| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Soils, Universidade Federal Rural do Rio de Janeiro-FAPERJ Seropédica, 23890-000, RJ, Brazil
b Natural Resource Ecology Laboratory, Colorado State University, Fort Collins, CO 80523
c Embrapa Agrobiologia, Caixa Postal 74.505, Seropédica, 23890-000, RJ, Brazil
d Embrapa Soja, Caixa Postal 231, 86001-970, Londrina, PR Brazil
e Embrapa Trigo, Caixa Postal 451, Passo Fundo, 99001-970, RS, Brazil
f Dep. of Agronomy and Range Science, University of California Davis, CA 95616
* Corresponding author (jwsix{at}ucdavis.edu)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: CT, conventional tillage • MWD, mean weight diameter • NT, no-tillage • NV, native vegetation • SOM, soil organic matter
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Under NT, the soil fertility and SOM becomes stratified, with the highest concentration of SOM in the topmost soil layer (Six et al., 2000a). The conservation of crop residues on the soil surface also leads to soil moisture conservation and a decrease in the fluctuations of soil temperature (Lal et al., 1978). These soil characteristics generally improve the biological activities in the surface layers of the soil (Cattelan et al., 1997; Hungria et al., 1997).
Conventional tillage destroys the original soil structure, breaking up the macroaggregates into microaggregates and this results in an increase in microporosity and may increase soil bulk density values (Tisdall and Oades, 1980; Carpenedo and Mielniczuk, 1990). Recently, Six et al. (2002) reviewed the literature on bulk density values under NT and CT and found that in tropical and subtropical soils there were rarely differences in this parameter between the two tillage systems. In contrast, bulk density was generally higher in no-tilled than in conventionally tilled temperate soils. These results indicate differences in soil structural responses to implementation of NT practices in temperate versus (sub)tropical soils. Six et al. (2002) also reviewed changes in soil aggregation under NT, and found a much higher increase in aggregation in the early years of NT adoption in tropical and subtropical regions compared with temperate regions. They suggested that this was a result of the specific soil mineralogical characteristics (i.e., high Fe- and Al-oxide contents and a kaolinite dominated clay mineralogy) of (sub)tropical soils such as Oxisols and Ultisols. However, this suggestion was based on a synthesis of only five publications reporting data for Oxisols. In addition, these five studies reported magnitudes of responses in aggregation to NT that differed greatly. Consequently, little is known about the effect of NT on soil aggregation and its relationship with C stabilization for Oxisols.
In addition, different crops have different effects on soil aggregation and C accumulation: perennial grasses due to their extensive root systems are more efficient in soil aggregation than annual crops cultivated under CT (Tisdall and Oades, 1980, 1982). Legume crops used as green manures in crop rotation systems have an important role in the introduction of N from biological N fixation into the soil–plant system and the maintenance of soil C. Recent reports have emphasized the importance of legume green-manure crops for the accumulation of SOM in crop rotations managed under NT in southern Brazil (Gonçalves and Ceretta, 1999; Amado et al., 2001; Sisti et al., 2004).
The objective of this study was to determine the effect of different crop rotations and tillage on soil C and N accumulation, the formation of water-stable aggregates, and the quantity of aggregate-associated C at two sites in southern Brazil both situated on Oxisols. We postulate that the specific mineralogical characteristics of Oxisols (i.e., high Fe- and Al-oxide contents and a kaolinite dominated clay mineralogy) alter the relationship between aggregate formation, soil C and N stabilization and management observed in temperate soils.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The second site was at the Experimental Station of the Embrapa Soybean Research Center in Londrina, Paraná State, located 23°23' S lat. and 51°11' W long., 566 m altitude. The area was originally forested (tropical rainforest) but has been under conventional agriculture since the 1980s, with two crops a year: soybean in summer and wheat in winter. However, from 1994 until implementation of the current experiment in 1997, management was changed to NT with a sequence of soybean/lupin (Lupinus sp.)–maize (Zea mays)/oats (Avena sativa)–soybean/oats.
At both sites, the soils were Oxisols with clayey texture (typic Haplorthox) (Table 1). Both sites had NV, CT, and NT treatments. However, the NV plots were located in adjacent areas near the field experiment plots. The CT plots at Londrina were disk plowed to a depth of 20 cm before planting. The CT treatment disturbed the soil less at Passo Fundo than at Londrina because only the winter cover crop was disk plowed, whereas the summer crop was direct drilled. In the NT treatments, both the summer and winter crops were direct drilled.
|
At Londrina, the crop rotations utilized were (Table 1): (R1)–soybean/lupin–maize/oats; (R2)–soybean/wheat–soybean/lupin, and (R3)–soybean/lupin–maize/wheat.
Soil Sampling and Aggregate Separation
In July 2001, soils from each treatment were sampled at two depths (0–5 and 5–20 cm). Samples were taken to evaluate the soil bulk density using Kopek rings (5.1-cm i.d. and 4.6 cm height) inserted into the side of a trench approximately 50 x 50 cm square. For these samples taken at the Londrina site the porosity of the soil was determined by placing the Kopek cylinders on a tension table. When the samples were taken, the litter layer (where present) was removed. Once in the laboratory, the moist soil was passed through an 8-mm sieve by gently breaking apart the soil and the samples then air dried, and stored at room temperature.
The method of aggregate separation was adapted from Elliott (1986). Briefly, aggregates were separated by wet sieving the air-dried soil through a series of three sieves (2000, 250, and 53 µm). The air-dried soil was quickly submerged in deionized water on top of the 2000-µm sieve, resulting in slaking of the soil. Aggregate separation was achieved by manually moving the sieve approximately 3 cm up and down for 50 times during a period of 2 min. After the 2-min cycle, the stable >2000-µm aggregates were gently back washed off the sieve into an aluminum pan. Floating organic material (>2000 µm) was decanted and discarded because this large organic material is, by definition, not considered SOM. The aggregates were oven dried (50°C), weighed and stored in glass jars at room temperature.
The MWD was calculated as:
 |
Where M is the proportion of the soil weight in the aggregate class with a size given in the subscript.
Carbon and Nitrogen Analysis
Carbon and N concentrations were measured with a LECO CHN-100 analyzer (Leco Corp., St. Joseph, MI) for the aggregate-size fractions. For appropriate comparisons between treatments and sites it was necessary to correct for the sand content in each aggregate-size class (Elliott et al., 1991). The sand content was determined by isolating, drying, and weighing the >53-µm fraction after shaking 5 g of aggregates in 15 mL of water with 12 glass beads for 18 h.
Calculation of Carbon and Nitrogen Stocks
To correct for differences in soil bulk density between tillage and/or rotation treatments, stocks of C and N in the soil to a depth of 20 cm were calculated using the procedure described by Angers et al. (1997) based on equal mass of soil in the soil profile.
Iron and Aluminum Analysis
Non-crystalline inorganic forms of Fe and Al were determined by the dithionite-citrate method described by Blakemore et al. (1987). All aggregate-size classes from the Londrina site were analyzed by taking a 1-g subsample, which was dispersed in 50 mL of a solution containing 20 g L–1 of sodium dithionite and 220 g L–1 sodium citrate. This suspension was shaken overnight and left to stand for 24 h. A 1-mL aliquot was then diluted to 50 mL and left to stand for 2 d longer. The supernatant was analyzed for Fe and Al by atomic absorption spectrophotometry (Zeeman Atomic Absorption Spectrometer, 4100 ZL, PerkinElmer, Wellesley, MA).
Statistical Analysis
In Londrina, the field experiment (2 tillage treatments x 3 rotations) was arranged in a randomized complete block (factorial) design with four replicates. The Passo Fundo field experiment (2 tillage treatments x 4 rotations) was arranged in a randomized complete block (split plot) design with three replicates with the tillage treatments in the main plots and rotations in subplots. At both sites, the areas of NV were outside of the experimental design and therefore not included in the statistical analyses. The data were initially analyzed using MSTAT-C software (Michigan State University, East Lansing, MI) for ANOVA to determine the effects of the main variables (tillage and rotation) on the measured parameters. Where it was necessary to compare C or N contents of different aggregate-size classes, these size classes were considered as further split-plots.
Since the experiment at Londrina was a factorial design with all treatments in the main plots, it was necessary to examine the effect of the different crop rotation on any parameter within the same tillage treatment. Therefore, the variance associated with the difference between the individual means was compared with that of the residual and the F test was applied.
For the split plot design of the experiment at Passo Fundo, we used the procedure described by Little and Jackson-Hills (1978). The calculation of the least significant difference between means (LSD – Student) was applied where:
 | [1] |
These procedures were performed on the software SIS-VAR, produced by the Federal University of Lavras (UFV), Lavras, Minas Gerais. Statistical significance was determined at the P < 0.05 level except if indicated differently.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
In the Londrina experiment, while residue C and N inputs from each crop were very different in different years (Sisti et al., 2004) the total C and N inputs for the different rotations and tillage treatments were very similar ranging from 45 to 46 Mg C ha–1, and from 1.34 to 1.66 Mg N ha–1 (Sisti et al., 2004).
Whole Soil Characteristics
At both sites, the total organic C and N in the 0- to 5-cm depth interval decreased in the order NV > NT > CT (Table 2). The stocks of organic C and N in the two depth intervals were significantly higher under NT compared with CT at both Passo Fundo and Londrina except for C in the 5- to 20-cm depth interval at Passo Fundo. There were no significant differences in organic C when the crop rotations were compared within tillage systems. The organic C levels in the 0- to 20-cm depth at the Passo Fundo and Londrina sites were remarkably similar.
|
Aggregate-Size Distribution
The MWD was generally higher under NT than under CT, and this difference was significant (P < 0.01) at Londrina, being 55 and 26% higher in the 0- to 5- and 5- to 20-cm depth intervals at this site, respectively (Table 3). In the NT of the Passo Fundo site, MWD was higher in the soil under the more diverse rotations (R2) than under the continuous wheat–soybean (R1) but this difference was only statistically significant in the 5- to 20-cm depth interval. At Londrina, the last four crops were soybean, lupin, maize, and oats in R1; for R2 this sequence was soybean, wheat, soybean, and lupin; and for R3 maize, oats, maize, and wheat. Across the treatments, the MWD seemed to bear no relationship to legume–cereal content of the rotations.
|
|
The proportion of macroaggregates in the NV areas was higher than in cropped areas. At Londrina, the macroaggregates constituted approximately 94% of total soil dry weight (0–5 cm) in the NV area, 21 and 36% higher than under NT and CT, respectively. In contrast, the macroaggregate proportion at Passo Fundo (0–5 cm) was 84% of total dry soil weight for NV and only decreased by 4% under NT, and 14% under CT.
Aggregate Carbon and Nitrogen Concentrations and C/N Ratio
At Passo Fundo, organic C and N concentrations were significantly higher in the three larger aggregate-size classes (>53 µm) under NT when compared with CT in the 0- to 5-cm depth interval (Tables 5 and 6) and this was reflected in the total C and N stocks at this depth interval (Table 2). No significant differences in C and N content of these aggregate classes were found between NT and CT in the 5- to 20-cm depth interval. In the 0- to 5-cm depth interval of the NT at Passo Fundo, there were generally lower organic C and N contents of the micro- and macroaggregates in the R2A rotation than in the other rotations.
|
|
At both sites, there were no consistent differences in C/N ratio of any aggregate-size class between rotations or tillage treatments. However, there was clearly a considerably lower C/N ratio in all aggregate-size classes (>53 µm) of the soil under the NV than those under NT and CT treatments at both sampling depths.
With regard to the C content of the aggregate-size classes, we did not observe any trends across aggregate-size classes, except the largest aggregate-size class in Londrina had a higher C content than the other aggregate-size classes (Table 5). The same was observed for the N content of the aggregates (Table 6). In addition, there were no significant differences in C/N ratio between aggregate-size classes (Table 7).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Even though the Londrina experiment had only been established 4 yr before the time of sampling for this study, it was already possible to detect significantly higher stocks of C and N in the 0- to 5- and 5- to 20-cm depth intervals in the soil under NT compared with CT (Table 2). This fast response might be related to the high clay content (720 g kg–1) and/or depleted soil C level due to at least 15 yr of CT practices before the installation of the experiment. Similarly, the long-term CT practices at Londrina probably resulted in the similar organic C and N concentration in Londrina and Passo Fundo, even though the higher clay content at Londrina than at Passo Fundo (630 g kg–1) would be expected to lead to greater organic C concentration at Londrina (Feller and Beare, 1997).
Aggregation
At both sites the data showed consistently that CT decreased the proportion of the largest macroaggregate class (>2000 µm) in comparison with NT management and there was a corresponding increase in the proportion of the 53- to 250-µm aggregate class. This breakdown of the larger aggregates by tillage was also evidenced by the lower MWD of aggregates under CT, which was statistically significant (P < 0.05) at both sites except in the 5- to 20-cm depth interval at Passo Fundo.
At Passo Fundo, the composition of crops in the three R2 rotations was identical (soybean twice and maize once as a summer crop, and wheat, oats, and vetch as winter crops), only the year in which the rotation started was different. However, the R1 rotation consisted of a continuous annual cycle of wheat–soybean. Our results suggest that the addition of vetch, maize, and oats to the wheat–soybean rotation, favored aggregate formation under NT in that the MWD of the aggregates and the proportion of macroaggregates in R2 was higher than in R1, especially in the 5- to 20-cm depth interval.
The observation that in both tillage treatments, the Londrina-R1 (sampled immediately after oats, preceded by maize) showed the highest MWD and proportion of macroaggregates suggests a positive influence of the cereal root systems on macroaggregate formation. Campos et al. (1999) found a similar tendency when comparing the effect of different cover crops on soil aggregation under NT: Oats promoted a greater aggregate stability when compared with lupin, vetch, or fallow. However, Paladini and Mielniczuk (1991) and Campos et al. (1999) observed that 3 yr of cultivation of legume cover crops with CT practices, but alternated with NT maize, promoted a higher proportion of macroaggregates.
The smaller differences in macroaggregation between the cultivated systems and the NV in Passo Fundo than in Londrina are probably due to different site histories. The long-term experiment at Passo Fundo was sampled 14 yr after its establishment, before this time the area was NV. In contrast, the Londrina experiment had only been established 4 yr before sampling but the area was previously under CT for at least 15 yr.
Aggregate Associated Carbon and Nitrogen
Within tillage treatments, the effect of crop rotation was observed in the largest aggregate class of CT. Under the R3 at Londrina, where there was a larger input of residues with high C/N ratio in the year of sampling, a higher concentration of C in the macroaggregates was observed as compared with the R2, which was under soybean following lupin at the time of sampling. Both of these legume crops have a low C/N ratio. At the Passo Fundo site, in rotation R2C (soybean/wheat–soybean/vetch–maize/oats) there was a higher input of residues of high C/N ratio in the year immediately before sampling than in the other three rotations (Sisti et al., 2004), where in the last year the crop sequence was either soybean–wheat (R1 and R2A) or soybean–vetch (R2B). However, there was no apparent tendency for the C/N ratio of the organic matter associated with the aggregates to be any lower in these latter treatments (Table 7). The results from the Londrina study were very similar; in the year immediately before sampling, in R2 the sequence was of two legumes (soybean followed by lupin) and for R1 and R3 the crops were both cereals (maize–oat and maize–wheat, respectively), but this did not result in consistent differences in C/N of all aggregate-size classes.
The lack of differences in C and N content and C/N ratio across aggregate-size classes indicate that, in contrast to 2:1 clay-dominated soils, the 1:1 clay plus oxide dominated soils we studied, do not express an aggregate hierarchy (Tisdall and Oades, 1982). This is in accordance with several studies (Elliott et al., 1991; Feller et al., 1996; Oades and Waters, 1991; Six et al., 2002), which reported the inexistence of aggregate hierarchy in 1:1 clay plus oxide dominated soils. The high Fe and Al contents, along with kaolinite, in the Londrina and Passo Fundo soils suggest that Fe and Al oxides serve as the principal agents of aggregation. Therefore, aggregate formation is less reliant on the binding action of SOM, resulting in no aggregate hierarchy being expressed (Oades and Waters, 1991; Six et al., 2000a, 2000b, 2002).
The lack of existence of aggregate hierarchy and the decrease in C and N content within an aggregate-size class in the order NV > NT 
CT precludes the direct linkage between C and N accumulation and aggregation as proposed by Elliott (1986) and Six et al. (2000a). These two studies, based on long-term tillage experiments in temperate soils, suggested that the loss of C with increasing cultivation intensity (i.e., NV < NT < CT) could be partly explained by the observed loss of C-rich macroaggregates and increase in C-depleted microaggregates. These observations were made in temperate soils dominated by 2:1 clays expressing an aggregate hierarchy. In our two Oxisols, we observed a decrease in macroaggregates and increase in microaggregates, but one aggregate-size class was not depleted in C compared with the other. Consequently, the direct link between aggregate loss and C loss is not shown in our data. Six et al. (2000b) also failed to observe this link in a soil with mixed mineralogy from Kentucky.
At Passo Fundo the C concentration in the aggregates under both tillage treatments and in both depth intervals were quite similar to the C concentrations found in the forest. This indicates that the soil aggregation and the C status after 14 yr of NT adoption are close to those under the NV. At the Londrina site, in the 0- to 5-cm depth, in all classes of aggregates, the C concentrations in the forest were higher than in the cultivated area, but similar in the 5- to 20-cm depth. Similar results were found by Elliott et al. (1991), Feller et al. (1996), and Six et al. (2000a) in kaolinitic soils.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Received for publication January 6, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS 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 | |||
