| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Canadian Forest Service, Natural Resources Canada, 5320 122nd St., Edmonton, AB, Canada T6H 3S5
b British Columbia Ministry of Forests, 3401 Reservoir Road, Vernon, BC, Canada V1B 2C7
* Corresponding author (bkishchu{at}nrcan.gc.ca)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
65 mg kg-1) at both S-sufficient and S-deficient sites, and were not different between the sites. Soluble inorganic SO4 concentration was significantly greater in B horizon soil at the S-sufficient sites than at the S-deficient sites. Soluble inorganic SO4 was significantly and positively correlated with organic SO4 (r2 = 0.24; P = 0.03) and organic C (r2 = 0.76; P < 0.001). Foliar SO4-S was significantly greater at the S-sufficient sites, and was significantly and positively correlated with B-horizon soluble inorganic SO4 (r2 = 0.84; P < 0.001), organic SO4 (r2 = 0.46; P = 0.032), and organic C (r2 = 0.72; P = 0.002). The highest R2 for a regression model between soil or foliar properties and response to N fertilization was for a model containing foliar N, foliar SO4-S, and B horizon organic C (R2 = 0.81; P = 0.004). Cycling of soluble inorganic SO4 through organic SO4 in mineral soil appears to be the process limiting S availability on S-deficient sites. Organic SO4 and soluble inorganic SO4 concentrations at the S-sufficient sites are maintaining foliar SO4-S at levels required for response to N fertilization.
Abbreviations: B.C., British Columbia • IC, ion chromatography
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Nitrogen deficiencies are widespread in B.C. forests and consequently fertilization research has emphasized N additions. Although response of lodgepole pine to N fertilization has been well documented (Weetman et al., 1988; Brockley, 1996, 2001b), the response is variable and is not reliably predicted by measures of stand N status, such as foliar N concentration and soil mineralizable N (Brockley, 1991; Weetman et al., 1992). On some sites in B.C., growth response of lodgepole pine to N fertilization appears to be limited by low S status, with N additions inducing or exacerbating S deficiencies (Brockley and Sheran, 1994). A large number of fertilizer research trials have been established to document response to N and N + S fertilization. At many locations in the central interior of B.C., combined N and S additions have significantly increased radial increment above that achieved with N alone (Brockley 2000, 2001b).
Foliar total S and SO4-S concentrations and N/S ratios are currently used to evaluate the S status of lodgepole pine stands and to predict their responsiveness to fertilization with N and N + S (Brockley 2000, 2001a). Sulfur in excess of that required to balance N in protein formation accumulates as SO4-S in foliage, and the usefulness of foliar SO4-S as a diagnostic and predictive tool was previously demonstrated in fertilization research with both radiata pine (Pinus radiata D. Don) and Douglas-fir [Pseudotsuga menziesii (Mirb.) Franco] (Kelly and Lambert, 1972; Turner et al., 1977, 1979; Lambert and Turner, 1998).
Although reasonably successful for predicting lodgepole pine response to fertilization (Brockley, 1991, 2000), these measures do not address the underlying cause of low S status on these sites. There is little information about the soil properties that determine the productivity of lodgepole pine in B.C., and in particular, there has been a lack of knowledge about the site and soil properties contributing to S deficiencies.
Most soil S is in organic forms. Within the organic S pool, two fractions have been described: organic SO4 and C-bonded S. McGill and Cole (1981) proposed a dual mechanism for S mineralization based on these fractions. Organic SO4 is mineralized extracellularly in response to microbial demand for S (biochemical mineralization), while the C-S bond is oxidized intracellularly for its energy release, analogously to N (biological mineralization).
The contribution of organic forms of soil S to S availability has not been well characterized. The nature of soil S fractions on lodgepole pine sites in B.C. and their relationship to growth response were investigated in this study. The objectives of this research were to identify the soil properties related to S availability on these sites, and to evaluate relationships among soil properties, foliar nutrition, and growth response to N fertilization.
|
MATERIAL AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
There is documented evidence of differential growth response to a combined application of N + S compared with N alone for stands at the six S-deficient sites (Brockley and Sheran, 1994; Brockley, 2000). Foliage from these stands indicated low S status (Brockley, 2000), with total S concentrations below 1.0 g kg-1, SO4-S concentrations below 80 mg kg-1, and N/S ratios approaching 14.6, above which S deficiency may exist (Kelly and Lambert, 1972). The occurrence of nutrient deficiencies other than N and S has not been systematically tested in these stands. However, based on foliar nutrient interpretative criteria developed for lodgepole pine (Brockley, 2001a), other nutrient deficiencies are not indicated, with the exception of a severe B deficiency induced by fertilization at the Blackwater Creek site (Brockley, unpublished data, 1996), resulting in relatively small growth responses to both N and N + S application (Brockley 2000).
Statistically significant growth responses to fertilization with N alone with no additional response obtained with a combined application of N + S were documented in three of the four S-sufficient stands (Brockley, 2000). The fourth site (Canal Flats) was inferred to be S-sufficient on the basis of favorable foliar S status and a large growth response following fertilization with N alone (Brockley, 2000).
Soils in both locations were Spodosols, Alfisols, and Inceptisols (Table 1). There was no previous information about the S properties of soils at these sites.
Soil Sampling
Three 1-m width by 1-m length by 1-m depth soil pits were dug in unfertilized areas at each of the ten study sites in July and August 1992 and August 1993. Surface organic soil (Oi, Oe, Oa horizons) and mineral soil to the BC or C horizon were described and sampled by horizon. Soil profiles were described according to B.C. Ministries of the Environment and Forests (1990) and Canadian Soil Information System criteria (Agriculture Canada Expert Committee on Soil Survey, 1983).
Samples were kept cool until returned to the lab (maximum 10 d) and air dried. Air-drying of organic samples may result in greater extractable-S values than obtained with fresh samples (David et al., 1982); however, all samples collected from soil profiles were treated consistently. Green tissue, woody debris, and cones were removed from forest floor material prior to grinding in a Wiley Mill with a 2-mm screen (Arthur H. Thomas Co., Philadelphia, PA). Mineral soils were passed through a 2-mm sieve. Subsamples of the <2-mm fraction were crushed to a smaller mesh size where required for analyses.
Soil nutrient concentrations were determined for individual horizons. To allow averaging of results for samples within and across sites, results were considered for three horizons or depth-weighted means of horizons present in all soil profiles: the Oe horizon, the average of the first three B horizons weighted by horizon thickness (except for the Canal Flats site) and the lowermost horizon in each profile (BC or C horizon). The mean depth to the lower boundary of the third B horizon was 54 cm for the S-deficient sites (n = 18) and 47 cm for the S-sufficient sites (n = 12). This portion of the profile reflects the stratum where processes of soil development are occurring at these sites. This depth also corresponded to the average depth at which there was a vertical transition in fine root abundance from plentiful (10 to 100 dm-2) to few (<10 dm-2) (B.C. Ministries of the Environment and Forests, 1990). The average depth to rooting transition was 49 cm over all sites (n = 30).
Soil Analysis
All soil data are expressed per unit mass of oven-dry soil (105°C for 24 h). Sulfur determinations are shown schematically in Fig. 1
. Total S was determined with a Leco SC-132 S Analyzer (Leco Corp., St. Joseph, MI) (David et al., 1989). Soluble S was extracted from forest floor samples only in 0.01 M NH4Cl and measured using an ARL 3560 ICP spectrophotometer (Applied Research Laboratories, Valencia, CA) (Kalra and Maynard, 1991). The soluble S fraction includes soluble inorganic S, organic SO4, and C-bonded S.
|
Total inorganic SO4 was taken as the total inorganic S fraction. Inorganic S forms more reduced than SO4 were assumed to be negligible in these soils. Soils were generally well to rapidly drained, with mottles indicative of gleying present only in lower horizons of six profiles.
Hydriodic-acid reducible S was determined colorimetrically in soil by a modified Johnson-Nishita procedure (Johnson and Nishita, 1952; Kowalenko and Lowe, 1972). Sulfur reduced by HI includes both organic- and inorganic-SO4 forms (Freney, 1961; Kowalenko, 1993b). Organic SO4 was estimated by subtraction of total inorganic SO4 from HI-S (Landers et al., 1983). The organic SO4 fraction contains both ester sulfates and sulfamates. Carbon-bonded S values were obtained by subtraction of HI-S from total S. Total organic S is the sum of organic SO4 and C-bonded S.
Soil pH was measured in distilled water and in 0.01 M CaCl2 (Kalra and Maynard, 1991). Total C and N were determined by a Leco CHN-600 Analyzer (Leco Corp., St. Joseph, MI). Organic C was determined by wet oxidation (Kalra and Maynard, 1991) in mineral soil samples with pH of 
6.5 (0.01 M CaCl2), since the presence of inorganic C was assumed. For all other samples, organic C was assumed to be equivalent to total C, although small quantities of charcoal may have been present in some forest floor samples. Element ratios were determined using organic C and total N or S concentrations.
Extractable NH4 and NO3 were extracted in 2 M KCl and measured on a Technicon auto analyzer (Technicon Instrument Corp., Tarrytown NY). Available P was determined by extraction in Bray-P1 solution, reduction by Murphy and Riley (1962) solution, and measured colorimetrically with a UV/Visible spectrophotometer (Felsted, Dunmow, Essex, England) (Kalra and Maynard, 1991). Exchangeable Ca, Mg, and K were extracted in 1 M NH4OAc at pH 7.0 and measured by ARL 3560 ICP spectrophotometry (Kalra and Maynard, 1991).
Soil Incubations for Mineralizable Nutrients
Soil samples were collected in August and September 1993 for estimates of mineralizable nutrients. One sample each of forest floor and mineral soil were collected from each of five unfertilized plots at the ten study sites. Forest floor material (Oe or Oe-Oa) was obtained by removing fresh litter and collecting the underlying organic material to the surface of the mineral soil.
Mineral soil was collected to a depth of 30 cm below the mineral surface. Soils were passed through a 4-mm sieve prior to bagging to remove large roots and coarse fragments. Samples were kept refrigerated (5°C) until the incubation experiment was established (maximum 15 d). Moisture content (105°C for 24 h) was determined for each sample. A subsample was extracted moist to estimate initial extractable nutrient concentrations.
A closed aerobic laboratory soil incubation was established to estimate net mineralizable-SO4, NH4, and NO3 (Maynard et al., 1983; Valeur and Nilsson, 1993). A constant oven-dry mass equivalent of moist soil was used for each sample incubation. Soils were placed in a 1-L plastic tub and brought to a moisture content equivalent to field capacity and 80% of field capacity for forest floor and mineral soil, respectively (Maynard et al., 1983).
Field capacity for each sample was determined by saturating a uniformly filled cylinder of soil and allowing the soil to equilibrate on a tension table under the tension of a 1-m water column for 24 h. The moisture content at field capacity was estimated as the moisture content of the soil following equilibration, and was determined by difference in mass after drying at 105°C for 24 h. This method provided a basis for establishing a moisture content for these soils in the disturbed conditions of the incubation. Each tub was covered with a thin polyethylene bag allowing O2 and CO2 exchange but minimal moisture loss (Eno, 1960; Gordon et al., 1987). The soils were incubated at 22°C for 32 d. Samples were reweighed after 14 d and moistened to the original mass.
Subsamples of fresh and incubated soils were extracted moist for initial and final nutrient concentrations, respectively. Extractable SO4, including organic and inorganic SO4 forms, was used to estimate mineralizable SO4 (Kowalenko and Lowe, 1975). Forest floor samples were extracted in 0.01 M NH4Cl (Maynard et al., 1987) and mineral soils in Ca(H2PO4)2 (500 mg P L-1) (Kalra and Maynard, 1991). Reducible S was determined using HI reduction in the extracts (Kowalenko and Lowe, 1972). Forest floor and mineral soil samples were extracted in 2 M KCl for determination of extractable NH4 and NO3 on an Alpkem RFA 300 Autoanlyzer (OJ Analytical, College Station, TX). Net mineralizable nutrient concentrations are calculated as final extractable values less initial values.
Foliar Sampling and Analysis
Current-year foliage of unfertilized trees was sampled in the fall of the year of soil sampling. Foliage was sampled from 15 trees per site and combined to create three composite samples per site. Total S in current-year foliage was determined on a Leco SC-132 S Analyzer, and SO4-S by extraction in 0.1 M HCl, HI reduction, and colorimetric determination (Johnson and Nishita, 1952). Total N was determined by H2SO4–H2O2 digestion and colorimetric determination on a Technicon II Autoanalyzer (Technicon Instrument Corp., Tarrytown NY) (Parkinson and Allen, 1975).
Statistical Analysis
Results of individual samples for the Oe horizons, the depth-weighted averages of the first three B horizons in each profile, and the lowermost horizons in each profile were averaged over the three soil pits within each site. Soil mineralizable nutrients and foliar nutrients from individual samples within a site were also averaged to obtain a site mean. Site means were averaged to obtain means for S-deficient and S-sufficient groups of sites. Differences in soil and foliar nutrients between S-deficient and S-sufficient groups of sites were determined using a nested analysis of variance with individual sites nested within the S-deficient or S-sufficient groups (
= 0.05) (Hicks, 1982). All data were tested for homogeneity of variance ( = 0.10) prior to analysis of variance. Where variances were heterogeneous, data were transformed, and analyses performed on the transformed data. Logarithmic, reciprocal, reciprocal of square root, cube, and arcsine transformations were used. Statistical analyses were done with SPSS 7.0 (SPSS Inc., 1996).
Bivariate correlations of chemical properties in B horizon soil were done with individual soil samples from the first three B horizons per soil profile. Net mineralizable nutrients in the forest floor and upper mineral soil were correlated with S constituents in the Oe and upper B horizons, respectively. Correlations with mineralizable nutrients were done with site means of Oe horizon and depth weighted B horizon properties as there were different numbers of samples for mineralizable nutrients and Oe and B horizons.
Simple linear regressions using foliar S properties as dependent variables and soil properties as independent variables were determined. Regressions were done between soil and foliage site means, as there were different numbers of soil and foliage samples. For all sites except Canal Flats, the depth-weighted mean of the first three B horizon per soil profile was used as the site mean for all B horizon soil properties. For the Canal Flats site, the depth-weighted means of soil properties determined independently of total S (HI-S, soluble inorganic SO4, total inorganic SO4, and organic SO4) were used as the profile mean for regression of soil and foliar properties. For total S and parameters dependent on subtraction from total S, only data from the first B horizon per pit at the Canal Flats site were used as the soil term in the regressions.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
65 mg kg-1 (Table 2), with a range of 60 to 112 mg kg-1. Soluble inorganic SO4 concentration in B horizon soil was significantly greater at the S-sufficient sites than in the S-deficient sites. Data from the first B horizon only in each profile at the Canal Flats site (S sufficient) was used in means of S-deficient and S-sufficient sites, since HI-S concentrations were greater than total S concentrations in a small number of lower B horizons at this site, and could not be subtracted from total S. The B horizon HI-S values at this site were the highest observed in the study (>130 mg kg-1). Soils at the Canal Flats site were calcareous (17% CaCO3 equivalent), and the observed high HI-S concentrations are most likely because of the inclusion of SO4 present as gypsum (CaSO4) or CaCO3–SO4 mineral complexes. Total S values in this study were approaching the detection limit of the analytical instrument (Prietzel et al., 1995) and there may have been some experimental error for all samples associated with these low concentrations. The discrepancy between total S and HI-S values in these samples was likely because of the combination of low total S, and high HI-S concentrations in calcareous soils from this site. Exclusion of samples with greater HI-S than total S values lowered the means of S-sufficient sites for HI-S and organic SO4 and decreased differences in these constituents. Total S concentration of C horizon soils did not differ significantly between S-deficient and S-sufficient sites (Table 2). Greater than 95% of total S in Oe, B, and C horizon soils was organic S.
Mineralizable Sulfate and Nitrogen in Forest Floor and Mineral Soil
There was no difference in net mineralizable SO4, NH4, or NO3 in the forest floor between S-deficient and S-sufficient sites (Table 4). Forest floor net mineralizable SO4 was positively correlated with Oe horizon organic SO4 (r2 = 0.84; P < 0.001), total S (r2 = 0.78; P = 0.001), forest floor soluble SO4 (r2 = 0.78; P = 0.001), C-bonded S (r2 = 0.64; P = 0.006), and negatively correlated with C/S (r2 = 0.50; P = 0.021).
|
Relationships Among Soil Properties
Chemical properties of the first three B horizons per soil profile were correlated to evaluate relationships between soluble inorganic SO4 as the S form available to plants, and other B horizon soil properties (Table 5). Soluble inorganic SO4 was significantly and positively correlated with organic SO4, organic C, and total N, but was not correlated with total S, organic S, C-bonded S, or total inorganic SO4. Organic SO4 and C-bonded S were poorly correlated with each other, while C-bonded S was positively correlated with organic S and with total S. Correlations among soil fractions must be interpreted conservatively since determination of some fractions is by subtraction of other measured fractions.
|
|
|
Relationships between foliar SO4-S and soil pH or exchangeable Ca were determined to identify whether properties associated with calcareous soils were involved in foliar S nutrition. These regressions were not significant.
In contrast to mineral soil, Oe horizon S properties were more strongly related to foliar total S than to foliar SO4-S. This may be explained by the development of forest floor from foliar litter, with Oe horizon properties reflecting foliar nutrition. There were significant, positive relationships between foliar total S and forest floor soluble SO4 (r2 = 0.45; P = 0.035) and soluble S (r2 = 0.47; P = 0.028). However, the slopes of the regression lines for these relationships are approaching zero, indicating little change in foliar total S with increasing forest floor S.
Relationships Among Soil Properties, Foliar Nutrition, and Growth Response
Linear regressions were done with site means of growth response (dependent variable) and B horizon soil and foliar nutrition variables (independent variables) (n = 10). Sixth-year growth response (diameter at breast height increment) to 200 kg N ha-1 relative to the control treatment was the growth response variable.
Variables were combined in multiple linear regressions, with variables added manually. Soil properties selected as independent variables had significant regressions with foliar SO4-S concentration. The variables tested were B horizon soluble inorganic SO4, organic SO4, and organic C concentrations determined following trial establishment and prior to the sixth-year growth response measurement and foliar N and SO4-S concentrations at the time of trial establishment.
Relationships were weak between individual soil properties and growth response to N fertilization. Regressions of the individual soil constituents organic SO4, soluble inorganic SO4, or organic C concentration with growth response each had r2 values <0.10, and regressions were not significant. Combining two and three soil constituents increased R2 values but regressions with multiple soil variables alone were not significant.
Regressions were improved when foliar N and S properties were used as independent variables. The highest r2 or R2 value for any single or combined foliar nutrition variables was obtained with foliar N and SO4-S concentrations (R2 = 0.73; P = 0.004). These are the independent variables currently used by the B.C. Ministry of Forests for predicting lodgepole pine response to N fertilization (Brockley, 2000). The foliar N term in the regression model was not significant and the regression equation containing foliar SO4-S is shown in Table 7.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Soluble inorganic SO4 concentration in B horizon soils was the property most strongly related to S availability as indicated by foliar SO4-S concentration. The relationship between soluble inorganic SO4 and foliar SO4-S is not unexpected, as soluble inorganic SO4 is the available S. Soluble inorganic SO4 is the most active pool of soil S, and is maintained by inputs of mineralized and desorbed S (Johnson and Henderson, 1979; Feger, 1995). Potential sources of soluble inorganic SO4 are mineralization of organic SO4 and C-bonded S, and the total inorganic SO4 pool.
Organic SO4 appears to be the pool contributing most directly to soluble inorganic SO4 and to S availability on the basis of two observations. First, soluble inorganic SO4 was significantly correlated with organic SO4 but not with C-bonded S or total inorganic SO4. Secondly, the relationship between foliar SO4-S and soil organic SO4 concentration was stronger than was observed for total inorganic SO4 or C-bonded S. The importance of an organic-S pool in S availability on these sites is supported by the strong correlations between soil soluble inorganic SO4 and organic C, and between soil organic C and foliar SO4-S concentration. A positive correlation between soluble inorganic SO4 and organic C has been found in other forest soils (Neary et al., 1987).
The major source of soil organic SO4 is production by microbial biomass (David et al., 1987). Incorporation of inorganic SO4 into organic SO4 in forest soils is enhanced by available C supply, either as exogenous C (Fitzgerald et al., 1983) or inherent soluble C (David et al., 1983, 1987). Organic-S production following clearcutting was shown to be a microbial process that may have diminished as labile-C sources decreased several years after harvesting (Spratt, 1997). The mineralization of organic SO4 to inorganic SO4 results from hydrolysis by extracellular sulfatase enzymes produced by bacteria, fungi, roots, and soil fauna (Fitzgerald, 1976; Houghton and Rose, 1976). There is also a strong positive correlation between sulfatase activity and soil organic C (Tabatabai and Bremner, 1970). Both the production and mineralization of organic SO4 are influenced by soil C, and cycling of organic SO4 is likely to be affected by soil organic matter status.
The specific fire history of these sites has not been examined here; however, fire has undoubtedly been a factor in their development. Past fires may be influencing S cycling through soil organic matter on these sites. Sulfur is volatilized to SO2 during the oxidation of soil organic matter and foliage (Tiedemann, 1987; Agee, 1993), and low soil C status may be related to loss of soil organic matter during fires (Johnson, 1992). Burning of forested sites was found to decrease soil arylsulfatase activity, which would be expected to influence organic-SO4 dynamics (Eivazi and Bayan, 1996; Staddon et al., 1998). The role of fire and other disturbance in S and soil organic matter dynamics on these sites requires further investigation.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
Received for publication December 5, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 | |||
