Agriculture Reference
In-Depth Information
A future approach could be inspired from the plant sciences, which clearly separate plant
taxonomy, phytosociology, and phytotechnology. Plant taxonomy aims at producing a system of
classiÝcation that best reÞects the totality of similarities and differences among plants. Phytosoci-
ology handles the distribution of plant communities, reÞecting the interdependence of the species
and the environmental relationships that inÞuence their distribution. Phytotechnology classiÝes
plants in terms of their use (cereals, tubers, Ýber, fruits, oil crops, etc.) across taxonomic boundaries
and on the basis of speciÝcally utilitarian differentiae. With regard to soil classiÝcation, a taxonomy
in a global context could consist of two levels, a Ýrst one comprising reference soil groups, which
would accommodate the major high-level classes of different national systems, and a second level
of qualiÝers, allowing for a more precise description and deÝnition of subdivisions of the reference
groups. Several qualiÝers could be used to classify a soil in a sequence that best suits the purpose,
or in accordance with an agreed-upon ranking. The development of such a World Reference Base
for Soil Resources started in the 1980s (Dudal, 1980; 1990) and has recently materialized in a
concrete proposal (FAO/ISRIC/ISSS, 1998; Deckers, 2000; Nachtergaele et al., 2000) that was
endorsed by the IUSS in 1998. It has allowed for a balanced overview of the worldÔs soil resources
and serves the purpose of small-scale soil resource inventories (FAO, 1991). The reference base
serves as a means of communication and as a bridge between national systems that, at lower levels
of generalization, can accommodate geographically speciÝc differentiae.
Increased attention should be given in soil surveys to the mapping of soil landscapes, and
highlighting temporal and spatial dynamics within a landform, rather than merely recording the
distribution of taxa as single units or as unqualiÝed associations. The spatial structure of the soil
cover does not usually correspond with the hierarchy in systems based on taxonomic differentiae.
Environmental issues require information of variation within the landscape, which is not provided
by the prevailing soil survey procedures. A World Soils and Terrain database (SOTER)
(UNEP/ISSS/ISRIC/FAO, 1995) is being established, containing digitized map units and their
attributes composed of sets of Ýles for use in relational database management systems. In addition
to soil and terrain boundaries, observation points should be recorded and georeferenced, making
full use of tools and techniques of information technology.
Interpretation of soil characteristics toward land use requirements should continue to rely on
soil-landscape models of variation. However, modern demands on soils require that data, as well
as the differentiae of taxa, be measured and recorded. The limits of the human mind are no longer
constraining; data storage and processing can now be handled digitally. Systematic surveys are not
required to obtain the necessary information. Georeferenced databases lend themselves to selecting
and grouping clusters of relevant data in function of demand. They allow for a Þexible weighting
of properties in accordance with purpose (Burrough, 1993). Hence the boundaries in technical
groupings often cut across taxonomic subdivisions.
Although the three purposes of soil classiÝcationÐtaxonomy, soil survey, and interpreta-
tionÐwould beneÝt from being handled more independently, it is imperative that the experience
gained with the relationship among soils, landform, and climate be blended with the new methods
of data processing. Blind mathematical approaches could lead to stratiÝcations of doubtful signif-
icance. A geographic dimension of the soil cover is essential for ensuring sound applications of
soil science.
Our soil classiÝcation has served us well. It has enabled us to explain and characterize soil
diversity in function of different sets of soil-forming factors. However, with increasing demands
for targeted soil information and with the advent of modern tools for data storage and processing,
soil classiÝcation now needs to be addressed more speciÝcally to the purposes that it is meant to
serve. The rationale of its hierarchy and of the selection and weighting of its differentiae has to be
clearly spelled out. Additional research is required on the relationships among soil characteristics,
their effects on plant growth, and their use for different purposes.
FAO (1998) initiated a database for identifying plant species appropriate for given environ-
ments and uses, and for assessing crop response to environmental and management factors. Soil
