[RFC] Feature Proposal – Aerodrome Classification (moderate rewrite)

General talk Tagging general discussion
, ,
1

This is a proposal for a system of classifying different types of aerodromes based on their size and prominence. Additional tags can be added to specify usage.

Proposal:Aerodrome_Classification

Summary:

The “aerodrome” tag is used to describe the infrastructure level of the aerodrome.
The “usage” tag is used to describe the usage at the aerodrome in terms of commercial vs general.

If an aerodrome is international, if it is private or if airsports are practiced there then the “access”, “port_of_entry” and “sport” tags can be used.

Please discuss this proposal on its wiki talk page or in this thread.

Edit: The proposal has been moderately rewritten following the advice given.

2

I have answered comments and modified the proposal.

I cut some of the aerodrome criteria for being unreliable and moved one to another section.

3

Repeating myself from the last time that this was mentioned in the forum, but I believe that it is straightforward as a data consumer to do that with existing tags and existing tag usage by mappers in OSM (which doesn’t always match what wiki authors would like).

As an example, see here. There are basically two parts to it:

  1. Detect a bunch of edge cases, mainly around disused.
  2. Use iata or icao to indicate a “large” (e.g. larger than just “General Aviation”) aerodrome or heliport.
  3. Use various tags to detect military aerodromes.

There are alas a whole bunch of edge cases - some very small airports are “international” because they are in very small countries. Some share runways with nearby military bases. Some are basically cargo hubs with a few passenger flights.

Moving on to the proposal itself, the lack of any mention of military or disused in the proposal sounds like a huge omission.

Also, a tag like aerodrome=intercontinental / continental / regional / local / basic sounds hugely subjective. As an example Jandakot in Perth, Australia is currently aerodrome=continental. I’m sure that some flights from there do go a long way, but a glance at the website confirms that it is very different from a regular passenger airport** - there’s no list of scheduled destinations, for example - just a list of airport tenants.

** it’s the only place I’ve been where the pre-flight instructions included “hold the back door open until we have taken off so that the flies are blown out of the plane” :slight_smile:

4 Likes
4

IATA and ICAO tags are nowhere near enough for classifying aerodromes.
IATA gives a code to any aerodrome which receives commercial traffic. There are dirt strips in Papua New Guinea with IATA codes.
Similarly, ICAO tags are not enough. It’s true that they are given to the most relevant aerodromes however in countries with very large territory like the US some local aiports do not get the chance to receive an ICAO code.

Road class tags are also highly subjective. It’s just the nature of tagging human creations.
Each tag has multiple criteria that explain the nature of each category. The names were chosen for the prevalent influence each aerodrome size exudes.

In the case of Jandakot specifically, it had 210K movements in the 2024 FY (regional), seems to have no terminal (local), has a large supporting infrastructure (continental), seems to have no connections (local), seems to be used by light aircraft (regional or local).
Jandakot would be classified as regional.

5

If I lived in the forests in PNG I’d probably think that they were quite important :slight_smile:

I’d agree with that, although it was tagged as continental when I looked in OSM. I think that most of the traffic is workers to mines etc.in Eastern WA and beyond. There is a lot of traffic, but you can’t just rock up and jump on a plane like you can in most of the places that people call airports.

I wonder if something verifiable like “has scheduled flights” would work better than a loose classification?

3 Likes
6

I do not deny that at all. However the airstrips in PNG and NY JFK most definitely should not be in the same category.

Except you might. A lot of these general aviation aerodromes provide joy flights for the public and if you are a private pilot many rent out airplanes for you to fly. They can also take you skydiving and the like.
I searched on Google and found 3 companies in Jandakot providing these services.

This is the difference between commercial aviation focused and general aviation focused aerodromes outlined in the proposal. In one you travel, in the other you enjoy leisure.

The presence of scheduled flights, like an IATA code, is not enough. There are many local aerodromes which are predominantly general focused which do have small scale scheduled flights. This would skew the aerodrome classes.

The reality is that aerodromes present a lot of variability and are not easily categorizable following only one variable.

7

And the classification for you, depends on the weight you give to the various variables.
So OSM should contain the variables, letting you make your own classification.
Typically a general map should rather focus on the variables mentioned by @SomeoneElse. For a general aviation map, the criteria would be different.
Is highway=motorway an important road? On a cycling map, it’s not.

3 Likes
8

Do you think I’m making up new concepts of existence for use solely in OSM? Refer to the “variables” by name and share your specific grievances with me here.

My proposal calls for size of terminal, size of associated industrial area, type of aircraft seen operating at the aerodrome and passenger and aircraft movement statistics. These are all very real and easily verifiable.

Let’s give the variables a name. They are the presence of IATA and ICAO codes. I’ll say this again: They are not enough to be useful. IATA gives out codes like they’re candy. If an aerodrome has any sort of commercial passenger service, no matter how small or infrequent, it gets an IATA code so bags can be routed properly.

How are you going to solve practical problems for an average user looking for airports on a map? A regional aerodrome and an international airport are next to each other and both have IATA codes so they can’t be distinguished. How do you know which aerodrome let’s you take a flight out of the country? What’s the point in even trying to distinguish them using the codes when they all just end up being lumped together anyway?

9

And? What does prevent from adding those data in OSM (as we do have population for places, we could have traffic_range=more_than_10M,between_1_and_10M, less_than_1M and keep the algorithm outside from OSM?

1 Like
10

You are free to add those as you like but I find it unnecessary when the “algorithm” amounts to a table look-up at worst. For most cases, it’s just associating that big buildings => big aerodrome.

11

Which is not part of your proposal!

12

What do you mean? It’s in the Tagging section, Main Tag subsection. I even mentioned the criteria here.

13

To be clear - those are the variables I’m currently using in my maps, and they work pretty well in UK/IE, but I’m sure more would be needed elsewhere. I also mentioned “The presence of scheduled flights”, but these were all just examples of “things you can actually go and survey, and later verify”.

The latest version of the proposal is better (it now mentions cargo). Still no mention of military or disused though.

Perhaps pick a few examples worldwide and say how you would decide the main tag when the suggested criteria actually conflict? An example might be East Midlands, which is a huge cargo hub for flights from all around the world, but handles mainly European “bucket and spade” passenger flights.

14

I believe you. Believe me when I say that your system was the first one I tried. You just run into too many exceptions.

Here’s a couple I’m aware of:

Lisbon Airport and Cascais Aerodrome in Portugal both have IATA codes but one is a major hub and the other is a local aerodrome with a subsidised local airline.

There are 6 civillian aerodromes around Paris with IATA codes. Only 2 of them are commercial airports, the rest are general aviation.

Then we have the case of the USA. The American general aviation market is extremely vibrant. There are hundreds of regional aerodromes chock-full of activity, most of them with IATA codes. And they can be located anywhere, in the middle of nowhere, near a city, near other aerodromes. In this situation it becomes very difficult to distinguish between commercial airports and regional aerodromes using IATA codes alone.

Thank you.

I saw military and disused as outside the scope of what I was trying to accomplish. If there is demand I can approach it but I’d rather it be with a separate proposal.

Most aerodromes do not fit a category perfectly. You need to make an averaged decision of which category to choose, but this is par for the course.

East Midlands would normally be a continental class aerodrome ignoring cargo but we can factor it in.

First, we can see large cargo terminals and industrial zone in the satellite imagery.

Second, we can take the total cargo transported annually and calculate a very rough passenger equivalent. East Midlands seems to have transported 375000000kg in 2024. If we divide by 85kg, the standard weight for an adult male for use in weight and balance for aircraft, we get a kind of human equivalent of 4.4 million. Added to real passengers (4.1M), you get 8.5M.

All added up, we have 8.5M equivalent passengers in 2024 (continental), multiple large terminals (intercontinental), large supporting infrastructure (continental) and use by large and medium jets (intercontinental).

East Midlands seems to be right at the edge between an intercontinental and a continental aerodrome, but it leans smaller. I’d classify it as continental.

15

Except if your intention is to make a half proposal, you should include military (disused could be managed with a life cycle prefix). Some airports may be dual use (civil and military) sharing the runways etc. but not the buildings of course. This one was military and CA, now military and GA. I would consider it a military with something like general_aviation=yes as the airport wouldn’t be maintained just for GA.

1 Like
16

How do you handle places with multiple “types” of use?

Our local airport (which is the 5th busiest in Australia) has:

Intercontinental traffic from Asia,

International (or whatever you want to call it?) traffic from New Zealand & various Pacific Island Nations,

Domestic traffic from all over Australia,

Regional General traffic, &

Recreational traffic!

What type of airport is it?

17

it is, these are major landmarks and also barriers, hence they should get more prominent display compared to other roads (and usually do).

18

Fair enough, I’ll do it.

1 Like
19

Unfortunately the information you gave me wasn’t the one I needed.

I apologize for the investigation but in your profile you say you’re from Gold Coast. Could I assume your local aerodrome is Gold Coast Airport?

BITRE says it had 6M passengers and and Services Australia says it had 78K movements in 2024 (continental), it has 1 large terminal (continental), supporting infrastructure is sizable (regional), aircraft type in use seems to be mixed but I see medium jets in use (continental).

Gold Coast Airport would be a continental aerodrome.

20

In my opinion this should be decided by mappers as all things are and not based on existing tags. I think there should be 3 airport types. International, Regional and single minimally used runways.