Current soil family particle-size criteria

For the purposes of classification at the family level, as currently defined in the CSSC (SCWG 1998), particle size refers to the whole soil composition including coarse fragments (>2 mm diameter). Family particle size is related to soil texture but differs in that soil texture refers specifically to the composition of the fine earth fraction (i.e. proportion of sand, silt and clay size particles ≤2 mm diameter) and is specific to individual horizons. Family particle size includes the fine earth fraction by mass (≤2 mm diameter) plus coarse fragment (>2 mm diameter) content by volume averaged for the entire control section of the profile. Exceptions are provided for soils with strongly contrasting particle sizes within the control section that affect soil properties which are not captured at higher taxonomic levels (SCWG 1998, p. 138–139). Modifiers for coarse fragment content are applied to soil texture classes for each horizon, but these are not separate texture classes (Day 1983). In practice, family particle size is typically assessed based on the textures within a profile weighted for the thickness of each horizon plus consideration of the overall coarse fragment content.

Within the current version of the CSSC (SCWG 1998), there are 11 classes/categories used to describe particle-size characteristics at the family level. Note however that only seven classes are presented in the family particle-size triangle (Fig. 1; SCWG 1998, p. 136) because the remaining four classes result due to increased coarse fragment content. These are: clayey-skeletal (includes very fine clayey and fine clayey base classes with ≥35 to <90% coarse fragments by volume), loamy-skeletal (includes coarse-loamy, fine-loamy, coarse-silty and fine-silty classes with ≥35 to <90 % coarse fragments), sandy-skeletal (includes sandy class with ≥35 to <90% coarse fragments) and fragmental (soils composed of ≥90% coarse (>2 mm diameter) fragments). Versions of particle size prior to 1973 had only three primary particle size classes: coarse, medium and fine which followed boundaries to form groups of soil textural classes plus skeletal versions for each and a fragmental group (NSSC 1968). The classes, adopted from US Soil Taxonomy, were originally intended to reflect an equal balance between agronomic and engineering influences (CSSC 1973). For example, the limit of 18% clay between coarse-loamy and fine-loamy classes reflects the change from non-plastic to plastic behaviour with increasing clay content (Handy and Fenton 1977). This is considered by engineers to be an important distinction. Similar breaks related to plastic and liquid behaviour occur at 35% and 60% clay content. There is also a difference between the coarse and fine silty and loamy classes, which relate to capillary rise and available moisture-holding capacity. These breaks were intended to allow groupings of soils with similar responses to management and to some extent, for engineering and related uses (SCWG 1998).

Fig. 1.

Current particles-size classes (left) and soil textural classes (right) based on SCWG (1998).

Impracticality of the current soil family particle-size classes

When comparing the current family particle-size classes with soil texture classes (Fig. 1), it quickly becomes evident that the boundaries between many of the family particle-size classes do not coincide with boundaries for textural classes, resulting in particle-size boundaries dividing soil texture classes. Exceptions are the breaks between clay and heavy clay textures corresponding to fine-clayey and very-fine-clayey particle sizes and the boundary between sandy and coarse-loamy particle-size classes coinciding with the boundary between sandy loam and loamy sand textures. There is also a difference between the engineering and agronomic definitions in the grain-size cutoff for sand and silt-sized particles where the cutoff for agronomic applications is typically 0.05 mm diameter while engineering applications use 0.074 mm (SSDS 2017; Schoeneberger et al. 2012). There are also differences between engineering and pedological definitions for subdivisions (fine, medium and coarse) of gravel (USDA 1987; Schoeneberger et al. 2012).

Although the current family particle-size classes were intended as a compromise to incorporate engineering criteria at the family level (CSSC 1973), different classification systems exist worldwide which are used for different applications (Garcia-Gaines and Frankenstein 2015), and sometimes there is no direct relationship between systems. Whereas pedologists are more concerned with the control section, engineers focus primarily on parent materials and materials below (Pawluk 1970). In fact, regulations require that topsoil (A horizon materials) must be stripped and stockpiled from sites prior to construction activities of interest to engineers. Engineering properties of the solum are consequently of little use in practice.

The utility of the current particle-size triangle was based on the premise that analytical data such as sand, silt and clay content (including values for the 0.074 mm division between sand and silt) and coarse fragment content were collected and readily available (i.e., published soil survey reports) in addition to agronomic attributes required for classification at the series level. Soil survey maps and reports are often lacking engineering test data such as Atterberg and plasticly indexes, particle-size distributions according to unified and or A.A.S.H.O. systems which are not commonly used by agronomists, and therefore these reports are of limited direct utility to engineers without additional interpretations or assumptions. Data required for particle-size classes are for the most part lacking and definitely absent in older soil survey reports before the advent of the soil family class. Grain-size distributions expressed as a graph also afford more useful information for engineers than do particle-size classes, and because such interpretations are multi-disciplinary, they should be handled separately (Pawluk 1970). Although there were some indifference and reluctance among soil survey committee members at the time to adopt family criteria, it was recognized that there was a need to sort and group soil series to point out discrepancies in the criteria used for soil series (CDA 1970).

Soil survey reports, with some exceptions, provide profile descriptions with soil texture classes for individual horizons with only some original data for grain-size analyses (percentages of sand, silt and clay). In Canada, some data related to engineering properties have been included in soil survey reports; however, they are often limited to single pedons for soil series with extensive areal distribution (e.g., Presant and Wicklund 1971; Coen and Holland 1976; Eilers and Halstead 1981; Luttmerding 1981; Wang and Rees 1983; Kingston and Presant 1989; Lamontagne et al. 2014) and are only available for a limited number of published soil surveys. Whereas textural classes see immediate use in classification at the series level, grouping of soil series into soil families for the purpose of applying management practices may only be considered sometime later. Any analytical data collected at the time of sampling may have been lost, or the original textural data may have been estimated from hand texturing. The result being that the current particle-size classes are difficult to accurately assess, particularly when based on older pedological data without additional data required for classification at the family level typically not collected initially. Consequently, the family classification has largely been ignored in the past, and agronomic management practices have been applied with little formalized guidance towards grouping of soils for targeting proposed management practices which was the original intent of the family taxon. Use of family classes to group soil series for application of appropriate management practices would have a benefit of streamlining and efficiently targeting field trials to help identify and target soils which may have been overlooked in the past or alternatively avoid duplication of research efforts on soils that are similar. This would also aid in identifying soils with similar properties at the family level for application of appropriate management practices targeting specific soils.

Another disadvantage of the current family classification is the use of long attribute-based family names, making their application very cumbersome. Although the attributes are highly descriptive, assignment of a common soil series name to designate each family has been recommended for convenience and brevity (SCWG 1998, p. 144). Adoption of a series name (e.g., Breton family in place of Orthic Gray Luvisol, fine-loamy, mixed, neutral, cold, subhumid family) for the family name based on a series that is most representative (i.e. greatest areal extent) or oldest or most recognized name within a given soil family class would be most appropriate. Formalized adoption of common series names in place of attribute-based names would help facilitate more widespread application and ease of use.

Particle-size classes

Proposed changes to soil particle-size classes are summarized in Figs. 2, 3 and 4. The modified ‘Base’ particle-size triangle (Fig. 2) is similar to s-type particle sizes used in Forest Ecosystem Classification (Sims et al. 1989) which align with soil texture (fine earth) classes. This triangle (Fig. 2) is proposed to replace the particle-size triangle on Figure 41, left side (p. 136) of CSSC (SCWG 1998). These new proposed particle-size classes are based on percentages by weight of clay, silt and sand using 0.05 mm (50 μm) as the division between sand and silt grain sizes. An exception is that very fine sand textures are lumped with the coarse-loamy particle class rather than the sandy particle-size class. Soils containing coarse fragments of all sizes >2 mm diameter exceeding 35% by volume (approximately 50% by mass) and classified as ‘skeletal’ are summarized in a new figure (Fig. 3), while those soils with ≥90% coarse fragments of all sizes >2 mm diameter by volume are classed as fragmental. Coarse fragments as defined within certain size and shape ranges are shown in Table 1 (Day 1983). To maintain consistency, the proposed particle-size class names and those for ‘skeletal’ and fragmental classes are the same as the current version of the CSSC (SCWG 1998), except that the boundaries have been shifted to coincide with soil texture boundaries (Fig. 3).

Fig. 2.

Proposed particle-size triangle for soils with <35% coarse fragments by volume. *Very fine sand is included with coarse-loamy particle-size class.

Fig. 3.

Proposed particle-size triangle for soil with ≥35% to <90% coarse fragments by volume (i.e., skeletal soils). *Very fine sandy-skeletal is included with loamy-skeletal particle-size class.

Fig. 4.

Proposed particle-size triangle describing terric layers within organic soils. *Very fine sand is included with the loamy particle-size class.

Table 1.

Modifiers applied to texture classes based on coarse fragment (>2 mm diameter) shape and size.

Textural and engineering soil classification systems are not directly translatable because the latter are not based purely on texture (grain size) but also use plasticity data which is reflected in clay mineralogy. It is unlikely that a precise translation between textural and engineering classifications will ever be made because their purposes differ (Handy and Fenton 1977). Engineering classifications are directed towards variations in soil behaviour relevant to engineering, and textural classifications are more concerned with pedological description. Consequently, soil data required for engineering purposes should be collected and handled separately and in addition to pedological data.

The distinction between coarse-silty and fine-silty particle-size classes has been combined in favour of a larger single silty particle-size class that encompasses silt and silt loam textures. The distinction between coarse and fine silty particle-size classes relates to differences in support strength and to a lesser extent moisture-holding capacity (Handy and Fenton 1977). Although the occurrence of pure silty materials (coarse-silty particle size) may be more common in the United States, they seem to be rare in the Canadian landscape, thus not warranting more than a single particle-size class. This reduces the number of ‘base’ particle-size classes (coarse fragment contents <35% by volume) from 7 to 6. Similarly, proposed particle-size classes for soils with coarse fragment contents ≥35% by volume (Fig. 3) also have boundaries coinciding with soil texture boundaries retaining particle-size class names from the current version of the CSSC. A new silty-skeletal class is proposed, and although rare, it is included for completeness (Fig. 3). Figure 4 provides the particle-size classes used to described terric layers found in organic soils. All 10 proposed family particle-size classes are summarized in Table 2. Table 3 provides a summary of proposed family particle-size classes along with corresponding modifiers used for soil texture classes for horizon descriptions which are based on coarse fragment content by volume.

Table 2.

Summary of proposed soil family particle-size classes.

Table 3.

Summary of soil texture class modifiers and family particle-size classes based on coarse fragment content.

Classification of soils with two or more parent materials in the control section with strongly contrasting particle-size classes must also be revised. Table 4 provides a listing of terms for particle-size classes for soils having strongly contrasting layers. The difference from the prior table (SCWG 1998, p. 139) is the combination of coarse-silty and fine-silty particle size columns into a single silty class and addition of the silty-skeletal class. The minimum significant thickness of a strongly contrasting layer is 15 cm, and the transition between layers is less that 12 cm thick as per the current version (SCWG 1998). Table 4 herein should be substituted for table 1 on page 139 of the CSSC for consistency with the proposed changes.

Table 4.

Strongly contrasting particle sizes.

Mineralogy classes in mineral soils

A minor change/addition to table 2 Key to Mineralogy Classes (p. 140) of the current version of the CSSC (SCWG 1998) is to add a ‘Mixed’ class under ‘Classes applied to soil families of any particle-size class’ (SCWG 1998, p. 140, table 2). This class is proposed as a mixed mineralogy class which is by far the most common mineralogy class compared with those currently listed yet was not included in previous versions.

Reaction classes in mineral soils

Table 5 is proposed as an amendment to reaction classes for mineral soils substituting for the current listing of pH reaction classes on page 141 of the current version of the CSSC (SCWG 1998). The amended table adds the more detailed reaction classes as reported in Day (1983) to aid in clarity for the reader.

Table 5.

Soil family reaction classes and corresponding field pH ranges.

Calcareousness classes in mineral soils

A minor change to family calcareousness classes is the addition of a non-calcareous class corresponding to acid and neutral pH classes (Table 6). This additional calcareous class is proposed for completeness and to be substituted for the current list on page 141 in the current version of the CSSC (SCWG 1998).

Table 6.

Soil family calcareous classes and corresponding field calcareous class ranges.

Depth classes in mineral soils

Table 7 provides a comparison between current and proposed depth classes for lithic and permafrost (cryic) contacts for mineral soils. Proposed additions are the inclusion of non-lithic and non-cryic classes for soils with lithic contacts or permafrost contacts at >100 cm depth. These new criteria are proposed for completeness but may be omitted from many descriptions. There are no changes proposed for lithic and permafrost for organic soils. It is proposed that the depth of the lithic and cryic contacts between very shallow and extremely shallow be changed from <20 cm to <25 cm for consistency with other world soil classification systems and inclusion of the Leptosolic order to the CSSC (see Warren et al. In press). It is also proposed for the time being that the extremely shallow lithic criteria be applied as necessary only as a phase to very shallow lithic families. Otherwise, because of the hierarchical structure of the CSSC, assignment of separate series names would necessitate differentiation at the series level for soil belonging to extremely shallow vs. very shallow families. This is recommended in an effort to minimize unnecessary proliferation of soil series names.

Table 7.

Soil family depth classes for mineral soils (lithic and cryic).

All cryic criteria need only be applied in the case of extremely cold and very cold soil temperature regimes. Otherwise, a non-cryic class is implied and not explicitly used as it is redundant for all temperature regimes warmer than very cold and therefore omitted from most family class names.

Implications for acceptance of the proposed revisions

Adoption of the proposed family particle classes would align with particle-size classes which are currently used in Forest Ecosystem Classification (CFEC 2010) and other classification systems in Canada (Sims et al. 1989). Adoption of proposed particle-size criteria will also facilitate use of historical soil survey data published prior to the introduction of the soil family taxon in the 1960s for classification of these soils at the family level. This would facilitate a means of grouping soils with similar properties at the family level for application of appropriate management practices targeting specific soils which was the original intent of implementing the soil family taxon. Use of family classes to group soil series data for application of appropriate management practices would have a benefit of targeting field trials to develop management practices for soils which may have been overlooked in the past or alternatively avoid duplication of research efforts on soils with similar properties. The proposed changes will also separate engineering properties from pedological properties. Engineering data should be included in future soil survey reports; however, it should be collected and tabulated specific to engineering uses separate from pedological classification.