Regional trends in the chemical and mineralogical properties of upland soils in humid Asia: With special reference to the WRB classification scheme

Abstract

Soils in humid Asia exhibit relatively incipient mineralogical characteristics because of the dominant steep slopes, crust movement and volcanic activity compared with many tropical soils on stable plains. In the present study, the relationship between the mineralogical and chemical properties of these soils was investigated with special reference to the World Reference Base (WRB) classification scheme. A total of 186 upland soil profiles were collected and the chemical and mineralogical properties of the B-horizon soils were analyzed for clay mineral composition, pH(H₂O), exchangeable cations, cation exchange capacity (CEC), total C content, particle size distribution and sodium-dithionite-extractable oxides (Fed and Ald). The majority of the soils were acidic. The CEC/clay of the soils derived from sedimentary rocks (excluding limestone) or felsic parent materials showed a clear regional trend, which was usually higher than 24 cmol_c kg⁻¹ (corresponding to Alisols if the argic horizon is recognized) under the udic and perudic soil moisture regimes in Indonesia and Japan, whereas it was predominantly lower than 24 cmol_c kg⁻¹ (corresponding to Acrisols) under the ustic soil moisture regime in Thailand. In contrast, soils derived from mafic volcanic rocks or limestone were more variable in clay mineral composition, CEC/clay and pH, and were often high in Fed. This trend is in accordance with the clay mineral composition, in that mica and kaolin minerals dominated under the ustic soil moisture regime in Thailand, whereas significant amounts of 1.4 nm minerals formed under the udic and perudic soil moisture regimes in Japan and Indonesia. In conclusion, of the 186 soils studied, only nine and eight soils are classified into Luvisols (or Lixisols) and Cambisols (eutric), respectively, whereas the majority (169 samples) are classified as acid soils, such as Andosols, Podzols, Alisols, Acrisols and Cambisols (dystric). The WRB classification is generally consistent with regional trends in the chemical and mineralogical properties of the soils and successfully describes the distribution patterns of the acid soils in humid Asia using the criteria of CEC/clay = 24 cmol_c kg⁻¹.

Key words:

• cation exchange capacity/clay
• clay mineral composition
• humid Asia
• soil pH
• WRB classification

INTRODUCTION

Recently a new classification scheme of the world's soils was proposed, the World Reference Base for Soil Resources (WRB) (International Society of Soil Science (ISSS) – International Soil Reference and Information Centre (ISRIC) – Food and Agriculture Organization 1998a; International Union of Soil Science (IUSS) – International Soil Reference and Information Centre – Food and Agriculture Organization 2006). One of the new ideas in the WRB is the introduction of clay activity (or cation exchange capacity [CEC]/clay) to classify soils in the highest category level (i.e. reference soil groups [RSG]), which is used to discriminate soils that have an argic horizon into Lixisols and Luvisols (high base saturation [> 50%] soils) or Acrisols and Alisols (low base saturation [< 50%] soils). Upland soils in humid Asia are predominantly acidic because of excessive precipitation and, hence, Acrisols and Alisols are in the major RSGs if argic horizons are recognized (ISSS-ISRIC-Food and Agriculture Organization 1998b). The regional distribution patterns of these two soils remains, however, a controversial issue in this region, and there is very little data on which to base any discussion. Although both Acrisols and Alisols are acidic and agricultural production is strongly restricted on these soils (Food and Agriculture Organization 2001), higher CEC/clay values in Alisols indicate that the levels of exchangeable and soluble Al achieve toxic quantities and cause more serious constraints for agricultural use in Alisols than in Acrisols. It is necessary to understand the regional distribution patterns of the two soils with reference to pedogenetic factors, such as geology and climate.

Soils in humid Asia exhibit relatively incipient mineralogical characteristics because of the dominant steep slopes, crust movement and volcanic activity on “young alpine fold belts (Food and Agriculture Organization 2001)” compared with soils developed on stable plains associated with the “Precambrian shield” in eastern South America or equatorial Africa. Watanabe et al. (2006) reported a regional trend in clay mineralogy, relating to pedogenetic factors, in upland soils in humid Asia as follows: (1) mica and kaolin minerals dominated under the higher pH conditions associated with the monsoon climate and the ustic soil moisture regime (Soil Survey Staff 2006) in Thailand, (2) significant amounts of expandable 1.4 nm minerals formed under the presence of dioctahedral mica as a precursor under the lower pH conditions of the udic and perudic soil moisture regimes in Japan and Indonesia, (3) in soils derived from mica-free parent materials (or mafic volcanic materials), kaolin minerals and smectite were predominant. Such mineralogical composition derived from geological and climatic conditions is considered to govern some physical and chemical properties of the soils and should be reflected in any soil classification.

In the present study, the relationship between the mineralogical and chemical properties of upland soils in humid Asia was further investigated. In addition, we discuss the WRB classification scheme for classifying soils in humid Asia in terms of the regional distribution patterns of the physicochemical and mineralogical properties of the soils.

MATERIALS AND METHODS

Soils

A total of 186 upland soil profiles with minimum disturbance (i.e. under forest or cropland with low-input management) were surveyed according to the distribution pattern of soils in respective regions in terms of the geology and climate (Table 1; Fig. 1). The clay mineralogy of 104 of the 186 samples was reported in our previous paper (Watanabe et al. 2006). These soils are grouped into seven categories based on pedogenetic conditions.

Figure 1  Study area with soil moisture regimes.

According to Fig. 1, a large part of East Asia is situated either under a ustic, udic or perudic soil moisture regime and, hence, the present study was considered to cover major soils in terms of soil moisture conditions. As soils derived from intermediate to mafic volcanic rocks and limestone are considered to exhibit different properties, judging from the results of a previous study on soil mineralogy (Watanabe et al. 2006), they are separated from the other soils and grouped as “mafic”. These soils were derived mainly from volcanic rocks in the Java and Sumatra Islands of Indonesia, and partly from limestone in northern Thailand and mafic intrusive rocks in Japan.

The sample groups of IDL and IDH comprise soils derived from sedimentary rocks (mostly sandstone and mudstone) or felsic materials in low (< 600 m) and high (> 600 m) elevation areas of Indonesia, respectively. Although both elevation areas would be classified as having a hyperthermic soil temperature regime according to US Taxonomy (Soil Survey Staff 2006), they are divided into two groups in the present study because the percolation of organic matter into the B-horizon soils is apparently more extensive at elevations above 600 m according to soil survey results, presumably because of the extremely humid conditions (i.e. perudic soil moisture regime) there.

The THL and THH soils were collected from Thailand, mostly from the northern mountainous region, which has a monsoon climate. The border of the two groups is approximately 800 m a.s.l., above which there is predominately evergreen forest, presumably because of decreasing water stress during the dry season. The soil temperature regimes of THL and THH are hyperthermic and thermic, respectively. Most of the parent materials of these soils are a wide variety of sedimentary rocks and granite.

Table 1 Outline of the sample soils

Sample group	No. samples^†	Sampling regions	Parent materials	Annual precipitation (soil moisture regimes)^‡	Mean annual temperature (soil temperature regimes)^‡
IDL	39 (10)	Low elevation area (< 600 m) of East Kalimantan, Java, and Sumatra islands of Indonesia	Sedimentary rocks (excluding limestone) and granite	> 1,500 mm (mostly udic)	> 22°C (hyperthermic)
IDH	10 (7)	High elevation area (> 600 m) of Kalimantan island of Indonesia	Sedimentary rocks (excluding limestone)	> 2,000 mm (perudic)	> 22°C (hyperthermic)
THL	40 (38)	Mainly low mountains (< 800 m) in northern Thailand	Sedimentary rocks (excluding limestone) and granite	1,200–1,600 mm, with distinct dry season (ustic)	> 22°C (hyperthermic)
THH	24 (22)	Mainly high mountains (> 800 m) in northern Thailand	Sedimentary rocks (excluding limestone) and granite		15–22°C (thermic)
JPS	30 (9)	South-temperate to sub-tropical zones of Japan	Sedimentary rocks (excluding limestone) and granite	Mostly > 1,500 mm, no dry season (udic or perudic)	> 15°C (thermic or hyperthermic)
JPN	20 (11)	North-temperate zone of Japan			8–15°C (mesic)
mafic	23 (7)	Mostly from Java and Smatra islands of Indonesia (partly from Thailand and Japan)	Volcanic rocks (intermediate to mafic) and limestone	> 1,500 mm (varying in soil moisture regimes: ustic, udic, and perudic)	> 15°C (thermic or hyperthermic)
Total	186 (104)
^†Clay mineralogy was reported in Watanabe et al. (2006) for soil samples numbered in parentheses.
^‡According to US Soil Taxonimy (Soil Survey Staff 2006).

The JPS and JPN soils were collected from warm and cool temperate forests in Japan and, therefore, the soil temperature regimes of these soils were thermic and mesic. Japan is situated under a humid climate and the soils collected have a udic or perudic soil moisture regime. The parent materials are predominantly sedimentary rocks (predominantly mudstone with minor occurrences of sandstone and shale) and partly felsic igneous rocks. Soils strongly affected by volcanic ejecta were not included in the samples. Analytical data on 18 soil profiles out of 50 are cited from Hirai (1995), in which almost the same analytical procedures were used as in the present study.

Analytical methods

Soil samples were collected from each pedogenetic horizon of the soil profile, air-dried and passed through a 2 mm mesh sieve for the following chemical and mineralogical analyses. Soil pH in water was measured with a glass electrode using a 1:5 soil : solution ratio. Cation exchange capacity and the content of exchangeable bases were measured after extracting with 1 mol L⁻¹ NH₄OAc at pH 7.0 and then with a 10% NaCl solution (Thomas 1982). The NH₄ ⁺ extracted with 10% NaCl solution was distilled after the addition of concentrated NaOH solution and collected into a 2% H₃BO₄ solution, followed by determination of NH₄ with HCl titration (0.01 mol L⁻¹). The contents of exchangeable bases (Na, K, Mg and Ca) in NH₄OAc solution were determined by atomic absorption spectrophotometry (AAS) (Shimadzu, AA-840-01, Kyoto, Japan). Exchangeable Al and H were extracted with 1 mol L⁻¹ KCl. Exchange acidity (Al + H) was determined by titration with 0.01 mol L⁻¹ NaOH to pH 8.3 using phenolphthalein as an indicator. Then, after the addition of 4% NaF solution to liberate OH⁻ from Al(OH)₃ precipitates, the exchangeable Al was determined by back titration to the same pH (8.3) with 0.01 mol L⁻¹ HCl. The content of exchangeable H was determined as the difference. Total C and N contents were measured with an NC analyzer (Sumigraph NC-800; Sumika Chemical Analysis Service, Ltd., Tokyo, Japan). Particle size distribution was determined using a combination of sieving and pipette methods, in which complete dispersion of silt and clay particles was achieved using a pH adjustment to 9–10 and supersonication, after pretreatment with H₂O₂ at 80°C to remove organic matter (Gee and Bauder 1986). The clay mineral composition was semi-quantified by the relative peak areas of mica (1.0 nm), kaolin minerals (1.0 and 0.7 nm) and expandable 2:1 minerals (1.4 nm) in X-ray diffractograms using Cu–Kα radiation (X-ray diffractometer, RAD–2RS; Rigaku, Tokyo, Japan). Clay mica and halloysite with a 1.0 nm diffraction peak were further distinguished based on the behavior of the peak after glycerol solvation of Mg-saturated specimens and sequential heating of K-saturated specimens at 350–550°C. Free oxides (Fed and Ald) were extracted with a citrate–bicarbonate mixed solution buffered at pH 7.3 with the addition of sodium dithionite (DCB) at 80°C (Mehra and Jackson 1960) and then the contents of Fe and Al in the extract were determined using multi-channel inductively coupled argon plasma atomic emission spectroscopy (ICP–AES) (SPS-1500; Seiko, Chiba, Japan) after filtering the extracts through 0.45 µm Millipore filters, Tokyo, Japan.

The CEC derived solely from mineral components (CEC_min) was calculated by eliminating the contribution of organic C using the following equation for each soil profile (4–7 soil samples):

The coefficient a was determined by multiple regression, followed by determination of CEC_min by subtracting the contribution of total C. The median values of a in different soil groups ranged from 0.16 to 0.22 cmol_c g⁻¹ C. In cases where the coefficients a and b could not be determined with reliability because of multicollinearity, the median value of a in each soil group was used instead. Then the CEC value per unit clay content (CEC_min/clay hereafter) was determined to be CEC_min divided by the clay content.

For the analysis of B-horizon soils, we used the data from each sub-horizon soil from the soil profiles: most profiles ranged from 30 to 50 cm in depth and corresponded to the upper argic or comparable horizon. The data obtained were statistically analyzed using SYSTAT version 8.0 software (SPSS 1998).

RESULTS AND DISCUSSION

Regional trend in the general physicochemical and mineralogical properties of the soils

The average values of selected chemical and mineralogical properties of the B-horizon soils from different regions, with anova results, are summarized in Table 2. Although clay mica was always the dominant component in the 1.0 nm minerals, judging from virtually no expansion of the 1.0-nm diffraction peak after glycerol solvation as well as a clear diffraction peak remaining after 550°C heating in our samples, there is a possibility that halloysite is a minor component. As shown in Table 2 and Fig. 2, the majority of soils were acidic. Base saturation (by CEC at pH 7) is usually below 50%. Among the B-horizon soils, soils from the ustic soil moisture regime in Thailand show significantly higher pH(H₂O) values. Some soils from mafic parent materials also exhibit higher base saturation above 50% (Fig. 2). According to Table 2, Alisols and Podzols are strongly acidic, followed by Andosols, Cambisols and Acrisols. As for clay content, the soils from Indonesia are less clayey than those from Thailand or from mafic parent materials. Total C content clearly increases as soil temperature decreases in the following order: IDL, IDH, THL (hyperthermic) < THH, JPS (thermic) < JPN (mesic).

Table 2 Average values of selected chemical and mineralogical properties of the soils from different regions

Sample group	No. samples	pH(H2O)		Clay		Total C^†		CECmin/clay		Fed^†		Clay mineral composition
												1.4 nm		1.0 nm		0.7 nm
		AVE	SE	AVE	SE	AVE	SE	AVE	SE	AVE	SE	AVE	SE	AVE	SE	AVE	SE
				(%)		(g kg^−1)		(cmolc kg^−1)		(g kg^−1)		(%)		(%)		(%)
Sorted by regions and parent materials
IDL	39	4.58	0.07 a	38.6	2.1 ab	5.4	0.5 a	32.4	1.5 bc	20.7	1.5 a	37	3 b	7	1 ab	56	3 cd
IDH	10	4.52	0.06 a	34.0	2.3 ab	6.5	1.0 ab	45.2	2.9 d	22.1	3.1 abc	37	5 b	18	4 abc	45	4 bc
THL	40	5.44	0.07 c	53.0	2.3 c	9.6	0.5 bc	22.2	1.3 a	37.3	2.6 cd	4	1 a	31	3 c	66	3 de
THH	24	5.50	0.09 c	47.2	3.2 bc	14.6	2.0 c	25.7	1.4 ab	40.8	5.0 bcd	12	2 a	17	4 b	71	4 ef
JPS	30	4.71	0.04 ab	35.1	2.4 a	12.9	2.6 bc	40.0	2.1 d	23.6	2.6 ab	58	4 c	11	2 ab	31	3 b
JPN	20	4.59	0.05 a	47.1	1.9 bc	33.4	3.0 d	39.9	2.2 cd	34.0	1.9 bcd	80	3 d	9	2 ab	11	2 a
Mafic	23	5.02	0.16 b	66.5	2.3 d	11.2	0.9 bc	31.4	2.9 bc	65.4	6.7 d	14	5 a	4	3 a	81	5 f
Sorted by reference soil groups
Andosols	4	4.66	0.06 ab	50.9	1.7 ab	46.0	5.8 c	31.2	4.8 ab	30.4	2.2 a	83	3 c	9	3 ab	8	1 a
Podzols	6	4.38	0.09 a	43.7	2.1 ab	33.3	3.6 c	43.4	2.5 b	42.0	1.6 a	75	5 c	15	5 ab	10	1 a
Acrisols	32	5.28	0.09 b	56.4	2.6 b	9.4	0.7 ab	18.1	0.8 a	39.1	4.1 a	6	1 a	25	3 b	70	4 c
Alisols	64	4.76	0.06 a	45.7	1.8 a	6.8	0.5 a	33.2	1.1 b	31.6	2.5 a	33	3 b	9	1 a	57	3 b
Luvisols and Lixisol	9	5.81	0.20 c	56.3	7.0 ab	9.9	1.3 ab	30.9	4.0 b	47.8	7.5 a	6	3 ab	15	7 ab	79	7 bc
Cambisols	71	4.95	0.07 ab	41.0	1.9 a	15.7	1.5 b	36.2	1.6 b	32.6	3.0 a	39	4 b	15	2 a	46	4 b
All	186	4.96	0.04	46.3	1.2	12.5	0.8	31.9	0.9	34.4	1.7	32	2	14	1	54	2
^†anova was applied after logarithmic transformation to normalize the dataset. Values with the same letters are not significantly different using a Tukey test (P < 0.05). AVE, average; SE, standard error.

Figure 2  Acidity and base saturation of the soils studied.

According to Fig. 3a, in which all the 1.0 nm minerals are supposed to be clay mica, the mineralogical composition exhibited the same regional trend as the previous report by Watanabe et al. (2006). The soils derived from mafic parent materials (mostly mica-free) have clay mineralogy dominated by kaolin minerals with small amounts of mica (Fig. 3a), and occasionally have higher amounts of iron oxides (Fig. 3b). Under the higher pH conditions associated with the monsoon climate and the ustic soil moisture regime in Thailand, mica and kaolin minerals dominated with lesser amounts of 1.4 nm minerals. In contrast, under the lower pH conditions of the udic and perudic soil moisture regimes in Japan and Indonesia, significant amounts of 1.4 nm minerals formed. Among them, most of the soils from Japan are dominated by 2:1–2:1:1 intergrade minerals that have appreciable amounts of interlayered materials, whereas those from Indonesia usually do not exhibit such interlayering (data not shown). These regional trends of clay mineral composition are also observed in the average values shown in Table 2, in which an increasing trend of 1.4 nm minerals in the order of THL, THH < IDL, IDH < JPS < JPN is clear. Essentially the same trend in clay mineralogy was also reported by Ohta et al. (1993) for soils in East Kalimantan and by Yoshinaga et al. (1995) for soils in northern Thailand.

Figure 3  Mineralogical composition of the soils. (a) Relative abundances of mica, kaolin and expandable 1.4 nm minerals in clay fraction. (b) Relative abundance of expandable 1.4 nm minerals in clay fraction and DCB-extractable Fe of soils.

The values of CEC_min/clay are plotted together with ECEC/clay (effective CEC [sum of exchangeable bases and Al] divided by clay content) or exchangeable Al/clay in Fig. 4a,b. The CEC_min/clay of the soils derived from sedimentary rocks (excluding limestone) or felsic materials showed a clear regional trend; that is, it was usually higher than 24 cmol_c kg⁻¹ (corresponding to Alisols if the argic horizon is recognized) under the udic or perudic soil moisture regimes in Indonesia (IDL and IDH) and Japan (JPS and JPN), whereas it was predominantly lower than 24 cmol_c kg⁻¹ (corresponding to Acrisols) under the ustic soil moisture regime in Thailand (THL). The values of THH soils are occasionally higher than 24 cmol_c kg⁻¹, presumably because the more percolating soil moisture conditions in high mountains result in similar conditions to the udic soil moisture conditions. Most of the cation exchange sites represented by ECEC seem to be occupied by exchangeable Al (Fig. 4b) and, therefore, the level of Al toxicity should be higher in the soils from the udic or perudic soil moisture regimes of Indonesia and Japan compared with those from the ustic soil moisture regime of Thailand. Similarly high CEC/clay values were reported for soils with a udic soil moisture regime, for example, 34.0 and 26.3 cmol_c kg⁻¹ in subsoil layers of upland soils in northern Sarawak, Malaysia (Tanaka et al. 2005), or average values of 34.7 and 41.2 cmol_c kg⁻¹ for 25 and 17 upland soils, respectively, in southern Sarawak, Malaysia (Tanaka et al. 2007). In contrast, for soils with a ustic soil moisture regime, lower values of CEC/clay were often reported. Toriyama et al. (2007) reported the CEC/clay values for subsoils of well-drained Acrisols in central Cambodia (60–100 m a.s.l.) to be approximately 10 cmol_c kg⁻¹. According to Watanabe et al. (2004), the clay mineralogy and CEC/clay values in upland soils in northern Laos (600–1,100 m a.s.l.) were essentially similar to our results in that illite and kaolinite were dominant in the clay fraction, and the values of CEC/clay of the B-horizon soils were almost equal to or lower than the threshold value (i.e. 24 cmol_c kg⁻¹). All these reports indicate that our results obtained in some limited regions in humid Asia can be applied to different countries with similar geological or climatic conditions.

Figure 4  Characteristics of the cation exchange capacity (CEC) of the soils. CECmin/clay, CEC value per unit clay content; ECEC/clay, effective CEC (sum of exchangeable bases and Al) divided by clay content.

The relationship obtained between CEC_min/clay and exchangeable Al/clay for Acrisols and Alisols (i.e. exchangeable Al/clay = –3.57 + 0.52 CEC_min/clay) is close to that reported in Food and Agriculture Organization (2001) for soils in Indonesia, the Caribbean region, Rwanda, Cameroon, Peru and Colombia; that is, Al/clay = –1.28 + 0.64 CEC/clay. Thus, the overall properties of Al retention and perhaps of the proportion of permanent and variable negative charges of the soils in the present study are considered to be similar to those in Acrisols and Alisols from other regions.

There is a unique characteristic of the soils derived from mafic parent materials compared to the soils from sedimentary rocks (excluding limestone) or felsic volcanic materials; these soils are more variable in pH (Fig. 2) or CEC_min/clay (Fig. 4a). According to Fig. 4b, the majority of the soils do not retain appreciable amounts of exchangeable Al, even when the value of CEC_min/clay is high. One reason for this must be the relatively higher pH and base saturation for some soils (Fig. 2). Another possible explanation is that variable negative charges derived from free oxide surfaces can contribute to the apparent increase of CEC at higher pH ranges (i.e. 7), and the actual CEC at the lower pH range at which Al can dominate is small. Variation in the chemical properties of soils derived from mafic materials (including limestone) in Java Island was also reported by Supriyo et al. (1992).

Relationship between clay mineral composition and cation exchange capacity in the soils

It is widely known that clay mineralogical properties of soils strongly affect soil physicochemical properties. In fact, the CEC_min/clay in the present study is influenced by the relative abundance of 1.4 nm minerals in the clay fraction (Fig. 5a). The following equation is obtained:

This equation suggests that the 1.4 nm minerals contribute to a CEC increase of 20.1 cmol_c kg⁻¹. As most of the 2:1 minerals in the Japanese soils are modified by hydroxy-interlayered materials (Hirai 1995; Kitagawa 2005), the CEC values of these soils could be estimated to be lower than the possible negative charge derived from the 2:1 lattice structure (Funakawa et al. 2003). On the contrary, among soils with the lowest amounts of expandable 2:1 minerals (i.e. Thai soils), the relative contribution of other mineral components increases, and the CEC_min attributable to the 1.4 nm minerals should be appreciably lower than the value of the intercept of the equation, that is, 25.5 cmol_c kg⁻¹. As a result, the actual contribution of expandable 2:1 minerals to CEC may be considerably higher than 20.1 cmol_c kg⁻¹ clay. Thus, the CEC of the upland soils in humid Asia is primarily determined by the relative abundance of expandable 2:1 minerals, in which exchangeable Al is dominant as represented by the equation: Exch. Al/clay = –3.57 + 0.52 CEC_min/clay.

Figure 5  Relationship between mineralogical composition and chemical properties of the soils. Relationships between relative abundances of expandable 1.4 nm minerals in clay fraction and (a) CECmin/clay and (b) pH(H2O) of soils.

Relationship between clay mineral composition and soil pH

Soil pH is an important index that affects a wide range of agronomical and environmental functions of soils (Robson 1989). In our present study, a negative correlation is observed between the relative abundance of 1.4 nm minerals and soil pH (Fig. 5b), and the following equation is obtained:

As most of the soils studied are acidic and base saturation is lower than 50%, the dominant cation retained on the cation exchange sites of soils is Al³⁺. The equation above indicates that the presence of expandable 2:1 minerals contributes to pH reduction by 0.008 units per clay percentage. In solution, the Al³⁺ ion behaves as a weak acid through a hydrolytic reaction: that is, Al³⁺ + OH⁻ = Al(OH)²⁺ (pKa = 5.0), and soils rich in exchangeable Al show a distinct buffer zone against OH⁻ addition during titration (Funakawa et al. 1993, 2008). The strong acidity far below pH 5.0 of some soils must be affected by the presence of a stronger acid(s), that is, H⁺, although quantitative analysis of exchangeable H⁺ is difficult, and the selectivity coefficient between Al³⁺ and H⁺ ions on the permanent negative sites of soils is rarely determined.

In contrast, the oxide surfaces represented by the Fed fraction are considered to increase soil pH by 0.00451 units per g Fed kg⁻¹ soil. As is well known, the oxide surfaces act as weak acids that have a zero point of charge (ZPC) of approximately 6–9 (McBride 1989). In lower pH regions, it might mitigate H⁺ in solution through the protonation reaction: M–OH + H⁺ = M–OH₂ ⁺, where M represents the metal ions (Fe, Al) at the surface of the soil particles, and thereby contributes to increased soil pH. Soils derived from mafic volcanic materials are often rich in the Fed fraction, which was originally inherent in the parent materials (Fig. 3b), and soil pH was relatively high (Table 2). In humid Asia, the major distribution of the mafic parent materials is generally limited to specific regions, such as the volcanic belts of Java and Sumatra Islands. In the humid tropics of other continents, for example, eastern Latin America or equatorial Africa, however, mafic materials cover wider regions. It can be said that acid mitigation by oxide surfaces is relatively limited in humid Asia, unlike in other continents. The general results obtained in the present study are considered to be specific for some regions of humid Asia, in which sedimentary rocks or felsic volcanic materials are the dominant parent materials of the soils.

Classification of the soils examined according to WRB and its relationship to mineral weathering conditions

Table 3 summarizes the classification of the soils examined according to WRB (IUSS-ISRIC-Food and Agriculture Organization 2006). The RSGs to which “mafic” soils are classified are diverse, including Alisols, Acrisols, Luvisols and Cambisols. More detailed surveys focusing on the relationship between parent materials or primary minerals and properties of the soils that developed there are necessary.

Table 3 Summary of the classification of the soils according to the World Reference Base

Sample group	No. samples	No. soils classified into different RSGs
		Andosols	Podzols	Alisols	Acrisols	Luvisols (or Lixisols)	Cambisols
							(Dystric)	(Eutric)
IDL	39	0	0	34	2	0	2	1
IDH	10	0	0	5	0	0	5	0
THL	40	0	0	5	20	3 (1)	9	2
THH	24	0	0	6	6	1	9	2
JPS	30	0	0	7	0	0	23	0
JPN	20	4	6	0	0	0	10	0
mafic	23	0	0	7	4	4	5	3
Total	186	4	6	64	32	9	63	8
RSGs, reference soil groups.

Most of the members of IDL and IDH belong to Alisols according to WRB, except for those on relatively steep slopes in high mountains that are classified into Cambisols (dystric) because of a lack of a clear argic horizon. Higher soluble Al³⁺ in Alisols may be a more serious constraint for agricultural production compared with Acrisols. In contrast, Acrisols are dominant in THL, whereas the proportion of Alisols and Cambisols (dystric) increased in THH. In Japan, the majority of the soils are Cambisols (dystric) and some from JPS are Alisols. In JPN, some soils are classified into Andosols or Podzols because of the relatively high influence of amorphous materials under cool temperate climates and/or the presence of nearby active volcanoes. It should be noted that no extremely weathered soils such as Ferralsols or Plinthosols were found in the present study, although there is still a necessity to extend future surveys to whole regions of humid Asia, especially to the relatively stable plains of Northeast Thailand, and to central Cambodia.

Thus, of the 186 soil samples, only nine and eight soils are classified into Luvisols (or Lixisols) and Cambisols (eutric), respectively, indicating that the upland soils in humid Asia are predominantly acidic. In our previous report (Watanabe et al. 2006), dioctahedral mica inherent to sedimentary rocks or felsic igneous rocks was thought to weather to form expandable 2:1 minerals, that is, vermiculitization, under the lower pH conditions associated with the udic or perudic soil moisture regime. In contrast, mica is relatively stable under the higher pH conditions associated with the ustic soil moisture regime, whereas other primary minerals, such as feldspars, are unstable and dissolve to form kaolin minerals and gibbsite. These processes should be further critically analyzed considering the possibility of selective dissolution of interlayered K⁺ ion of mica under acidic conditions from a clay mineralogical viewpoint (see Fanning et al. 1989). Our finding in the present study suggests that the regional distribution patterns of Acrisols and Alisols in humid Asia are strongly related to pedogenetic conditions through clay mineral formation. In turn, such a close relationship suggests that the soils in humid Asia are predominantly on the course of pedogenesis under the present bio-climatic conditions, probably unlike tropical soils on plains in other continents.

Conclusions

Soil chemical properties, such as CEC/clay and pH, are primarily derived from the clay mineralogy of the soils, reflecting weathering conditions in terms of geology and climate. The CEC/clay of the soils derived from sedimentary rocks (excluding limestone) or felsic parent materials showed a clear regional trend; that is, it was usually higher than 24 cmol_c kg⁻¹ (corresponding to Alisols if the argic horizon is recognized) under the udic and perudic soil moisture regimes in Indonesia and Japan, and it was predominantly lower than 24 cmol_c kg⁻¹ (corresponding to Acrisols) under the ustic soil moisture regime in Thailand. In contrast, soils derived from mafic volcanic rocks or limestone were more variable in clay mineral composition, CEC/clay or pH, and often high in Fed. This trend is in accordance with the clay mineral composition, in that mica and kaolin minerals dominated under the ustic soil moisture regime in Thailand and significant amounts of 1.4 nm minerals formed under the udic and perudic soil moisture regimes in Japan and Indonesia. The WRB classification is generally consistent with regional trends in the chemical and mineralogical properties of soils and successfully describes the distribution patterns of acid soils in humid Asia using the criteria of CEC/clay = 24 cmol_c kg⁻¹.

ACKNOWLEDGMENTS

We gratefully acknowledge that analytical data on some Japanese forest soils were supplied from the doctoral thesis of Dr Hideaki Hirai, Utsunomiya University.

Notes

Present address: Graduate School of Urban Environmental Sciences, Tokyo Metropolitan University, Tokyo 192-0397, Japan.