Agriculture Reference
In-Depth Information
Some of the deÝnite improvements
of the 1991 classiÝcation system over the 1977 system
include the following:
¤
Good provision for the classiÝcation of most soils of the semiarid and arid regions. From a practical
land-use point, the clear separation between soils with Soft carbonate and Hardpan carbonate
horizons, respectively, is very good. There are still some outstanding issues regarding soils of the
semiarid and arid regions, however, as will be pointed out later.
¤
The much Ýner subdivision of soils with Podzol B horizons. Although these soils are of limited
geographic extent and of little agricultural value, they are very important for the forestry industry.
¤
RedeÝnition of the E horizon to force soil surveyors to determine their color in the dry state.
¤
Some important new family criteria, such as recognition of the yellow E horizons, provision for
the separation of bleached A horizons from nonbleached A horizons, provision for thin Humic A
horizons, and the distinction between structured B horizons with subangular/Ýne angular blocky
structure and those with medium/coarse angular blocky structure.
¤
Improved structures for the key to soil forms and the family matrix tables to make them easier to use.
The 1991 system has a number of deÝciencies, some minor and some serious. These are a
few of these deÝciencies:
¤
The change in deÝnition of the Vertic A horizon. In the 1977 system, it had to have one or more
of the following: clearly visible slickensides or cracks wider than 25 cm throughout at least half
of the thickness of the horizon when it is dry, or self-mulching properties at the surface. In the
1991 system, it must have at least one of the following: clearly visible slickensides or a plasticity
index greater than 32 (using the standard SA Casagrande cup to determine liquid limit) or greater
than 36 (using the British Standard cone to determine liquid limit). According to P.A.L. Le Roux
(University of the Free State, personal communication), there are many soils with higher plasticity
values than these that are not Vertic A horizons. Van der Merwe (2000) found clearly that plasticity
index cannot be used as criterion to differentiate between Vertic A and Melanic A horizons. She
also showed that the statement in the discussion on Melanic A horizons in the 1991 classiÝcation
topic, that the difference between Melanic A and Vertic A horizons in regard to swell-shrink
properties can mainly be ascribed to differences in clay mineralogy, is not true (Van der Merwe
et al., 2001).
¤
The biggest deÝciency of the 1991 system is that the deÝnition of new soil series was left hanging
in the air. For many soil forms this is not a problem, because their family deÝnitions cover just
about everything that is of practical importance that can be used in taxonomic classiÝcation. The
remaining criteria that can be used in soil surveying are those that are normally used as phase
criteria (Van Wambeke and Forbes, 1986). In these forms, there is actually no ÑspaceÒ for series
below the soil family. Unfortunately, these include almost exclusively the Ñlow qualityÒ soil forms,
which are unsuitable, or at best only marginally suitable for crop production, either rainfed or
irrigated. The shallow Glenrosa form had 20 in series in the 1977 system, and has 16 well-deÝned
families in the 1991 system (Tables 16.4 and 16.5). The problem lies with the Ñgood qualityÒ soils,
for which the family deÝnitions are very wide and do not allow good land suitability evaluation,
especially with regard to management requirements, as will be pointed out. The Hutton form, for
example, had 27 noncalcareous series in the 1977 system, but has only six families in the 1991
system (Tables 16.2 and 16.6).
ADVANCES IN SOIL CLASSIFICATION IN SOUTH AFRICA SINCE 1991
From 1991 to 1996, nothing much happened with regard to soil classiÝcation in South Africa.
The main emphasis during this period was supposed to be on the development of new soil series.
For various reasons, this became bogged down completely. One reason was that the majority
feeling in 1983 was already that series should be deÝned as natural soil bodies regarding soil
texture, and not as artiÝcial entities, subdivided into arbitrary textural classes. But it was never
