This technical log is part of the most important elements in this GIS project. It goes beyond being able to add rasterised paper maps, and is about ensuring that the correct projection on a digital map is performed, by using the correct Coordinate Reference System (CRS). Depending on the CRS choice, and actually selecting the correct CRS for the digital map data being displayed is critical to successful dimensioning tasks when creating vector points, polylines, or polygons in a GIS project files vector layers.
Given distance and point locations is part of this thesis project, this set of technical logs is about exploring CRS, and their relationships, selections, and potential problems.
Maps are broken down into different projections (as noted in my gis.tex and gis.pdf technical logs). Earth is a sphere, it is a complex three dimensional shape. The different projections of earth are related to the fact that maps display a two dimensional, flat representation of a sphere, or selected portions of a sphere. As such, any two dimensional map will distort our visual representation because we are flattening out a convex surface.
GIS systems can overcome this flattening, providing a true representation on screen using wonderfully clever (well, I think its clever!) mathematical conversions, but the imperative here is to understand from which of the dimensional forms the imported OSM digital map is derived from.
This is done using a conversion algorithm, developed as part of the coordinate reference system, and applying it to the digital map imported from online or other resources. More on this shortly.
This exercise was about observing the effects of using the wrong and the correct coordinate reference systems and how they alter or modify the imported digital map’s on-screen projection view, and more importantly to see how the various local CRSs could adversely affect the OSM map of Belgium (and, for that matter, any digital map) as they are loaded into a mapping project – in this case the my “First Run Mapping” project. It is about becoming used to seeing the effects of using the correct coordinate reference system, and is as much about observing how they affect a map and also being able to notice that a CRS being used might be incorrect.
Click Settings > Custom CRS.
The Custom Coordinate Reference System Definition dialog window opens.
Or you can set the CRS in Project Properties.
Click Project > Project Properties. Click the CRS panel button on the left.
Figure 1: The Project Properties window, with the CRS panel selected.
Here in this dialog you can add any of a full range of CRS related to various regions of the world.
The first point to note was the NAD27 / Alaska Albers CRS had been loaded (this I believe now I accidentally loaded in with the OSM map – I cannot be sure. In any case, it proves an interesting point in this exercise – knowing it is clearly the wrong CRS geographic location by the name alone – and another point too.
In the panel below the “Selected CRS” line, there are a series of coordinates. The latitude and longitude coordinates mentioned are incorrect for the region of Belgium (taken from the ETRS89 / Belgian Lambert 2008 – EPSG:3812:
+proj=lcc +lat_1=49.83333333333334 +lat_2=51.16666666666666 +lat_0=50.797815 +lon_0=4.359215833333333 +x_0=649328 +y_0=665262 +ellps=GRS80 +towgs84=0,0,0,0,0,0,0 +units=m +no_defs
Where the Alaskan coorindates from the NAD27 / Alaska Albers – EPSG:2964 are:
+proj=aea +lat_1=55 +lat_2=65 +lat_0=50 +lon_0=-154 +x_0=0 +y_0=0 +datum=NAD27 +units=us-ft +no_defs
This was easy to determine. Longitude west of Greenwich is measured from 0° to -180° and Alaska (lon_0=-154) is a long way west of Greenwich. Longitude east of Greenwich is measured from 0° to 180°. Belgium is not very far east of Greenwich and the lon_0=4.359215833333333 measurement is indicative of this.
• When I begin a new project I will verify how this Alaskan CRS was made present.
To find a suitable CRS (remembering this is an exercise in selection rather than verifying which CRS to use - which will come as a later technical log given its importance) from the embedded list in the CRA panel, I entered data into the Filter row for “Belg” in this example (intending to capture as many CRS datasets as possible. Results are below in Figure 2.
Figure 2: Filtered CRS listings for ‘Belg’
To apply a specific CRS, you select it, and click ‘Apply’.
Figure 3: Selecting and applying a CRS – once selected, click the apply button.
• A function of the software (probably a non-critical bug) is that you must now click ‘OK’, have the window exit, and wait for the changes to fully display in the dialog window. However the Apply click re-renders the map to the new CRS algorithm. Then reopen the window to see the results displayed correctly.
Upon clicking ‘Apply’, I watched the map get altered according to the CRS mathematical recalculation of the local coordinates. The bottom left panels loaded monitor showed progress.
This is where the application of a CRS can be made visible on a digital map like the OSM map of the world. For the purpose of this exercise, I have selected, as Figure 4 shows, a range of CRS definitions. I progressively applied each and recorded a screen shot of all of them as shown below.
Figure 4: The CRS choices made during this exercise.
Figure 5: ETRS89 / Belgian Lambert 2008 – EPSG:3395
Figure 6: ETRS89 – EPSG:4258
Figure 7: WGS 84 / World Mercator – EPSG:3395.
Figure 8: NAD27 / Alaska Albers – EPSG:2964. Note the very different result here! The rotation is in the order of 270°, and graphically demonstrates the wrong CRS can have dramatic results. We would say OOPS here?
This demonstrates the mercator projections deficiencies in consistent conversion as the eliptic point on the earth changes according to the region being georeferenced. Alaska is a long way north of Belgium and also a considerable distance from the datum points for the mercator, as is Belgium. The conversion here clearly shows the consequences of using the wrong CRS conversion. The error here was expected.
Figure 9: NAD83(CSRS98) – EPSG:4140.
Each CRS chosen has different effects on the map. For the time being, I left the ETRS89 / Belgian Lambert 2008 – EPSG:3395 CRS selected (Figure 5).
While this might not be the correct choice for the OSM map, I will cover this element in the next technical log!
This technical log is part of the most important elements in this GIS project. It goes beyond being able to add rasterised paper maps, and is about ensuring that the correct projection on a digital map is performed, by using the correct Coordinate Reference System (CRS). Depending on the CRS choice, and actually selecting the correct CRS for the digital map data being displayed is critical to successful dimensioning tasks when creating vector points, polylines, or polygons in a GIS project files vector layers. Given distance and point locations is part of this thesis project, this set of technical logs is about exploring CRS, and their relationships, selections, and potential problems. Maps are broken down into different projections (as noted in my gis.tex and gis.pdf technical logs). Earth is a sphere, it is a complex three dimensional shape. The different projections of earth are related to the fact that maps display a two dimensional, flat representation of a sphere, or selected portions of a sphere. As such, any two dimensional map will distort our visual representation because we are flattening out a convex surface. GIS systems can overcome this flattening, providing a true representation on screen using wonderfully clever (well, I think its clever!) mathematical conversions, but the imperative here is to understand from which of the dimensional forms the imported OSM digital map is derived from. This is done using a conversion algorithm, developed as part of the coordinate reference system, and applying it to the digital map imported from online or other resources. More on this shortly. This exercise was about observing the effects of using the wrong and the correct coordinate reference systems and how they alter or modify the imported digital map’s on-screen projection view, and more importantly to see how the various local CRSs could adversely affect the OSM map of Belgium (and, for that matter, any digital map) as they are loaded into a mapping project – in this case the my “First Run Mapping” project. It is about becoming used to seeing the effects of using the correct coordinate reference system, and is as much about observing how they affect a map and also being able to notice that a CRS being used might be incorrect. Click Settings > Custom CRS.
Figure 1: The Project Properties window, with the CRS panel selected. Here in this dialog you can add any of a full range of CRS related to various regions of the world. The first point to note was the NAD27 / Alaska Albers CRS had been loaded (this I believe now I accidentally loaded in with the OSM map – I cannot be sure. In any case, it proves an interesting point in this exercise – knowing it is clearly the wrong CRS geographic location by the name alone – and another point too. In the panel below the “Selected CRS” line, there are a series of coordinates. The latitude and longitude coordinates mentioned are incorrect for the region of Belgium (taken from the ETRS89 / Belgian Lambert 2008 – EPSG:3812: +proj=lcc +lat_1=49.83333333333334 +lat_2=51.16666666666666 +lat_0=50.797815 +lon_0=4.359215833333333 +x_0=649328 +y_0=665262 +ellps=GRS80 +towgs84=0,0,0,0,0,0,0 +units=m +no_defs Where the Alaskan coorindates from the NAD27 / Alaska Albers – EPSG:2964 are: +proj=aea +lat_1=55 +lat_2=65 +lat_0=50 +lon_0=-154 +x_0=0 +y_0=0 +datum=NAD27 +units=us-ft +no_defs This was easy to determine. Longitude west of Greenwich is measured from 0° to -180° and Alaska (lon_0=-154) is a long way west of Greenwich. Longitude east of Greenwich is measured from 0° to 180°. Belgium is not very far east of Greenwich and the lon_0=4.359215833333333 measurement is indicative of this. • When I begin a new project I will verify how this Alaskan CRS was made present. To find a suitable CRS (remembering this is an exercise in selection rather than verifying which CRS to use - which will come as a later technical log given its importance) from the embedded list in the CRA panel, I entered data into the Filter row for “Belg” in this example (intending to capture as many CRS datasets as possible. Results are below in Figure 2.
Figure 2: Filtered CRS listings for ‘Belg’ To apply a specific CRS, you select it, and click ‘Apply’.
Figure 3: Selecting and applying a CRS – once selected, click the apply button. • A function of the software (probably a non-critical bug) is that you must now click ‘OK’, have the window exit, and wait for the changes to fully display in the dialog window. However the Apply click re-renders the map to the new CRS algorithm. Then reopen the window to see the results displayed correctly. Upon clicking ‘Apply’, I watched the map get altered according to the CRS mathematical recalculation of the local coordinates. The bottom left panels loaded monitor showed progress. This is where the application of a CRS can be made visible on a digital map like the OSM map of the world. For the purpose of this exercise, I have selected, as Figure 4 shows, a range of CRS definitions. I progressively applied each and recorded a screen shot of all of them as shown below.
Figure 4: The CRS choices made during this exercise.
Figure 5: ETRS89 / Belgian Lambert 2008 – EPSG:3395
Figure 6: ETRS89 – EPSG:4258
Figure 7: WGS 84 / World Mercator – EPSG:3395.
Figure 8: NAD27 / Alaska Albers – EPSG:2964. Note the very different result here! The rotation is in the order of 270°, and graphically demonstrates the wrong CRS can have dramatic results. We would say OOPS here?
This demonstrates the mercator projections deficiencies in consistent conversion as the eliptic point on the earth changes according to the region being georeferenced. Alaska is a long way north of Belgium and also a considerable distance from the datum points for the mercator, as is Belgium. The conversion here clearly shows the consequences of using the wrong CRS conversion. The error here was expected.
Figure 9: NAD83(CSRS98) – EPSG:4140. Each CRS chosen has different effects on the map. For the time being, I left the ETRS89 / Belgian Lambert 2008 – EPSG:3395 CRS selected (Figure 5). While this might not be the correct choice for the OSM map, I will cover this element in the next technical log!