Assessing the Australian soil classification using cladistic analysis.
Before the introduction of the Australian Soil Classification (ASC; Isbell 1992, 2002), two schemes were in operation in Australia: the Handbook of Australian Soils (Stace 1968) and The Factual key (Northcote 1971). In the former, Stace (1968) described 43 Great Soil Groups differentiated by central concepts of morphology, distribution and land utility, which, although easy to understand, were not explicitly defined (Mazaheri et al. 1995). The classification therefore had numerous misidentifications (McKenzie and Austin 1989). To overcome these, Northcote (1971) developed a key, marking a shift from classifying external factors towards intrinsic soil morphological properties (Northcote 1984). The bifurcating hierarchical design, which used mutually exclusive soil morphological properties, most notably the Principle Profile Form (PPF), removed uncertainty (Fitzpatrick et al. 2010).
The ASC is a general-purpose soil classification that uses clearly definable soil morphological and non-morphological properties. It is similar to Northcote (1971) in that it functions as a sequential key, with mutually exclusive criteria. Its usefulness lies in its ability to identify soils, rather than in capturing shared relationships between them. However, the scheme's lack of mutually exclusive characteristics has led to groups, such as the Hydrosols, having the same characters used that define other groups. Situations such as this arise from artificial classifications and highlight the need for the application of alternative methods to create a natural classification. The benefit of natural classifications is that they are independent of the interpretation of the classifier and represent only what occurs in nature (Wilkins and Ebach 2014). Moreover, they are predictive in that they can accommodate soils from other classifications, or the discovery of new soils.
The importance of natural classifications lies in what can be inferred from them. If taxa (units of classification) are more closely related to each other than they are to any other taxon, they form a natural group or homology (Table 1). For example, a natural soil group (homology) is evidence of a common history, process or mechanism, namely soil genesis. There is a common consensus that soils undergo traceable and explainable development or evolution (i.e. soil genesis), making them amenable to a natural classification assessment (Wilkins and Ebach 2014). The problem, however, has always been the lack of a reliable methodology to discover natural classifications.
Biological systematics offers an avenue in this endeavour, namely, cladistics (from the Greek klados, i.e. 'branch'). The concept began with Phylogenetic Systematics (Hennig 1966), a method that uncovers relationships between taxa based on shared derived character-states. The simple premise that two taxa are more closely related to each other than they are to a third taxon underpins cladistics (Schuh and Brower 2010). The power of cladistics lies in amalgamation of parts of taxa in the form of characters and their states, to create a branching diagram (cladogram) supported by character-states at nodes. Cladistics has proved successful in the natural grouping of birds and dinosaurs (Clark et al. 2002) in geology (Echeverry et al. 2012) and in astronomy (Fraix-Bumet et al. 2006).
Here, we use cladistics to examine whether existing Soil Orders, Suborders and Great Groups in the ASC form a natural classification, by conducting a cladistic analysis using TNT version 1.0 (Tree analysis using New Technology; Goloboff et al 2000). A natural classification occurs when clades (relationship between Soil Groups, Suborders and Orders) are formed in the analysis. A secondary aim is to identify the informative shared characters in the Suborders and Great Soil Groups, based on their distinguishing morphological and non-morphological character-states, and the impact of nonmorphological, continuous and colour characters on the ASC.
Material and methods
Australian Soil Classification
The ASC (Isbell 2002) is a hierarchical key and consists of five tiers: Orders, Suborders, Great Groups, Subgroups and Families. Table 2 shows the 14 Soil Orders. The Suborders are based on subsoil colour (i.e. Red, Brown, Yellow, Grey, Black), with Great Groups differentiated on a range of characters (Table 3). The Subgroups are a continuation of Great Groups. The Family criteria are common to all Soil Orders, including A-horizon thickness, A1 gravel and texture as well as B horizon texture and depth.
To assign an unknown soil to a taxa, the user employs a series of morphological (e.g. soil colour) and non-morphological characters. By using this approach, and considering all tiers, an unknown profile can fit in a maximum of 53 845 soil taxa. This reduces to a maximum 11076 at the fourth tier (i.c. Subgroups), 565 at the third tier, 78 at the second tier (i.e. Suborders), and 14 for the Soil Orders. Typically, soil is classified into Soil Orders.
One way in which these Soil Orders are increasingly being related to each other is shown in Fig. 1a, which is a plot of climate (i.e. rainfall) and parent material, and in lower parts of a landscape (Gray and Murphy 2002). This is because climate is seen as one of the most powerful soil-forming factors through its influence on the rate of chemical weathering of minerals from parent material. Figure 1 a can be represented as a cladogram to recognise relationships, based on shared homologies of climate and parent material, by the branching of taxa (Fig. 16).
Although it is inappropriate to classify soil solely on the basis of parent material, climate and topography (Basinski 1959; Marbut 1951), it does illustrate the potential of cladistics in identifying homologies (i.e. natural groups) based on a common ancestry (i.e. parent material) and relationship with climate as a function of morphological and non-morphological characters. For example, one homologous character of Vertosols and Ferrosols is their mafic parent material. However, Ferrosols form in higher rainfall and in situ and elevated areas associated with the south-east Australian Cenozoic Volcanic Sequence, whereas the Vertosols form as weathered and eroded products in semi-arid areas to the west and on the lower lying and expansive clay alluvial plains of the Murray-Darling Basin. A similar explanation can be provided for the siliceous Soil Orders, whereby, Podosols form from aeolian sands and along high-rainfall coastal areas of south-eastern Australia, whereas Tenesols form in the arid areas of central Australia.
Characters and coding
Forcladistic analysis, the morphological and non-morphological characters used to identify the soil taxa need to be coded into binary or multiple character-states. For most characters, coding is 'absent' (0) or 'present' (1). For continuous characters, it involves discrete classes. For example, exchangeable sodium percentage (ESP) or Character 77 (C.77), indicating non-sodic (<6), sodic (6-15%), moderately sodic (15-25%) and extremely sodic (>25%), are designated character-states of C.77:0 (absence) and C.77: 1, C.77:2, and C.77:3 (presence), respectively. The characters, character-states and coding are listed in Table 3.
The exponential growth in potential soil taxa with each additional tier makes it impractical to include all tiers. This exclusion is supported by the reasoning that important characters are located in the upper tiers (Isbell 2002). For these reasons, the cladistic analysis is limited to the Order, Suborder and Great Group tiers. The analysis was also undertaken with the removal of the Anthroposols, because their characters are unnaturally formed. The base cladogram (BC or cladogram 1) includes all taxa from 13 Soil Orders, with the 113 characters described in Table 3.
Measures used within cladistic analysis: consistency index, retention index and confidence levels
When interpreting cladistic analysis, all cladograms are evaluated in terms of a consistency index (CI), retention index (RI) and confidence levels. CI and RI are measures aimed at determining how well data fit a cladogram (Rexova et al. 2003). CI identifies the amount of character-state conflict, demonstrating how well a character fits onto a cladogram. CI is calculated by the number of steps on the cladogram (M) divided by the actual number of steps that could be on the cladogram if perfectly resolved (S), multiplied by 100: CI = M/S x 100 (Kitching et al. 1998). A non-conflicting cladogram of three Soils Orders (Orders A, B and C) and one soil character (0 = absence, 1 = presence; Orders A = 0, B and C = 1) will recover the soil cladogram (A,(B,C)), resulting in one step, supported by character-state 1. The cladogram can be calculated as M/S x 100 = 1, meaning data fit the tree perfectly, without conflict. If, for instance, another soil order D and another soil character (0 = absence, 1 = presence; Orders A = 0, B, C and D= 1) are added, we would have enough data to form two steps; however, only character 1 forms a step, namely B, C and D form the soil cladogram (A,(B,C,D)). Therefore, M = 1, S = 2, CI = 0.5, Soil Orders A and D = 0, B and C = 1, would form two steps (CI = 1).
Retention index measures how many shared characterstates in a given cladogram are retained on the cladogram; that is, the RI measures how well the character-states explain the cladogram. It is defined as (G - S)/(G - M), where G represents the greatest number of times a character-state can occur (Kitching et al. 1998). For example, if the maximum number of times a character-state can occur is three (G = 3), and the minimum number of occurrences is one (M= 1) and it occurs twice (S = 2), this results in (3 - 2)/(3 - 1) = 0.5. As the example illustrates, RI reflects the characters occurrence as a shared character-state between soil taxa, with less regard for character conflict. The closer RI is to 1, the more occurrences of shared character-states, which represents a better fitting cladogram.
All cladograms are subjected to bootstrapping, which enables the determination of confidence levels of the Soil Orders. The confidence levels of Soil Orders is determined by randomly removing characters and re-running the analysis to see if the Soil Orders still group together. These characters are returned to the analysis and new characters are removed and the analysis re-run (Siebert 1992). This removal of characters occurs 100 times, meaning the confidence level is on a scale of 0-100. A high confidence level indicates inclusive groups (clades) that are supported by multiple characters. The higher the confidence levels, the more confident one can be that the Soil Order identified is natural.
Coding of matrices: base and ASC tiers
The list of characters shown in Table 3 represents those used in the BC or cladogram 1. Table 4 summarises the cladogram names, number of possible taxa and characters of the other matrices. To determine the effect of removing morphological and nonmorphological characters to distinguish the Soil Order tier, cladogram 2 includes characters from Suborder and Great Group tiers only. Cladogram 3 is aimed at assessing the influence of characters in the Suborder tier, whereas cladogram 4 considers the effect of removing all characters from the Great Group tier.
Coding of matrices: continuous, non-morphological and soil colour characters
Cladistics works most effectively when either absence-presence and/or other qualitative characters are used (e.g. C.52, clear and abrupt textual B horizon). Continuous characters such ESP are problematic because the scale values for various states are arbitrary (Pimentel and Riggins 1987) and because they result in polymorphisms, which are poorly resolved nodes at ends of trees causing conflict in the analysis (Nixon and Davis 1991). To assess the effect of polymorphism, all continuous characters are removed in cladogram 5.
Cladistic analysis relies upon the use of characters, that is to say, intrinsic features that are observable in a taxon (Kitching et al. 1998), for example, C.52 (clear and abrupt textual B horizon). Applying this cladistic requirement of applying only morphological characters to the ASC means that characters such as C.39 (profile saturated for at least 2-3 months) are not used. Non-morphological characters describe the environmental conditions under which taxa are found, rather than using morphological properties. To identify the effect of nonmorphological characters, we removed them in cladogram 6.
The use of soil colour to identify relationships between taxa is problematic for two reasons. The first is the arbitrary nature of how colour is determined. The colour chart used is determined by hue, so if it is 5 YR or redder there is a choice of three colours, whereas if the hue is yellower there is a choice of four colours (Isbell 2002). Second, the same colours appear multiple times in Soil Orders (i.e. Vertosols, Kurosols, Sodosols, Chromosols, Ferrosols, Dcrmosols, Kandosols), namely red, brown yellow, grey and black. The frequent use of colour makes it a conflicting character. To establish its impact, it is removed in cladogram 7. Cladogram 8 is aimed at determining the combined effect of removing continuous, non-morphological and colour characters.
Coding of matrices: Soil Orders
To determine the naturalness of each Soil Order, and the impact each has on the others, we excluded each in turn. Table 4 summarises the 13 matrices developed, which include cladograms 9-21. After the initial BC analysis, a final cladogram was created; cladogram 22 saw the removal of the seven Soil Orders that registered no confidence level on the BC analysis.
Cladistic analysis of matrices
In all 22 matrices, an out-group is created, rooting the cladogram. To achieve this, an extra soil taxon (OG) is added, which comprises all Os. The 22 matrices are run through a parsimony algorithm using the cladistics program TNT version 1.0 (Goloboff et al. 2000). The analysis is run using the ratchet setting within the New Technology Search. For all analyses, the ratchet is set at 1000. For each analysis, a combined consensus cladogram (including cladogram length), CI, RI and confidence levels (bootstrapping) are generated.
Description of base cladogram
A well-resolved, natural classification would be discernible in a cladogram if clades of each taxon exist on a radiating node away from a common node (see Fig. 1 b). This is not the case in Fig. 2, which shows the BC (i.e. cladogram 1) that considers all three tiers and morphological and non-morphological characters. This is because most Soil Orders emanate from one mother node. This means that the BC has poor structure and few relationships can be inferred between the different Soil Orders. This lack of structure indicates that the ASC is not a natural classification.
The result does show that the Organosols, Podosols, Vertosols, Calcarosols, Kandosols, Rudosols and Ferrosols are identified as separate clades, and indicates that the tiers and morphological and non-morphological characters used to identify these Soil Orders are equally effective in a cladistic analysis. Although not completely identified as a separate cladc, some Tenosols are resolved, with the majority grouping into a clade based on C.58 :1, C.58 :2, C.58 :3 (i.e. A1 horizon overlies calcrete pan, or partially weathered rock, or unconsolidated materials, respectively).
The remaining Soil Orders, Hydrosols, Dermosols, Sodosols, Chromosols and Kurosols, do not form separate clades. The reason is that their defining characters occur multiple times. The lack of grouping means that the characters used to identify these Soil Orders are not as exclusive as those used by Organosols, Podosols and Vertosols.
One of the most problematic is the Hydrosols, because it groups out into five different clades (e.g. Hydrosols 1, 2, etc.) with Kurosols, Sodosols (2), Chromosols (2), Kandosols, Rudosols and Tenosols (2, 3) (Fig. 2). These groupings result from the use of several morphological and non-morphological characters in different tiers. For example, the morphological characters C.52 (clear textural B horizons) and C.53 (acidity of the major part of upper 0.2 m of B2) used to identify Kurosols, Sodosols and Chromosols are also used within the Great Group tier to identify Hydrosol taxa. A thorough revision of each of the Soil Orders to accommodate the Hydrosols will result in a more natural classification.
The grouping of the Sodosols (1), Chromosols (1) and Dermosols (2) is due to the shared Great Group characters C.71 : 1 (upper 0.2 m of the B2 horizon has a strong blocky or polyhedral structure), C.72 : 1 (very weak adhesion between peds), C.73 : 1 (usually high salt contents resulting in weak dry strength), and C.74: 1 (some B2 horizons may be weakly subplastic). The fact that four characters group them together adds considerable strength to their relationship, allowing cladistics to isolate them from other soil taxa and group them into a single clade. This indicates the need to use multiple characters when defining taxa, and not to rely on one.
Base cladogram: confidence levels, consistency index and retention index
We undertook a bootstrapping of the BC in order to give a level of confidence for the 13 Soil Orders. Table 5 shows Organosols (64), Podosols (58), Vertosols (61), Calcarosols (29), Kandosols (2) and Rudosols (33) all registered confidence levels. The remaining seven Soil Orders had far lower values. Given the significantly higher confidence levels (>50) of the Organosols, Podosols and Vertosols, it can be concluded these Soil Orders are the only natural groups.
The greater confidence levels are based on their numerous unique characters. Specifically, Organosols only contain three characters that are found in other Soil Orders (i.e. characters 4, sulfuric materials occur in upper 1.5 m; 5, sulfidic materials in upper 1.5 m; 6, at least some part of the B horizon is calcareous). The Podosols only have one shared character C.17 (long-term saturation in B horizon), whereas the Vertosols share character 17 with Podosols, as well as sharing the five colour characters with seven other Soil Orders (e.g. Kurosols and Sodosols). The remaining characters in the Organosols, Podosols and Vertosols are unique to their respective Soil Order.
Values of CI and RI are calculated for the BC, to determine the occurrence of character-state conflict and shared character-state forms, respectively. Table 6 shows that the BC generated a low CI (0.196) and relatively high RI (0.753). The high RI indicates characters used to identify relationships between different soil taxa. However, the reoccurrence of the same character-states (i.e. low CI) is indicative of the artificial nature of the ASC.
The Soil Orders that registered low or no confidence, such as the Chromosols, Ferrosols and Dermosols, have highly conflicting characters in the Suborder and Great Group tiers. These Soil Orders rely exclusively on one or two characters located in the Order tier to group them into a single clade. For example, the Ferrosols rely exclusively on the sole characterstate C.94 : 1 (major part of B2 horizon has a free iron content >5% in fine earth). The removal of this character means that Ferrosols no longer group together with confidence. This is due to the remaining characters in the Suborders and Great Groups being found in multiple Soil Orders. The example of the Ferrosols illustrates the need for Soil Orders to be defined by multiple unique character-states in all tiers of the classification, not just the upper most tier.
Removal of ASC tier characters
The removal of the 26 characters that define the Soil Order tier and from the BC (cladogram 2) resulted in an increase in both the CI (0.282) and R1 (0.763). The increase indicates less characterstate conflict and more shared character-states, resulting in a better fitting classification. However, the removal of the Soil Order tier characters means none of the 13 Soil Orders form clades with any confidence, with all Soil Orders registering 0 (Table 5), revealing that the 13 Soil Orders rely exclusively on the characters in the Order tier in order to group.
Without Order-tier characters, the taxa grouped out into smaller clades in cladogram 2 (not shown) and based on their Suborder or Great Group characters. Of the multiple clades, only five scored confidence levels >50, including the grouping of all 13 Podosol Suborder and Great Group taxa into three groups based on their Suborder characters: Aerie Podosols (91), Semiaquic Podosols (79) and Aquic Podosols (54).
There was also a grouping of five Sodosols, five Chromosols and five Dermosols with a confidence level of 67 and based on the Great Group C.71 : 1 (upper 0.2 m of the B2 horizon has a strong blocky structure), C.72 : 1 (very weak adhesion), C.73 : 1 (usually high salt contents) and 74:1 (B2 horizons may be weakly subplastic). It is also worth noting two taxa from the Calcarosols and Rudosols grouped together with a confidence level of 63 based on C.85: 1 (dominantly consist of gypsum crystals). Although these results are unsurprising, the groupings described show the strength of various Suborder and Great Group characters and identify them as useful characters that should be retained.
Removal of the Suborder characters (i.e. cladogram 3) leads to a higher CI (0.373) than for the BC. This reduction in character-state conflict is to be expected because seven of the Orders (i.e. Vertosols, Kurosols, Sodosols, Chromosols, Ferrosols, Dermosols, Kandosols) use the same five colour characters, with their removal reducing their reoccurrence, and improving the fit. The RI (0.680) is also lower than for the BC. This indicates that these Suborder characters are identifying relationships between soil taxa. An example of a valuable character that was removed is multi-state character 3 (75% volume of organic material is peat; Fibric (1), Hemic (2), Saparic (3)), because its removal resulted in a decrease in confidence level of the Organosols (53) compared with the BC.
The removal of the Suborder characters saw significant increases in the confidence levels of the Podosols (89), Vertosols (93) and Kandosols (20), however. The removal of the Subgroup characters also resulted in the Ferrosols (19) registering a confidence level. The increase in the confidence levels for the Vertosols and Kandosols and the emergence of the Ferrosols is due to the use of colour, and because this is removed, there is reduced character-state conflict. This allows Soil Orders to group out with greater confidence.
Table 6 shows that the removal of the Great Group characters had the greatest impact on the CI relative to BC. This is because cladogram 4 (Fig. 3) had a CI of 0.602, which is more than triple that of the BC (0.196). This large increase in CI highlights the high level of character-state conflict that the Great Groups characters contribute when all tiers and all morphological and non-morphological characters are considered (i.e. cladogram 1). This is because ~50% of the characters found within the Great Group tier occur within 2-8 different Soil Orders. For example, C.69: 1 (red-brown hardpan either in or directly underlying B horizon) and 70:1 (B horizon that is not calcareous and overlies a calcrete pan) occur multiple times in the Sodosols, Chromosols, Calcarosols, Dermosols, Kandosols, Rudosols and Tenosols. This indicates that characters located in the Great Group tier require the greatest amount of revision, with the aim to increase significantly the number of unique characters used in this tier.
The removal of the highly conflicting Great Group characters also results in significantly higher confidence levels. The Organosols (89), Calcarosols (72), Kandosols (37) and Rudosols (56) all increase by at least 22 relative to the BC. There is also the emergence of the Hydrosols (57), Ferrosols (5) and Dermosols (9). The increase in confidence levels for the majority of the Soil Orders demonstrates the benefits and need to revise the characters used in the Great Groups.
The only Soil Orders to register lower confidence levels from the removal of the Great Group characters are Podosols (42) and Vertosols (39). This reduction may be due to the high percentage of unique characters in the Great Group tier. For the Podosols, this includes characters 20 (with only a Bs horizon) and 21 (with only a Bhs horizon), and for the Vertosols, characters 33 (surface that is moderately self-mulching) and 37 (massive or weakly structured surface).
Removal of characters types: continuous, non-morphological and soil colour
The removal of the four continuous characters (i.e. cladogram 5), including C.77 (ESP in upper 0.2 m of B2), C.53 (acidity in upper 0.2 m of B2), C.68 (base status of upper B2 horizon) and C.81 (soil with hard calcrete fragment), resulted in a slightly higher CI (0.288) than for the BC. This result is expected and consistent with the removal of polymorphism, which decreases character-state conflict. The removal of the four continuous characters also results in a higher RI (0.780). This means that the characters remaining identify more relationships between soil taxa. However, whereas both CI and RI are higher, the confidence levels remain the same (i.e. Organosols 64) or decrease. The only increase in confidence is for the Kandosols (12) with the Ferrosols (2) just emerging.
The removal of the 12 non-morphological characters (i.e. cladogram 6) produces a lower CI (0.189). By comparison, the RI (0.755) remains unchanged. The removal of the nonmorphological characters has a mixed effect on confidence levels of Soil Orders. Specifically, the Podosols (67) and Kandosols (19) increased, and the Organosols (53), Vertosols (51), Calcarosols (15) and Rudosols (21) decreased. There is no discernible reason because the majority of these Soil Orders have limited reliance on non-morphological characters.
Although not optimal or desirable in cladistics, it appears that the 12 non-morphological characters are beneficial to the ASC (Pimentel and Riggins 1987). In a revision, one would need to examine whether these non-morphological characters manifest themselves in some observable morphological properties. This could mean changing the non-morphological characters to morphological characters, marking a move away from characters based on environmental conditions towards characters using observable morphological properties.
The removal of the five colour characters (i.e. cladogram 7) produces a higher CI (0.313). This is to be expected because the five colours used occur multiple times and their removal gives rise to lower character conflict (i.e. RI = 0.724). These results indicate that retaining colour is beneficial in grouping soil; however, they need not be used so frequently. This conclusion is reinforced by the confidence levels of the Soil Orders, which changed little, except for the Vertosols (79) and Ferrosols (31), which saw the greatest increase in confidence levels. This increase mirrors their respective increases in confidence levels when the Suborder characters (including colour) are removed. The other Soil Order that was expected to increase significantly with removal of colour, but did not, was the Kandosols (8). The reason is unknown; removal of colour in cladograms 3 and 8 resulted in increases in Kandosol confidence levels.
The removal of continuous, non-morphological and colour characters (i.e. cladogram 8), unsurprisingly, leads to a higher CI (0.358) than removal of these characters independently; however, the RI (0.754) remained relatively unchanged. The removal saw a large increase in confidence levels of Vertosols (94) and Kandosols (25) and strong emergence of Ferrosols (29); however, the remaining Soil Orders showed no change. These results indicate that colour is more important than continuous and non-morphological characters and mean that revision of colour is crucial.
Removal of Soil Orders
The effects of the removal of entire Soil Orders on the CI are shown in Tabic 6. The removal of the Organosols (0.190), Podosols (0.175), Vertosols (0.192) and Rudosols (0.188) results in a slight decrease in the CI relative to the BC (0.196). This is because these four Soil Orders have a high level of unique characters and their removal increases characterstate conflict. This is consistent with the fact that these Soil Orders have high confidence in the BC.
The converse is true when the Kurosols (0.205), Sodosols (0.205), Ferrosols (0.202), Dermosols (0.222) and Tenosols (0.199) are removed, because the removal of their conflicting character-states results in an increase in CI. This is consistent with the lack of confidence levels that the five Soil Orders register in the BC. Removing them improves the classification. This is not the case with regard to removing Chromosols (0.173) and Hydrosols (0.168), which surprisingly leads to a poorer classification in spite of character-state conflict and their lack of confidence in the BC.
With respect to the RI, there was no clear observable pattern with the removal of individual Soil Orders relative to the BC. In this regard and with regard to low variability in the CIs and RIs generated from cladograms 9-21, these results indicate the limited effect that the removal of any single Soil Order had on the BC.
In order to determine what the effect might be of the removal of the seven Soil Orders that registered no confidence levels in the BC (i.e. Hydrosols, Kurosols, Sodosols, Chromosols, Ferrosols, Dermosols and Tenosols), cladogram 22 shows that CI (0.360) and RI (0.809) are substantially higher. This indicates a better fitting classification. Their removal also results in greater confidence levels for the remainder: Organosols (86), Podosols (61), Vertosols (70), Calcarosols (63), Kandosols (50) and Rudosols (85). The removal of the seven Soil Orders created a cladogram with less character conflict, more shared characterstates between soil taxa, and more confident grouping of the remaining six Soil Orders.
Potential revision of types of characters
The removal of continuous (cladogram 5, CI 0.288), nonmorphological (cladogram 6, 0.189), and colour (cladogram 7, 0.313) characters indicates that one option to improve the fit of the ASC could be achieved in the revision of the continuous and colour characters. This is because their removal not only increased the fit of the classification (higher CI and RI) but also saw a greater number of Soil Orders generate higher confidence levels. However, the usefulness of colour in determining chemical and physical properties such as element content (i.e. iron), water saturation and content of organic material is well established (Davey et al. 1975).
The usefulness of colour has resulted in its employment in most soil-classification systems, including those of the USA (Soil Survey Staff 2010), New Zealand (Hewitt 1992) and Great Britain (Avery 1973). Therefore, complete removal of colour characters appears problematic given its diagnostic value. Nevertheless, the appropriateness of using all five colours in all seven Soil Orders could be reassessed. For example, the Ferrosols use all five of the soil colours in Suborders, although most Ferrosols are either red or brown (Isbell 2002). Removing the other colours will reduce the number of taxa and reduce the amount of character-state conflict.
Potential revision of ASC tiers
The results obtained from the cladograms where the characters Orders (cladogram 2, CI 0.282), Suborders (cladogram 3, CI 0.373) and Great Groups (cladogram 4, CI 0.602) are removed indicated higher levels of character-state conflict in Suborders and Great Groups, and this represents an opportunity to revise the characters in these tiers. In this regard, greater use of characters based on properties arising from parent material, as is the case in the US Soil Taxonomy (Bockheim et al. 2014), might be appropriate. This is because the use of characters (i.e. homology) creates an initial starting point for divergence of soil taxa (Arnold and Eswaran 2010), and in the case of the ASC, different Soil Orders. This approach was also suggested by Coffey (1912) and Brevik (1999), who argued that characters based on parent material increase the chance of identifying traceable linages.
Potential revision of Soil Orders
The poor resolution in the BC (Fig. 2) highlights the inability of the characters and structure used in the ASC to identify relationships between Soil Orders. The cladistic analysis did identify some natural Soil Orders. Organosols, which key out first, are rare in Australia because the continent is dominated by semi-arid and arid climatic regimes. As such, this Soil Order is confined to very cold and wet areas near Mt Kosciuszko and in north-western Tasmania.
Although the Podosols are found in similarly high-rainfall areas, owing to the hotter coastal climates and highly siliceous parent material, the organic matter is broken down quickly, creating chelates with iron, which leads to concentration of organic matter (i.e. 9:1 Bs horizon) and iron oxides (i.e. 10:1 Bh horizon) into B horizons. Conversely, Vertosols are very clayey (>35%) and found in semi-arid areas where 2:1 clay minerals (e.g. smectite) form, enabling shrinkage and swelling and distinct morphological characters (i.e. C.26:1, slickensides of lenticular peds).
This identification of natural Soil Orders in the ASC marks the first step to understanding the relationships between the Soil Orders. Interestingly, Isbell (2002) accurately captured soil at either end of the spectrum of parent material and rainfall (Fig. la), with the Podosols representing extremely siliceous soil with high average rainfall, and the Vertosols representing mafic soils with low average rainfall (Gray and Murphy 2002).
The remaining Soil Orders, with the exception of Organosols, Hydrosols and Rudosols, fall between these. The issue here is in the differentiation of these remaining Soil Orders. The problem is that the Chromosols, Ferrosols and Dermosols have no unique characters defining the soil taxa in their Suborders and Great Group tier. This lack of differentiating characters makes it impossible to discern the different Soil Orders and in turn the relationships between them. In any review, unique characters need to be used to distinguish these and the other Soil Orders in terms of their genesis.
The ASC (Isbell 2002) is an artificial key, the preferred model for classifications of soil around the world. This is because they can be used by a wide range of people. The issue considered here is not whether the ASC is a bad key, rather, whether it is a natural classification that can reflect soils genesis and thereby provide a set of unique character-states for each soil relationship. In natural classifications, the emphasis is on shared derived characteristics (soil-genesis relationships), not unique describers for each taxon (identification). The cladistic method reported here is a first attempt at testing the naturalness of the ASC.
From the cladistic analysis of the ASC, the following conclusions can be made. The ASC is not a natural classification; however, the Organosols, Podosols and Vertosols are natural Soil Orders. It could also be determined that because of the highly conflicting character-states in the Suborder and Great Group tiers, there is low or no confidence in soil Orders such as the Chromosols, Ferrosols and Dermosols because they rely exclusively on one to two characters located in the Order tier to group them into a single clade. A revised ASC based on relationships derived from soil genesis will provide pedology with a natural classification, one that would require little revision over time. Moreover, the revision of the ASC is also the first step in testing other soil classification systems around the world.
In terms of seeking areas where improvements in the ASC can be obtained, the revision of the Great Group tier characters is ideal. In identifying poor characters, we found that removal of continuous characters and soil colour is an area of promise in terms of revising the ASC. Improvements might also be made if soil colour is used more parsimoniously in seven of the 14 Soil Orders, because this is shown to achieve a reduction in the number of potential taxa and reduced character-state conflict.
We thank David Edwards, John S. Wilkins and two anonymous reviewers for their critical and constructive comments, which have materially improved the readability and scientific merit of the manuscript.
Arnold RW, Eswaran H (2010) Conceptual basis for soil classification: Lessons from the past. In 'Soil classification: A global desk reference'. (Eds H Eswaran, R Ahrens, TJ Rice, BA Stewart) pp. 27-42. (CRC Press: Boca Raton, FL, USA)
Avery BW (1973) Soil classification in the soil survey of England and Wales. Journal of Soil Science 24, 324-338. doi:10.1111/j. 1365-2389.1973. tb00769.x
Basinski JJ (1959) The Russian approach to soil classification and its recent development. Journal of Soil Science 10, 14-26. doi: 10.1111/j.13652389.1959.tb00662.x
Bockheim JG, Gennadiyev AN, Hartemink AE, Brevik EC (2014) Soilforming factors and Soil Taxonomy. Geoderma 226-227, 231-237. doi: 10.1016/J.geoderma.2014.02.016
Brevik EC (1999) The soil science--geology connection. NDGS Newsletter Vol. 26, pp. 1-4.
Clark JM, Norell MA, Makovicky PJ (2002) Cladistic approaches to the relationships of birds to other theropod dinosaurs. In 'Mesozoic birds: Above the heads of dinosaurs'. (Eds LM Chiappe, LM Witmer) pp. 31-60. (University of California Press Ltd: Oakland, CA, USA)
Coffey GN (1912) A study of the soils of the United States. Bureau of Soils Bulletin No. 85, pp. 23-38. US Department of Agriculture, Washington, DC.
Davey BG, Russell JD, Wilson MJ (1975) Iron oxide and clay minerals and their relation to colours of red and yellow podzolic soils near Sydney, Australia. Geoderma 14, 125-138. doi: 10.1016/0016-7061 (75)90071-3
Echeverry A, Silva-Romo G, Morrone JJ (2012)Tectonostratigraphicterrane relationships: A glimpse into the Caribbean under a cladistic approach. Palaeogeography, Palaeoclimatology, Palaeoecology 353-355, 87-92. doi: 10.1016/j.palaeo.2012.07.007
Fitzpatrick RW, Powell B, McKenzie NJ, Maschmedt DJ, Schoknecht N, Jacquier DW (2010) Demands of soil classification in Australia. In 'Soil classification: A global desk reference'. (Eds H Eswaran, R Ahrens, TJ Rice, BA Stewart) pp. 77 100. (CRC Press: Boca Raton, FL, USA)
Fraix-Bumet D, Choler P, Douzery EJ, Verhamme A (2006) Astrocladistics: a phylogenetic analysis of galaxy evolution I. Character evolutions and galaxy histories. Journal of Classification 23, 31-56. doi : 10.1007/s00357-006-0003-5
Goloboff P, Farris S, Nixon K (2000) TNT (Tree analysis using New Technology) Version 1. Cladistics.com TNT. Available at: www. cladistics.com/aboutTNT.html
Gray JM, Murphy BW (2002) Parent material and soil distribution. Natural Resource Management 5, 2-12.
Hennig W (1966) 'Phylogenetic systematics.' (University of Illinois Press: Urbana, IL, USA)
Hewitt AE (1992) New Zealand Soil Classification. Scientific Report No. 19. DSIR Land Resources, Lower Hutt, New Zealand.
Isbell R (1992) A brief history of national soil classification in Australia since the 1920s. Australian Journal of Soil Research 30, 825-842. doi: 10.1071 /SR9920825
Isbell R (2002) 'The Australian Soil Classification.' Revised edn (CSIRO Publishing: Melbourne)
Kitching I, Forey P, Humphries C, Williams D (1998) 'Cladistics: The theory and practice of parsimony analysis.' (Oxford University Press: New York)
Marbut CF (1951) 'Soil classification: Life and work of C.F. Marbut.' (Artcraft Press: Columbia, MO, USA)
Mazaheri SA, Koppi AJ, McBratney AB (1995) A fuzzy allocation scheme for the Australian Great Soil Groups Classification system. European Journal of Soil Science 46, 601-612. doi: 10.1111/j. 1365-2389.1995. tb01356.x
McKenzie NJ, Austin MP (1989) Utility of the factual key and soil taxonomy in the lower Macquarie Valley, NSW. Australian Journal of Soil Research 27, 289-311. doi:10.1071/SR9890289
Nixon KC, Davis JI (1991) Polymorphic taxa, missing values and Cladistic analysis. Cladistics 7,233-241. doi: 10.1111/j. 1096-0031.1991,tb00036.x
Northcote KH (1971) 'A factual key for the recognition of Australian soils.' (Rellim Technical Publications: Glenside, S. Aust.)
Northcote KH (1984) 'Soils, soil morphology and soil classification: training course lectures.' (Rellim Technical Publications: Glenside, S. Aust.)
Pimentel RA, Riggins R (1987) The nature of cladistic data. Cladistics 3, 201-209. doi: 10.1111 /j. 1096-0031.1987,tb00508.x
Rexova K, Frynta D, Zrzavy J (2003) Cladistic analysis of languages: IndoEuropean classification based on lexicostatistical data. Cladistics 19, 120-127.
Schuh RT, Brower AVZ (2010) 'Biological systematics: principles and applications.' 2nd edn (Cornell University Press: New York)
Siebert DJ (1992) Tree statistics; trees and 'confidence'; consensus trees; alternatives to parsimony; character weighting; character conflict and its resolutions. In 'Cladistics: A practical course in systematics'. (Eds PL Forey, CJ Humphries, IJ Kitching, RW Scotland, DJ Siebert, DM Williams) pp. 72-88. (Oxford University Press: New York)
Soil Survey Staff (2010) 'Keys to Soil Taxonomy.' 11th edn (USDA Natural Resources Conservation Service: Washington, DC)
Stace HCT (1968) 'A handbook of Australian soils.' (CSIRO Publishing: Melbourne)
Wilkins JS, Ebach MC (2014) 'The nature of classification: relationships and kinds in the natural sciences.' (Palgrave Macmillan: New York)
Gregory P. L. Miltenyi (A), Malte C. Ebach (A), and John Triantafilis (A,B)
(A) School of Biological, Earth and Environmental Sciences, UNSW Australia, Kensington, NSW 2052, Australia.
(B) Corresponding author. Email: firstname.lastname@example.org
Table 1. Cladistic terms proposed by Wilkins and Ebach (2014) Term Cladistic term Definition Character- Synapomorphy A shared derived character state between at least two taxa indicating the presence of a shared common ancestor Formism Monophyly; The grouping of taxa that are (formology) natural group more closely related to each other than to any other taxon (indicating a relationship between them) Aformism Non-monophyly The grouping of taxa that are not more closely related to each other than to any other taxon Radogram Cladogram; A visual representation of the branching relationships between the taxa diagram; analysis. It is the final Consensus result from the cladistic tree; Tree analysis Rade Clade A group of taxa defined by a shared character Character-state Homoplasy The reoccurrence of the same conflict character-state in multiple places throughout the radogram Table 2. Australian Soil Orders and their defining characters (Isbell 2002) Soil Order Ranking Defining characters No. of in Soil soil taxa Key in Soil Order to 3rd tier Organosols OR 2 >0.4 m of organic 18 material within the upper 0.8 m Podosols PO 3 Bs horizon (iron 13 compounds); Bhs horizon (organic-aluminium and iron compounds); Bh horizon (organic-aluminium compounds) Vertosols VE 4 Clay field texture or 24 [greater than or equal to] 35% clay through the solum; when dry, open cracks [greater than or equal to] 5 mm occur; slickenside and/or lenticular peds occur Hydrosols HY 5 Saturated for at least 64 2-3 months in most years Kurosols KU 6 Clear and abrupt textural 35 B horizon; B2 horizon strongly acid Sodosols SO 7 Clear and abrupt textural 55 B horizon; B2 horizon sodic and not strongly acid Chromosols CH 8 Clear and abrupt textual 70 B horizon; B2 horizon not sodic and not strongly acid Calcarosols CA 9 Calcareous through the 43 solum; carbonate accumulation pedogenic Ferrosols FE 10 B2 horizon has a free 25 iron oxide content >5% Fe Dermosols DE 11 Major part of B2 horizon 70 structure more developed than weak Kandosols KA 12 Maximum clay content in 64 B2 horizon exceeds 15%; B2 horizon massive or only weak structure or grade Rudosols RU 13 Minimal developed A1 21 horizon; presence of <10% of B horizon material in fissures in the parent rock; A1 horizon apedal or only weakly structured; little to no textural or colour change in the solum Tenosols TE 14 Other soils 54 Total = 556 Table 3. List of the soil characters used in the base cladogram Order characters, bold; Suborder characters, italic, Great Group characters, plain Cladogram Character Characters used in the Character- base cladogram states 2 0 >0.4 m of organic material Absent (0), within the upper 0.8 m present (1) (0.4 can be cumulatively taken)# 2 1 Organic materials Absent (0), extending from the surface present (1) to a min. depth of 0.1 m that directly overlies rock or other hard surfaces (e.g. gravel, humus)# 3 2 75% of the volume of Absent (0), organic material is peat@ Fibric (1), Hemic (2), Saparic (3) 4/6/8 3 More or less freely Absent (0), drained and never present (1) saturated for >7 days, and contain organic materials that occur as in definition (ii) for Organosols (extending from the surface to min. depth of 0.1 m) 4 4 Sulfuric materials within Absent (0), the upper 1.5 m of the present (1) profile 4 5 Sulfidic materials within Absent (0), the upper 1.5 m of the present (1) profile 4 6 At least some part of the Absent (0), B or the BC horizon present (1) calcareous 4 7 Major part of the organic Absent (0), material not calcareous present (1) but not strongly acid 4 8 Major part of the organic Absent (0), material strongly acid present (1) 2 9 Bs horizon: visible Absent (0), dominance of iron present (1) compounds# 2 10 Bh horizon: organic- Absent (0), aluminium compounds# present (1) 3/6/8 11 Free drainage, no Absent (0), restriction to through- present (1) drainage in the B horizon or whole substrate@ 3 12 B horizon weakly coherent Absent (0), and porous@ present (1) 3 13 Often boringly coloured Absent (0), and lack evidence of present (1) seasonal reduction@ 3/6/8 14 Short-term saturation in Absent (0), the B horizon@ present (1) 3/6/8 15 Saturation not sufficient Absent (0), to reduce or remove present (1) significant amounts of the accumulated iron@ 3 16 May be great accumulation Absent (0), of organic compounds and present (1) less iron in the zone of maximum saturation@ 3/6/8 17 Long-term saturation in Absent (0), the B horizon@ present (1) 3/6/8 18 Saturation sufficient to Absent (0), reduce most iron compounds present (1) and move them out of the B horizon; hence, Bh horizons are usually prominent@ 4 19 With pipey B Horizon Absent (0), present (1) 4 20 With only a Bs Horizon Absent (0), present (1) 4 21 With only a Bhs Horizon Absent (0), present (1) 4 22 With only a Bh horizon Absent (0), present (1) 4 23 With a Bh and Basi horizon Absent (0), present (1) 2 24 Clay field texture or Absent (0), [greater than or equal to] present (1) 35% clay throughout the solum except for thin, surface crusty horizons [less than or equal to] 0.03 m thick# 2 25 When dry, open cracks Absent (0), occur at some time in most present (1) years, these are >5 mm wide and extend upward to the surface or to the base of any plough layer, self- mulching horizon or thin, surface crusty horizon# 2 26 Slickensidc and or Absent (0), lenticular peds occur at present (1) some depth in the solum# 3 27 May be indicated by the Absent (0), presence of mottling and present (1) grey colours@ 3/7/8 28 Dominant colour class in Absent (0), the major part of the present (1) upper 0.5 m of the B2 Horizon is Red@ 3/7/8 29 Dominant colour class in Absent (0), the major part of the present (1) upper 0.5 m of the B2 Horizon is Brown@ 3/7/8 30 Dominant colour class in Absent (0), the major part of the present (1) upper 0.5 m of the B2 Horizon is Yellow@ 3/7/8 31 Dominant colour class in Absent (0), the major part of the present (1) upper 0.5 m of the B2 Horizon is Grey@ 3/7/8 32 Dominant colour class in Absent (0), the major part of the present (1) upper 0.5 m of the B2 Horizon is Black@ 4 33 Soils with a surface that Absent (0), is moderately to strongly present (1) self-mulching; when the soil is dry the self-mulching layer should be [greater than or equal to] 10 mm think 4 34 Initial drying may form a Absent (0), thin surface flake which present (1) readily disintegrates to a mulch on further drying 4 35 A pedal (stronger than Absent (0), weak grade, commonly present (1) blocky or polyhedral) A horizon that is either not or only weakly self-mulching, and there is no surface crusty horizon 4 36 After wetting and drying, Absent (0), may form a thin, 5-10-mm present (1) surface flake, which cracks into irregular polygons 0.03-0.1 m diameter 4 37 Massive or weakly Absent (0), structured surface crusty present (1) horizon [less than or equal to] 0.03 m thick, often of lighter texture (lower clay content) than underlying pedal clay (blocky or polyhedral), which is not self-mulching 4 38 Massive or weak blocky Absent (0), (usually >0.05 m peds) A present (1) horizon and no surface crusty horizon 2/6/8 39 Profile saturated for at Absent (0), least 2-3 months in most present (1) years# 3/6/8 40 Inundated by saline tidal Absent (0), water@ infrequent water table (1), infrequent storms (2), frequent (3) 3/6/8 41 Fresh water inundation@ Absent (0), uncommon (1), common (2) 3/6/8 42 Soils of the saline playa Absent (0), lakes@ present (1) 3/6/8 43 Saline groundwater table Absent (0), (salinised soils)@ present (1) 3/6/8 44 Other soils with a Absent (0), seasonal or permanent mottled (1), water table@ whole coloured (2) 4 45 Dominated by organic Absent (0), material to a depth of 0.5 present (1) m 2 46 Calcareous through the Absent (0), solum# present (1) 4 47 Dominant sediment size to Not depth of 0.5 m described (0), clay (1), sand (2), silt (3) 4 48 A gypsic horizon occurs Absent (0), within the upper 0.5 m of present (1) the profile 4 49 Colour of major part of Absent (0), the upper 0.5 m of the mottled (1), profile whole colour (2) 4 50 Major part of the upper Absent (0), 0.5 m of the profile present (1) consists of materials dominated (>50%) by halite crystals 4 51 B horizon containing or Absent (0), directly underlain by present (1) ferricrete, a petroferric horizon or a petroreticulite horizon 2 52 Clear or abrupt textural B Absent (0), horizons# present (1) 2/5/8 53 Acidity of the major part Not of the upper 0.2 m of the described B2 horizon (or the major (0), strong part of the entire B2 acid (1), horizon if <0.2 m thick)# not sodic and not strong acid (2), sodic and not strong acid (3) 2 54 B2 horizon with structure Absent (0), more developed than weak present (1) throughout the major part of the horizon# 2 55 Major part of the B2 Absent (0), horizon massive or has present (1) only weak grade of structure# 2 56 Max. clay content in some Absent (0), parts of the B2 horizon present (1) >15% (c.g. heavy sand loams)# 2 57 A peaty horizon# Absent (0), present (1) 2 58 At or A2 horizon directly Absent (0), overlies a calcrete pan, A1 horizon hard unweathered rock or (1), A2 other hard material; or horizon (2), partially weathered or A humose, decomposed rock or melacic, or saprolite; or melanic unconsolidated mineral horizon, or materials# conspicuously bleached A2 horizon (3) 2 59 A tenic B horizon or a B2 Absent (0), horizon with [less than or present (1) equal to] 15% clay (SL), or a transitional horizon (ClB) occurring in fissures in the parent rock or saprolite, which contains 10-50% of the B horizon material (including pedogenic carbonate)# 2 60 A ferric or bauxitic Absent (0), horizon >0.2 m thick# present (1) 2 61 A calcareous horizon >0.2 Absent (0), m thick# present (1) 2 62 Minimal development of an Absent (0), At horizon# present (1) 2 63 Presence of [less than or Absent (0), equal to] 10% of B horizon present (1) material (including pedogenic carbonate) in fissures in the parent rock or saprolite# 2 64 Soils are apedal or only Absent (0), weakly structured in the present (1) At horizon and show no pedological colour changes apart from the darkening of an A1 horizon# 2 65 Texture or colour change Absent (0), with depth unless present (1) stratified or buried soils are present# 4 66 Exchangeable Ca/Mg ratio Absent (0), <0.1 in the major part of present (1) the B2 horizon 4 67 Major part of the upper Absent (0), 0.2 m of the B2 horizon is present (1) sodic 4/5/8 68 Base status in the major Absent (0), part of the B2 horizon dystrophic (1), mesotrophic (2), eutrophic (3) 4 69 Red-brown hardpan either Absent (0), within or directly present (1) underlying the B horizon 4 70 B horizon that is not Absent (0), calcareous and which present (1) directly overlies a calcrete pan 4 71 Upper 0.2 m of the B2 Absent (0), horizon (or the B2 horizon present (1) if it is <0.2 m thick) has a strong blocky or polyhedral structure in which average ped size is usually in the range 5-20 mm 4 72 Very weak adhesion between Absent (0), peds (when dry it is very present (1) easy to insert a spade into the upper B2 horizon) 4 73 Usually high salt contents Absent (0), resulting in weak dry present (1) strength and a bulk density of1.3 t [m.sup.-3] or less 4 74 In some soils the B2 Absent (0), horizons may be weakly present (1) subplastic 4 75 Fine earth effervescence Absent (0), (1 m HCI) throughout the present (1) solum 4 76 Major part of the upper Absent (0), 0.2 m of the B2 horizon is present (1) mottled 4/5/8 77 ESP in the major part of Absent (0), the upper 0.2 m of the B2 6-15 (1), horizon 15-25 (2), >25 (3) 2 78 Soils with strongly Absent (0), subplastic upper B2 present (1) horizons# 4 79 Major part of the B2 Absent (0), Horizon is strongly present (1) subplastic 4 80 Soils in which the Absent (0), carbonate is evident only present (1) as a slight to moderate effervescence (1 m HCI) and/or contain <2% soft, finely divided carbonate 4/5/8 81 Soils with a calcareous Absent (0), horizon containing hard 0-20% (1), calcrete fragments and/or 20-50% (2), carbonate nodules or >50% (3) concretions and/or carbonate coated gravel 4 82 Soils with a calcareous Absent (0), horizon containing >20% of present (1) mainly soft, finely divided carbonate 2 83 Carbonate accumulate must Absent (0), be judged to be pedogenic# present (1) 3 84 Soils in which the Absent (0), profile, with the possible present (1) exception of the A horizon, is calcareous, either loose or only weakly coherent both moist and dry, and consists dominantly of sand-size fragments of shells and other aquatic skeletons@ 3 85 Dominantly consist Absent (0), ofgypsum crystals, which present (1) are sand-sized or finer@ 4 86 Argic horizon within the B Absent (0), horizon present (1) 4 87 Major part of B horizon Absent (0), has a grade of structure present (1) that is stronger than weak 4 88 Overlie a hard siliceous Absent (0), pan present (1) 4 89 Directly overlie hard rock Absent (0), present (1) 4 90 Directly overlie partially Absent (0), weathered or decomposed present (1) rock or saprolite 4 91 Directly overlie marl Absent (0), present (1) 4 92 Directly overlie Absent (0), unconsolidated mineral present (1) material 4 93 Other soils with a Absent (0), calcareous horizon present (1) 2 94 Major part of B2 horizon Absent (0), has a free iron content present (1) >5% Fe in fine earth fraction# 4 95 Bulk density appears Absent (0), relativity low present (1) 4 96 Soil with a thick iron- Absent (0), pan occurring within or present (1) directly underlying the B horizon 4 97 Soils with massive to Absent (0), weakly structured (10 mm present (1) subangular blocky parting to finer granules) B horizons that are very porous with a weak consistence strength when moist 3 98 Soils that are highly Absent (0), saline@ present (1) 3 99 Frequently stratified but Absent (0), do not have permanent or present (1) seasonal water table and do not show any evidence, such as mottling, or episodic wetting by groundwater@ 3 100 At least the upper 0.5 m Absent (0), of the profile is not or present (1) only slightly gravelly (<10% >2 mm) throughout, either loose or only weakly coherent both moist and dry, and the texture is sandy (up to 10%)@ 3 101 Aeolian cross-bedding may Absent (0), be present but there is present (1) little if any evidence of other stratification or buried soils@ 3 102 At least the upper 0.5 m Absent (0), of the profile consists not gravel dominantly of (1), unconsolidated mineral distinct not material@ gravel (2), gravel (3) 4 103 Other soils that have Absent (0), formed in tephric present (1) materials that may be visibly stratified 4 104 Ferric horizon and which Absent (0), overlies ferricrete, present (1) petroreticulite or petroferric horizon 3 105 Soils which have a peaty, Absent (0), humose, melacic or melanic present (1) horizon@ 3 106 An unbleached A2 horizon Absent (0), may be present between the bleached (1) dark surface horizons and the substrate materials@ 3 107 Other soils with peaty, Absent (0), humose, melacic or melanic present (1) horizons@ 3 118 A conspicuously bleached A Absent (0), 2 horizon is present@ present (1) 3 109 Soils with a ferric or Absent (0), bauxitic horizon (nodules present (1) or concretions) [greater than or equal to] 0.2 m thick and occupies >50% of the solum depth; the solum depth excludes cemented layers@ 3 110 Soils with a calcareous Absent (0), horizon (consisting of present (1) >20% pedogenic carbonate) [greater than or equal to] 0.2 m thick@ 4 111 A tenic B horizon Absent (0), present (1) 4 112 Andie properties and have Absent (0), formed in basaltic tephric present (1) materials that may be visibly stratified Note: bold indicated with#. Note: italic indicated with#. Table 4. List of soil matrices used in the study Cladogram name No. No. of No. of soil taxa characters Base cladogram (BC) 1 556 113 BC without Orders characters 2 556 87 BC without Suborders characters 3 150 57 BC without Great Groups characters 4 72 37 BC without continuous characters 5 556 109 BC without non-morphological characters 6 556 103 BC without colour characters 7 287 108 BC without continuous, 8 287 94 non-morphological and colour characters BC without Organosols 9 538 107 BC without Podosols 10 543 99 BC without Vertosols 11 532 103 BC without Hydrosols 12 492 100 BC without Kurosols 13 521 112 BC without Sodosols 14 501 110 BC without Chromosols 15 486 112 BC without Calcarosols 16 513 109 BC without Ferrosols 17 531 112 BC without Dermosols 18 486 113 BC without Kandosols 19 492 110 BC without Rudosols 20 535 105 BC without Tenosols 21 502 101 BC without Hydrosols, Kurosols, 22 183 72 Sodosols, Chromosols, Ferroso], Dermosols, Tenosols Table 5. Confidence levels of the Soil Orders generated by the bootstrapping of different cladistics analyses with altered character BC, base cladogram Soil order BC Orders Suborders Great Continuous Groups Organosols 64 -- 53 89 64 Podosols 58 -- 89 42 57 Vertosols 61 -- 93 39 47 Hydrosols (HY) -- -- -- 57 -- Kurosols (KU) -- -- -- -- -- Sodosols (SO) -- -- -- -- -- Chromosols (CH) -- -- -- -- -- Calcarosols 29 -- 23 72 26 Ferrosols (FE) -- -- 19 5 2 Dermosols (DE) -- -- -- 9 -- Kandosols 2 -- 20 37 12 Rudosols 33 -- 27 56 28 Tenosols TE) -- -- -- -- Soil order BC without: Colour Continuous, HY, K.U, Non- non- SO, CH, morphological morphological FE, DE, TE and colour Orders Organosols 53 67 56 86 Podosols 67 51 65 61 Vertosols 51 79 94 70 Hydrosols (HY) -- -- -- -- Kurosols (KU) -- -- -- -- Sodosols (SO) -- -- -- -- Chromosols (CH) -- -- -- -- Calcarosols 15 17 23 63 Ferrosols (FE) 12 31 29 -- Dermosols (DE) -- -- - -- Kandosols 19 8 25 50 Rudosols 21 27 23 85 Tenosols TE) -- -- -- Table 6. Cladogram lengths and calculated consistency index (Cl) and retention index (Rl) of the cladogram, including the base cladogram, cladogram without Suborder and Great Groups characters and unsuitable characters Cladogram Radogram CI RI length 1 Base cladogram (BC) 680 0.196 0.753 2 BC without Order characters 369 0.282 0.763 3 BC without Suborder characters 185 0.373 0.680 4 BC without Great Group characters 83 0.602 0.781 5 BC without continuous characters 530 0.288 0.780 6 BC without non-morphological 619 0.189 0.755 characters 7 BC without colour characters 402 0.313 0.724 8 BC without continuous, non- 279 0.358 0.754 morphological and colour characters 9 BC without Organosols 657 0.190 0.752 10 BC without Podosols 679 0.175 0.740 11 BC without Vertosols 642 0.192 0.751 12 BC without Hydrosols 666 0.168 0.720 13 BC without Kurosols 645 0.205 0.752 14 BC without Sodosols 623 0.205 0.748 15 BC without Chromosols 765 0.173 0.656 16 BC without Calcarosols 628 0.205 0.754 17 BC without Ferrosols 652 0.202 0.757 18 BC without Dermosols 599 0.222 0.764 19 BC without Kandosols 599 0.217 0.760 20 BC without Rudosols 653 0.188 0.750 21 BC without Tenosols 607 0.199 0.762 22 BC without Hydrosols, Kurosols, 222 0.360 0.809 Sodosols, Chromosol, Ferrosols, Dermosols and Tenosols
|Printer friendly Cite/link Email Feedback|
|Author:||Miltenyi, Gregory P.L.; Ebach, Malte C.; Triantafilis, John|
|Date:||Oct 1, 2015|
|Previous Article:||Change in water extractable organic carbon and microbial PLFAs of biochar during incubation with an acidic paddy soil.|
|Next Article:||Management options for water-repellent soils in Australian dryland agriculture.|