Agriculture Reference
In-Depth Information
excursions where long discussions were held on Soil Taxonomy in and next to large pits, testing
keys of taxonomic systems. To some, this emphasis on classiÝcation was problematic, as little
attention was paid to quantitative process-studies, to soil-landscape analysis, and to land use and
its effect on soils.
Still, as long as classiÝcation systems were being developed, one could support such a focus.
Now, however, systems have reached a certain degree of maturity, and it is wise to Ýrst signiÝcantly
advance our understanding of dynamic soil processes before focusing on the development of yet
another revision of existing classiÝcation systems.
SOIL SURVEY INTERPRETATION, OLD AND NEW
Soil survey interpretations have been a prime vehicle for introducing soil classiÝcation to a
wide variety of soil users. Particularly in the United States, interpretations of soil maps for uses
ranging from agricultural suitability to suitability for on-site application of septic tank efÞuent have
been quite succesful over the years. They demonstrated the implications for land use of having a
wide variety of soils occurring in landscapes. Standard soil survey interpretations were made in
terms of deÝning relative suitabilities or limitations for a wide variety of land uses. Increasingly,
more detailed and quantitative information was needed to answer more speciÝc questions, and
better monitoring methods and computer simulation models have been used to provide such
information (e.g., Bouma and Jones, 2001). At Ýrst, questions associated with water movement in
soils received the most emphasis, but applications were later extended to plant growth and movement
of solutes (e.g., Bouma et al., 1980). Unfortunately, use of simulation modeling in the context of
soil survey interpretation has remained limited in the United States. Elsewhere, simulation has been
widely used in the context of land evaluation, which was often based on soil maps deÝning regional
soil patterns. As models became more advanced, lack of appropriate data to feed the models formed
an increasing problem. The advance of pedotransferfunctions, which relate available soil survey
data (that cannot be used as such) to data that can be used in modeling has been an important
justiÝcation for using soil survey data (e.g., Wosten, 1997; Wosten et al., 1999). Also, simulations
could be improved by a functional characterization of a given soil pedon. In other words, pedological
soil horizons were compared to see whether they were different in terms of physical (Wosten et
al., 1985) or chemical (Breeuwsma et al., 1986) behavior. When different pedological horizons
showed comparable behavior, they were lumped into thicker functional horizons. Thus fewer
functional horizons could be used in the simulations, making the work more efÝcient. Functional
characterization has also been applied recently to farmersÔ Ýelds, deÝning management units for
precision agriculture (Van Alphen and Stoorvogel, 2001).
Soil survey interpretations or land evaluation, whether qualitative or quantitative, are provided
for soil taxa, often at the soil series level. This has remained attractive over the years because the
occurrence of soil taxa in the Ýeld is expressed on soil maps through the map legend. As of old,
interpretations relate to areas of land which is crucial for all applied work. Of course, much
variability is involved here, which explains all the work that has been done on spatial and temporal
variability in soil science, to be discussed next.
VARIABILITY IN SPACE AND TIME: DISTINGUISHING MEANS AND PURPOSES
Classical soil survey interpretations stratify landscapes in mapping units that are each represented
by a Ñrepresentative proÝle,Ò for which assessments are made that are implicitly expected to apply to
the entire units, ignoring spatial variability within the units and assuming sharp boundaries among
them. Geostatistics, as recently reviewed by Heuvelink and Webster (2001), has been developed and
perfected as a tool during the last decades to deal with variations in space and time. These authors
