Abstract

Key Words

available potassium, potassium adsorption curves, mangrove mud, acidification, potassium deficiency, South Alligator River

Introduction

Potassium is an essential macronutrient for all plants. Plant-available K includes the K⁺ ions in the soil solution and exchangeable K adsorbed on the soil colloid surfaces. The latter form of K⁺ is readily released to solution by natural equilibrium processes in the soil, or by exchange when extracted with salt solutions. The concept of a nutrient potential was suggested by Schofield (1955) as a measure of the work a plant must do to remove nutrients from the soil. Woodruff (1955a, b) related classical thermodynamics to soil exchangeable K⁺, and Ca²⁺ plus Mg²⁺ release to the soil solution for determining the free energy of K-Ca exchange equilibria in soils. Woodruff (1955a, b) defined the energy of exchange as a measure of the chemical potential of K in the soil relative to the chemical potential of Ca in the same soil. Subsequently, Arnold (1962) showed that the difference in free energy between K and divalent ions, such as Ca and Mg, was well correlated with K uptake by ryegrass and could define the K status of soils, provided other soil factors were examined. The ability of a soil to supply K to plants is characterised by both the total amount of nutrient present (quantity, Q) and the energy level at which it is supplied (potential, P). The K⁺ potential (ΔG_K) is a free energy measure of the soil nutrient availability, expressed as a ratio of the relative activity and exchange between K⁺ and Ca²⁺ plus Mg²⁺.

Just as the soil moisture characteristic relates the free energy of soil water (i.e. moisture potential) to the amount of water held (i.e. soil moisture content), K⁺ potentials are similarly related to the total amount or quantity (Q) of exchangeable K on the soil colloid, and to its removability or intensity (I), measured by ΔG_K via the K adsorption curve (Beckett 1971). Hence, the analogous concept of the moisture characteristic is applied to determine the amount of nutrient available to a plant, called the nutrient capacity of the soil. Barrow et al. (1965) defined the nutrient capacity as the ability of a soil to resist change in potential. The soil with the least change in potential for a given change in adsorbed nutrient, in this case K, with minimal slope on the adsorption curve, is referred to as having the greatest capacity for that nutrient. Water soluble K and exchangeable K are readily available to plants and provide most of the available K measured by the K⁺ potential, but the availability of K is affected by the supply characteristics of the soil, which influence K⁺ uptake. Changes in exchangeable K must be balanced by opposite changes in other exchangeable cations, notably Ca and Mg. The relationship between quantity and potential (“adsorption curves”) has commonly been used in agricultural soils to predict K availability to crops, but has had only limited applications in natural vegetation studies under different soil conditions (Jafari 1994). The K⁺ potential is a significant means of examining K availability in the stratigraphy of acid sulfate soil (ASS) floodplains and enables some understanding of the morphodynamics, pedogenesis, and vegetation changes in the South Alligator River floodplain during the Holocene.

These sulfidic materials, known as acid sulfate soils (ASS), when first deposited from estuarine or brackish water, are often K-rich clays that support widespread mangrove swamps. The large organic inputs from the mangroves, combined with dissolved sulfate and generally reduced conditions, allows the accumulation of sulfide minerals, mainly as cubic iron pyrite (Dent 1986; Melville et al. 1991). These sulfidic materials oxidise slowly when exposed to air by evapotranspiration under natural conditions (or rapidly with human disturbance such as with excavation or drainage), forming sulfuric acid. At the pedological time scale, the annual export of K in estuaries during the monsoonal wet in northern Australia, such as measured by Woodroffe et al. (1986) and Hart et al. (1987), depletes K from the ASS of the estuarine floodplain and exports it from the floodplain surface soils to offshore sinks. This paper aims to use K⁺ potentials to explain the role of K availability in depositional and oxidation processes in the ASS development of the South Alligator River floodplain.

Methods

Soil sampling and profile descriptions

Soil samples were collected at 0.1 m depth intervals from various depths in ASS profiles on the floodplain of the South Alligator River, Kakadu National Park, in northern Australia. Field descriptions and pH measurements using a hand-held TPS MC-81 pH–Conductivity–Salinity meter were undertaken at these intervals for each soil profile. During collection, precautions were taken to avoid contamination and oxidation prior to transport to the laboratory where the soils were oven-dried at 80-85°C for at least 24 h then passed through a 2 mm sieve in preparation for analysis.

Soil solution chemical analyses

Soil samples were selected for laboratory analysis from various depths in the soil profile, based on horizon or layer differences. Characterisation of the profiles was further made in duplicate with pH measurements in H₂O and 1 M KCl extracts (UNICAM 9455 pH/ISE meter), electrical conductivity (EC) measured in H₂O extracts (TPS 900-C Conductivity–Salinity meter), and soluble basic cations measured using atomic absorption spectrophotometry (AAS) in H₂O extracts (UNICAM 929 SOLAAR AA spectrometer). All extracts used a 1:5 soil to solution ratio, after the soil had been equilibrated for 0.5 h on an end-over-end shaker. Means and standard errors were calculated for all duplicate extracts.

Potassium adsorption characteristics of soil samples were determined by the methods detailed in Keene and Melville (1999), first described in Beckett (1964a, b). Samples of soil were equilibrated in duplicate with extracting solutions containing a constant ionic strength background of 0.01 or 0.10 M CaCl₂ and increasing amounts of KCl. The concentration of the ions Ca²⁺, Mg²⁺, Na⁺ and K⁺ in cmol_c/kg was measured for each extract using AAS, based on standard methods of soil analysis. Single ion activity coefficients (γ_i) were calculated for the ions using the Davies equation, a further modification of the extended Debye-Hückel equation. The activity (a_i) of an ionic species i in solution is related to its concentration (c_i) in mol/L through its single ion activity coefficient (γ_i), and calculated using the relationship a_i = c_iγ_i (Sparks 1995). The K⁺ potential, ΔG_K, in J/mol for each extract was calculated according to the equation:

∆G_K = RT ln{a_K/(a_Ca+Mg)^½}

in which R is the molar gas constant, T is the absolute temperature, and a_K/(a_Ca+Mg)^½ is the activity ratio in (mol/L)^½. The ratio of activities describes the intensity (I) of K in a soil at equilibrium with its soil solution (Woodruff 1955a; Arnold 1962; Beckett 1964a). Curvilinear regression analysis was used to predict the relationship between the K⁺ potential, ΔG_K, and change in exchangeable K⁺, ΔK, and fit the data to a quadratic model commonly referred to as adsorption curve. Exchangeable K content was measured in 0.01 M CaCl₂ following analysis of the soil solution with no addition of K.

Soil solid phase chemical analyses

The distribution of total sulfur in the profile, mainly as the sulfide mineral pyrite, as well as total carbon were determined gravimetrically following dry combustion and analysis of the soil samples by a Leco CNS Analyser. This procedure is based on standard methods of soil analysis, using whole soil samples oven-dried at 60°C and finely ground to <250 µm size fraction.

Mineralogical analyses of the soil samples were determined by x-ray powder diffractometry, and examined on treated oriented aggregate samples of the clay fraction and on an “as received” basis for the identification of clay minerals and mineral phases, respectively. These procedures are based on standard methods of soil analysis, with whole soil samples oven-dried at 60°C and finely ground to 80 µm size fraction.

Results and discussion

ASS profile characteristics

The distribution of the various materials that typify the sedimentary environment of the South Alligator River deltaic-estuarine floodplain has been described in detail by Woodroffe et al. (1985, 1986). Alluvial floodplain deposits of black organic cracking clays, heavily marked and modified by oxidation, overlie a freshwater-marine transitional zone and blue-grey, estuarine saline mangrove muds at depth. Much of these floodplain soils exhibit a profile form comprised of five distinct soil layers, described as:

1. organic topsoil A horizon;

2. oxidised mottled B layer;

3. transition zone of seasonally oxidised sulfidic materials;

4. reduced sulfidic C layer; and

5. pre-Holocene basal sediments.

Figure 1 shows the uppermost four distinct soil layers, which include an organic topsoil A horizon, a clearly defined oxidised and mottled B layer, a seasonally oxidised sulfidic layer and a reduced sulfidic C layer. The seasonally oxidised sulfidic layer, referred to as the transition zone, was of varying thickness and lay between the sulfuric B layer and the underlying sulfidic materials of the C layer. The base of the oxidised zone was generally distinct and the lower limit of oxidation referred to the oxidation front of the soil profile as perceived by the absence of mottles and an increase in pH_F values at the base of the transition zone.

The upper and lower floodplain deposits of the South Alligator River were characterised by an organic A horizon as the uppermost soil layer with a matrix colour (moist) of very dark grey (10YR 3/1) in a light medium clay-textured soil containing abundant root material. This layer often had distinct orange mottles (5YR 5/8) of Fe oxides associated with fine roots, large organic matter content evident as total C (Figure 1) and a massive cracking pattern, characteristic of these surface clays. A heavily oxidised and strongly acid B layer of 0.5 to 1 m thick overlay a transitional zone forming a partially oxidised diffuse boundary of similar thickness. The upper 1 to 2 m of the floodplain deposits were described as very dark to dark grey (2.5Y 3/1 to 5Y 4/1) medium clays, with up to 60% pale yellow and orange red mottles of jarosite and Fe oxides, respectively. Identified as an actual acid sulfate soil (AASS) or sulfuric horizon, this layer had a pH_F ranging from below 3.70 (mean pH_F 4.66) and minimal total S content due to oxidation (Figure 1). The transition zone was a gradational horizon, with a matrix colour (moist) light brownish or olive grey (5Y 6/2) and a medium-heavy clay texture, exhibiting partial oxidation and some reduced components of the subsoil. There was evidence of oxidation, with heavy mottling generally distributed along cracks and relict root channels, concurrent with an increasing pH_F from 4.04 to 6.87 (mean pH_F 5.44) and reduced total C and S contents (Figure 1). The materials beneath the oxidised zone were generally different from those at the surface. Identified as a potential acid sulfate soil (PASS) or sulfidic material, the sulfidic C layer was characterised by estuarine saline mangrove muds or undifferentiated estuarine sediments. The reduced C layer was typically a silty clay-textured material, with a matrix colour (moist) of dark grey to light blue-grey (5Y 4/1 to 5B 5/1), often containing abundant organic material. Much of the South Alligator River plains are underlain by blue-grey muds containing wood fragments and shown by pollen analysis to have been deposited beneath mangrove forests (Woodroffe et al. 1985, 1986), which was reflected in the large total C contents in this layer (Figure 1). Significantly, this layer showed no evidence of oxidation products, with large total S content and no jarosite or Fe mottling occurring within the sulfidic materials and a pH_F ranging from 5.25 to 7.35 (mean pH_F 6.77) (Figure 1).

Figure 1. Field pH and total sulfur and carbon contents (g/kg) of a South Alligator River soil profile. Reference lines show the boundary between identified acid sulfate soil layers.

The results from XRD analysis enabled a general interpretation of the clay mineral composition of the South Alligator River floodplain soils. The dominance of kaolinite in the clays was characteristic of these tropical and weathered acid soils. Figure 2 reveals the pattern of decreasing kaolinite from the organic and oxidised surface soils to the transition zone. Below the oxidation front, kaolinite increased in the sulfidic C layer up to 60% of the clay fraction. However, the soils were also characterised by similar amounts of smectite. The presence of expandable 2:1 clay minerals such as smectite reflected the shrink swell nature of the surface Vertosols. Relative amounts of smectite in the clay fraction increased from the organic topsoil to the oxidised B layer and subsequently decreased through the transition zone to the sulfidic C layer for these soils. Illite was a considerably smaller proportion the clay fraction and relatively constant over depth, increasing from around 10% in the organic and oxidised surface soils up to 20% in the sulfidic C layer.

Figure 2. Mineral composition of the clay fraction for a South Alligator River soil profile. Reference lines show the boundary between identified acid sulfate soil layers.

Quantity/potential relationships as K adsorption curves

From measurements of the concentration of cations in the equilibrium solution, values of ΔG_K (J/mol) and ΔK (cmol_c/kg) were calculated and illustrated for soils of the South Alligator River floodplain. The ΔK values were determined from the analysed concentration of K after equilibration, relative to the initial concentration. Negative values of ΔK are where the soil has adsorped K from the equilibrating solution. The ΔG_K values at ΔK=0 represent the K⁺ potential that existed in the soil at the time of sampling. Values of ΔG_K closer to zero represent K held less strongly by the soil and therefore more available for plant uptake. The slope of these adsorption curves around ΔK=0 (Barrow et al.’s, 1965, “nutrient capacity”) is also important because it describes the relative change that will occur in the availability as K is added to the soil (e.g. by fertiliser causing a less negative ΔG_K), or removed by plant uptake or leaching (causing more negative ΔG_K). The K adsorption curves depicting Q/P relationships for the selected soils are shown in Figure 3, which indicated that potentials diverged as K⁺ was removed from the exchange complex. In terms of K⁺ exchange equilibria, the Q/P relationships showed that there were significant differences in K⁺ potentials between the stratigraphic units of the profile, and in the slopes of the adsorption curves.

Figure 3. Relationship between K⁺ potential and the change in exchangeable K for a South Alligator River floodplain soil profile. The curvilinear lines are regression fits to the data. The vertical reference line shows the equilibrium value of ΔG_K when there is no exchange by K (i.e. ∆K=0).

From Figure 3, the K adsorption curves for sulfidic materials had K at less negative (more available) ΔG_K values. The lack of a distinct Q/P relationship was evident at 1.9-2.1 m depth, with the regression only accounting for 76% of the variation in ∆G_K. Defined as a mangrove mud, this blue-grey silty clay contained abundant organic material, which probably affected the exchange of K⁺ in the soil. The cation exchange for this soil may have been at least in part due to the large amounts of organic matter (see Figure 1). However, this soil had a large capacity for K, as shown by the greater range of ∆K values with minimal change in potential. It was evident that much less work was required in order to remove K⁺ from this organic silty clay than the underlying mangrove muds between 2.1-4.2 m depth. For these silty clays containing sparse organic material, with potential declining as K was removed, K⁺ was held strongly to the clay and more work was required to extract K⁺. This soil may have been influenced by oxidation or some reworking of the basal sediments into the overlying stratigraphic units (Woodroffe et al. 1986), which would account for the more negative K⁺ potential in the mangrove muds. Thus, there existed a significant difference in K⁺ potentials for the mangrove muds of the sulfidic layer, with less negative ∆G_K values in the blue-grey silty clay containing abundant organic material.

The K in the surface soils, in contrast to that in the sulfidic subsoil materials, had much more negative (less available) ΔG_K values. For these soils of the organic topsoil and the oxidised B layer comprising the upper floodplain clays, reductions in the concentration of K in the solution phase related to a lower ability to extract K⁺ from the soil, which was reflected in the more negative K⁺ potentials (Figure 3). For the organic and oxidised surface soils, the removal of only 0.35 cmol_c of K/kg of soil reduced the K⁺ potential by 1886 J/mol to –14574 J/mol. By contrast, in the sulfidic material at 2.1-4.2 m depth, the removal of a similar amount at 0.26 cmol_c of K/kg caused the K⁺ potential to decrease by only 611 J/mol to –11446 J/mol. Therefore, the oxidised surface soils have K held much more strongly and a much reduced ability to supply K to plants. The mangrove muds on the other hand represent a significant K store.

K availability defined by the equilibrium K⁺ potential

Following the interpolation of the potential of soil K describing availability, K⁺ potentials at soil equilibrium have been summarised for the stratigraphic layers of the South Alligator River floodplain soils. The ΔK axis values can be rescaled to exchangeable K by adding the difference in the ΔK value for each equilibration solution point on the adsorption curve, to the mean exchangeable K concentration of the equilibrated solution containing no added K. These exchangeable and soluble K contents (cmol_c/kg), and ΔG_K values at ΔK=0 (J/mol) for selected soil samples analysed are shown in Table 1.

Table 1. Mean equilibrium K⁺ potential, ∆G_Kº, for the ion pair K-(Ca+Mg) and selected mean soil properties for a representative ASS profile on the South Alligator River floodplain.

Depth (m)	Field pH	EC1:5 (dS/m)	Soluble Na+ (cmolc/kg)	Soluble K+ (cmolc/kg)	Exch K+ (cmolc/kg)	C/S ratio	Illite (%)	∆GKº (J/mol)
								mean	se
0.0-0.1	4.57	0.64	2.84	0.04	0.12	34.48	9	-14510
0.2-0.7	5.12	2.28	9.23	0.10	0.13	15.6	10	-14175
0.8-1.0	5.77	3.43	13.26	0.15	0.26	5.97	13	-13213
1.0-1.6	6.56	4.75	17.26	0.22	0.28	2.78	13	-12933
1.9-2.1	7.23	10.69	33.79	0.73	0.85	0.67	23	-10891
2.1-4.2	7.10	9.76	31.94	0.67	0.23	1.01	19	-18290

These results indicate that the undifferentiated transitional sediments and the underlying organic muds exhibited a considerably less negative equilibrium K⁺ potential than the other stratigraphic layers, which reflected a greater availability of K. A less negative potential was evident for the mangrove muds beneath the floodplain surface, however a large ∆G_Kº value of –18290 J/mol for the blue-grey silty clay containing sparse organic material between 2.1-4.2 m depth was different from all other values within this stratigraphic unit. The mottled organic floodplain clays that characterise the upper floodplain were found to have a very negative K⁺ potential of –14175 J/mol, suggesting that these oxidised surface soils were K deficient. Table 1 also shows that there was a clear similarity in profile trends between the exchangeable K content and the ΔG_K value at ΔK=0, probably due to the clayey nature of these materials. At all depths, exchangeable K content was greater than the soluble K content, except for the underlying mangrove muds between 2.1-4.2 m depth, where leaching from the upper parts of the profile may have had a greater influence.

Processes controlling K availability in ASS profiles

With significant differences in the K adsorption curves and equilibrium K⁺ potentials for these acid sulfate soils, profile variations in other soil properties were examined and are shown in Table 1. It is evident that the more negative ∆G_Kº values of the oxidised upper floodplain surface soils were associated with small pH_F values, large organic C, and small illite contents. The underlying mangrove muds were characterised by large pH_F values, large EC and soluble cations, and large illite contents, and this reflected less negative ∆G_Kº values for all soil samples analysed with the exception of the sediment between 2.1-4.2 m depth (Table 1).

Kaolinite content decreased from the underlying mangrove muds to the oxidised floodplain surface clays, concomitant with decreasing soluble basic cation concentration as salts were leached from the profiles. Smectite content of the oxidised floodplain clays and transitional materials increased from the mangrove muds, as shown in Figure 2. More negative equilibrium K⁺ potentials were associated with a greater smectite content of the soils, indicating a deficiency of K. The relative abundance of smectite in the oxidised floodplain clay surface reflected the shrink swell capacity of these expanding 2:1 clay minerals. The formation of ASS has been associated with an increase in montmorillonite, and the presence of the 1:1 layer kaolinite is often at the expense of 2:1 clay minerals in ASS (van Breemen 1973). The most important K-containing clay mineral is the 2:1 layer silicate illite, which is considered responsible for K fixation and release (Rich 1972; Goulding 1987). Illite content of the clays increased with depth (Table 1 and Figure 2), which reflected increased K content and greater K availability in the mangrove muds. Relatively smaller amounts of illite were present in the oxidised floodplain clay surface, suggesting a K deficiency in these soils. These results suggest that the acid weathering of K-bearing silicate minerals such as illite has increased amounts of smectite and kaolinite, and decreased the availability of K by providing much of the K⁺ in the oxidation product jarosite in the AASS layer as well as releasing K⁺ to the soil solution (van Breemen 1973; Fanning and Fanning 1989; Fanning et al. 1989).

During the oxidation of sulfide minerals, the acidity generated reacts with and changes the host clay minerals from K-rich illitic clays to K-poorer smectitic clays. This transformation was facilitated by the very small activity of K in the soil solution, relative to the activity of replacing cations such as Ca or Mg (Fanning and Fanning 1989). The release to the soil solution of plant-available K from silicate minerals by acid attack during pedogenesis provides the K necessary to form jarosite in the upper sulfuric horizon. The formation of jarosite under acid conditions involves a high preference for K⁺ derived from the soil exchange complex, from the soil solution and from K-bearing minerals, mainly mica and feldspars (van Breemen 1973; Fanning and Fanning 1989; Fanning et al. 1989). For the South Alligator River floodplain soils, K deficiency was associated with the formation of jarosite, with its crystal structure serving as a K sink for most of the K released from silicate minerals by acid attack, and because of the very small activity of K⁺ in the soil solution (van Breemen 1973; Fanning and Fanning 1989; Fanning et al. 1989). This reduces the amount of K that is readily available for plant growth in the upper floodplain clays. As this jarosite subsequently hydrolyses, the K is released to enable its upward leaching and export from the soil profile via diffusion into flood waters during the wet season and, to a lesser extent, by leaching. Upward leaching and uptake by plants of K is promoted by soil ripening, which includes the further oxidation of jarosite, decreasing the available K level in the soil solution and contributing to the decrease in K availability from the ASS profile.

Drainage systems and agricultural practices that export increased amounts of K can exacerbate the natural K deficiency that is likely to exist in these ASS landscapes. Estuarine sediment deposited across the floodplain during the wet season had subsequently been oxidised because of seasonal water table fluctuations and decreasing tidal influence over the long term. Beneath the floodplain surface, partly ripe estuarine sediments represented a sedimentary and pedogenic transition from the reduced estuarine blue-grey clays of the PASS layer to the oxidised floodplain clays of the upper AASS layer. Mottles were produced because of the oxidation of pyrite from the underlying mangrove sediments, with the formation of Fe products such as jarosite occurring during pedogenesis. Organic rich alluvium, the so-called “black soil” of the region, comprises the dark brown to black clays that veneer the floodplain surface and overlie ripe AASS. These materials are not of immediate upland fluviatile origin but are marine-sourced sediments deposited in wet season freshwater flows with large organic inputs. Unoxidised “big swamp” subsoil sediments (Woodroffe et al. 1985), dominated by kaolinite and increased illite contents, exhibited a large availability of K as shown by the less negative ∆G_Kº values (Table 1). There was a general depletion of K throughout the oxidised floodplain clay surface, which was dominated by smectitic clays.

Conclusion

Results from the K adsorption curves for the acid sulfate soils of the South Alligator River floodplain have shown that the oxidised upper part of the sediment profile has decreased equilibrium values of K⁺ potential. The oxidation of sulfide-bearing minerals and associated acidification during pedological development depleted K from the clays of the estuarine floodplain surface, and resulted in changes to the clay mineral structure. Leaching and uptake by plants of K from this oxidised, mottled layer decreased the K level and have contributed to the decrease in K availability. Potassium deficiency was associated with the formation of the sulfide mineral oxidation product jarosite, which acts as a continuous sink for K in the upper sulfuric horizon, and reduced the amount of K that was readily available for plant growth. This contrasted with the sulfidic sediments of the subsoil, which exhibited considerably less negative equilibrium K⁺ potential values and revealed a significantly greater availability of K. Under acid conditions, the weathering of illitic clays of the sulfidic layer releases exchangeable K to the soil solution, transforming the K-rich illite into the K-deficient smectitic clays of the upper oxidised layer and depleting K from the clays of the floodplain surface by the formation of jarosite, and by leaching and plant uptake of K. These pedogenic changes in clay mineralogy are accompanied by natural hydrological and oxidation processes in an estuarine environment with markedly wet and dry seasons that cause upward leaching and export of K from these ASS landscapes.

References

Arnold PW (1962) The potassium status of some English soils considered as a problem of energy relationships. Proceedings of the Fertilizer Society 72, 25-43.

Barrow NJ, Ozanne PG, Shaw TC (1965) Nutrient potential and capacity. I. The concepts of nutrient potential and capacity and their application to soil potassium and phosphorus. Australian Journal of Agricultural Research 16, 61-76.

Beckett PHT (1964a) Studies on soil potassium. I. Confirmation of the ratio law: measurement of potassium potential. Journal of Soil Science 15, 1-8.

Beckett PHT (1964b) Studies on soil potassium. II. The 'immediate' Q/I relations of labile potassium in the soil. Journal of Soil Science 15, 9-23.

Beckett PHT (1971) Potassium potentials - a review. Potash Review 5, 1-41.

Dent D (1986) 'Acid sulphate soils: a baseline for research and development.' ILRI Publication 39. (International Institute for Land Reclamation and Improvement: Wageningen, The Netherlands)

Fanning DS, Fanning MCB (1989) 'Soil morphology, genesis, and classification.' (John Wiley and Sons: New York)

Fanning DS, Keramidas VZ, El-Desoky MA (1989) Micas. In 'Minerals in soil environments'. (Eds JB Dixon and SB Weed) SSSA Book Series No. 1. pp. 551-634. (Soil Science Society of America: Madison, WI)

Goulding KWT (1987) Potassium fixation and release. In 'Proceedings of the 20th Colloquium of the International Potash Institute'. pp. 137-154. (International Potash Institute: Berne)

Hart BT, Ottoway EM, Noller BN (1987) Magela Creek system, northern Australia. I. 1982-83 wet season water quality. Australian Journal of Marine and Freshwater Research 38, 261-288.

Jafari M (1994) Saltbush growth and soil properties from southeastern Australian semi-arid rangelands with dieback and salinity problems. PhD thesis, The University of New South Wales, Sydney, Australia.

Keene AF, Melville MD (1999) Role of nutrients in the delimitation of the grey mangrove, Avicennia marina, in estuarine sediments. In 'Geodiversity: readings in Australian geography at the close of the 20th century'. (Eds JA Kesby, JM Stanley, RF McLean and LJ Olive) Special Publication Series No. 6, pp. 45-53. (University of New South Wales, Australian Defence Force Academy: Canberra)

Melville MD, White I, Willett IR (1991) Problems of acid sulphate soils and water degradation in Holocene pyritic systems. In 'Applied Quaternary Studies'. (Eds G Brierley and J Chappell) Papers presented to a workshop at the Australian National University July 2-3, 1990. pp. 89-94. (Australian National University: Canberra)

Rich CI (1972) Potassium in soil minerals. In 'Proceedings of the 9th Colloquium of the International Potash Institute'. Potassium in soil. pp. 15-31. (International Potash Institute: Berne)

Schofield RK (1955) Can a precise meaning be given to available soil phosphorus? Soils and Fertilizers 18, 373-375.

Sparks DL (1995) 'Environmental soil chemistry.' (Academic Press: San Diego, CA)

van Breemen N (1973) Soil forming processes in acid sulphate soils. In 'Acid sulphate soils'. (Ed H Dost) Proceedings of an international symposium on acid sulphate soils. ILRI Publication 18. Vol 1, pp. 66-128. (International Institute for Land Reclamation and Improvement: Wageningen)

Woodroffe CD, Thom BG, Chappell J (1985) Development of widespread mangrove swamps in mid-Holocene times in northern Australia. Nature 317, 711-713.

Woodroffe CD, Chappell JMA, Thom BG, Wallensky E (1986) 'Geomorphological dynamics and evolution of the South Alligator tidal river and plains, Northern Territory.' Mangrove Monograph No. 3. (Australian National University North Australia Research Unit: Canberra)

Woodruff CM (1955a) The energies of replacement of calcium by potassium in soils. Soil Science Society of America Proceedings 19, 167-171.

Woodruff CM (1955b) Ionic equilibria between clay and dilute salt solutions. Soil Science Society of America Proceedings 19, 36-40.