Magnesium fertiliser dissolution rates in pumice soils under Pinus radiata.
Magnesium (Mg) deficiency is common in a number of forest regions in the world. In New Zealand it has been linked to a condition in Pinus radiata called upper mid-crown yellowing (Payn 1991; Beets and Jokela 1994). In Europe it has been linked to `new type forest decline' or `crown thinning' in stands of Norway spruce (Picea abies) (Landmann et al. 1997). In north-eastern North America, Mg deficiency has been found in declining stands of sugar maple (Acer saccharum March.), yellow birch (Betula alleghensis), red maple (Acer rubrum), balsam fir (Abies balsamea), and mountain maple (Acer spicatum) (Landmann et al. 1997). Research by Will (1961) and Hunter et al. (1986) demonstrated that severe Mg deficiency in P. radiata could be corrected by the application of Mg fertiliser. The effectiveness of most Mg fertilisers in correcting Mg deficiency largely depends on their rates of dissolution in soils. An ideal Mg fertiliser for forestry is one which releases Mg slowly but in adequate amounts to satisfy the Mg needs of the tree over many years. More recently, trials have been initiated by New Zealand Forest Research to test the effectiveness of Mg fertiliser in improving the Mg status of P. radiata. Two of these trials on pumice soils in Kaingaroa Forest have shown that calcined magnesite (calmag) applied as 2-5 mm chips at a rate of 150 kg Mg/ha is very effective in increasing soil exchangeable Mg concentration within 2 years of application (Mitchell et al. 1999). Further analysis of soils from the above-mentioned trials indicated that in the almost 2 years since calmag fertiliser application, approximately 90% of the fertiliser had dissolved. This, however, does not provide any information on the amount of fertiliser dissolved at various times since application, or any indication of the rate of dissolution of any other Mg fertiliser products currently available.
Dissolution of Mg fertilisers is expected to be influenced by several factors such as the composition of the fertiliser, soil acidity, soil moisture, surface area of the fertiliser particle, and diffusion of dissolved constituents away from the fertiliser and protons to the surface of the fertiliser.
Several of the commonly used Mg fertilisers (MgO, Mg[CO.sub.3], Ca[CO.sub.3].Mg[CO.sub.3]) are alkaline and have similar reactions in soils to limestone. Therefore, any models developed to describe the dissolution of limestone (Ca[CO.sub.3]) could be expected to also apply to the dissolution of these fertilisers. To describe limestone dissolution, Elphick (1955) developed the hypothesis of equal reduction, which postulates that the rate of particle-diameter reduction is proportional to the surface area of the limestone particles. From Elphick's work, Swartzendruber and Barber (1965) developed mathematically a cubic equation and tested this hypothesis. They found that the rate of dissolution of limestone particles followed the cubic equation but the specific fertiliser dissolution rate constant ([micro]g limestone dissolved per [cm.sup.2] surface area of limestone particles per day) increased with decreases in particle size. They explained the variation in specific fertiliser dissolution rate constant as probably due to the skewed nature of particle size distribution within a size class.
Laboratory incubation studies conducted on a range of Mg fertilisers of varying particle size, mixed with 2 pumice soils (pH 4.6-5.1), have shown that the rate of dissolution was in the sequence: Epsom salts [is greater than or equal to] fine calcined magnesite [is greater than] coarse calcined magnesite, coarse granmag (partially acidulated and granulated calcined magnesite), fine dolomite [is greater than] fused magnesium phosphate (Loganathan et al. 1999). In another incubation study, Heming and Hollis (1995) found that the rate of dissolution of Mg fertilisers added to 5 soils from the south of England (pH 6.2-8.2) was in the sequence: kieserite granules (1-3 mm) [is greater than] calcined magnesite powder ([is less than] 1 mm) [is greater than] calcined magnesite granules (1-3 mm) [is greater than] magnesian limestone (65% [is less than] 1.5 mm and 35% [is less than] 150 [micro]m).
To our knowledge, however, no long-term studies on determining the rate of dissolution of Mg fertilisers under field conditions have been published. This is probably because the commonly used laboratory method of determining the rate of dissolution of Mg, by measuring the increase in dissolved Mg in soils, cannot be applied to the field situation where losses of dissolved Mg from the site of application by plant uptake and leaching could lead to errors in determining the amount of dissolved fertiliser. Recently, a method has been developed to determine the rate of Mg dissolution by measuring the amount of undissolved fertiliser Mg remaining in soils and subtracting it from the amount of Mg applied (Loganathan et al. 1999). This method is expected to accurately determine the rate of Mg dissolution under field conditions.
The objectives of this study were: to compare the rates of dissolution of a range of Mg fertilisers, with and without litter, in a pumice soil under radiata pines; to investigate the effect of different forms of Mg fertilisers on soil exchangeable Mg and pH; and to test whether Mg fertiliser dissolution can be explained by a cubic model based on the hypothesis of equal reduction.
Materials and methods
Fertilisers and their characteristics used in this study are shown in Table 1. Calcined magnesite (calmag, MgO) is made by calcining magnesite (Mg[CO.sub.3]) to increase the Mg content and to produce a hardened coarser material which is more suitable for blending with other fertilisers and for aerial application. Two particle sizes (1-2 mm and 2-4 mm) of this fertiliser were chosen for this study to investigate the effect of particle size on the rate of dissolution.
Table 1. Characteristics of the fertilisers used Granule size or grade Fertiliser Symbol (mm) Calcined magnesite (calmag 1, MgO) CM1 1-2 Calcined magnesite (calmag 2, MgO) CM2 2-4 Granulated calmag 20% acidulation GM 2-4 (granmag, Mg[SO.sub.4].MgO) Dolomite (Ca[CO.sub.3].Mg[CO.sub.3]) Dol Forestry(C) Epsom salts (Mg[SO.sub.4].7[H.sub.2]O) Eps <0.25 Total Liming value(A) Mg (CaO equiv. (%) per kg Fertiliser fertiliser) Calcined magnesite (calmag 1, MgO) 51 1.389 Calcined magnesite (calmag 2, MgO) 51 1.389 Granulated calmag 20% acidulation 34 1.111 (granmag, Mg[SO.sub.4].MgO) Dolomite (Ca[CO.sub.3].Mg[CO.sub.3]) 11 0.612 Epsom salts (Mg[SO.sub.4].7[H.sub.2]O) 10 0 Liming value(B) Price for (CaO equiv, bagged per kg fertiliser Fertiliser Mg applied) (NZ$/kg Mg) Calcined magnesite (calmag 1, MgO) 2.72 1.0 Calcined magnesite (calmag 2, MgO) 2.72 1.0 Granulated calmag 20% acidulation 3.27 1.8 (granmag, Mg[SO.sub.4].MgO) Dolomite (Ca[CO.sub.3].Mg[CO.sub.3]) 5.56 2.2 Epsom salts (Mg[SO.sub.4].7[H.sub.2]O) 0 6.5
(A) Buckman and Brady (1960); for Granmag, 80% of the liming value of CM1 (and CM2) was considered since the 20% Mg[SO.sub.4] component has no liming value.
(B) Liming value per kg of fertiliser divided by the fractional weight of Mg in the fertiliser.
(C) Particle size distribution: 0.06-0.25 (mm) 37%; 0.25-0.5 10%; 0.5-1 18%; 1-2 34%; 2-3 1%.
Granmag is a fertiliser made by 20% acidulation of finely divided calmag with [H.sub.2][SO.sub.4], thereby increasing the water solubility of a portion of the fertiliser Mg and producing a granular material more suitable for blending with other fertilisers and for aerial application (Loganathan et al. 1999). Granmag contains both fast release Mg as Mg[SO.sub.4] and slow release Mg as MgO.
Forestry grade dolomite fertiliser consists of a range of particle sizes from very fine (0.06 mm) to coarse (3 mm). A breakdown of the particle size distribution is shown in Table 1.
When deciding which fertiliser is best, the fertiliser cost per unit weight of Mg should be considered along with the dissolution characteristics. Of the fertilisers used in this study, Epsom salts was the most expensive per unit weight of Mg and calmag the least expensive (Table 1).
Field trial site and soil description
This study was established in Kaingaroa forest (near Rotorua, New Zealand) in the buffer zones of the control plots of Forest Research (FR) trial FR190/5 in compartment 1079. The site is a second rotation stand of 20-year-old P. radiata located in northern Kaingaroa Forest and had no history of fertiliser use for at least 20 years. The soil in this trial site belongs to the order Pumice Soils in the New Zealand soil classification (Hewitt 1993) and is a member of the Kaingaroa series. It is classified as Typic Udivitrand in US Soil Taxonomy. These soils have developed from rhyolitic flow-tephra deposits from the latter stages of the Taupo eruption (Rijkse 1988), 1800 years BP. The soil (0-10 cm) had a pH in water (1:2.5 w/w soil: water ratio) of 5.2 and organic carbon content of 4.7%. It had exchangeable Mg of 0.6, Ca of 3.9, and K of 0.6 [cmol.sub.c]/kg and effective CEC (exchangeable Mg + Ca + K + Na + Al) of 6.9 [cmol.sub.c]/kg.
Field trial establishment
Split plots (2 m by 2 m) with and without litter removed (1 m separating with- and without-litter plots) were located equidistant from the surrounding trees (Fig. 1). The trial consisted of 5 fertiliser treatments (Table 1) and a control (no fertiliser) treatment. Each treatment was replicated 4 times and randomly allocated to the various plots to give a total of 48 plots. Each replicate was located in a different FR trial control plot to give 4 blocks of 12 plots. As-received fertiliser materials (except for the forestry grade dolomite and Epsom salts) were passed through sieves of specific openings to obtain the specific particle sizes used in this study (Table 1). The forestry grade dolomite and Epsom salts were applied in the as-received form. Fertilisers were applied at a rate equivalent to 200 kg Mg/ha.
[Figure 1 ILLUSTRATION OMITTED]
Soil sampling and analysis
One day after fertiliser application, on 20 September 1996, soil samples were collected from each of the plots by taking 8 soil core (2.5-cm-diameter) samples per plot to a depth of 5 or 10 cm. Soil samples have continued to be collected on a regular basis at 3, 6, 12, 18, and 27 months after application of Mg fertilisers. The samplings at 1 day and 3 and 6 months were collected to a depth of 5 cm; the remaining samplings were to a depth of 10 cm as it was believed that the fertiliser may have moved down the soil profile with increased time. Soil samples were air-dried, ground to pass through a 2-mm sieve, and analysed for dissolved Mg and undissolved fertiliser Mg as per the sequential extraction method of Loganathan et al. (1999). In this method the soil samples were extracted with 0.25 M Ba[Cl.sub.2]-0.2 M Triethanolamine (TEA) at pH 8.2, to determine apparent Mg dissolution, followed by extraction with 0.5 M HCl to determine partially dissolved Mg and finally with 2 M HCl to determine undissolved Mg. The amounts of apparent Mg fertiliser dissolution and partially dissolved and undissolved fertiliser Mg in the soils were calculated by subtracting the amounts of Mg extracted by the respective chemicals in the control (no fertiliser) treatment from those in the soils treated with Mg fertilisers. The true dissolution of Mg is determined by subtracting the partially dissolved and undissolved Mg from Mg added in the fertiliser. Soil pH in water (1:2.5 w/w) was also determined (Blackmore et al. 1987).
Results were tested for significant differences (P [is less than] 0.05) between treatment means and for fertiliser source, litter, and time of sampling interactions by analysis of variance using the statistical package SAS for Windows, Version 6.12 (SAS 1996).
The liming value, expressed in terms of the CaO equivalent, was calculated for the calmag, granmag, and dolomite fertiliser treatments by dividing the molecular weight of CaO by the molecular weight of the liming material (MgO or Ca[CO.sub.3].Mg[CO.sub.3]) in the fertiliser (Table 1) (Buckman and Brady 1960). This value was used to calculate the liming value of the fertiliser in CaO equivalents expressed per kg Mg applied. This allows fertilisers of differing composition to be compared on a common basis.
Cubic model of fertiliser dissolution
Swartzendruber and Barber (1965) derived the following cubic equation for the dissolution of limestone assuming that the dissolution follows the equal-reduction hypothesis of Elphick (1955):
(1) (1 - u) 1/3 = 1 - ct
(2) c = 2k/[Rho][D.sub.g]
and u is the fractional mass dissolution, k is the specific limestone dissolution rate constant, [Rho] is the panicle density and [D.sub.g] is the geometric mean panicle diameter, and t is time. The model further assumes that the initial mass of limestone is present in the form of spheres of uniform size, density, and composition, and that the rate of limestone mass dissolution is directly proportional to the total instantaneous surface area of the limestone spheres. If the equal-reduction hypothesis is valid, a plot of [(1 -u).sup.1/3] against t should be a straight line of slope -c with a Y intercept of 1. Furthermore, a plot of c v. 1/[D.sub.g] should be a straight line through the origin. The specific limestone dissolution rate constant, k, can then be calculated from the slope of this line.
As dolomite has chemical properties similar to limestone particles, Eqn 1 is used in this paper to describe the dissolution of this fertiliser.
Like limestone dissolution, the rate of oxidation of elemental sulfur ([S.sup.0]) in soils is also proportional to the surface area of the sulfur panicles. During the last decade, cubic equations have been used successfully to describe the rate of [S.sup.0] oxidation (McCaskill and Blair 1989; Chatupote 1990; Watkinson and Blair 1993; Hedley et al. 1995). As the dolomite used in this study had a range of panicle size fractions, specific fertiliser dissolution rate constants for this fertiliser, as well as for the others (calmag and granmag), were determined using an approach based on the [S.sup.0] oxidation model derived by Chatupote (1990), which considers oxidation of [S.sup.0] fertiliser with a range of particle size fractions.
The model of McCaskill and Blair (1989) assumed that [S.sup.0] particles were spheres and defined [S.sup.o] oxidation rate in terms of the decrease in particle radius ([Delta]r) for each time increment ([Delta]t):
(3) [Delta]r/[Delta]t = k/[Rho]
where k is the specific [S.sup.0] oxidation rate constant and [Rho] is the density of [S.sup.0] particles. The radius of particles after [t.sub.n] days ([MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]) is given by:
(4) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
where [r.sub.t0] is the initial particle radius, [S.sub.t0] is the initial amount of [S.sup.0] and [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is the amount of [S.sup.0] remaining after [t.sub.n] days. As [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], Eqn 4 can be rewritten as follows:
(5) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
This equation is similar to Eqn 1 of Swartzendruber and Barber (1965) used for limestone dissolution. Chatupote (1990) and Hedley et al. (1995) modified this equation and used it with Eqn 3 to develop a computer-based iterative procedure for the calculation of k. The iterative procedure requires the user to input an initial guess for a probable k to fit the [S.sup.0] oxidation data. This probable k is used to calculate [Delta]r/[Delta]t for the observed data using Eqn 3. The amounts of [S.sup.0] remaining ([MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]) at the end of each day over the trial period are calculated and those relevant to the sampling dates are compared with the observed [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] values. Using the method of least squares iteration, a new value for the probable k is selected and the iteration continues until changes in [Delta]r/[Delta]t fall below 0.01% of the r value. The k value at this iteration step is considered to be the k for the observed oxidation. The computer program was further modified by Chatupote (1990) and Hedley et al. (1995) to take account of [S.sup.0] particles consisting of a range of particle sizes. The initial geometric mean radius of the particles in each size class was input into the iteration. The constraint here is that the boundaries of each particle class should not differ more than 2-fold. The iteration procedure then subtracts [Delta]r from the initial geometric mean radius in each size class to obtain a new mean radius for that size class. This is carried out on a cumulative basis for each day of the trial period. The sum total of all [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] for all particle sizes is then calculated and compared with observed values and the k calculated in the same way as for a single particle size class described above. This computer program was used to calculate the Mg fertiliser specific dissolution rate constants in terms of [micro]g fertiliser dissolved per [cm.sup.2] of surface area of fertiliser particles per day for the field and laboratory trials.
Using the k values the model was also used to predict the percent of fertiliser dissolved at various periods of the trial according to the procedure used for determining percentage of [S.sup.0] oxidation (Hedley et al. 1995).
Laboratory incubation of dolomite fertiliser in soils
A laboratory incubation study was undertaken to investigate the effect of particle size on the rate of dissolution of dolomite fertiliser and to test the validity of the cubic model in explaining the rate of dissolution of the different particle size fractions of dolomite. As-received forestry grade dolomite was passed through sieves of varying fineness to determine its particle size distribution (Table 1). A soil sample (0-10 cm) taken from the same site as the field dissolution study was air-dried and ground to pass through a 2-mm sieve, and subsamples were used for the incubation study. Three replicates of each size fraction of dolomite fertiliser were mixed with 60 g (oven dry) soil at a rate equivalent to about 200 kg Mg/ha (10 cm depth of mixing, 0.7 g/[cm.sup.3] bulk density). The soil and fertiliser mix was incubated at 60% (w/w) moisture content (approximately field capacity) at 20 [+ or -] 2 [degrees] C. Percentage dissolution (Loganathan et al. 1999) was determined at 28, 76, and 120 days after mixing of fertilisers.
Results and discussion
Dissolution of Mg fertiliser infield trial
The percentage of apparent dissolution of fertiliser Mg and partially dissolved and undissolved (residual) fertiliser Mg, as determined by the sequential extraction method of Loganathan et al. (1999), 6 months after fertiliser application is shown in Fig. 2. The weakness of the conventional method (Heming and Hollis 1995) of determining Mg dissolution in field trials by only measuring the increases in exchangeable Mg over the control (no fertiliser) treatment can be illustrated using these data. For example, in the Epsom salts treatments the increase in exchangeable Mg due to fertiliser application (Ba[Cl.sub.2] TEA-Mg in fertilised soil minus those in unfertilised soils) is only 20-25%, suggesting that between 75 and 80% of the fertiliser remains undissolved and therefore Epsom salts have a relatively slow rate of dissolution. This is not true because Epsom salts are highly soluble in water and would have completely dissolved soon after application (Loganathan et al. 1999). However, if we consider the percentage of residual Mg fractions, which is [is less than] 5%, the data would confirm that [is greater than] 95% of Epsom salts had dissolved.
[Figure 2 ILLUSTRATION OMITTED]
The recovery of approximately 25% of Mg in Epsom salts treatment in dissolved and undissolved fractions suggests that the remaining 75% of applied Mg had been lost from the soil layer sampled largely by leaching and some by plant uptake. In contrast to Epsom salts, the sequential extraction method was able to recover nearly 100% of the added Mg for most of the less soluble fertiliser treatments. These results indicate that losses due to leaching for the less soluble fertilisers are small.
Considering the undissolved Mg fractions, in general, significantly (P [is less than] 0.05) more of the calmag I fertiliser dissolved compared with the calmag 2, granmag, and forestry grade dolomite fertilisers up until 18 months into the trial (Fig. 3). As expected, nearly 100% of the soluble Epsom salts dissolved in 1 day after application. At 27 months after fertiliser application, between 91 and 100% of the Mg in Epsom salts, calmag 1, and granmag had dissolved. The corresponding value for calmag 2 was 82-85% and forestry grade dolomite was 65-72%. However, of these only dolomite differed significantly (P [is less than] 0.05) from Epsom salts in terms of percent dissolution after 27 months.
[Figure 3 ILLUSTRATION OMITTED]
Comparison of the rates of dissolution of calmag 1 and calmag 2 showed that increases in particle size decrease the rate of dissolution. However, although dolomite had a lower geometric mean particle diameter (0.42 mm) than either calmag 1 (1.41 mm) or calmag 2 (2.83 mm), it dissolved much more slowly than either of these two fertilisers. This reflected its much lower solubility. This is in agreement with the calculations of Lindsay (1979), who used equilibrium constants (K [degrees]) for the chemical reactions involved in the dissolution of these two fertilisers (Eqns 6, 7, and 8) to demonstrate that over a wide range of pHs and log([CO.sub.2] concentration), the concentration of Mg dissolved from MgO (calmag) was about [10.sup.10] times that from dolomite:
(6) MgO + 2[H.sup.+] = [Mg.sup.2+] + [H.sub.2]O log K [degrees] = 21.74
(7) Ca[CO.sub.3].Mg[CO.sub.3] + 4[H.sup.+] = [Ca.sup.2+] + [Mg.sup.2+] + 2[CO.sub.2](g) + 2[H.sub.2]O log K [degrees] = 18.46
(8) Ca[CO.sub.3] + 2[H.sup.+] = [Ca.sup.2+] + [CO.sub.2](g) + 2[H.sub.2]O log K [degrees] = 9.74
When fertilisers are applied to soil after removing litter, they have direct contact with the soil. This helps the soil exchange complex to act as a sink for the dissolution products of the fertilisers (Mg and Ca) and also provides protons for reacting with the fertilisers, thereby increasing the rate of dissolution compared with when fertilisers are applied onto the litter layer. The results of this study, however, showed that there was no significant difference (P [is less than] 0.05) in the amount of Mg fertiliser dissolved between plots with and without litter (Fig. 3). Only for the dolomite treatment at the 12-month sampling was the difference between with and without litter plots close to statistical significance at P = 0.05. There was no significant (P [is greater than] 0.05) interaction between litter treatments and time of sampling or between litter treatments, fertiliser source treatments, and time of sampling.
Soil exchangeable Mg in the field trial
All fertiliser treatments except Epsom salts significantly increased soil exchangeable Mg over the control 12 and 27 months after application (Table 2). The order of increase in soil exchangeable Mg generally follows the rate of dissolution of the fertilisers except for the highly soluble Epsom salts treatment, where the dissolved Mg may have been lost by leaching (Table 2, Fig. 3). In addition, the pH of soils treated with Epsom salts is lower than the pH of soils treated with other fertilisers (Table 2, Fig. 4) and therefore the negative charge density of the soil colloids in the Epsom salts treatment was lower. This would have given rise to lower amounts of Mg retained by the soil and hence greater leaching of Mg in the Epsom salts treatment. These results suggest that soluble Mg fertilisers may not be effective in providing an increased supply of plant-available Mg for a long period in pumice soils at Kaingaroa Forest. The cost per unit weight of Mg is also much higher for soluble Mg fertilisers (e.g. Epsom salts, Table 1) and this further discourages their use in forestry.
[Figure 4 ILLUSTRATION OMITTED]
Table 2. Fertiliser dissolution rate constants calculated using Chatupote's (1990) model and soil exchangeable Mg and pH (soil depth 0-10 cm) at 12 and 27 months
Dissolution rate constant for the with-litter plots ([micro]g/ Exchangeable Mg Fertiliser [cm.sup.2].day 12 months 27 months symbol of fertiliser)(A) ([cmol.sub.c]/kg) CM1 587 2.62 2.58 CM2 426 1.73 1.92 GM 385 2.48 2.18 Dol 18 1.44 2.18 Eps n.a. 0.73 1.04 Control n.a. 0.65 0.61 l.s.d. (P = 0.05) 0.58 1.08 Fertiliser Soil pH symbol 12 months 27 months CM1 5.48 5.52 CM2 5.51 5.53 GM 5.51 5.67 Dol 5.52 5.68 Eps 5.33 5.41 Control 5.10 5.30 l.s.d. (P = 0.05) 0.22 0.20
n.a., not applicable.
(A) Particle density of 3.5 g/[cm.sup.3] for CM1, CM2, and GM, and 2.8 g/[cm.sup.3] for Dol was used in model calculations.
Fertiliser effects on soil pH in the field trial
The results show that calmag 1, calmag 2, granmag, and dolomite fertiliser treatments have significantly (P [is less than] 0.05) increased soil pH over the control treatments from 6 months onwards (Table 2, Fig. 4). This is due to the alkalinity of these fertilisers. Generally, the increases in soil pH reflect the rate of fertiliser dissolution except for the dolomite treatment (Table 2, Fig. 3). Despite the lower percentage dissolution, the dolomite treatments resulted in similar increases in soil pH to the calmag and granmag treatments. This is due to the higher liming value of dolomite relative to its Mg content (Table 1). For the sampling immediately after application, the Epsom salts treatment resulted in a significant (P [is less than] 0.05) decrease in measured soil pH compared with the calmag, granmag, dolomite, and control treatments. This is a consequence of measuring soil pH in [H.sub.2]O. The Epsom salts would have increased the ionic strength to cause a reduction in measured soil pH. With increases in time this effect disappeared because most of the dissolved Epsom salts was probably lost by leaching.
Modelling fertiliser dissolution rates
Plots of [(1-u).sup.1/3] v. time (Eqn 1) (Swartzendruber and Barber 1965) for the range of dolomite size fractions tested in the laboratory incubation study are shown in Fig. 5a. Although there are a limited number of samplings and very little of the coarser size fractions (1-2 and 2-3 mm) has dissolved, the data points seem to fit to straight line relationships, as predicted by the cubic model. The results suggest that, within each size class, the rate of dissolution of dolomite over time was controlled by the changing surface area of the particles (Fig. 5a).
[Figure 5 ILLUSTRATION OMITTED]
But, when several size classes (different [D.sub.g] values) were considered together and when c v. 1/[D.sub.g] (Eqn 2) was plotted (Fig. 5b), instead of the data points falling on a straight line passing through the origin as the equation specified, they curved upwards as 1/[D.sub.g] increased (particle size decreased). The slopes of straight lines drawn from the origin through each data point increased consistently with increases in 1/[D.sub.g], instead of remaining constant. For example, the slope for the 2-3 mm particles is 10% of that for the 0.5-1.0 mm particles. Thus, as Swartzendruber and Barber (1965) reported, the specific fertiliser dissolution rates calculated from the slopes and particle density of dolomite (2.8 g/[cm.sup.3]) (Eqn 2) increased with decreases in particle size (Table 3).
Table 3. Dissolution rate constants ([micro]g/[cm.sup.2].day of fertiliser) for dolomite size fractions investigated in the laboratory incubation study as calculated by (A) the computer based cubic model for [S.sup.0] oxidation of Chatupote (1990) and (B) the cubic model for limestone dissolution of Swartzendruber and Barber (1965)
Specific fertiliser dissolution rate constants Dolomite size fraction (mm) A B 0.06-0.25 239 226 0.25-0.5 119 129 0.5-1.0 49 50 1.0-2.0 22 24 2.0-3.0 18 18
Specific fertiliser dissolution rate constants (Table 3) for each size fraction as calculated by the computer model of Chatupote (1990) also increased with decreases in particle size. The rate constants calculated by the models of Chatupote (1990) ([k.sub.1]) and Swartzendruber and Barber (1965) ([k.sub.2]) agreed very well with a regression equation of [k.sub.1] = 1.05*[k.sub.2] - 4.85 ([R.sup.2] = 0.9945, number of data sets 5).
The variability of k with fertiliser particle size observed for both Swartzendruber and Barber (1965) and Chatupote (1990) models' predictions could be due to a number of reasons. The chemical composition (25% Ca and 11% Mg) and mineralogy (only calcite and dolomite identified by X-ray diffraction) of the dolomite were found not to be different between particle size classes and therefore they cannot explain the variability in k. In their investigation, Swartzendruber and Barber (1965) explained the variation of k with particle size as possibly because of the geometric mean of the particle sizes ([D.sub.g]) not being a representative measure of the range of diameters of the limestone particles within each size fraction. The variability of k with fertiliser particle size could also be due to the rate of dissolution of the fertilisers not obeying zero-order kinetics assumed in the models. Factors other than surface area of the fertilisers, such as soil pH, soil pH buffering capacity, and Ca and Mg concentrations in soil solution and the exchange complex, may be changing differentially as the particles dissolve for the different particle size classes. For example, consider 2 fertiliser particle size classes where the sum total surface area for the small and large fertiliser particles is the same. Then, for the smaller particles, the sum total soil volume surrounding the fertiliser particles, where diffusion gradients of the fertiliser dissolution products (Mg, Ca) and protons between soil solution and the fertiliser particle surface (diffusive volume) exist, is proportionally larger than that for the larger fertiliser particles. If we assume that the fertiliser particles are spherical and diffusion gradients in the soil extend to a distance of 1 mm from the surface of the fertiliser particles, then a simple calculation would show that the diffusive volume in the soil for fertiliser particles of 0.25 mm diameter is 26 times higher than that for 2.0 mm diameter particles. Therefore, there would be proportionally a greater number of sinks for dissolution products and a larger source of protons available in the soil to smaller fertiliser particles compared with larger particles, causing a higher specific dissolution rate for the smaller particles.
Future work on the modelling of Mg fertiliser dissolution should consider both fertiliser and soil properties. A model similar to the mechanistic model developed by Kirk and Nye (1986) and modified by Bolland and Barrow (1988) to explain phosphate rock dissolution could be developed.
The order of fertiliser dissolution in the field trial was generally reflected in the specific dissolution rate constants generated by the cubic model (Chatupote 1990) (Fig. 6, Table 2). Specific fertiliser dissolution rate constants for the with-litter plots ranged from 587 [micro]g/[cm.sup.2].day of fertiliser for calmag 1 to 18 [micro]g/[cm.sup.2].day of fertiliser for dolomite. There was little difference in the rate constants calculated for calmag 1 and calmag 2 (587 and 426 [micro]g/[cm.sup.2].day of fertiliser, respectively), suggesting that the rate at which a fertiliser dissolved expressed in terms of per unit of surface area was predominantly independent of fertiliser particle size within 1-4 mm size range. Granmag dissolution up to 6 months was higher than that predicted by the computer model. This is due to the fast-release of Mg from the readily soluble Mg[SO.sub.4] component in granmag. The overall dissolution rate constant of 385 [micro]g/[cm.sup.2].day of fertiliser was, however, similar to that of the calmag fertilisers, reflecting the dissolution characteristics of the unacidulated MgO in the granmag fertiliser (Table 2, Fig. 6). Dolomite had the lowest dissolution rate constant of 18 [micro]g/[cm.sup.2].day of fertiliser because of differences in its chemical composition compared with calmag and granmag fertilisers (Eqns 6, 7, and 8).
[Figure 6 ILLUSTRATION OMITTED]
In pumice soils under P. radiata the rates of Mg fertiliser dissolution within 27 months follow the sequence: Epsom salts [is greater than] calcined magnesite (1-2 mm) [is greater than] granmag [is greater than] calcined magnesite (2-4 mm) [is greater than] forestry grade dolomite (0.06-3 mm; geometric mean 0.42 mm). The rate of dissolution depends on the particle size of the fertiliser: rate of dissolution increases with decreases in particle size. The rate of fertiliser dissolution was little affected by whether fertiliser is applied directly on to the soil surface (litter removed) or on to the forest litter layer.
The alkaline, slowly soluble Mg fertilisers (calcined magnesite, granmag, and dolomite) have a significant liming effect on the soil. Dolomite with its highest liming value relative to its Mg content has a proportionally greater effect on soil pH than the other fertilisers, despite a slower overall dissolution rate. The slowly soluble Mg fertilisers are more effective in increasing soil exchangeable Mg for a long period than the easily soluble and more expensive (per unit weight of Mg) Epsom salts and therefore they are more suitable fertilisers for P. radiata.
A computer program based on an [S.sup.0] oxidation cubic model can explain the rate of dissolution of Mg fertilisers within a narrow fertiliser particle size range but fails when a wide range of particle sizes is considered. The specific fertiliser dissolution rate constant ([micro]g/[cm.sup.2].day of fertiliser) increases with decreases in particle size, suggesting that the rate of dissolution depends on factors other than surface area when particle size varies widely.
This research was financially supported by Fernz Chemicals (NZ) Ltd, through the Upper Mid Crown Yellowing Research Group. We thank Fletcher Challenge forests for allowing access to the field installations. We acknowledge the assistance in the field of Mr Doug Graham and Ms Caroline Anderson from NZ Forest Research Institute Ltd, Rotorua. We also thank Dr M. J. Hedley and Mr M. Bretherton from Institute of Natural Resources, Massey University, Palmerston North, for their assistance with the modelling section and Dr Jim Barrow of Australia for critically reviewing the manuscript.
Beets PN, Jokela EJ (1994) Upper mid-crown yellowing in Pinus radiata: some genetic and nutritional aspects associated with its occurrence. New Zealand Journal of Forestry Science 24, 35-50.
Blackmore LC, Searle PL, Daly BK (1987) Methods for chemical analysis of soils. New Zealand Soil Bureau, Scientific Report 80. Lower Hutt, NZ
Bolland MDA, Barrow NJ (1988) Effect of level of application on the relative effectiveness of rock phosphate. Fertiliser Research 15, 181-192.
Buckman HO, Brady NC (1960) `The nature of and properties of soils.' 6th edn. (The Macmillian Company: New York)
Chatupote W (1990) An investigation of some factors influencing the rate of oxidation of elemental sulphur fertilisers. PhD Thesis, Massey University, NZ.
Elphick BL (1955) Studies in use of agricultural limestone: II. New Zealand Journal of Science and Technology 37A, 156-173.
Hedley MJ, Chatupote W, Heng LK, Bolan NS, Loganathan P, Bretherton MR (1995) An elemental sulphur oxidation model for all seasons. In `Fertilizer requirements of grazed pasture and field crops: macro- and micro-nutrients'. (Eds LD Currie, P Loganathan) pp. 170-181. Occasional Report No. 8. (Fertilizer and Lime Research Centre, Massey University: Palmerston North, NZ)
Heming SD, Hollis JF (1995) Magnesium availability from kieserite and calcined magnesite on five soils of different pH. Soil Use and Management 11, 105-109.
Hewitt AE (1993) `New Zealand soil classification.' (Manaaki Whenua-Landcare Research New Zealand Ltd: Lincoln, NZ)
Hunter IR, Prince JM, Graham JD, Nicholson GM (1986) Growth and nutrition of Pinus radiata on rhyolitic tephra as affected by magnesium fertiliser. New Zealand Journal of Forestry Science 16, 152-165.
Kirk GJD, Nye PH (1986) A simple model for predicting the rates of dissolution of sparingly soluble calcium phosphates in soil I. The basic model. Journal of Soil Science 37, 529-540.
Landmann G, Hunter IR, Hendershot W (1997) Temporal and spatial development of magnesium deficiency in forest stands in Europe, North America and New Zealand. In `Magnesium deficiency in forest ecosystems'. (Eds RF Huttl, W Schaaf) pp. 23-64. (Kluwer Academic Publishers: Dordrecht, The Netherlands)
Lindsay WL (1979) `Chemical equilibria in soils.' (John Wiley: New York)
Loganathan P, Payn TW, Mitchell AD, Tillman RW (1999) A sequential extraction method for the determination of dissolution of magnesium from fertilisers applied to pumice soils. Communications in Soil Science and Plant Analysis 30, 199-211.
McCaskill MR, Blair GJ (1989) A model for the release of sulphur from elemental S and superphosphate. Fertilizer Research 19, 77-84.
Mitchell AD, Loganathan P, Payn TW, Tillman RW (1999) Effect of calcined magnesite on soil and Pinus radiata foliage magnesium in pumice soils of New Zealand. Australian Journal of Soil Research 37, 545-560.
Payn TW (1991) The effects of magnesium fertiliser and grass on the nutrition of P. radiata planted on pumice soils in the Central North of New Zealand. PhD Thesis, University of Canterbury, NZ.
Rijkse WC (1988) Soils of Kaingaroa Plateau, North Island, New Zealand. New Zealand Soil Bureau, DSIR District Office Report RO14. (Rotorua, New Zealand)
SAS Institute Inc. (1996) `SAS for Windows-Version 6.12.' (SAS Institute Inc.: Cary, NC)
Swartzendruber D, Barber SA (1965) Dissolution of limestone particles in soil. Soil Science 100, 287-291.
Watkinson JH, Blair GJ (1993) Modelling the oxidation of elemental sulphur in soils. Fertilizer Research 35, 115-126.
Will GM (1961) Magnesium deficiency in pine seedlings growing in pumice soil nurseries. New Zealand Journal of Forestry 4, 151-160.
Manuscript received 13 July 1999, accepted 10 December 1999
A. D. Mitchell(A), P. Loganathan(A), T. W. Payn(B), and R. W. Tillman(A)
(A) Soil and Earth Sciences, Institute of Natural Resources, Massey University, Palmerston North, New Zealand.
(B) New Zealand Forest Research (Institute Ltd), Rotorua, New Zealand.
|Printer friendly Cite/link Email Feedback|
|Author:||Mitchell, A. D.; Loganathan, P.; Payn, T. W.; Tillman, R. W.|
|Publication:||Australian Journal of Soil Research|
|Date:||May 1, 2000|
|Previous Article:||Accessibility of subsoil potassium to wheat grown on duplex soils in the south-west of Western Australia.|
|Next Article:||Recent advances in the application of [sup.13]C and [sup.15]N NMR spectroscopy to soil organic matter studies.|