Discussion: Mapping multiple levels on sprint maps

The following link is quite interesting. It shows the work of eight mappers on the same piece of sprint terrain. What I find intriguing is the variety of means used to represent or not represent what looks like an area of vertically overlapping terrain on the north end of what most mappers show as the largest building.

http://orientering.forbundetonline.no/nyheter/Side...

The bush map looks the most correct.

Have you been there tRicky?

Seen one bit of bush, seen them all.

Wow, this is really interesting!

The Vancouver Sprint Camper in me found the top one (SJ) to be the one that I understood the most. Fortunately, Bing has 45 degree aerial images from all four directions, so you can really see all of the decisions that the mappers faced on this one. There's also some Google Street View.

I'll also note that looking at Google's aerial imagery, the largest building looks like there's a new addition exactly where all of the non-SJ-mappers say where there's a big building. So it could be that SJ made his/her map before that building was built, and everyone else made their maps afterwards.

I've got a multi-level ISSOM map I need to make this year, and I'll definitely be studying this. Thanks for sharing!

Doesn't this sort of call into question the entire notion of trying to map multiple levels using orienteering symbology? I mean if capable mappers can't agree what to do, how is a runner supposed to figure this out at speed?

I like sprint orienteering, but stuff like this is not fun and not fair.

I sort of agree with you Clem. I faced a situation a little like this and went for the ODT solution... making the less useable level of the overlap into an impassable feature.
And what have I missed in not finding the Bing location?

Whoops, sorry Neil. I forgot to link it.

http://binged.it/NysMpa

Use the curvy arrows around the compass to rotate 90 degrees.

If you select "aerial" under the Bird's Eye tab, you'll get the aerial photos, same as Google.

Amazing the differences - even that little square block between the NW and N buildings has been mapped almost completely different from each mapper. On some it's shown as a low wall, others individual trees and one is just OOB.

Thanks for that link. It changes my opinion as to my preferred version.

So which do you think? I think SJ mapped it whilst the building existed but has mapped the ability to be able to run across the top of it.

I'm with you Clem... I'm sitting here at my computer and even after several minutes, I'm still not sure how to decipher those maps. (Of course, the runner has advantage of being able to see terrain too... Sock's image helped)

My reading of ISSOM is that course setter and mapper need to collaborate on multi-level and select only one primary running level to focus on. This isn't so easy to do for a general-purpose, multi-use map.

Something about ISSOM just drives you to overmap. I'm looking at all the subtle jogs and bulges and overhangs on those maps... and by the end I'm thinking that as a runner, I wouldn't pay half a second's attention to any of that, and the brutalist final map is looking pretty attractive.

But the brutalist final map doesn't make much sense of the overpass IMHO.

Man, I want to say that SJ mapped it before the new building was built, because what he's mapped and the Bing bird's eye view makes perfect sense to me.

But at looking at the Google Earth imagery dates, that new part of the building was there in April 2012, and this mapping exercise was done in 2013.

Perhaps SJ doesn't live near the map, and did a lot of it from just the online images?

Also, some of the interpretations I think are wrong. Look at the tunnel that comes in from the north edge of the map. On ODT's version, the tunnel opens up on the south end without the thick black line. Reading that, I assume that I can just magically run from one level to the next without any obstacle there.

Perhaps SJ doesn't live near the map, and did a lot of it from just the online images?

...or that my comment above holds true in that he mapped the upper level of the building as passable :-) Drawing a sprint map with no fieldwork is a crime punishable by death IMO.

he mapped the upper level of the building as passable

Possible, I suppose. But what are the chances that they put a brand new building on top of the top level of parking garage, and they leave existing objects from the top level there (the three X objects and the mini building).

Another possibility is that the SJ map, since it's listed first, is the original map from before the new building and the mapping exercise.

The bing image I looked at appears to match the map with the features you have mentioned. To me the other mappers just said 'that's a building' (despite it really being a carpark) and mapped it as impassable. Another one of those perspective things I guess, that and there's nowhere to go once you get up there may have made it an easy choice for the other seven.

How's work tRicky?

what's work Tooms?

I'd seen the article last week. Without the bing link, it made no sense. Now I can see the only person who mapped it right was the SJ map.

The car park is runnable - not an impassable building.

The final map is wrong - it ignores useful detail like canopies/overhangs on the edges of the building.

The detail is all part of the point of ISSOM.

It's not up to the mapper to know what will be passable or not on competition day sometime in the future. If the carpark OOB, then all that can be seen from the running level is a building. SJ can be quickly converted if the carpark isn't accessible, (and if I were planning I'd like to use the carpark). All the others would require the planner to remap, so boo to them.

I'm interested to note that in 20 comments nobody mentioned the contour interval, which I guess is wrong on all but ID (unless ID has 1.25m). I don't think people pay attention to climb in sprint routechoice.

While the runner will probably be able to make sense of things when he/she gets to the area, the problem is with route planning. If you don't know what is up/down, passable/not, when you are planning your route, you are either losing time hesitating, or worse, going to be lucky or not. Of course, at major sprints now there seems to be a tradition of doing hardcore due diligence on the terrain beforehand (even making your own maps). I suppose the expecation is that this is part of the game for serious competitors. But, the essence of orienteering (IMO) is that a competitor shows up to unknown terrain to run a competition, and gets a map, which is in the lingua franca of our sport. If you have to do other things to fairly negotiate a course, or win the competition, that's contrary to the spirit.

@j-man hardcore due diligence on the terrain beforehand
And the corresponding WOC-planning due diligence of adding new impassible bits on the morning of the event.
Me, I went off on a bearing at Anza Borrego last week without knowing what to expect - that didn't work out too well either ;)

I'm interested to note that in 20 comments nobody mentioned the contour interval,

I did notice that all of the contours were different from each other, which surprised me a little bit. I'd be curious as to why that is, and how each of the mappers got their contour information.

I also noticed that some of the mappers clipped the contour lines to show the flatness of the parking lot on the east side of the largest building (with the descending stairs). Other mappers left the contour lines as is, which doesn't accurately depict the flatness.

I'd also be curious to know how much time each of the mappers spent on this exercise, too.

I think tRicky is explaining it exactly right Mr Pinksocks - the building was there when he mapped it and he has mapped the roof level as runnable.

This really is a nasty problem and it is a clear example of why the ISSOM states: "The cartographic representation of more than one level is in general impossible.
Hence only the main ‘running’ level should be represented on the map".

So in this case SJ must have felt that level was the main funning level and the course planner will have to be very careful not to run legs close to the carpark on the lower level.

But regardless, jman is not entirely correct in saying the problem is mostly with RouteChoice - ask Louise ;-) There is a comparable situation somewhere in Canada (I forget where) and if you are running and find yourself at the side of the carpark, with SJ's map you will be in danger of having no idea at all where you are - you are looking for an impassable wall will a car park on the other side, but you are looking at a building. Maybe experience will help you. But I would not use this for a "serious" sprint competition.

How's work tRicky?

It's near the end of the month. Obviously I have little to do until next week :-) (except planning how to jump over a wall to access a car park)

So in this case SJ must have felt that level was the main funning level and the course planner will have to be very careful not to run legs close to the carpark on the lower level.

For the most part, the upper carpark IS the main running level - it's the only running level, since downstairs it's a solid building. I find it interesting that SJ used the underpass symbol to represent the lower canopy, rather than either leave it off or fudge the edge of the impassable wall to allow the canopy on the outside.

From the Bing link, it looks like that carpark is double-level, meaning under the lower canopy it's triple-level. One hopes there's no underpasses going east-west, because...yikes.

IMO SJ's treatment of the overpass on the northern side is the most correct, although I would be tempted to use ID's technique of thicker (but not impassable) lines to indicate the overpass. All the others compromise one level or the other for no benefit - HG, ODT blocked off the upper level, JVL managed to compromise both levels and I have no idea what LOK was trying to achieve.

I think tRicky is explaining it exactly right Mr Pinksocks - the building was there when he mapped it and he has mapped the roof level as runnable.

AZ, have you looked at any of the images? SJ's map shows exactly what you see in Bing 45 degree view, which is a multi-level car park with the runnable roof as the primary running layer.

But if you look at aerial images from 2012 and onward, there appears to be a completely new layer on top of that car park, which suggests one large building (more like what the other 7 mappers show).

If SJ made the map with that new building in place, then it completely falls apart. If it's still runnable on that layer, the new building should be shown as a really large canopy (light gray), instead of open-air pavement.

It seems more reasonable to me that SJ made the map before that new building was added atop the car park, and everyone else made the map after (since the exercise was done in 2013, after the building went up).

I only viewed the link given above - I have no idea where this place is to try and find it. I didn't realise there was more recent stuff.

Neil, here's how you get to the more recent stuff:

If you select "aerial" under the Bird's Eye tab, you'll get the aerial photos, same as Google.

It's in Skien, Norway, per the article. I had to surf around a bit to find the exact spot. Good thing it's not a big town.

Got it, now I see what you're on about :-)

The article looked to be in a foreign language so I didn't read it.

I am surprised nobody said anything about the N–S tunnel that is just a hair east of the center of the map (under the passenger drop-off zone). Six of the mappers pretend it's not there. ID shows it as correctly as possible. It's curious that SJ apparently attempts to show it too (if not, why the contour detail in the middle of what on top is a flat parking lot?), but either omits the tags on the 512.1 Bridge symbol, which makes it indistinguishable from 521.1 Impassable wall, or changes his mind and decides that it's out of bounds and forgets to move the contours to some place sensible.

None of the maps show the grass island just E of the passenger drop-off zone correctly, or maybe it's new.

There's so much on this map what we can play the "I'm suprised that nobody..." game all day!

Vlad, I don't think there's a tunnel there, at least per this angle from Google Street View. (I hope that link works for everyone, it should on Chrome, at least).

So if we accept that ISSOM cannot represent dual levels of terrain (other than simple underpass situations), then how does one deal with terrain like this? Not map it? Map it so that second levels are ignored, or ensure courses avoid giving runners the need to interpret these situations? I think options 2 and 3.

How does one deal with terrain like this?

The mapper does the best he can. Then the course planner takes over. Just as with forest maps, if the CP is thinking to himself - "Gosh this bit is very complex and open to different interpretations" then maybe it's best to avoid it. It is always wise for a CP to ask himself - how would I feel experiencing this as an orienteer? Does it look like I expected it to based on the map? Would I feel 'screwed around' if I had to find a control near here? etc. Then it is a personal choice. Personally I am fairly conservative. I would not want to be in the position of having competitors frustrated with me even if I could prove that 'technically' I was correct.

One interesting possibility for multi level areas is to use the "hide" function in ocad or other mapping programs to include multiple levels of runability in a single map, with specified directions for which symbols to hide or unhide when a particular level is used.

You could even have a map exchange in the same event, with an early leg running across the parking lot roof (mapped as the primary level) then a later leg going through the tunnels on a different version of the map with the tunnels shown as the primary level. Now that sounds like fun!

Once we evolve to running with electronic maps, you could toggle which level you want to look at on your e-map during the race for more clarity.

Or, you could just solid-model the entire terrain and get whatever view you wanted.

Welcome to orienteering! Please have your indemnity form and notarised GIS degree handy at registration!

:p

Whatever anyone draws, it will be wrong.

Upnorth is on the right track. ISSOM talks about the need for cooperation between setter and mapper - they need to work together to make the map fit the courses.

You also need to ensure that runners cannot inadvertently arrive at a level that is not mapped - that means blocking off (tapes/barricades) access to those levels.

If in doubt about whether the map can fairly depict a particular area, avoid it or make it out of bounds.

Most of the examples I see of confusion about levels is due to trying to add a degree of trickiness to the course - which is unnecessary.

30 years from now when we're still using paper maps...

Welcome to orienteering! Please have your indemnity form and notarised GIS geographic history degree handy at registration!

I taped off an area at the last sprint event I set and people still ran through it.

sad - but at least then they have clearly broken the rules and can be disqualified (or told it's their own fault the map didn't make sense in the OOB area) and you can rest easy in the knowledge you have done your best.

Apparently olive green on the map doesn't mean much to some people if there's a nice garden bed right there to run through, even with an impassable fence behind it ;-) (both obstacles proved to be non-deterrents)

I once set a course that went through a pedestrian tunnel under a freeway. Put a control right at the entrance to the tunnel. Several people opted to go up the embankment, over a cyclone fence, across the freeway, over another cyclone fence and down the embankment on the other side.

Got it, now I see what you're on about :-)

Me too.

But I find it hard to believe these guys mapped it at a different time. For one thing what would be the point? And for another then the article would be extremely misleading.

I think that probably after realizing how terrible this was for Sprint orienteering they fixed it by putting up a building to make the map consistent, no matter which way you came at it.

But seriously, assuming all the mapping was done at the same time (before the new building), if you map from the NE then you do like SJ (emphasize the runability of the roof), and if you map from the other directions you do as the others did (emphasize the building-ness) - no?

Hm, except that the mapping meeting was in 2013 and you say the imagery showing the new building was from 2012... so now I'm starting to understand PinkSocks' thinking (which is actually a relief)

Did all eight mappers get paid for their work?

@mikeminium - changing the map to fit the course is exactly what ISSOM recommends. If there's a different "main running level", you need a different map - I've seen it done.

All these maps were surveyed at exactly the same time, during the Norwegian Mapping Conference which ran alongside the Norwegian Sprint champs in Skien.

I took part and spent the same time as the rest of the mappers looking at this area, but I did not deliver my (partial) results to be included in the linked article.

IMHO the HG entry gives the best overall view of the area, the SJ one shows the rooftop parking in a way to takes away to much information about stuff at lower levels.

The mapping standards are such that it is effectively impossible to show more than 2 vertical levels, and in this case there is a major tunnel (very runnable!) that is partly shown as the entrance to an underground parking on the JVL and SJ maps.

This tunnel leads straight north, going underneath the small underpass shown near the north end!

Terje, how much time were the mappers given? I just wonder, because there are some obvious mistakes (JVL's northern building, contours across the parking lot), and other features where some people had wildly different interpretations (impassible wall vs passible wall vs no wall at all), and I'm wondering if they would have converged to a more similar treatment with more field checking time.

They may have been trainees too?

(and who is SJ, and what is his rationale for how he mapped the parking lot exactly like an old Bing image ;-)

Probably explains why Norway employed a Czech mapper for the 2010 WOC sprints!

Ya I've never run on a multilayered map before. I really wanna try it!

We only had about an hour, most of us spent that time discussing how to map the area around that multi-level parking entrance, so I am not too surprised that several of those who submitted their efforts must have spent some time with Bing/Google images in order to fill out the rest.

With a bit more time to calibrate the mapping we would obviously have ended up with much more similar interpretations, but afaik only HG (who had previously mapped the neighboring championship map) have spent much time on sprint maps, the rest of us tend to work on forest areas.

Re the wall treatment: This is the usual Sprint map conundrum: Do you map low walls/steps and if you do, should they be passable or impassable?

This is pretty much up to the mapper, and sometimes the course planner want to have lots of impassable walls in order to make route choice more complicated.

Pretty much what I expected - the eight mappers only had an hour to map so what we are seeing is a first draft. Most of the maps would be ok to run on apart from the last which I think is too generalised. A really accurate sprint map only comes from lots of different versions and several different people (mapper(s), controller, IOF controller) all going out and making/asking for incremental refinements and discussing contentious areas and looking at control sites - something only done for large events. Split levels are a pain - some levels on the Oceania/World cup 2013 maps went through 15-20 different versions before everyone was mostly happy. We even roped off / made out of bounds one area where we didn't want runners ending up on an upper level accidentally.

Nice. Remember reading this (was it Veivalg?). If I got to choose which one to run on, it would be ODT's. Definitely my favourite mapping style :)

Or, you could just solid-model the entire terrain and get whatever view you wanted

WOC 2050 Indoors even:
3D holographic projection coming from a digital "paper"