Agriculture Reference
In-Depth Information
is slightly beyond the concept of a taxonomic class. Either the soil can be classiÝed at the lowest
taxonomic level that it Ýts, or it can be correlated to the closest available series and any properties
that are outside the range of the selected series can simply be disregarded in the taxonomic
description (presumably to be handled as inclusions in mapping). In most instances, soil survey
practitioners have followed the second alternative, although not always. For example, the soil survey
of the Flathead National Forest (Martinson and Basko, 1998) correlated the soils at the family level,
and constructed map unit names using phases of higher categories for brevity. Two examples are
Fluvents, alluvial fans
Despite the passage of
18 years since the decision was made to restrict series ranges to the limits of higher categories,
Alexander (1983) and Borst (1983), each reacted to Guthrie by arguing that series should be allowed
to encompass a range reÞecting the properties observed in the Ýeld, even if they cross higher class
limits. Nettleton et al. (1991), describing what they called the Ñtaxadjunct problem,Ò reported that
about half of the pedons analyzed by the National Soil Survey Laboratory have one or more
properties outside the family limits for the series identiÝed in the Ýeld. They suggested that we
classify the central concept of the soil and allow the described range in characteristics to cross the
boundaries of higher taxa (one of the original alternatives discussed 28 years earlier at the NCSS
conference). Holmgren (1988) suggested that an operational deÝnition should have been developed
to depict the procedure for describing a pedon in a spatial context. He contends it was a mistake
to develop an operational deÝnition for the pedon itself. While he did not propose an operational
deÝnition, it seems that he envisioned a set of multiple observations, related to each other geo-
graphically (similar to Ñsatellite samplingÒ), to characterize the pedon. In this way, there would be
multiple possible outcomes that could be realized as one observes a pedon repeatedly. A polypedon,
then, would have some ÑprobabilityÒ or ÑexpectationÒ of meeting a particular classiÝcation.
Today we still talk of pedons, although rarely do we truly identify, describe, and sample this
three-dimensional body. More often than not, the pedon becomes a spoil pile next to the pit and
we describe a two-dimensional soil proÝle. Perhaps we need a procedure to Ñdissect a pedonÒ rather
than simply Ñdigging a pit.Ò The polypedon concept has been largely ignored because it is too
difÝcult to locate itÔs boundary in the Ýeld, and itÔs very concept is contradictory and relies on
circular reasoning (Soil Survey Division Staff, 1993).
and
Typic Eutroboralfs, silty till substratum, hilly.
Taxonomic Purity of Map Units
A number of pedologists have looked at the taxadjunct problem in the context of the composition
of map unit delineations to estimate the proportion of pedons not Ýtting within the taxonomic class
for which the map unit is named. McCormack and Wilding (1969) studied a portion of northwestern
Ohio that had been mapped before the introduction of Soil Taxonomy. It was considered to have
very complex soil patterns, resulting in a high degree of variability. Overall, they found that 74%
of the more than 200 pedons observed Ýt the expected Order, while just 17% Ýt the expected Family
and Series. An additional 26%, however, were close enough to be considered taxadjuncts. Despite
the low taxonomic purity, they concluded that only a few of the delineations observed were
improperly mapped.
Nordt et al. (1991) studied the taxonomic makeup of four map units in east Texas. The soils
observed were Crockett (Ýne, montmorillonitic, thermic Udertic Paleustalfs), Rader (Ýne-loamy,
mixed, thermic Aquic Arenic Paleustalf), Robco (loamy, mixed, thermic Aquic Arenic Paleustalfs),
and Spiller (Ýne, mixed, thermic Udic Paleustalfs). The taxonomic purity of the map units, following
a pattern similar to that found by McCormack and Wilding (1969), is highest at the Order, Suborder,
and Great group levels, and drops with progressively lower taxa, with the percent of pedons
classifying in the named series ranging from just 11% for Rader to 49% for Crockett. In addition
to the taxonomic purity, they also considered the interpretive purity of the map units. When soils
considered to be similar to the named series were added, the percentage of Ñinterpretive purityÒ
was estimated to be 48 for Rader, 52 for Robco, 81 for Spiller, and 86 for Crockett. The Rader
