Lat / Lon Conversion

[colin.green14]

Hi

I am visiting Ypres area in October, and have been plotting the last steps of my Grandfather using tmapper to convert Trench Map co-ordinates to lat / lon.  Have now purchased Linesman (Ypres DVD) and I find that the latitude and longitude's are different.  Is there an explanation or is it me, and if it is me what am I doing wrong.

 

Regards

 

Colin

[MaxD]

I haven't got both systems to cross check but an example on tmapper compared with Google Earth agrees within decimal points of seconds.  What is the difference between the two systems?

 

Max

[MaxD]

Had a look at the Linesman site and noticed the following note, which looks rather like a caveat:

 

Please note that the yardage grids detailed above are arbitrary and overlaid over a metric topographical map and are not continuous.  It is, therefore, not possible to produce a simple mathematical transform to consistently convert First World War map references into a modern lat/ long or WGS 84 references from the grids.

 

I have found using the side by side gimmick on NLS Maps works extremely well (at no cost) and the map coverage is greater than tmapper appears to be.

 

Max 

[Howard]

Early British maps were redrawn from the nasty French 1:80,000 series and were very inaccurate, the actual geometry of the map is distorted and no good fit can be achieved. The RE re-surveyed the British front and later (1915+) maps were much better. If a map is geo-referenced (fitted) to a modern map using calculated points you can expect errors of +/- 20 yards because that is the limit of lat/long positional accuracy of even the re-surveyed maps**. If maps are geo-referenced using a fit of features shown on both old and new maps the accuracy will be much better, many Linesman maps were laboriously fitted that way. Doing it yourself using Google Earth relies on such a method and local accuracy will be down to a few metres provided that one can select features that really have not moved, i.e. rebuilt churches are on the old foundations, road centres and junctions have not moved etc. Do not expect every wiggle of every road to fit, at the time they were skecthed in by eye.

 

If you look carefully at geo-referenced maps, they fit where they touch, some bits look remarkably accurate, others bits less so. Re-fitting the dodgy bit then pushes the good bit off. Fitting Great War maps to modern maps is an interesting but sometimes frustrating task, sometimes one must accept defeat. Not all is lost however, if you take bearings from fixed points to a point of interest the accuracy achievable can be surprising, that is because map makers used fixed points in setting out their maps so providing a church spire is where it was in the war, using it for a bearing is still useful and may give a better position than a fitted map with a lat/long value.

 

Howard

 

**According to Col. Jack in his Report on Survey on the Western Front

Edited 3 August , 2018 by Howard
Link added

[WhiteStarLine]

I'm wondering if Colin is referring to the overlays on tmapper, rather than the conversions.  The conversions of single and bulk points should be as accurate as any others given they use the Bonne forward and inverse calculations of Snyder similar to Muninn Project and presumably Linesman.  I used tmapper on a tablet to navigate to an underground brigade HQ near the River Somme and loaded the predictions into Google Maps / GPS to navigate to the remains of the entrance shaft.  Ditto for an attack on the Hindenburg Line, ironically placing the tablet down next to a rusted shell case.

 

For each calculated lat / lon, tmapper also has direct links to Muninn Project and the Scottish National Library web site so you can compare your level of comfort with the conversions with 2 other sites.  The latter has the best georeferenced maps known.

 

The overlays on tmapper are sub par and if Colin has purchased Linesman, then this will get him exactly where he wants to go.

[Howard]

Speaking to Guy Smith of Linesman years ago, he (at least initially) georeferenced the maps to GPS traces recorded whilst driving around the battlefields. This approach works well but is very tedious and requires judgement, which roads have moved etc. It is much the same technique as fitting a map to Google Earth using their software. Maps fit where they touch.

 

Years ago I put in a lot of hours to use Synder’s ellipsoidal equations to fit the lat/long values given on many 1:40,000 maps to modern maps and had some success but there is the technical problem that the Figure of the Earth values used at the time are not explicitly stated in any the sources I found, this gives rise to a doubt in the calculations. To get around that I wrote a spreadsheet with the Snyder equations and tabulated the calculated WGS84** results against observed ones then changed the Figure of the Earth values until a best fit was achieved, I had hoped to determine which (standard) Figure values were used. Sadly that did not work so I still cannot state with any confidence which Figure of the Earth was used at the time. Had it worked there would still be the +/- 20 yard problem.

 

The kinds of positional error also needs explanation, a surveyor will fix certain points like church spires, road junctions etc. then sketch in the wiggly bits by various means, e.g. a compass or theodolite traverse or even sometimes by eye. They also used aerial photographs but these have their own peculiar problems especially if the land has any slope, i.e. the problem of orthorectification. The upshot is that whilst points may georeference nicely, the wiggly bits often will not. There is also the problem that roads re-laid after the war have moved, that is especially true of tracks. When roads were widened to suit modern traffic, were they widened on one side (so the road centre moves) and have the junctions been re-aligned to suit modern safety standards, many have been. All this means that some judgement is required whether an old map “fits” an new one. Then of course trenches were shelled, re-dug and reinforced so their absolute position changed, sometimes by quite a lot. This Google Earth file shows the problem, it was geo-referenced by eye and fits in some places very well, in others there are errors of +/- a whole road width or more. No amount of fiddling can make all points fit at once. The map makers at the time had the same trouble, fitting an aerial photo to a map, tracing the trenches then drawing them on an updated map. The GE map fragment shows the modern crop marks of trenches match fairtly well with the maps but has ploughing moved the crop marks?

 

To get the best fit one still has to fiddle the results. Some maps simply will not georeference, the early maps of Loos being one case, I gave up with those, so did Guy. Later maps that came from the re-survey are far better and overlaying part of a whole sheet on a modern map works quite well. Fitting a whole sheet often shows errors.

 

To be really certain that a point on a modern map is where you think it is, it is best to start with a georeferenced map then to draw lines between fixed points and known bearings to intersect with one’s point. It takes time but is quite interesting and when one visits has that added certainly of success. Relying only on a GPS (that has its own positional errors) is sadly not a means to fix a point with certainty.

 

Howard

 

**A point on the ground does not have a single lat/long value, it depends on which datum is used. There are many to choose from. A GPS out of the box uses WGS84 but that is not the same datum as used in the Great War. To convert from the old one to WGS84 requires Snyder’s equations and knowledge of the size and shape of the earth (the Figure of the Earth), different cartographers used different values and they did not always say which. It all gets very confusing! I am still trying to find which ones the French used for their Lambert maps.

[colin.green14]

Hi,

 

Due to a family emergency have not been here since I posted.  Thanks for all the replies, will read all soon and digest.  Linesman is the way to go, just curious as to differences.

 

Regards

 

Colin 

[Howard]

One of the myriad of error sources that the amazing people in the Great War had to cope with is updating trench positions when the Germans would unsportingly shoot at surveyors on the ground. Aerial photos were used. They were taken in light aircraft bouncing around in the turbulence, printed and the grid drawn on them. Then techniques such as differential dividers were used to plot trench lines back to the lithographic drawing then used to print the new overlay. As you can see from this fragment, such photos are not always vertical so the grid is drawn as far as possible to match the geometry on the ground, here the grids are not parallel. As new trench lines are marked on the map each determination of position will have a slight change in scale, the skill of the draughtsman will be key. Then it gets printed on the base map, hopefully with each print run for separate colours exactly in registration. If you look at map corners you will see they are often not in good registration. The upshot of all this is not to put too much faith in “exact” positions, especially trenches between map versions.

 

In view of the all the troubles and the technical nature of the task I remain in absolute awe of the people who made the maps in the Great War. Their achievement was outstanding.

 

Howard

[MaxD]

Howard

 

Just to say that your contributions to this thread have been interesting and informative in the extreme.  As an erstwhile RA surveyor they have given much more depth to the scant knowledge I had already of the work of our Sapper forebears.

 

Thank you.

 

Max

[Howard]

7 hours ago, MaxD said:

Howard

 

Just to say that your contributions to this thread have been interesting and informative in the extreme.  As an erstwhile RA surveyor they have given much more depth to the scant knowledge I had already of the work of our Sapper forebears.

 

Thank you.

 

Max

Max, thank you for that. The Great War is such a huge subject that I had to take an interest in one limited area or my brain would pop, apart from family history of course. Some look at kit, others at tactics, medical practice, railways etc., I decided to look at maps and mapping and have gone beyond just 1914-1918, even so far as finding disappointing errors in the famous 6” OS maps once thought to be the epitome of  “accuracy”. Reading J B Hartley’s work on the OS I find that the idea of accuracy in any map is a somewhat slippery concept, an idea I had not entertained before starting out.

 

All very interesting and useful, even down to finding just how amazingly hard it is to draw a map by hand when comfortably at home with good light and a fresh cup of tea. I then imagine drawing such a map with men’s lives at risk if I muck it up, drawing by candle light in a flea infested dugout with rats running around and shells raining down. When I have trouble fitting a map to my GPS, I think of that, part of my own remembrance.

 

Howard

Edited 9 August , 2018 by Howard

[Smithmaps]

Gents,

To clear something up which is raised on this thread.

 

I created LinesMan, so I know what I am talking about.

 

The Grid overlay on British Trench maps is a 1000 yard square overlay, printed square from corner to corner and has nothing whatsoever to to with geographical referencing. No part of this grid was used in the creation of LinesMan.

 

The grid overlayed by the British Army, was simply there to help the users of the map share locations. 

The 1000 yard squares do not fit exactly, and various maps had to have the grid cut off, in order to make it fit.

 

As you may know, a map is a flat representation of a curved surface, and there are numerous mathematical ways that this can be represented.

A square grid is noting of the sort, and it is in my view, impossible to do a transform between the grid and a Lat Long, that will ever be accurate. Only ball park.

 

LinesMan is the only system, which takes into account the original Belgian Bonne map projection with some heavyweight mathematics, and though the maps are presented square, without adjustment in the tool, the lat long grid behind the maps is skewed appropriately to give accuracy of position over the entire map.

 

I have personally proved that accuracy by driving the entire Western Front.

 

The only inaccuracies to be found, are due to poorly drawn early maps, or indeed the state of the ground post war requiring roads to be put back in slightly different places, and indeed widened to double instead of single carriageway.

 

Google Earth uses a totally different mathematical projection, and the maps will never fit if overlayed the way they are displayed, except in very small areas. 

It is like laying an egg shape over a circle, clearly they are not the same. 

Try it yourself in LinesMan.

Put up a modern map on one side of the screen, and a trench map on the other. Draw a line round the trench map border using overlay, and see what shape displays on the modern side. It will be a distorted rectangle. This is done automatically within the tool, and maintains the same reference for two points between old and new maps.

 

I have total confidence that LinesMan will give you the best and most accurate Lat Long conversion that is currently possible, but it is only as good as the map that was drawn by a Tommy a hundred years ago. We have nothing more to go on.

Guy

[Smithmaps]

PS.

The whole point of LinesMan, was to convert Trench Maps, drawn in a Belgian Bonne Projection, into a GPS ready WGS84 coordinate system, which will work with a modern GPS.

This was done mathematically from source, and the same conversion was used for every map across the whole Western front.

Initially a very small constant inaccuracy was identified, and this was tuned out at source for the final product.

 

This conversion was then tested in the field by returning GPS tracks across huge areas of the battlefield from end to end. This proved that on average, the map registration is correct. Slight variations occur due to post war differences, or indeed badly drawn maps.

 

At no point have GPS traces, or coordinate points been collected or used to tune individual maps.

In this way, the fact that the early maps turn out to be incorrect, is in my view, an important historical fact, as it contributed to inaccuracy of artillery at that point in the war.

 

The variation between maps of differing dates can be directly compared, and the effect on accuracy of artillery bombardment even measured.

It is very interesting to see directly, how the mapping department of the RE improved over time, and the 1918 maps are simply stunning, given the tools that they had to work with.

Guy