of India where they mapped a whole sub-continent using basic triangulation techniques - took them 70 years. They measured the distances of baselines using chains - more stable than rope.

I have this image of the resulting map on my office wall - one of the great feats of mapping https://commons.wikimedia.org/wiki/File:1870_Index....

I think I picked up the fixed-compass-on-board from the same source as Steve Key, working with Bakken & Helgesen.
I always had a separate non-sighting baseplate compass, as much for the pacing scale, as longer one handed bearings, which is more of a necessity in my local "French Creek" style terrain.
Only in the last decades, when doing mostly revision work, have I gotten to rely almost exclusively on board oriented bearings, B&H style.

I carry a broken piece of a ruler for the pacing. I recently discovered the joy of printing the basemap at some random scale where my pace count is conveniently a round number, so I don't have to do any arithmetic in my head. Wasn't an option in the days of fieldchecking on a mylar sepiatone copy of the original base.

I print out a scale in paces and tape that to a ruler on a string. I have several depending on the fw scale (eg forest, sprint or school map) . Pre digital days I did the same but hand drew the pace scale. It was never a good idea to rely on mental arithmetic - that was just asking for avoidable errors - I don't have the brain for that, especially when tired/cold/hot/stressed etc etc.

Never occurred to me to adjust the scale to match my paces - clever!.

I just make a point of taking one-meter steps when doing fieldwork, so the arithmetic is easy. Of course, this option is not practically available to everyone.

Pat Dunlavey also used that approach. Too hard for me to be consistent doing that.

Just like @ebone I also developed a very accurate 1m stride length which I used for all field work, as well as any time when I needed to estimate some distance fairly accurately. I typically got within a few percent of the real distance, which meant that any object placed this way would be close enough to exact that nobody would notice any actual error. Now 40 years later I can still turn on that particular stride and feel quite confident that it will give a well under 10% error.

These days I use lidar for everything, using 1m contours plus a slope image to locate the exact placement of larger paths and tracks, as well as small streams. When I find boulders I can often locate the boulder shadow on a high-res ortho photo, and use that to fix any 50-150 cm errors.

Gord - On reading this today I immediately went to my files and miraculously discovered I still had two Brownsburg Village maps from the 1977 North Americans with my routes drawn. Is this one of the affected maps? For all of my significant errors (seven) I was off left. I think your "off their compass headings to the right" is backwards; if I set the angle on the map with grid north then run with that setting, I will be off left where mag north is left of grid north. This was my first ever "A Meet"; maybe with the correct north lines I would have done better than 8 out of 23 in M35; I know I would have done better if I had not been wearing blue jeans on a wet day -- never again!

For all my mapping I used a settable hand compass with pace scale taped on its edge (no need for a separate ruler or mental calculations) and a compass taped on my board to keep it oriented; I lost a sighting compass in the woods very early on and never replaced it.

Sorry Tony but the error was caught and corrected before the Brownsburg Village map came into use. I think it was from another photogrammetry order, also.

But you I think are right about the left. Give me some slack. It was almost 50 years ago and several thousand miles away.

Darn, must have been me that was off. Sure, at our age we need all the slack we can get.

maybe with the correct north lines I would have done better than 8 out of 23

maybe with the correct north lines everyone else might have done better also.

But it is a shame that the bad experience with blue jeans in the rain discouraged Tony so much that he never came back, and therefore never introduced his children or grandchildren to orienteering, either. Maybe one of them might have turned out to be good at it.

:-)