This blog post is currently a work in progress, updated as I learn more about the subject.
A number of people have asked, and I've been planning for a long time to create a description/tutorial on how I've been making topographic maps using freely available data, and open source software.
A lot of credit to the processes outlined goes to various information sources online, too many for me to remember or list here, but I'll try to link to sources where possible.
Why Learn Map Theory?
Before creating your own maps, if you intend to make accurate maps that you wish to use for navigation, I think it's important to have at least a basic understanding of map theory. It's entirely possible to follow a step by step process on how to make a map outlined further in this document, but without an understanding of how maps actually work, you will struggle with:
- Fixing problems with your map.
- Bringing in different sources of map data.
- Taking what you learn here and applying it to your own specific use cases.
Many of the different technical attributes of a map (Projection, Datum, etc) are commonly referenced using identifiers as part of a database. If you see something like ESPG:3785 in your mapping journeys, this is a reference to item 3785 in the European Petroleum Survey Group's database. ESPG:3785 happens to be the Web Mercator map projection. Geographic data is almost always accompanied with these identifiers which enable software to interpret the data correctly.
It should also be mentioned that there are also some other databases such as OpenLayers, ESRI and OSGEO.
About Map Projections
As the early explorers and scientists discovered, the world is not flat, but is in fact of a spherical shape. This greatly increased the complexity of representing the known world on a flat map. To address this problem of how to represent a spherical (and bumpy/lumpy) surface on a flat, rectangular map, map projections were invented.
There are an infinite number of ways you could choose to map one shape to another, and therefore, there are a huge number of map projections huge number of map projections in existance already. Every projection has unique characteristics, they distort the Earth's surface in different ways to make t fit on a flat surface. Which projection you choose to use will depend heavily on how you plan to use the map. You can (and I recomend to) read more on this topic at Wikipedia: Map Projection.
There are a number of common (and important) map projections which you will often encounter during the process of making a map, so I've provided a brief description of each here.
The Mercator Projection is probably the most common, and well known projection. A Mercator projection of Earth is very recogniseable, and everyone will be familiar with it.
If you plot your course on a Mercator map, and you are following a constant compass bearing (excluding the effects of magnetic variation/deviation), this will turn out to be a straight line. This feature of the Mercator Projection makes it very useful for navigation maps intended to be used with a compass for navigating from point to point because you can draw a line between these points, and measure the bearing/angle on the map, this can followed on your compass to reach the same point in the real world. For comparison, if you attempted the same thing with say, a Gnomonic Projection, you would not end up at your intended destination!
Web Mercator is the map projection used by most online mapping applications such as Google Maps, OpenStreetMap and others. It is a variation on the standard Mercator Projection.
Universal Transverse Mercator
The Universal Transverse Mercator projection system is actually not a single projection, but a composite of many Mercator projections, one for each "zone", specifically optimised for that area in order to reduce distortion. UTM is now very common for hiking maps, and is the projection I have been using for the maps that I make.
Included in the UTM system is the concept of UTM coordinates. These are an alternative to using Latitude and Longitude to represent a position on the map. Instead, you specify the UTM zone, and then the distance (in meters) East, and the distance (in meters) North from the origin of the zone. These are also known as "eastings" and "northings".
An example (from Wikipedia) of this coordinate system could be:
17N 630084 4833438
The first word/number 17N in this coordinate is used to represent the UTM zone that the point is located in. Unfortunately there are currently two methods in use for specifying the zone.
The second word/number 630084 is the distance east from the zone origin in meters.
The third word/number 4833438 is the distance north from a reference point relating to the zone. When in the Northern Hemisphere, this is the distance north from the zone origin on the equator. When in the Southern hemisphere it is the distance north from roughly the south pole.
17N 630084 4833438
The 17 designates that the zone is UTM zone column 17. The N designates Northern hemisphere.
17T 630084 4833438
The 17 designates that the zone is UTM zone column 17. The T designates which latitude row in the column that the zone is in. In a sense the T is partially redundant because part of that information is also encoded in the northing number.
A more detailed/accurate description of the coordinate system is available on the wikipedia page.
About Earth Shape Models
In order to project a map of earth onto a flat surface, you need to have a notion about what exact the shape of the earth is in the form of a mathematical model. The most basic model for the shape of the earth is that of a sphere. However, Earth is an irregular shape, and not a perfect sphere. This means that projections using a sphere as the model will, in certain areas, be much less accurate and therefore less useful.
Luckily, most of the irregularities are small and localised, so a better (than a sphere) model of the Earth's shape can be built using an ellipsoid. These are also known as an Earth Ellipoid or Reference Ellipsoid. There are many many different refererence elipsoids which have been constructed and defined, but the most common one in use today is defined in the WGS 84 standard, but the ITRF datum is gaining popularity due to it being an open and improving standard.
A Datum or Coordinate Reference System (CRS) is a coordinate system. There are horizontal datums with models used for locating a point in terms of latitude and longitude, and there are vertical datums with models used for locating a point in terms of altitude. You could use a single model to do both, but there are benefits to using something like a simple ellipsoid for horizontal location, and a more complicated (and accurate) shape like a geoid for the vertical location. The models are also callibrated with respect to a particular reference frame or position.
A reason why you might want a different datum is that an ellipsoid can be callibrated/made more accurate for a certain area of coverage than a more general model like WGS 84. In Australia, for topographic maps, we now use the GDA94 datum, which in turn uses the GRS 80 reference ellipsoid. Sometimes the difference between the same latitude and longitude represented with different datums can be hundreds of meters, unacceptable to choose the wrong one if you're trying to find a campsite in the dense fog with your gps.
There's some nice presentations here and here outlining the differences between GDA94, ITRF and WGS84. The TLDR is that as time goes on and the tectonic plates shift, the differences between these coordinate systems increases. As far as I know, WGS84 is a static datum, meaning it has no model for how the surface of the earth moves as the tectonic plates slide around. GDA94 is also a static datum, but this works well for Australia, because mainland Australia is located on a single plate, and positions relative to each other in GDA94 on mainland Australia stay the same. Going to a country like New Zealand, a datum which is dynamic, and takes into account the velocity of the tectonic plates is required in order for the datum to maintain its accuracy for longer periods of time.
Coping with the changes due to tectonic plate shifting with respect to measurements taken using GPS is an interesting/advanced topic, there's a great document about it here at: Coping With Tectonic Motion. The basic problem is that if you measure a location (such as a mountain peak) accurately in 1994, and then return there 20 years later, you could obtain a result which is several meters different. Not huge problem for Topographic maps in the short term, but definitely something that surveyors need to worry about.
Which Shape Model to Use?
I have used GDA94 for my Australian maps, however, the next GDA, GDA2020 is on the horizon? Using GDA94 requires that any input data using a different CRS needs to be transformed in order to be able to plot it on the GDA94 map.
Before you begin making any map, you need a source for the data you're going to use in the map. By data, I mean, the source of information for all the features to be included in the map. The more detailed, and more accurate your source of information, the better your map can be (provided you spend the time figuring out how to present this information in a sensible manner). There are many freely available sources of geographical data online, but you can also use data you generate/provide yourself.
There are a number of different data types which can be used in maps, having an understanding of these data types is important so you can know what to look for when you set out to construct a map.
Raster data is stored in regular sized blocks, essentially like pixels in a digital photograph. For raster data to be useful, it also needs to be accompanied with some callibration information to allow GIS software to know where on the planet the data is located, and what the size and location of the "pixels" are.
GeoTIFF is a very common container/format for raster data which packages the pixels along with the map callibration data, to allow you to accurately position this data in a coordinate system.
There are a number of good sources of geography data which cover large portions of the planet. Some of the ones I have used are listed.