QGIS 3 Coordinate and projection systems
A short introduction to coordinate systems and map projections. A huge subject—but this is all you need.
Coordinate systems vs.map projections
When vector data is gathered, it will be located by a coordinate system. This tells DIS software ‘where’ the data is located.
Projections are how the data is shown to you. Ever seen the maps that represent land‐area correctly, even if they distort shape? Or maps of the world that look like split orange‐peel? Those are decisions about how to ‘project’ map data. And there are decisions beyond this. QGIS can project data in 3D, producing spectacular map‐images of, for example, a bicycle route moving across hillsides.
Why do we discuss this?
Sophisticated land‐map software cares about map coordinate systems and projections. And QGIS is sophisticated land‐map software. So, not only is this a big subject, but it can cause hassle from the start.
An introduction to coordinate systems (CRS)
When vector data is gathered, it will be gathered in accord with a coordinate system. However, deciding ‘where’ a location is is difficult. The ‘earth’ is not stable. Continents drift, magnetic fields change, the earth’s surface flexes. These features of the earth add offsets year by year. Also, a coordinate system may be optimised for accuracy in one particular area, or be devised to work with GPS systems. And coordinate systems may vary depending who or which organisation gathered the data.
For different purposes, and to match between maps, transforms are supplied. While these solve the problem of matching maps from different systems, they add a layer of complexity.
Coordinate system accuracy
You may not be concerned about great accuracy. For example, continental drift is only a few centimetres per year. Earth movement is usually only 1m a year. If you are only looking to make a map of your local town, that’s irrelevant. In one hundred years the map may be less than a metre wrong. Another example, coordinate systems vary in precision. GPS has been increasing in precision, but until recently could only target within maybe ten meters. If a map was made from GPS data, you can never make it more accurate with another tranform.
So it’s worth knowing that pursuing transformation precision may not make an improvement to map‐data. Not if the initial data is imprecise or inaccurate. However, there is one case where this may become relevant, which is if you want to join maps with different coordinate systems. For example, an English map to a French map. If an appropriate transformation is not used for both maps, the two map datasets will not splice—there will be broken lines and maybe joins with gaps.
How much do coordinate systems matter to this guide?
Since the guide only covers data from the Ordnance Survey, from one source, the data will be to the grid reference system ‘British National Grid 1936’. And, since I only target a town or two, great accuracy is not important. QGIS will choose an appropriate transform to get you moving—OGIS includes many common transforms. Unfortunately, my packaged, not‐the‐latest QGIS did not have the latest transform. The increase in accuracy is from 2m to 1m which, I’d say for these purposes, is irrelevant. But QGIS will nag about this, which irritates. To get rid of the nags, I’ve included information about how to upgrade.
QGIS handling of Coordinate Rendering Systems (‘CRS’)
We do need to know about this. If your QGIS load is easy e.g. from packaging, get QGIS running, then come back here.
In a QGIS, look at ‘Project > Properties > CRS’. Or‘Settings > Options > CRS’. Or there’s a shortcut button in the QGIS status bar. Any way, you’ll find transforms available. And you will find how customisable QGIS is. It can set the CRS based on the first import. CRS can be overridden per layer (‘Right Click the layer > Set CRS > Set layer CRS’). If no CRS is stated, QGIS will take the CRS from the project setting, ‘Project > Properties > CRS’.
In short, the QGIS coordinate reference system is massively customisable by layer, project, and by allocation from assumptions made on import. Also, QGIS now has many common transforms built‐in, and extensive facilities for reporting on, and importing, new transforms.
A (non‐)introduction to projections (PCS)
You have a dataset, in our case a vector dataset from the Ordnance Survey with a well‐defined coordinate system devised for many purposes. But that is an abstract representation of a curved surface. How do you wish to present this, especially on a flat surface?
How much do projections matter to this guide?
I’m not covering them, simple as that. OGIS will choose something adequate. Be aware that, if you’re interested in a 3D bicycle route‐map, this subject is a large extra area of research.
Refs
QGIS info on projections,
End of coordinate and projection systems
Or, enough of them—because there is years of practice. Let’s finally get to install and data downloading.
Refs
Want to know what the acronyms mean? Or want to plunge into the intricacy of map point plotting? A British Ordnance Survey article from 2016 on the state of ETRS89,
https://www.ordnancesurvey.co.uk/documents/gps/updated-transformations-uk-ireland-geoid-model.pdf
British Ordnance Survey official guide on coordinate systems. Length is a term of university work,
https://www.ordnancesurvey.co.uk/documents/resources/guide-coordinate-systems-great-britain.pdf