Soil Mapping Units (SMU) and Soil Typological Units (STU):

Conceptually, Soil Mapping Units (SMU) are geographic objects that have a clear and well defined geometric shape and location.

Some SMUs can be pure in which case they are composed of only one Soil Typological Unit (STU) covering 100% of the SMU. Only in that case, STUs can be considered as having a known shape and location. But at small scales (1:100,000, 1:250,000 and smaller) that case is quite rare. Examples of pure SMUs are given on figure 1 by SMUs 2 and 4.

Figure 1: organising soil information in a geographical information system.

In most cases SMUs are not pure, i.e. they represent soil associations, also called soil complexes. This is due to the fact that some soil types (i.e. STUs) can be identified (i.e. semantically defined by their name, texture, and all other descriptive characteristics) but cannot be located precisely on the map (i.e. their shape and location cannot be determined with a better precision than their SMU "container"). Examples of complex SMUs are given on figure 1 by SMUs 1 and 3.

This system is setup to account for the difference in resolution between the semantic and the geometric information which commonly occurs in soil surveying: soil types are identified and described with a higher level of precision than their geometry. The database structure reflects this system to avoid as well as possible any loss of information provided by the soil scientist (another common strategy is to simply ignore non dominant soil types, making a much more simple database of only pure SMUs at the expense of a great information loss).

Moreover, despite the unability to delineate STUs, the soil scientist is able to evaluate the proportion of the area of each SMU that is covered by each STU it is made up with. It is the role of attribute %AREA to store that information. It is a small step forward to improve the gap in semantic vs geometric resolution. In some soil databases such as the 1:250,000 scale database of Europe, further steps can be taken in an attemp to better describe the "organisation" or arrangement of STUs within SMUs. But in the Soil Geographical Database of Europe at scale 1:1,000,000 only %AREA is asked for.

Note also that STUs may be present (i.e. contained) in more than one SMU. For example, two different SMUs could be made up of the same STUs but in different proportions. An example of a similar case is illustrated in figure 1 by STU 10 which fully (100%) covers SMU 4 but only is part (30%) of SMU 1.

Dominant value maps and associated purity maps:

Maps are views of a geographical database. In particular, soil thematic maps are views of a soil geographical database. Maps allways convey only part of the information available in the database. When the structure of the database is made complex, decisions have to be taken on how to present these views. In particular, when SMUs are complex, decisions have to be taken on what to present in views representing STU characteristics.

The example presented in figure 1 is used hereafter to demonstrate the use of dominant value maps and their associated purity maps. You can click on any of the maps to get an enlarged view.

Figure 2: mapping the SMU identifier.


Figure 3: mapping the STU identifier and its % within each SMU.

In figure 3b, each circle slice is proportional to % area of each STU within the SMU.
In figure 3c, each stripe width is proportional to % area of each STU within the SMU.


Figure 4: mapping the dominant STU and its % area within each SMU.

In figure 4b, each SMU is fully colored with the color assigned to the dominant STU. Thus the map reader will get an exagerated impression of what really is in the SMU. Mapping the "purity" (figure 4c) allows the reader to temperate that impression.


Figure 5: mapping the STU texture class and its % within each SMU.

In figure 5a, texture classes are listed for each STU within each SMU along with the STUs' % area. Whereas in figure 5b, % area of textures are summarised by SMU.
In figure 5c and 5d circle slices, respectively stripe widths, are proportional to % area of each texture within the SMU.
Get a picture equivalent to figure 5c for Europe just to be convinced that this type of map is inapropriate with too many features to be shown (more than ~30). And get a picture equivalent to figure 5d to see that it can be very difficult to read with a large number of features.

Figure 6: mapping the texture class of the dominant STU and its % area within each SMU.

In figure 6b, each SMU is fully colored with the color assigned to the texture class which characterises the dominant STU within each SMU. Figure 6c shows the associated purity map, which is the same as the STUs purity (figure 4c)
Get pictures equivalent to figures 6b and 6c for Europe.
Figure 7: mapping the dominant texture class and its % area within each SMU.
In figure 7b, each SMU is fully colored with the color assigned to the texture class which is dominant within each SMU. This map gives a much more accurate view of the actual texture class of the SMU. This type of view is called a dominant value map and is the most commonly used to present views of STU characteristics from our soil databases.
Nevertheless its user still needs to be warned that it might not be fully accurate. Thus figure 7c shows the associated purity map.
Get a picture equivalent to figure 7b for Europe. This map is all together the most exhaustive, the most readable, and the most accurate representation of an STU characteristic. But it has to be paired with the equivalent to figure 7c.

Figure 8: mapping a particular texture class.

In figure 8b, presence or absence of a specific value for the theme is shown, whether it is dominant or not in each SMU. Whereas in figure 8c, one can get an idea of how much of that specific value there is within each SMU.
These maps more accurate but less exhaustive than the previous one. They are good for answering question such as: where is there such thing?
Get pictures equivalent to figures 8b and 8c for Europe.


Most of the maps showing Soil Typological Unit (STU) characteristics output from our soil geographical databases are dominant value maps and have an associated purity map provided as a warning to the reader.

This applies to characteristics which are qualitative (e.g. soil names, texture classes, etc.). In the case of quantitative characteristics (e.g. soil granulometry, soil depth, etc.), one could compute SMU summary statistics (e.g. averages, % area weighted averages, etc.), provided this makes sense, and then map the result.