SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Mendham, D. S.
Right arrow Articles by Baker, T.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Mendham, D. S.
Right arrow Articles by Baker, T.
Agricola
Right arrow Articles by Mendham, D. S.
Right arrow Articles by Baker, T.
Related Collections
Right arrow Soil Fertility and Productivity
Right arrow Other Soil Management
Right arrow Forest Soils
Soil Science Society of America Journal 66:959-968 (2002)
© 2002 Soil Science Society of America

DIVISION S-7—FOREST & RANGE SOILS

Soil Analyses as Indicators of Phosphorus Response in Young Eucalypt Plantations

Daniel S. Mendham*,a, Philip J. Smethurstb, Greg K. Holzc, Robert C. Menaryd, Tim S. Grovee, Chris Westonf and Tom Bakerg

a Cooperative Research Centre for Sustainable Production Forestry and CSIRO Forestry and Forest Products, Private bag No. 5, Wembley, WA, Australia 6913
b Cooperative Research Centre for Sustainable Production Forestry and CSIRO Forestry and Forest Products, GPO Box 252-12, Hobart, Tasmania, Australia 7001
c Gunns Ltd., P.O. Box 63 Ridgley, Tasmania, Australia 7321
d Dep. of Agricultural Science, University of Tasmania, GPO Box 252-54, Hobart, Tasmania, Australia 7001
e CSIRO Forestry and Forest Products, Private Bag No. 5, Wembley, WA, Australia 6913
f Institute of Land and Food Resources, The University of Melbourne, Royal Parade, Parkville Victoria, Australia 3052
g Centre for Forest Tree Technology, P.O. Box 137, Heidelberg, Victoria, Australia 3084

* Corresponding Author (Daniel.Mendham{at}csiro.au)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The use of P fertilizer in Eucalyptus plantations can result in significant and economically viable increases in timber production. Soil analyses could potentially indicate sites requiring more P fertilizer than is currently applied and prevent excess P fertilizer application on sites where little or no response would be obtained. The ability of several soil P analyses to predict first year growth responses of E. nitens and E. globulus to P fertilizer was assessed in 24 previously established field experiments situated in southeast and southwest Australia on a range of soil types. Soil P analyses that were assessed included an intensity-based analysis (CaCl2 extracted P), quantity-based analyses (total P, bicarbonate extracted P, and Bray No. 2 P), and quantity–intensity relationships (P adsorption curves). An excellent relationship was found between CaCl2-extractable P (range: 26–162 µg kg-1, R2 = 0.83) and first year growth response to P applied at planting for 21 of the 24 field experiments. Quantity-based P analyses, such as bicarbonate P (range: 2–63 µg g-1, R2 = 0.43), Bray No. 2 P (range: 0.1–15 µg g-1, R2 = 0.30), acid-extractable P (range: 0.6–11 µg g-1, R2 = 0.37), and total P (range: 0.038–3.5 mg g-1, R2 = 0.39) did not correlate as well with plant growth response to P application. However, quantity-based P analyses may be useful if specific calibrations were developed for a limited range of soils. Inclusion of P adsorption data in multiple regressions of soil P concentration against relative yield generally did not improve the relationships. The excellent relationship between plantation response to P fertilizer and soil P intensity suggested that such analyses may be useful for managing P fertilizer in eucalypt plantations across the wide range of sites where eucalypts are currently being planted.

Abbreviations: EPC, equilibrium phosphate concentration • PBC, P buffer capacity


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
EUCALYPT PLANTATIONS are an excellent source of high quality pulp and paper resources. Currently, large areas are being planted to eucalypts, both in Australia (National Forest Inventory, 2000) and throughout the world (Food and Agriculture Organization, 2000). Optimizing plantation productivity is critical for ensuring that maximum economic benefit is gained from eucalypt plantations. Fertilizer application can increase tree growth rates (Schönau and Herbert, 1989), shorten rotation times, and improve economic returns (Gerrand et al., 1993). Eucalypt plantations in Australia are established on soils with a wide range of P status, and recommendations for P fertilizer rates are generally based on previous land use or empirical experiments on a few specific soil types and without reference to soil analyses (Birk, 1994; Cromer, 1996). Responses to P fertilizer in eucalypt plantations generally occur early (within the first year), and the magnitude of the response generally diminishes if fertilization is delayed after planting (Schönau and Herbert, 1989). An effective soil analysis indicator would enable correction of P deficiency at establishment, and allow an early growth response to be captured.

Soil P analyses have traditionally been grouped into those that measure soil P intensity and those that measure soil P quantity. Soil P intensity refers to P immediately available in solution, while quantity-based analyses extract both the soil solution P and a proportion of the labile P that is adsorbed to the solid phase. Soil P quantity-based analyses (e.g., extracts of Bray and Kurtz, 1945) have been well correlated with Pinus sp. response to P fertilizer in a number of countries, including the USA (Kushla and Fisher, 1980), Australia (Hopmans et al., 1978), New Zealand (Ballard, 1974), and South Africa (Payn and Clough, 1988). However, relationships between soil P quantity analyses and plant growth responses are generally limited to the specific soil types that they are developed for. These relationships require extensive empirical calibration for accurate diagnosis of P deficiency on a given soil type. Soil P intensity analyses (e.g., soil solution P and CaCl2-extractable P) have been little used for predicting response to P fertilizer in plantation forestry, but they have shown promise for predicting P response in agricultural crops over a wide range of soil types (Fox, 1981). Specific crops that have shown a good relationship between P intensity and response to P fertilizer include subterranean clover (Trifolium subterraneum L) (Dear et al., 1992), soybeans [Glycien Max.(L.) Mer.] (Moody et al., 1983), and other pasture species (Ozanne and Shaw, 1967). It was postulated by Holford (1997) that successful analyses of soil P for plant-growth response would need to incorporate both an intensity and a quantity component.

Short-term pot experiments with Eucalyptus nitens (Mendham, 1998) suggested that indicators of both soil P quantity (e.g., bicarbonate P), and intensity (e.g., CaCl2-P) may be useful for predicting P requirements in eucalypt plantations. Soil type-specific relationships were found for bicarbonate P. In contrast, relationships with CaCl2-extractable P (CaCl2-P) had less sensitivity to soil type, and therefore may have potential to be used to test for P deficiency in a range of soil types without the need for extensive empirical calibration.

The objective of this study was to compare the efficacy of several quantity and intensity-based soil P analyses at predicting plantation response to P fertilizer in the field. We examined relationships between soil P analyses and early response (growth at 1 yr) of Eucalyptus nitens and Eucalyptus globulus in 24 P fertilizer-rate experiments in southern Australia that were established by several different research groups and under widely differing soil type and climatic conditions.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Sites
Details of the experimental sites, including site location, soil classification, surface soil field texture and selected chemical properties, previous land use, and relative tree yield are shown in Table 1. All sites selected for use in this study consisted of an established P rate experiment that had a randomized complete block design and three to four replicates per fertilizer treatment. A range of P fertilized treatments (2–8 fertilizer rates, including a zero-P control) were installed at planting. All plots at each site were fertilized with a constant rate of major nutrients. Application rates of P differed between sites, but where there were more than two fertilizer rates, they were generally concentrated at the lower end of the range to enable good definition of the P-response curve. The main features of the experimental design at each site are shown in Table 2, including the plantation species, age at sampling, number of replicate plots sampled, number of P application rates, maximum rate of P applied, method of P application, and the number of trees measured per plot. A total of 75 plots were sampled for soil, and 363 plots were used in the assessment of tree growth response to P fertilizer. Seedlings for the experiments were raised in a nursery for 6 to 9 mo, then transplanted into the field in June or July of the year of establishment. Seedling growth was measured at 1 yr after transplanting.


View this table:
[in this window]
[in a new window]
  		Table 1. Location and site details of experiments used in this study.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Experimental design parameters at each site.

 
The study sites had been established independently by four research groups. This study was an opportunity to use these sites to improve our understanding of P relations in eucalypt plantations across a wider geographic area than would otherwise be possible. It was recognized that different tree establishment techniques and climatic factors may have influenced absolute tree growth, so only relative responses were compared. The broad range of conditions used during plantation establishment and growth was one of the strengths of the study, making results potentially applicable across a correspondingly broad range of situations.

Eucalypt stocking density at all sites was 1000 to 1100 trees ha-1. Nitrogen, K, and trace elements were applied at each of the sites at rates considered nonlimiting to growth. Phosphate was applied as triple superphosphate at all sites except for Boola, Glencoe, Kuark, Maryvale, and West Bemm, where single superphosphate was used. Beebes, Boorara, and Carpenters were deep ripped along planting lines prior to transplanting seedlings, while seedlings were planted into mound-ploughed rows at all other sites. The mounds were 3.0- to 3.5-m between tops, ~180 cm wide, and 50 cm above the original land surface when formed. Weeds were controlled at all sites. Soil sampling was conducted between 1 and 11 yr after tree planting. We assumed that P levels in the control plots (no P added) were in equilibrium and would not have changed markedly since plantation establishment. This is a valid assumption; soils had either never been fertilized (in the ex-native forest sites), or had been fertilized minimally at planting of the previous Pinus plantation, more than 20 yr prior to establishment of the eucalypts. The only exception was Potters, which was an ex-agricultural site and had a history of regular fertilization (N, P, and possibly K) prior to establishment of eucalypts.

For statistical analyses and presentation of the results, the soils were separated into Oxisol and non-Oxisol soil types. This separation was made because Oxisols have a high capacity to adsorb P, and their response characteristics may be quite different from other soils with lesser P adsorbing capacity. Oxisols also comprised approximately half of the study sites, and it was useful to compare P responses within a single soil order.

Soil Sampling and Analyses
Soil samples from all 24 sites were taken from the 0- to 10-cm depth range in the uncultivated area between the mound-ploughed rows of the control (no P added) plots. A single sample was collected from each plot by compositing 20 soil cores along a random transect of 20 to 50 m (transect length was dependent on plot size).

All soil analyses were conducted on the <2-mm fraction of air-dry soil. All extracts were analyzed for orthophosphate on a flow injection analyzer (QuikChem 800, Lachat Instruments, Milwaukee, WI) using the colorimetric P determination method of Murphy and Riley (1962). Extracts were vacuum filtered prior to analysis using 0.45-µm cellulose acetate filters. Calcium chloride-extractable P (Rayment and Higginson, 1992) was determined in soil extracts after shaking 5 g of soil end-over-end for 17 h with 50 mL of 5 mM CaCl2. Phosphate adsorption curves (Rayment and Higginson, 1992) were determined after shaking 5 g of soil end-over-end in 50 mL of a 10 mM CaCl2 extracting solution with five to seven initial levels of P for 17 h at 25°C. Initial P levels were between 0 mg P L-1 and 1000 mg P L-1. The relationship between P adsorbed by the soil and P remaining in the extract was described by the Freundlich equation: y = a x1/b, where y was P adsorbed by the solid phase (mg kg-1), x was concentration of P in the liquid phase (mg L-1), and a and b were fitted values. Equilibrium phosphate concentration (EPC) and P buffer capacity (PBC) were respectively calculated as the intercept and slope of a linear regression relating P adsorbed to the log of the concentration remaining in solution (Rayment and Higginson, 1992). The units of PBC were mg kg-1 (log10mg L-1) -1. Soils used to assess P adsorption curves were bulked across replicates resulting in one sample for each fertilizer treatment at each study site, so site heterogeneity in P adsorption could not be estimated. Bicarbonate-extractable P (Rayment and Higginson, 1992) was measured in 0.5 M NaHCO3 (pH 8.5) soil extracts after filtration and shaking 1 g of soil in 100 mL of extractant end-over-end for 17 h at 25°C. Acid-extractable P (Rayment and Higginson, 1992) was determined after shaking 1 g of soil for 17 h at 25°C in 100 mL of 5 mM sulfuric acid. Bray No. 2 P (Bray and Kurtz, 1945) was determined after shaking 5 g of soil with 50 mL of 0.03 M NH4F/0.1 M HCl for 40 s. Total P (Rayment and Higginson, 1992) was determined after digestion of 0.2 g of ground soil in 8 mL of a 33.3-g salicylic acid L-1 H2SO4 mixture.

Growth Response to Phosphorus Fertilizer
Response to P fertilizer was assessed as growth at 1 yr after seedlings were planted in the field. The 1-yr growth data was used because that age of measurement was common to all sites. Additionally, growth responses measured at 1 yr are usually maintained to the end of the rotation in eucalypt plantations (Schönau and Herbert, 1989). Growth was assessed as stem conical volume, based on tree heights and diameters. Only tree height was measured at some sites, so conical volumes were calculated based on an allometric relationship derived from the height (h) and diameter (d) of 1-yr old trees from five Tasmanian sites (R2 = 0.93, La Sala, personal communication, 1998). Growth responses have been published for the Boola, Maryvale, and Glencoe sites (Judd et al., 1996) and for the Carpenters site (Grove et al., 1991).

Maximum growth (i.e., growth not limited by P) at each site with more than two P-rate treatments was determined by fitting a Mitscherlich function of the form: y = a - becx, where y was growth (volume or aboveground dry matter), x was the rate of P applied, and a, b, and c coefficients were fitted values. The a coefficient was the asymptote of the regression, which was interpreted as maximum growth. Where only two rates of P were applied, growth in the high P treatment was set equal to maximum growth. Studies with two P fertilizer rates used spot application near the tree base, which would be highly accessible to tree uptake. Studies where the fertilizer was strip applied had more rates of P application (Table 2), and the response curve was well defined. Relative yield (RY) was defined as follows:

[1]
where SV was the stem conical volume at 1 yr in the control (no P added) and maximum (P fertilized) treatments.

Statistical Analysis
The Genstat statistical package (version 5.42, Lawes Agricultural Trust, Rothamsted Alliance, Harpenden, UK) was used to assess the effects of soil P analysis on relative yield using linear and nonlinear regression (depending on the shape of the response). Multiple linear regression analysis was employed to determine the influence of soil P buffering capacity on significance of the P analysis results. Variability within sites was indicated by standard errors of the mean for each soil P analysis (except P adsorption curves) and tree growth.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Relationships Between Phosphorus Measures
Relationships between CaCl2-P and P quantity indices of P availability and EPC are shown in Fig. 1 . Results from the Potters site were circled in each of the graphs because it was the only ex-agricultural site, i.e., it had a history of regular P fertilization prior to establishment of the plantation. Glencoe CaCl2-P was not included in these relationships because of its unusually high value of 1350 µg kg-1 CaCl2-P. Values of CaCl2-P were similar from all three replicates at the Glencoe site and a high EPC was also found (Table 3). Quantity based P extractions from the Glencoe site were within the normal range, while total P was one of the lowest at 0.05 mg g-1. The unusually high value of CaCl2-P at Glencoe was probably because organic, rather than inorganic processes dominated the P dynamics at that site. Organic processes are likely to predominate where PBC is low, and soils from Glencoe had the second-lowest PBC of the sites investigated. We postulate that a flush of P released from litter decomposition caused the high CaCl2-P values at that site. For the remaining 23 sites, significant relationships were observed between CaCl2-P and quantity-based P availability indicators: bicarbonate P (P < 0.01), Bray No. 2-P (P < 0.01), acid-extractable P (P < 0.05), and total soil P (P < 0.01), although these relationships accounted for <57% of the variation in test values (Fig. 1). While relationships between CaCl2-P and other P analyses may not have been linear at the lower or upper end of soil P status, fitting nonlinear relationships to the data was not justified because of variability in the data. The relationship between EPC and CaCl2-P was significant (P < 0.05, Fig. 1e). The scatter in Fig. 1e was partly attributable to the fact that both analyses were close to the detection limit, so the error in measurement was relatively high. Also, the EPC is calculated from all points in the adsorption curve, which can have a significant influence on the x-axis intercept of an adsorption curve. The relationship between CaCl2-P and PBC was also significant (R2 = 0.61, Fig. 1f). Good relationships (R2 >= 0.61) were also found among several of the quantity based indicators of P availability (data not shown).



View larger version (39K):
[in this window]
[in a new window]
  		Fig. 1. Calcium chloride-P relationship with (a) bicarbonate P, (b) Bray No. 2-P, (c) acid-extractable P, (d) total soil P, (e) equilibrium phosphate concentration, and (f) P buffer capacity for 23 sites (Glencoe excluded). The Potters site is circled in each graph, and the R2 of the relationship without Potters is in parentheses after the R2 of the relationship with all 23 soils. Error bars show standard error of the mean.

 

View this table:
[in this window]
[in a new window]
  		Table 3. Mean, standard deviation, and range of each soil P analysis across the 24 sites. Freundlich coefficients were derived from a graph of solid phase P (mg kg-1) versus solution phase P (mg L-1) for each soil

 
Phosphate buffer capacity was significantly correlated with EPC (Fig. 2a , P < 0.05), but none of the P quantity indicators were significantly correlated with PBC (Fig. 2b–d), either with or without Potters included in the regression. Equilibrium phosphate concentration ranged from 1.9 to 29.4 µg L-1, with the exception of the Glencoe site, for which EPC was 184 µg L-1 (Table 3). This EPC value was over six-fold higher than the next highest concentration (29.4 µg L-1, at Westfield). Phosphate buffer capacity ranged from 4.69 to 1214 mg kg-1[log10µg L-1]-1. The range of PBC across all sites in the current study was similar to that found by Singh and Gilkes (1991) for 97 southwest Australian soils, and by Moody et al. (1988) for 26 Queensland soils.



View larger version (26K):
[in this window]
[in a new window]
  		Fig. 2. Relationship between P buffer capacity and (a) equilibrium phosphate concentration, (b) Bray No. 2-P, (c) acid-extractable P, and (d) bicarbonate P. The Potters site is circled in each of the graphs. Error bars show standard error of the mean.

 
We found a significant increase in EPC with increasing PBC (Fig. 2a). However, poor relationships were found between PBC and soil P quantity indicators in the current study (Fig. 2b–d). Both Dalal and Hallsworth (1977), and Moody et al. (1988) found little correlation between PBC and bicarbonate P for soils with a wide range of phosphate levels. Dalal and Hallsworth (1977) also found no relationship between PBC and CaCl2-P, while Moody et al. (1988) found a weak inverse relationship between CaCl2-P and PBC. The positive correlation between EPC and PBC found in the current study may have been because P intensity (CaCl2-P and EPC) is likely to be proportional to P quantity within a single soil type (Holford and Mattingly, 1976), and most soils in our study were Oxisols (Fig. 2). The negative relationship found by Moody et al. (1988) was probably because of the agricultural nature of those soils, i.e., some would have accumulated large amounts of P through repeated fertilizer application. This would cause a shift in the equilibrium position on the P adsorption curve (Holford and Mattingly, 1976) with increased solution concentrations, resulting in reduced equilibrium buffer capacity. In the current study, soils (except Potters) had not been fertilized for at least 20 yr. The Potters site was the only ex-agricultural site, with a history of regular fertilization. This pasture soil (Potters) had high P quantity values, but a relatively low PBC. The equilibrium point of the unfertilized soils in this study was likely to be at the lower end of the adsorption curves where PBC is higher.

The order of extraction strength between analyses increased from CaCl2-P < Acid P < Bray No. 2 P < Bicarbonate P < Total P. This is approximately the same order of extraction strength found in other studies (Holford, 1983; Moody et al., 1988), but P extracted by the Bray No. 2 reagent is often higher than reported in this study, because the extraction time during the Bray procedure varies from 40 s to 10 min (Stewart et al., 1990).

Phosphorus Measures and Yield Response
Relative yield varied from 1.1 to 89% (Table 1), and significant relationships were found between all soil P quantity indicators and relative tree yield (Fig. 3) . Acid-extractable P described 37% of the variation in relative yield (P < 0.05), bicarbonate P described 43% (P < 0.05), and Bray No. 2-P described 30% (P < 0.05). Critical concentrations (i.e., that required for 90% of maximum relative yield) of each of the quantity indicators were: 2.4, 21.3, and 5.3 µg g-1 for acid-extractable P, bicarbonate P, and Bray No. 2-P, respectively. However, because of variability in P-availability data, the asymptote for those regressions was between 71 and 76% relative yield, rather than at a relative yield of 100%. We also fit a Mitscherlich model to the lower boundary of the points in Fig. 3, because maximum growth response to P fertilizer may decrease if tree growth was limited by some factor other than P at each site. This effect would cause lower maximum growth, so a relatively higher yield in the control treatment (when expressed as a proportion of the maximum growth). Hence the boundary line was fitted through the lower points. Critical concentrations of available P calculated from the adjusted boundary lines were 17.0, 70.6, and 23.7 µg g-1 for the acid-extractable P, bicarbonate P, and Bray No. 2-P, respectively.



View larger version (27K):
[in this window]
[in a new window]
  		Fig. 3. Relationship between relative yield and quantity-based indicators of P availability. Error bars show standard error of the mean.

 
It is generally accepted that nutrient availability relationships with quantity-based P analyses are specific to soil type (Holford, 1997). Even within the Oxisol soil order, there was high variability in response to P-quantity indicators. The critical values suggested by the boundary line analysis were higher than would generally be expected for agricultural crops (Reuter et al., 1995), probably because of the high P buffering capacity of the Oxisols. Unfortunately, most of our sites showed a large growth response to P fertilizer. This reduced the power of the relationships at the upper end of the scale. The levels required for 90% of maximum growth may have been better defined if sites with higher fertility had been included.

Critical soil test values for optimal tree growth have rarely been quoted in forestry, despite significant correlations between soil P availability indicators and growth (Ballard and Pritchett, 1975; Kadeba and Boyle, 1978; Hopmans et al., 1978). Ballard (1974) found that 12 µg g-1 Bray No. 2-P was optimal for growth of Pinus radiata in 16 New Zealand studies he conducted. This was also within the range of optimal concentrations (5.3–24 µg g-1) found in our experiment. The extraction time during the Bray No. 2 procedure significantly affects the amount of P extracted from soils (Stewart et al., 1990). Our Bray No. 2 test results were comparable with Ballard (1974), because the same procedure was followed.

Extensive studies of agricultural soils have shown that bicarbonate P can be a useful predictor of field crop response to P fertilization. In reviewing the results of more than 580 South-Australian field experiments, Reuter et al. (1995) found typical critical bicarbonate P values of 21 ± 1 (standard error) µg g-1 for wheat (Triticum aestivum) (306 experiments), 18 ± 3 µg g-1 for barley (Hordeum vulgare L.) (41 experiments), 13 ± 6 µg g-1 for potato (Solanum tuberosum L.) on soils with low P-adsorption capacity, and 46 ± 10 µg g-1 for potatoes on medium to high P adsorbing soils. The critical level of bicarbonate P for wheat in Western Australian soils was between 10 and 100 µg g-1, again depending on soil type (Bolland et al., 1994). The critical value of bicarbonate P for eucalypts in this study (24 sites) was between 20.6 and 70.6 µg g-1. The large proportion of Oxisols with a high P adsorbing capacity may have elevated the critical bicarbonate P to a higher level in this study than if it was measured on a group of soils with low P adsorption (Reuter et al., 1995). Moody et al. (1997) found a critical bicarbonate P for maize (Zea mays L.) of 20 to 32 µg g-1 in 17 experiments on high P-adsorbing Oxisols.

No significant trend was found between eucalypt growth response and PBC (Fig. 4a) , and only a weak correlation (R2 = 0.39 for all sites, or R2 = 0.48 with Boola and Maryvale excluded) was found between relative yield and total P (Fig. 4b). Conversely, a highly significant relationship was found between relative yield and CaCl2-P (Fig. 4c). With the exclusion of the Boola and Maryvale (circled) CaCl2-P described 83% of the variance in growth response. When Boola and Maryvale were included in the regression, 61% of the variance was explained (equation: y = 0.0067x - 0.17). The critical concentration of CaCl2-P required for 90% of maximum growth predicted by the linear regression (excluding Boola and Maryvale) was 155 µg kg-1.



View larger version (21K):
[in this window]
[in a new window]
  		Fig. 4. Relative yield in relation to (a) P buffer capacity, (b) total P, and (c) CaCl2-P (Boola and Maryvale circled). Error bars show standard error of the mean.

 
Boola and Maryvale had higher relative yield than other sites with similar CaCl2-P levels. The cause of this apparent anomaly may have been reduced maximum growth because of the relatively low rate (maximum of 50 kg ha-1) of P, of which half was applied 2 mo after planting, and the rest at ~8 to 9 mo. Greater maximum growth may have been obtained at those sites if higher rates of fertilizer had been applied immediately after planting. Schönau and Herbert (1989) cite several examples where later fertilization was less effective than fertilization soon after planting.

While soil P intensity analyses have not been used in forestry, they have been useful for predicting P deficiency in a number of agricultural crops, including soybean (Moody et al., 1983), subterranean clover (Dear et al., 1992), maize, groundnut (Apios americana Medik.), and potato (Fox, 1981). The optimum EPC found by Moody et al. (1983) for soybean was ~0.4 µM (equivalent to 120 µg kg-1 CaCl2-P), while the optimum solution concentration for growth of subterranean clover was ~3 µM (equivalent to 900 µg kg-1 CaCl2-P, Dear et al., 1992; Ozanne and Shaw, 1967). The optimum CaCl2-P concentration found for eucalypts in the current experiment (155 µg kg-1) was similar to that found for soybean by Moody et al. (1983).

Integrating both soil P buffering capacity and P quantity indicators increased the correlations between soil test results and growth of ryegrass (Lolium L.) (Gunary and Sutton, 1967; Holford and Mattingly, 1976) and wheat (Dalal and Hallsworth, 1976; Holford and Cullis, 1985). In our study, however, inclusion of indicators of PBC (or Freundlich coefficients) in multiple-linear regression analyses with P quantity indicators only marginally improved the relationships with relative yield, if at all (Table 4). This suggests that PBC is not as important for governing supply to eucalypts as has been found for some agricultural crops. The strongest relationship we found was between CaCl2-P and relative eucalypt yield at 1 yr of age (Fig. 4c). This suggests that CaCl2-P would be a useful indicator of P deficiency at new planting sites. Calcium Chloride-P was better correlated with relative yield than all combinations of soil P quantity, intensity or buffer power.


View this table:
[in this window]
[in a new window]
  		Table 4. Effect of including P buffer capacity (PBC) and Freundlich coefficients (individually and together) in multiple-linear regressions with P analyses on R2 coefficients and significance of the relationships between soil P analysis and relative yield.

 
One of the potential problems with intensity-based indices of nutrient availability is that they can be more sensitive to seasonal variation than quantity based indices. Most variation probably occurs in soils with very low P-adsorbing capacity, where organic processes govern nutrient cycling. Soil organic processes are sensitive to the changes in temperature and moisture that occur both within and between seasons. However, nutrient cycles that are governed by inorganic processes (such as P in soils with reasonably high P buffering capacity) are much less influenced by changes in temperature and moisture within the normal range of seasonal variation. Smethurst et al. (2001), for example, found little significant seasonal variation in soil solution P at three sites in Tasmania. The CaCl2-P data at Glencoe was a good example, that site had very low PBC, it was sampled at a late age (well after P cycling from litterfall would have commenced), and the CaCl2-P at that site was not related to other P indicators or plantation growth response to P within the same ranges found for all of the other sites. For soils other than those with very low buffering capacity, the CaCl2-P analysis may be a useful soil test for identifying sites with potential P deficiency.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Soils from 24 established P fertilizer experiments with E. nitens and E. globulus in southern Australia were assessed using several indicators of P availability. Soil was sampled between 1 and 11 yr after planting, and the relationship between soil P test results and relative yield at 1 yr was evaluated. Although rotation length is generally much longer than 1 yr (typically 10–15 yr), studies have shown that the largest responses to P fertilizer in eucalypts occur when the plantations are fertilized soon after planting, and the gains at 1 yr are maintained to the end of the rotation (Schönau and Herbert 1989). Measures of soil P quantity, including total P, bicarbonate P, acid-extractable P, and Bray No. 2-P were correlated with relative yield in response to P fertilizer additions across the broad range of soils used in this study, but the correlations were relatively low (R2 values between 0.30 and 0.43). Inclusion of buffer capacity measures with soil P quantity measures did not improve the relationships. Critical values of soil P quantity indicators (total, bicarbonate, acid, and Bray No. 2-P) were difficult to define, because of high variability between sites. These P quantity analyses may not be useful in a wide range of applications without further calibration for specific soil types. Calcium chloride-extractable P (an indicator of soil P intensity) was highly correlated with plant growth response to P fertilizer at 21 out of the 24 sites examined. The critical value for CaCl2 extractable P was 155 µg kg-1 (R2 = 0.83). This method may be useful in deteriming the P fertilizer requirements of Eucalyptus spp. plantations prior to planting.


   ACKNOWLEDGMENTS
 
This work was supported by Gunns Ltd. and the Cooperative Research Centre for Sustainable Production Forestry. Additionally, North Forest Products, Australian Paper Plantations, The Centre for Forest Tree Technology, State Forests of New South Wales, Norske Skog, and WA Plantation Resources supported several of the field sites used in this study. The authors thank Karen Faunt for provision of data for the Cussacks and Richardsons sites. Keryn Paul, Randall Falkiner, Jennifer Knoepp, and 2 anonymous referees also provided helpful comments on the manuscript.

Received for publication September 20, 2000.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Mendham, D. S.
Right arrow Articles by Baker, T.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Mendham, D. S.
Right arrow Articles by Baker, T.
Agricola
Right arrow Articles by Mendham, D. S.
Right arrow Articles by Baker, T.
Related Collections
Right arrow Soil Fertility and Productivity
Right arrow Other Soil Management
Right arrow Forest Soils

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome