Reclamation of a Badarkhali hot spot of acid sulfate soil in relation to rice production by basic slag and aggregate size treatments under modified plain–ridge–ditch techniques

Abstract

A field experiment was conducted for the reclamation of a Badarkhali hot spot of acid sulfate soil manipulated by flash leaching followed by basic slag (BS) at 10 t ha⁻¹ (BS₁₀) and 20 t ha⁻¹ (BS₂₀) and aggregate sizes (A) of soil less than 20 mm (A₂₀) and less than 30 mm (A₃₀) treatments under two different techniques (Tech 1: pyrite layer at top, jarosite layer at middle and topsoil at the bottom of the ridge; Tech 2: topsoil at top, pyrite layer at middle and jarosite layer at the bottom of the ridge). Responses to two cultivars of rice (Pizam [local cultivar] and BR 14 [high yielding cultivar]) with the treatments were evaluated. The initial soil had a very low pH(H₂O) 4.0 and a high electrical conductivity (EC) of 1.4 m S⁻¹, and the pyrite content was 68 g kg⁻¹. The exchangeable Mg content of the soil was approximately twice that of Ca and the Al content was at a highly toxic level. The average soil data of all the treatments, except for the control plots (where no amendment was applied), after harvesting of rice increased by 1.1 units for soil pH and 17–524% for the contents of N, P, Ca and Mg, while the concentrations of Fe, Al, Na, Cl⁻ and SO₄ ^2– decreased by 30–94% compared with the initial soil. The maximum growth and yield of rice grains (4.4 t ha⁻¹) were obtained by the Pizam compared with the BR 14 (4.0 t ha⁻¹) in the A₂₀BS₂₀ treatment in the ridges of Tech 2. The lowest grain yields of 0.02 (BR 14) and 0.07 t ha⁻¹ (Pizam) were recorded for the control plots. The other treatments also resulted in significantly (P ≤ 0.05) improved performance on rice production. The highest N, P, K, Ca and Mg contents in the shoots of BR 14 and Pizam rice were obtained under the A₂₀BS₂₀ treatment followed by the A₂₀BS₁₀ ≥ A₃₀BS₂₀ treatments. Application of A₂₀BS₂₀ under Tech 2 is the most appropriate reclamation option and the local Pizam is the most suitable rice for this soil.

Key words:

• acid sulfate soil
• aggregate size
• basic slag
• growth performance
• reclamation

INTRODUCTION

The projected increase of hungry mouths obviously requires strenuous efforts worldwide to grow more food and to extend the acreage under crop production. However, the availability of land for growing crops is limited and increasingly marginal and problem soils will need to be used. Among problem soils, acid sulfate soils (ASSs) have lasting, severe environmental degradation. It is estimated that these soils affect approximately 100 million hectares (M ha) of land worldwide (Sheeran 2003). Van Mensvoort and Dent (1998) claimed approximately 24 M ha of ASSs, of which approximately 0.7 M ha are located in different pockets of inundated coastal areas in Bangladesh. In these areas crop production is very low; and some lands are unproductive even though the soils have high agricultural potential (Khan 2000). The nature and characteristics of these pockets of ASSs varied from place to place and within pockets owing to differences in mangrove vegetation and the accumulation of sediments (Khan et al. 2007b). Runoff and leachate from ASSs can adversely affect aquatic communities, agricultural, engineering works and beneficial use of the environment (Orndorff and Daniels 2002). The ASSs generate H₂SO₄ that brings their pH from 6–7 to below 4, and sometimes to as low as 2. This acid leaks into drainage and floodwaters, corrodes steel and concrete, and attacks clay-liberating elements in toxic concentrations. In ASS areas, massive fish kills and ulcerative diseases have been reported by several scientists (Lin and Melville 1994; Sammut et al. 1996). Losses resulting from fish kills from such situations in the coastal plains of Bangladesh accounted for approximately US$3.4 million during 1988–1989 (Callinan et al. 1993).

Conventional reclamation of ASSs through liming and flash leaching is not sustainable because soil acidity produced by 1% oxidizable sulfur requires approximately 30 t of CaCO₃ per hectare (van Breemen 1993). In general, ASSs contain 1–5% oxidizable sulfur and the use of more than 10 t lime per hectare has an antagonistic effect on micronutrient levels as well as on the balance of basic cations in plants (Khan et al. 1994, 1996). Neutralization of ASSs with lime and/or leaching leads to the deterioration of the soils, related ecosystems, and to permanent soil acidification (Khan and Adachi 1999). Takai et al. (1992) reported that nutrient deficiency is an important factor when reclamation and improvement practices are performed in ASSs. Khan et al. (2006) reported that the application of basic slag (BS) in ASSs significantly increased soil pH, Ca and Mg with an associated decrease in Na, Fe and Al concentrations over time. Basic slag as a by-product from the steel industry in Bangladesh can be collected almost free of charge. This BS has a very high pH of 9.6 and contains 20.8% Ca, 9.8% Mg and 12.8% SiO₂ (Khan et al. 2006). Basic slag has been used on a small scale for the reclamation of ASSs in Bangladesh since 1985 and to date there is no evidence of any harmful effects. Reclamation of ASSs may be difficult but essential because of the formation of acidity through the natural phenomenon of oxidation–reduction processes in coastal soils, which receive saline seawater tides of approximately 0.2–0.6 m high (approximately 180 times per year). Potential ASSs with high pH (e.g. 6–7) does not mean that the soils are safe because they may create H₂S, Fe and some organic acid problems (Moore et al. 1990). The previous techniques of plain–ridge–ditch or raised beds that have been used for the reclamation of ASSs in Vietnam, Thailand, Indonesia and the Netherlands (Dent 1986; Dent and Mensvoort 1993; Dost and van Breemen 1982), where pyritic materials were taken from the sub-soils, were covered by non-pyritic topsoils and followed by flash leaching by rain or irrigation. However, in the present study, modified ridge–ditch techniques, where different sizes of aggregates were used for proper oxidation of pyritic materials and BS was used for the neutralization as well as the recovery of the basic cations lost during leaching, the placement of pyritic materials in different patterns was considered on the basis of tidal influences and the economic conditions of the farmers.

Tri et al. (1993) revealed that construction of a raised bed is the most important land management practice for ASSs. However, this practice has not had enough success, and its success depends on several factors of the soils, their position and environmental conditions (Agency for Agricultural Research and Development, The Land Water Research Group 1992; Khan 2000). Successful reclamation of ASSs may result in the development of productive fields for crop growth. Although poor soil reclamation may lead to the creation of unfavorable soil conditions for crop growth and the formation of actual ASSs, the real problem is in the coastal tidal flat plain areas (Khan et al. 2007a). Against this background, the objectives of the present field study were to evaluate the effects of BS, aggregate size and different techniques on the physico-chemical amendments of ASS in relation to rice production.

MATERIALS AND METHODS

Experimental site and conditions

A field experiment was conducted in a fallow land at a Badarkhali hot spot of ASS in the coastal old mangrove floodplain area of the Cox Bazar district of Bangladesh (Fig. 1). The lands are usually inundated by saline seawater tides of approximately 0.2–0.6 m high (approximately 180 times per year). On the basis of the tidal conditions, the height of the ridges constructed on the plain land under the techniques used was 0.6 m. However, during the experiment, the study site was protected from natural tides by a dike (1.5 m high) around the experimental site. The site enjoys a “tropical monsoon climate” and has three main seasons, namely, the monsoon or rainy season, the dry or winter season and the pre-monsoon or summer season. The monsoon season extends from June to October and is warm and humid. During this period, this locality receives above 85% of the total annual rainfall. The dry season extends from November to February and has the lowest temperature and humidity of the year. The pre-monsoon season extends from March to May and has the highest temperature and evaporation of the year. This area in Bangladesh was once occupied for centuries by dense mangrove forest. Today approximately 95% of the areas have been cleared for agricultural cultivation. As a result, the potential ASSs have become actual ASSs with very poor yields and low productivity. Eight soil series of ASSs were studied in the field on the basis of land type, land use, hydrological conditions and depths of the acid-forming layers. Among these soils, the Badarkhali hot spot of ASS at Purbapukuria was selected for further studies in relation to rice production.

Figure 1  Site map of the study area (oval) in Cox Bazar, Bangladesh.

Experimental design and field preparation

The experiment was set up in a completely randomized block design with three replicates. There was an approximately 1.5 m wide and 1.0 m deep drain around the experimental field, and an approximately 0.5 m and 0.3 m boundary around each main plot and subplot, respectively, for protecting the individual plot from contamination from the other treatments. The field of approximately 0.75 ha was divided into 12 main plots (four for plain, four for ridge and four for ditch) each measuring 10 m × 5 m in size. In each main plot of plain or ridge, there were a total of 48 sub-subplots (treatments: BS = 2, A = 2, rice = 2, Tech = 2 and replicates = 3, i.e. 2 × 2 × 2 × 2 × 3 = 48) and each sub-subplot had an area of 1 m × 1 m. The ditches were not considered for rice cultivation because they were newly excavated and they also did not count as a cropped area. The ditches were approximately 10 m × 5 m wide and 0.4–0.6 m deep. There were six separate control subplots in the plain land (two rice cultivars × 3 replicates) or check plots, where there was no treatment of A, BS and techniques. However, these control plots were compared with the treatments and techniques. These plots were irrigated, mainly by using rainwater (yearly rainfall > 4,000 mm) collected from the local rain-fed channels through the irrigation pump. Pond water was also used for irrigation during the dry period. Saline water intrusion and drainage water were controlled through dikes and flap gates. Basic slag was collected from a steel industry and then ground to less than 1 mm in size to apply in the field. Rates of BS₁₀ and BS₂₀ (basic slag at 10 and 20 t ha⁻¹, respectively) were incorporated into the topsoil (depth: 0–20 cm) by broadcasting during ploughing. For in situ neutralization of acidity arising from jarosite and pyrite layers in the modified plain–ridge–ditch techniques (Fig. 2), the same doses of BS were incorporated into the subsoil (0.2–0.4 cm) every 0.2 m using a soil slitter (5 cm diameter).

Modified plain–ridge–ditch techniques

Plain

The area of each plain plot (floodplain land) was 10 m × 5 m and the particle size of the soil in the plain plot was manually ground into two aggregate sizes of approximately less than 20 and 20–30 mm (A₂₀ and A₃₀, respectively). The ground soils were then processed under the sun and open air for maximum oxidation within 2 days. The plain land was considered to be reclaimed by the application of flash leaching (washed 5–7 times within 1 week until the pH was > 4.5 and the electrical conductivity [EC] values were reduced) followed by the application of the BS treatments (BS₁₀ and BS₂₀ as suggested by Khan et al. 2002). The drainage waters from the plain plots under the modified plain–ridge–ditch techniques were disposed of into the nearby ditches where the water was treated with BS as required to raise the water pH to 5.5.

Ridge

Ridges consisted of raised beds of approximately 0.6 m soil, stacked on floodplain soil (plain land), which was made by raising different layers of soil through excavation and arranged as shown in the Fig. 2 (Tech 1 and Tech 2). These 0.6 m high beds were constructed to facilitate leaching of acidity and salinity from the soil. The area of the top of each ridge was similar to that of the plain plot. Particle sizes of the bed materials were ground manually into two aggregate sizes of A₂₀ and A₃₀ (aggregate sizes of soil less than 20 and 20–30 mm, respectively). Smaller sizes of aggregates were used to understand the effects of oxidation of sulfidic materials (pyrite/jarosite) and their quick drainage from this heavy-textured ASS. Each plot on the ridge of the different techniques was brought into flash leaching followed by the application of BS treatments as done for the plain plot. The drainage water into the nearby ditches was treated with BS to raise the water pH to 5.5. The arrangements of soil layers in the different techniques were as follows:

• Technique 1: The top layer (surface soil) was extended to an approximately 0.2 m thick layer (first layer: 0–0.2 m) primarily onto the adjacent plots and was ground into different aggregate sizes (A₂₀ and A₃₀) as per the treatment requirement and was kept open to the air for oxidation within 2 days. Then, the second layer (0.2–0.4 m layer: sub surface soil with jarosite material) was extended to an approximately 0.2 m thick layer onto the processed first layer. The soil of the second layer was also ground to the desired sizes of aggregates and kept for oxidation within 2 days. Finally, the third layer (0.4–0.6 m layer: deeper soil containing pyrite material) was placed at the top of the previously stacked layers. This third layer was also considered for the preparation of different sizes of aggregates and for oxidation within 2 days. After that, the prepared 0.6 m raised beds underwent extensive leaching (washed 5–7 times within 1 week) with rain and pond water. Thereafter, the topsoil of the ridge was subjected to the different rates of BS application as per the treatments designed for the experiment.

• Technique 2: The preparation and processing of land including aggregate size, the duration of oxidation through air drying and the experimental treatments were similar to those stated for Tech 1; the only difference was the arrangement of the soil layers. In Tech 2, within the ridge (raised bed of 0.6 m), the jarosite layer was placed at the bottom (third layer: 0.4–0.6 m), then the pyrite layer was placed in the middle (second layer: 0.2–0.4 m) followed by the surface soil onto the top (0–0.2 m) of the ridge.

Figure 2  Modified plain–ridge–ditch techniques used in the field experiment for the reclamation and improvement of acid sulfate soil in relation to rice production.

Ditch

Adjacent to each ridge, there was a ditch of approximately 10 m × 5 m with a depth of approximately 0.4–0.6 m as a result of the excavation of soil for the preparation of ridges. These ditches were constructed as a reservoir of acid drainage waters, and these acidic waters were neutralized by the use of BS during the dry season. However, during the rainy season, these acidic wastes were automatically diluted by rain and/or controlled by runoff waters.

During August 2003, two varieties of rice (a high-yielding variety [HYV BR 14] and a local variety [Pizam]) were planted in the Badarkhali ASS. BR 14 was introduced by the Bangladesh Rice Research Institute and was recommended as a salt-tolerant cultivar for coastal salt-affected soils (Bangladesh Agriculture Research Council 1997). In contrast, the local Pizam is used by farmers as a saline-acid tolerant rice cultivar for this Cox’ Bazar area in Bangladesh. The topsoil (0–0.2 m) in each plot (plain and ridge) was fertilized with N, P and K at rates of 100, 80 and 60 kg ha⁻¹ as urea, triple super phosphate (TSP) and muriate of potash (MP), respectively, as a basal dose. All the TSP and MP and one-third of the urea were applied just 1 day prior to transplantation, whereas the remaining two-thirds of the urea was applied as top dressing 30 and 60 days after transplantation. The plots were allowed to receive natural rain and pond waters whenever necessary to maintain favorable conditions (maximum saturated to field-moist conditions) for rice. Thirty-day-old healthy and uniform seedlings were transplanted at a rate of four plants per hill. The distances between the hills were 15 cm. Pests were controlled by the use of the insecticide “Nogos” when required.

Soil analysis

The bulk samples obtained from the soil were stored for a couple of days under field-moist conditions (by putting the soil samples into polyethylene bags in an air-tight box) just prior to laboratory analysis. The sub-samples were air-dried and crushed to 2 mm before analysis. After treatment with 1 mol L⁻¹ CH₃COONH₄ (pH 5.0) and with 30% H₂O₂ to remove free salts and organic matter, respectively, the particle size distribution was determined using the pipette method (Day 1965). Soil pH was measured for the oven-dried soil in a soil : water (1:2.5) suspension using a Corning pH meter Model-7 as described by Jackson (1973). The pH of oven-dry soil was important with respect to the measurement of acidity (Khan et al. 1993). To determine the saturation extract of the soil, the EC (Richards 1954), water-soluble Na⁺ and K⁺ (flame photometry: Black 1965), water-soluble SO₄ ^2– and Cl⁻ contents (Jackson 1973); Ca²⁺, Mg²⁺, Fe³⁺ and Al³⁺ (atomic absorption spectrometry [AAS]: Hesse 1971) were determined. Organic matter content was determined (Nelson and Sommers 1982) using wet combustion with K₂Cr₂O₇. Available N (1.3 mol L⁻¹ KCl extraction, Jackson 1973), available P (0.001 mol L⁻¹ H₂SO₄ and pH 3 extraction Olsen et al. 1954) were determined. The cation exchange capacity was determined by saturation with 1 mol L⁻¹ CH₃COONH₄ (pH 7.0), ethanol washing, NH₄ ⁺ displacement with acidified 10% NaCl, and subsequent analysis by steam distillation (Kjeldhal method: Chapman 1965). Exchangeable Na⁺, K⁺, Ca²⁺ and Mg²⁺ were extracted with 1 mol L⁻¹ CH₃COONH₄ (pH 7.0) and determined by flame photometry (Na⁺, K⁺) and AAS. Exchangeable Al³⁺ (1 mol L⁻¹ KCl, Thomas 1982) and Fe³⁺ (1 mol L⁻¹ CH₃COONH₄: pH 4.8, Black 1965) were determined by AAS.

Plant analysis

The number of tillers, straw and grain yields of rice were determined at maturity. Composite samples of shoot dry matter were analyzed for N content using the micro-Kjeldahl method (Jackson 1973); P content was determined by spectrometry (Jackson 1973); K content by flame photometry (Black 1965); S content by turbidometry (Jackson 1973); and Ca and Mg contents by AAS (Hesse 1971) in HNO₃–HClO₄ acid (2:1) digest. The level of significance of the different treatments was determined using Duncan's New Multiple Range Test (DMRT) and least significant difference (LSD: Zaman et al. 1982).

RESULTS AND DISCUSSION

Pre-harvested and post-harvested soils

The Badarkhali ASS (0–20 cm depth) had a silty clay loam texture, an initially low pH of 4.0, a pyrite content of 68 g kg⁻¹, low base saturation (34%), high EC (1.4 S m⁻¹), and high exchangeable Fe³⁺ and Al³⁺ contents of 1.57 and 8.09 c mol kg⁻¹, respectively (Table 1). The contents of basic cations (Na⁺, K⁺, Ca²⁺ and Mg²⁺) in the initial soil were low to medium, while acidic cations (Al³⁺ and Fe³⁺) were very high in relation to the amounts found elsewhere (Donahue et al. 1987; Schlichting and Blume 1966). The pH value of the average soil data of all the treatments at post-harvesting of rice was found to have increased from 4.0 to 5.1 (i.e. by 1.1 units) and was higher compared with the control (i.e. initial value), while the EC value of the soil decreased to 0.33 S m⁻¹ (76% decrease compared with the control; Table 1). The contents of N, P, Ca and Mg in the average soil data at post-harvesting were found to have increased (increased over control [IOC], i.e. initial value) by 17 to 524% IOC. The contents of exchangeable Al³⁺, Fe³⁺, Na⁺, Cl⁻ and SO₄ ^2– in the soil were found to have decreased by 30–94% IOC (Table 1). The results also indicated that the physico-chemical properties of the ASS were strongly influenced by the application of leaching followed by the BS and aggregate size treatments in different reclamation and management techniques.

The post-harvested soil data (Fig. 3) of pH, exchangeable K⁺, Ca²⁺, Mg²⁺, Fe³⁺ and Al³⁺ contents were found to be affected significantly (P ≤ 0.05) by the application of BS and aggregate size treatments in the different techniques. Application of A₂₀BS₂₀ attained the highest soil pH value of 5.8 in the ridges of Tech 1 and 5.9 in the ridges of Tech 2 during post-harvesting followed by the A₂₀BS₁₀ (pH 5.3 in Tech 1 and 5.4 in Tech 2: Fig. 3) treatment in the ridges. The lowest soil pH of 4.0 was recorded in the control treatment (where there was no treatment of A, BS and techniques). The contents of exchangeable Al³⁺ and Fe³⁺ during post-harvesting were found to decrease sharply with the treatments and the decrements were more pronounced in the soil of the ridges of the techniques (Fig. 3). The highest amount of exchangeable Al³⁺, 8.09 c mol kg⁻¹, was recorded for the control plots, and this value was decreased to 0.14 for Tech 1 and 0.12 for Tech 2 by the A₂₀BS₂₀ preceded by the A₃₀BS₂₀ and A₂₀BS₁₀ treatments. The decrement was more pronounced with Tech 2. Among the treatments, the application of BS ranked first followed by the aggregate size treatments and techniques for the increments of soil pH and nutrient status of the soil, which might be because of the basic nature of BS (pH 9.6) as well as its release of some elements, mainly Ca and Mg, into the soils. The results agreed with the findings of some researchers. Khan (2002) reported that ASSs released a very large amount of Al, for example, 10 mmol L⁻¹. However, a very low concentration of Al can be hazardous. Concentrations of 1–2 mmol L⁻¹ Al are toxic to most crops and approximately 2 mmol L⁻¹ Al is toxic to rice. Fish are most susceptible and fish deaths occur at 0.5 mmol L⁻¹ Al. Standards and potable water mostly range from 5 to 1,450 µg L⁻¹ Al (2 × 10⁻⁴ to 6 × 10⁻² mol m⁻³ Al: Sittig 1994). Khan et al. (2006) reported that the application of BS to ASSs significantly increased soil pH, Ca and Mg with an associated decrease in Na, Fe and Al concentrations over time. However, it has been shown that an increase in pH as a result of the application of slag fertilizer caused a decrease in redox potential (Eh) (Nozoe et al. 1999; Ponnamperuma et al. 1967). However, the studied site is in a coastal belt that receives daily tides (0.2–0.6 m high) more than 180 times per year. Therefore, the chance of a decrease in Eh and its associated problems might be reduced in this study area. However, further study is necessary to evaluate these factors.

Figure 3  Influence of the two different techniques, and the basic slag (BS) and aggregate size (A) treatments on selected soil properties studied under field conditions. LSD, least significant difference.

Table 1 Some selected properties of the initial Badarkhali acid sulfate soil (0–20 cm depth) and the average soil (0–20 cm) data of all treatments post-harvest of the rice grown under field conditions

Soil properties	Initial soil	Post-harvest soil	IOC (%)
Textural class	Silty clay loam
Bulk density (Mg m^−3)	1.14	1.19	4
Moisture at field capacity (vol. %)	43	46	7
Soil pH (dry, 1:2.5)	4.0	5.1	28
Pyrite content (g kg^−1)^†	68	31	–54
EC (S m^−1)	1.4	0.33	–76
Organic carbon (g kg^−1)	19.7	19.9	1
Available N (1.3 mol L^−1 KCl: m mol L^−1 kg^−1)	5.14	6.01	17
Available P (0.001 mol L^−1 H2SO4, pH 3: mmol L^−1 kg^−1)	0.13	0.23	77
CEC (1 mol L^−1 CH3COONH4 pH 7.0: c mol kg^−1)	20.2	20.6	2
Base saturation at pH 7.0 (%)	34	73	116
Exchangeable cations:
Sodium (flame photometer: c mol kg^−1)	3.7	2.42	–35
Potassium (flame photomet.: c mol kg^−1)	0.41	0.68	66
Calcium (AAS: c mol kg^−1)	0.82	5.12	524
Magnesium (AAS: c mol kg^−1)	1.87	6.27	235
Aluminum (AAS: c mol kg^−1)	8.09	0.45	–94
Iron (AAS: c mol kg^−1)	1.57	0.76	–52
Water-soluble ions:
Sodium (flame photometer: c mol kg^−1)	3.28	2.23	–32
Potassium (flame photometer: c mol kg^−1)	0.21	0.41	95
Calcium (AAS: c mol kg^−1)	0.91	3.47	281
Magnesium (AAS: c mol kg^−1)	3.78	4.89	29
Aluminum (AAS: c mol kg^−1)	1.29	0.11	–91
Iron (AAS: c mol kg^−1)	0.33	0.42	27
Chloride (0.05 mol L^−1 AgNO3: c mol kg^−1)	3.19	2.23	–30
Sulfate (BaCl2: c mol kg^−1)	3.83	2.11	–45
^†Pyrite (FeS2) content was calculated from the total content of Fe {(Fe content/46.7) × 100, that is, FeS2 was considered to contain 46.7% Fe} in the acid sulfate soils. AAS, atomic absorption spectrophotometer; IOC, increased over control (initial value).

Tiller production

The maximum numbers of tillers per hill of 16 and 17 for BR 14 and 14 and 16 for Pizam were recorded with the application of A₂₀BS₂₀ in Tech 1 and Tech 2, respectively, followed by the treatments of A₂₀BS₁₀ (Fig. 4). The effects of the treatments in the soil in relation to the production of tillers were: Tech 2 > Tech 1, BS₂₀ > BS₁₀ and A₂₀ > A₃₀ (Fig. 4). In most cases, the productions of tillers by the two rice cultivars were found to have significant (P ≤ 0.05) positive effects with basic slag and aggregate size treatments and techniques. Most of the treatments were found to have exerted more positive effects in increasing tiller production on soil in the ridges using both the techniques, but the effect was more pronounced under Tech 2 (Fig. 4). The BR 14 rice cultivar ranked first compared with the local Pizam with respect to tiller production, although some tillers of the BR 14 rice cultivar were unproductive. Khan et al. (1996) also observed similar effects of the application of BS on the vegetative growth of rice cultivated in two saline–acid sulfate soils.

Figure 4  Influence of the two different techniques, and the basic slag (BS) and aggregate size (A) treatments on the growth and yield performance of rice grown in acid sulfate soil under field conditions. LSD, least significant difference.

Yield performance of rice

The results obtained from the different treatments on the growth performance of the two rice cultivars grown under the modified plain–ridge–ditch techniques in the Badarkhali ASS significantly (P ≤ 0.05) increased the straw and grain yields, and the improvement was more pronounced with Tech 2 (Fig. 4). This might be because of a smaller requirement of BS regarding neutralizing the acidity in the soil under Tech 2 (the pyrite or acidic layer was in the middle portion of the ridge). In Tech 1 (the acidic layer was on the top of the ridge), the oxidized layer required more BS in increasing the pH level (from 5 to 6) for optimum crop growth. The lowest quantities of straw and grain yields were recorded in the control treatment. It is noted that the complete oxidation of one mole pyrite will give 4 moles (+) of acid or 2 moles of sulfuric acid (Jaynes et al. 1984) as: FeS_2(solid=s) + 15/4O_{2(aqueous=aq)} + 7/2H₂O_(liquid=l) → Fe(OH)_3(s) + 4H⁺ _(aq) + 2SO₄ ^2– _(aq). In contrast, jarosite is an intermediate product of oxidation of pyrite as: FeS_2(s) + 15/4O_2(aq) + 5/2H₂O_(l) + 1/3 K⁺ _(aq) > KFe₃(SO₄)₂(OH)_6(s) + 3H⁺ _(aq) + 4/3SO₄ ^2– _(aq). These reaction mechanisms confirmed the facts stated above. Fageria and Baligar (1999) reported that rice was the most tolerant among the tested crops and produced maximum dry matter at a pH of 4.9. The present data on average pH value ranged from 4.0 to 5.9 (Fig. 3), where the higher rate of BS₂₀ was found to increase the soil pH to the highest level that resulted in maximum yield of rice under both techniques. The highest yields of grain and straw were recorded for the local Pizam (4.4 t ha⁻¹ grain; 4.7 t ha⁻¹ straw) compared with the HYV BR 14 (4.0 t ha⁻¹ grain; 4.1 t ha⁻¹ straw) under the treatment of A₂₀BS₂₀ in Tech 2, and this was greater than the same treatment in Tech 1 (Fig. 4). These results are also supported by Toure (1982), who reported that the yields of straw and grain of rice were increased after liming of ASSs in Senegal. The sizes of soil aggregates also influenced the growth of rice in the presence of the BS treatments. The A₂₀ aggregated soils yielded the highest production of rice, which might be because of the maximum oxidation of pyrite, release of Al and acidity from the soil, which in turn helped to increase the pH levels (Fig. 3), resulting in better availability of plant nutrients in the soil. Westerhof (1998) revealed that exchangeable bases and CEC had a positive correlation, while exchangeable Al was negatively correlated with the amount of soil in the micro-aggregate and primary particle fractions as studied in the field experiments. The plain lands in the present study had similar sizes of aggregates and treatments, but failed to obtain the expected production of rice. The general trends of the treatments on the yield of rice were A₂₀BS₂₀ > A₃₀BS₂₀ > A₃₀BS₁₀ > A₂₀BS₁₀ in the ridges for both the rice cultivars and the techniques. This sequence was attained after the release of acidity and its neutralization, where A₂₀ releases more acid than A₃₀, but BS₂₀ neutralizes more acid than BS₁₀, that is, the smaller aggregate sizes require more BS for neutralization. The higher release of acid and greater neutralization by BS₂₀ lead to a favorable condition for plant growth, while the case was reversed for A₃₀ and BS₁₀, that is, reduced acid release and reduced neutralization also resulted in suitable conditions for plant growth. However, in the case of A₂₀ and BS₁₀, greater release and less neutralization exerted unfavorable conditions for plant growth. Khan et al. (2006) reported that the larger aggregate sizes (A₃₀) with ground water at 50 cm below the surface resulted in a better response to rice under a pot experiment, which might be because of the maximum washout (leaching) of acidity and salinity, and the influence of capillary rise of sweet ground water (tap water pH 6.8) under the pot experiment. In contrast, under field conditions this capillary rise of saline–acidic ground water took place more for the larger aggregate (A₃₀) size of soils, which might have adverse effects on the rice cultivars, resulting in lower yields.

Apart from the treatments and techniques, the average grain yield increments of rice were approximately 7% higher for the local Pizam compared with the HYV BR 14 rice. The HYV rice failed to provide a better yield against an adverse environment, that is, soil salinity and acidity. However, the local Pizam cultivar exerted better growth and yield performance because of its gradual improvement of tolerance/adaptive capability under salinity and acidity conditions or relevant environmental hazards. Similar results were also reported by Kabir (2005), who showed that some HYV rice cultivars were not very effective in obtaining yields of rice comparable with the local varieties. The best harvest indexes of rice grown on the studied soil were also obtained in the treatment of A₂₀BS₂₀ > A₂₀BS₁₀ under both Tech 2 followed by Tech 1 (Fig. 4). Anderson et al. (1987) reported that the long-term use of calcium silicate slag was beneficial for the growth of sugarcane, rice and rice–sugarcane rotation crops grown on Everglades Histosols. The application of BS was reported to be effective because of the increase in the soil pH and the release of some elements, such as Ca and Mg, into the growing media as well as a large amount of Si, which is beneficial to rice growth (Khan et al. 1996, 2006). The effect of a smaller aggregate size (A₂₀) compared with a larger (A₃₀) size was more effective for rice production, which might be because of the maintenance of more favorable conditions associated with the initial fast waste-out of acidity and salinity (Khan et al. 2006).

Table 2 Influence of the different techniques, and the basic slag and aggregate size treatments on the tissue (shoot) concentrations (g kg−1) of nutrients (element basis) in rice plants grown in Badarkhali acid sulfate soil

Treatments		Technique 1
		Nitrogen		Phosphorus		Potassium		Calcium		Magnesium		Sulfur
		BR 14	Pizam	BR 14	Pizam	BR 14	Pizam	BR 14	Pizam	BR 14	Pizam	BR 14	Pizam
	Control	4.2 f	5.3 f	1.6 f	2.1 e	13.2 d	14.1 d	1.4 f	1.8 e	2.1 e	2.6 f	1.9 a	2.1 a
	A20BS10	8.1 de	8.6 de	2.1 e	3.0 d	15.9 c	17.6 c	3.6 d	3.7 d	3.4 d	4.0 d	1.5 bc	1.4 bc
Plain	A20BS20	9.6 bc	10.1 bc	3.3 bc	3.9 b	17.6 b	19.2 ab	4.1 cd	4.3 c	4.2 bc	4.7 bc	1.4 c	1.2 cd
	A30BS10	7.9 e	8.2 e	2.5 d	3.4 c	16.2 bc	17.4 c	3.1 e	3.5 d	3.2 d	3.6 e	1.6 b	1.5 b
	A30BS20	8.6 de	9.1 dc	2.9 cd	3.5 c	17.1 bc	18.2 bc	3.7 d	4.1 cd	4.1 c	4.4 c	1.5 bc	1.3 c
	A20BS10	10.1 b	8.9 d	3.4 bc	3.8 b	17.9 b	19.0 ab	4.9 bc	5.4 b	4.5 b	4.8 b	1.3 cd	1.2 cd
Ridge	A20BS20	11.3 a	11.7 a	4.4 a	4.6 a	19.7 a	20.6 a	5.8 a	6.1 a	5.1 a	5.6 a	1.1 d	1.0 d
	A30BS10	9.2 c	9.8 c	3.2 c	3.5 c	16.6 bc	17.9 bc	3.8 d	4.1 cd	3.4 d	4.1 cd	1.6 b	1.4 bc
	A30BS20	10.2 b	10.8 b	3.6 b	3.9 b	17.8 b	18.6 b	4.8 bc	5.1 bc	4.3 bc	4.3 cd	1.3 cd	1.2 cd
LSD at 5%		0.9	0.9	0.4	0.4	2.0	2.0	0.5	0.5	0.4	0.4	0.2	0.2
r-value		0.99**	0.96**	0.95**	0.93**	0.95**	0.94**	0.95**	0.87**	0.91**	0.86**	–0.81**	–0.96**
Technique 2
	Control	4.2 e	5.3 e	1.6 f	2.1 e	13.2 c	14.1 c	1.4 f	1.8 f	2.1 f	2.6 g	1.8 a	2.1 a
	A20BS10	8.4 cd	8.7 d	2.5 e	3.1 d	16.8 bc	18.3 b	3.7 de	4.0 e	3.5 d	3.9 e	1.4 c	1.4 c
Plain	A20BS20	9.8 bc	10.3 b	3.9 bc	4.2 b	18.4 ab	19.7 ab	4.4 c	4.5 d	4.4 b	4.8 c	1.1 d	1.2 d
	A30BS10	8.1 d	8.4 cd	2.8 d	3.2 d	16.7 bc	17.8 bc	3.2 e	3.6 de	3.2 de	3.5 f	1.6 b	1.6 b
	A30BS20	9.1 c	9.5 bc	3.4 c	3.6 c	17.4 b	18.5 b	3.8 de	4.0 e	4.2 c	4.6 c	1.3 cd	1.4 c
	A20BS10	10.5 b	10.9 b	4.1 b	3.7 c	18.9 ab	19.5 ab	5.2 b	5.3 b	4.4 bc	4.9 b	1.3 cd	1.2 d
Ridge	A20BS20	11.6 a	12.3 a	4.9 a	5.3 a	20.2 a	20.8 a	6.3 a	6.4 a	5.2 a	5.9 a	1.0 e	0.9 e
	A30BS10	9.6 bc	9.9 c	3.5 c	3.6 c	17.6 b	19.1 ab	4.1 d	4.2 de	3.5 d	4.3 d	1.3 cd	1.4 c
	A30BS20	10.6 b	10.6 b	3.9 bc	4.1 b	18.5 ab	19.7 ab	5.1 b	4.9 c	4.5 b	4.4 d	1.3 cd	1.1 de
LSD at 5%		1.0	1.0	0.4	0.4	2.0	2.0	0.5	0.5	0.4	0.4	0.2	0.2
r-value		0.98**	0.96**	0.90**	0.88**	0.97**	0.98**	0.93**	0.92**	0.91**	0.88**	–0.88**	–0.95**
A20 and A30, aggregate size less than 20 and 20–30 mm, respectively; BS10 and BS20, basic slag application at rates of 10 and 20 t ha^−1, respectively. Means followed by the same letter in a column or row are not significantly different at P < 0.05. *P < 0.05;
**P < 0.01.

Mineral nutrition of rice

Nitrogen, phosphorus and potassium

The highest N, P and K contents in both the BR 14 and Pizam rice shoots at maturity were mostly obtained in the A₂₀BS₂₀ treatment followed by the A₂₀BS₁₀ ≥ A₃₀BS₂₀ treatments under Tech 1 and Tech 2 (Table 2). These increments were more pronounced in the rice grown on the ridges under both techniques. Air-drying during the preparation of the ridges might affect ammonification and alkalization might also be associated with the increase in N content. The local Pizam exhibited the best responses to the uptake of these nutrients in rice plants under Tech 2. As expected, the lowest N, P and K contents in both rice plant cultivars were recorded in the control treatment (Table 2). The application of BS and aggregate size treatments under the different techniques significantly (P ≤ 0.05) increased the contents of these nutrients in the rice shoots, presumably because of reduced soil acidity and salinity as well as the release of N and P by microbial decomposition and associated soil chemistry with the treatments. Nitrogen (r = 0.99** and 0.98** for BR 14 and 0.96** and 0.96** for Pizam in Tech 1 and Tech 2, respectively), P (r = 0.95** and 0.90** for BR 14 and 0.93** and 0.97** for Pizam, respectively) and K (r = 0.95** and 0.97** for BR 14 and 0.94** and 0.98** for Pizam, respectively) contents in both rice cultivar shoots at maturity showed significant positive relationships with the grain yields of these rice cultivars and established the effectiveness of the treatments.

Calcium, magnesium and sulfur

The lowest contents of Ca and Mg in both rice cultivar tissues were recorded in the control treatment (Table 2). The Ca and Mg contents in the rice shoots increased significantly with the application of BS and aggregate size treatments under the different techniques, which created a favorable environment for the uptake of Ca and Mg by the rice plants. Sulfur content at maturity of both the rice plants grown in the control plots was high (1.8–2.1 g kg⁻¹) and was above the critical S content (0.6 g kg⁻¹; Yoshida and Chaudhry 1979). However, the S content in the rice shoots significantly (P ≤ 0.01) decreased with the application of BS and aggregate size treatments under the different techniques. The A₂₀BS₂₀ treatment recorded the maximum decrement in S content in both rice cultivars grown on the soil of the ridges under both techniques (Table 2). The S content in both the rice shoots showed a significant (P ≤ 0.01) negative relationship, while Ca and Mg contents showed a significant (P ≤ 0.01) positive relationship with the grain yields of both types of rice grown under the different techniques. The S content in the rice shoots was high because of the high concentration of sulfate (Table 1) in the growing media and did not show a positive relationship with grain production.

Conclusion

We can conclude that the application of BS₂₀ ranked first, followed by A₂₀ > A₃₀, for the reclamation and improvement of Badarkhali acid sulfate soil. The results obtained from this study can be used as a site-specific management for this type of soil. The significant (P ≤ 0.05) positive improvements of growth, yield, nutrition of rice and soil properties were more pronounced under Tech 2 than under Tech 1 using a modified plain–ridge–ditch system. The application of BS and aggregate size treatments using different techniques not only increased soil pH, but also improved the ionic balance between Ca and Mg, and markedly decreased the Fe and Al contents in the soil and plants.

The application of BS in the ASS was an effective measure and was also available at a reasonable price, which not only enabled the soil to be reclaimed, but also improved the growth and yields of the rice to optimum levels. Moreover, the modified plain–ridge–ditch techniques and the use of different sizes of aggregates contributed to the physico-chemical amendments of the soil as well as to the improved rice production. The local variety (Pizam rice) showed better growth and yield performance than the high-yielding BR 14 rice cultivar. However, for a cost–benefit analysis of these treatments in relation to acid neutralizing capacity in different fields over a long time, further studies on different soils and crops under variable climatic conditions need to be carried out.

ACKNOWLEDGMENTS

This study was carried out under the financial and technical support (1998–2005) of the Volkswagen (Ref: 1/73 802, dated 03-08-98) and the Alexander von Humboldt (AvH) foundations, respectively. We are also grateful to the German academic exchange service (Deutscher Akademischer Austausch Dienst) for providing two Sandwich Scholarships for the PhD programs of S. M. Kabir and M. M. A. Bhuiyan.