Readers' Rating:

Rate this article

Send to a friend

THE SUSTAINABILITY OF AGRICULTURAL SYSTEMS depends, to a great extent, on the maintenance of soil properties within levels of variation that would allow their restoration and would not affect either crop production or the environment. Feasibility of sustainable agriculture is based on the knowledge of the effect of management practices on soil properties and how they affect soil-crop relationships (Francis and Clegg, 1990) in order to make sound management decisions. Crop rotation is a key practice within agricultural systems since it affects soil productivity, crop behavior, and need for inputs, and is closely related to the economic sustainability of the system (Power and Follett, 1987; Francis and Clegg, 1990). However, long-term conventional tillage-based cropping generally diminishes soil productivity.

Soil organic matter (SOM) varies due to the way different management practices alter mineralizationhumification processes and to the influence they have on the amount of C returned to the soil (Stevenson, 1986; Oades, 1984). Cropping including conventional tillage (Lamb et al., 1985; Havlin et al., 1990), long and/or frequent fallow periods (Campbell and Zentner, 1993), or low-residue-return crop combinations (Havlin et al., 1990; Peterson et al., 1993; Campbell and Zentner, 1993) lead to a decrease in SOM. Changes in SOC dynamics as a response to management practices are asso ciated with changes in the relative amount and in the activity of SOM fractions (Doran and Smith, 1987).

Conventional cropping initially produces a more intense mineralization of labile fractions, leaving the more recalcitrant fractions as remnant. Microbial biomass N and microbial biomass C (MBC) are very sensitive to changes in soil management and diminish with the years under cropping (Angers et al., 1992; Dalai and Mayer, 1986c). The use of MBN and MBC as early indicators of management-induced changes in soil organic components has been frequently suggested (Angers et al., 1992; Follet and Schimel, 1989). Likewise, variation in soil organic components is a determinant of soil physical condition (Doran and Smith, 1987). Conventional cultivation destroys macroaggregate structure and produces loss of structural stability due to the greater exposure of formerly protected SOM fractions (Tisdall and Oades, 1982) and the consequent loss of labile fractions (Cambardella and Elliott, 1993; Golchin et al., 1994), which are responsible for binding microaggregates into macroaggregates (Tisdall and Oades, 1982; Cambardella and Elliott, 1994).

Inclusion of pastures in the rotation restores soil properties affected by cropping (Campbell, 1978; Haynes et al., 1991). Pastures produce an increase in soil organic components because of a greater production of aboveground and root biomass (Tisdall and Oades, 1982; Haynes et al., 1991). Pasture root systems, especially those of grasses, play a key role in this recovery since they provide excellent distribution and mixing of large quantities of organic materials into the soil (Oades, 1984). Continuous growth and death of dense root systems is associated with increases in microbial biomass, intense production of binding agents (Tisdall and Oades, 1982; Haynes and Francis, 1993; Haynes et al., 1991), and greater proportions of labile SOM fractions (Cambardella and Elliott, 1992). In addition, absence of soil disturbance under pasture not only facilitates soil physical and biochemical environment improvement, but also stimulates recovery of soil fauna (Oades, 1984).

Management of production systems in the southeastern Buenos Aires Province (Argentina) traditionally comprised alternation of no longer than 4-yr periods of conventional cash crop production with 3- to 5-yr periods under pasture for extensive livestock raising. This management contributed to maintaining agricultural soil physical and chemical properties almost unaltered and, consequently, to conserving high fertility and productivity.

In the last two decades, however, greater profitability of cash crops made the producers switch to a more intense cropping without rotating with pastures (Darwich, 1991). Soil physical and chemical-biochemical properties are being degraded (Echeverria and Fer rari, 1993) and soil erosion and nutrient deficiency problems are becoming more frequent. This process of deterioration is worsened by the fact that conservation tillage systems have not been completely adopted by producers.

We propose that it is feasible to define crop-pasture rotations with a maximum and a minimum duration of the periods under conventional cropping and under pasture to keep the variation of some soil properties within limits of change that would contribute to sustainable management in the area mentioned above. To demonstrate this hypothesis, SOC, LPC, MEN, and ASI changes in a long-term crop-pasture rotation experiment conducted on a Typic Argiudoll at Balcarce, Buenos Aires Province, Argentina, were analyzed.


MATERIALS AND METHODS

A long-term crop rotation experiment has been conducted since 1976 at Balcarce, Argentina (37°45'S, 58°18'W, 138 m above sea level, 870-mm mean annual rainfall, 13.7°C mean annual temperature) on a fine, mixed, thermic Typic Argiudoll with 2% slope (no erosion). Before the beginning of the experiment, this field had been under pasture for several years. The surface horizon of this soil has a pH of 6, loamy texture, 33.1 cmolc kg^-1 cation-exchange capacity, and 5.0 mg kg^-1 Bray and Kurtz P. The experiment was carried out as a randomized complete block design with a split-plot treatment arrangement, three replicates, and 10 by 35 m experimental units.

Treatments assigned to main plots were referred to as rotations and comprised continuous cropping (Rotations 5 and 6) and crop sequences alternated with pastures (Fig. 1). Croppasture rotations included combinations of approximately 50:50 (Rotations 2 and 7) and 75:25 (Rotations 1, 3, and 4) ratios of time under cropping to time under pasture. Pastures lasted 2 yr (Rotations 1,4, and 7) and 4 or 5 yr (Rotations 3 and 2, respectively). Crops included were spring wheat (Triticum aestivum L.), potato (Solanum tuberosum L.), corn (Zea mays L.), sunflower (Helianthus annuus L.), oat (Avena sativa L.), and oat + hairy vetch (Vida villosa L.) as green manure.

Pastures included grasses and legumes usually employed in the area (orchardgrass [Dactylis glomerata L.], bulbous canary grass [Phalaris tuberosa L.], tall fescue [Festuca arundinacea Schreb.], perennial ryegrass [Lolium perenne L.], white clover [Trifolium repens L.], red clover [Triflolium pratense L.], and alfalfa [Medicago sativa L.]) at recommended rates, with predominance of grasses. Pastures were not grazed but were periodically cut and the material removed to simulate grazing. Seedbeds for either agricultural crops or pastures were conventionally prepared (moldboard plowing, disking, and harrowing). Tillage operations were started no less than 3 mo before the planting date of each crop. Nitrogen fertilization levels (O and 90 kg N ha^-1) were assigned to subplots.

Either crops or pastures were fertilized at planting with superphosphate according to soil analysis.

Soil samples were taken from experimental units in fall (March-April) every year between 1981 and 1993. Between 1981 and 1989, samples were taken only from non-N-fertilized subplots. Samples were taken from the surface layer (0-0.15-m depth) and composed of 15 to 20 subsamples. Samples for biochemical analyses were air dried, ground to pass a 2.0-mm-opening sieve, and stored until analysis. Samples for aggregate stability determination were taken only from non-N-fertilized subplots in 1982 through 1986, 1988, 1989, and 1992. Samples were hand disrupted, passed through an 8.0-mm-opening sieve, air dried, and stored until analysis.

Soil organic C was determined on all samples, reground to pass a 0.5-mm-opening sieve, by the Walkley-Black procedure (Nelson and Sommers, 1982).

To determine a labile organic fraction, the densimetric fractionation method proposed by Richter et al. (1975) was used.

Soil samples were reground to pass a 0.18-mm-opening sieve and suspended in a bromoform-ethanol solution adjusted to density 2.0 Mg m^-3 to separate the light fraction. The suspension was centrifuged at 4300 X g for 120 s. The supernatant was filtered through fiberglass filter paper (Whatman GFA) and analyzed for LPC as described above for SOC.

The chloroform fumigation-extraction method (Brookes et al., 1985) was used to determine MEN. Since samples had been stored air dry, prior to fumigation they were preincubated for 72 h at field capacity and 25°C. Two subsamples of each sample were weighed; one of them was chloroform fumigated for 24 h, then extracted with K2SO4. The extract was digested with H2SO4 and analyzed for mineral N through microdistillation (Bremner, 1965). The other subsample was not fumigated and extraction and analysis for mineral N was done as described above. Microbial biomass N (mg kg^-1) was calculated as follows:

MEN = F- 770.47

where Fis mineral N content (mg kg^-1) of fumigated subsamples after fumigation, T is mineral N content (mg kg^-1) of nonfumigated subsamples, and 0.47 is the approximate proportion of the MEN that is recovered by this method (Ferrari et al, 1992).

Light-fraction C and MEN were determined on soil samples from non-N-fertilized experimental units corresponding to Rotations 2, 3, and 5 (Fig. 1) in 1983, 1985, 1987, 1989, and 1991. Because of the lack of enough stored soil to run the analyses, there were some missing results of these two variables.

Fig. 1. Crop rotation scheme applied since the beginning of the experiment.

Table 1. Results of the analysis of variance for four soil variables in a crop-pasture rotation experiment at Balcarce, Argentina.

To enlarge the image, click on it.



Aggregate stability was determined as the difference of mean weight diameter (MWD) between dry sieving and wet sieving in water (De Boodt et al., 1961). Air-dry samples were sieved through a three-sieve nest (4.80-, 3.36-, and 2.00-mmopening sieves). A proportion of each dry fraction obtained was wetted and sieved under water through a six-sieve nest (4.80-, 3.36-, 2.00-, 0.84-, 0.50-, and 0.30-mm opening sieves) for
30 min. Results were expressed as ASI which was calculated as follows:

ASI = (0.26/P)100

where 0.26 is the average of MWD difference of samples taken from similar soils not disturbed by tillage, and P is MWD difference between dry and wet sieving of samples taken from the experimental units.

The model proposed by Bartholomew and Kirkham (1960) was used to describe the relationship between the results obtained for each variable and either years under cropping or years under pasture:

V_t = V_e - (V_e - V_o) exp (-rt)

where V, is the value of the variable under analysis (SOC [g kg^-1], LPC [g kg^-1], MEN [mg kg^-1], ASI [unitless]) at time t; V_e is the value of the variable at equilibrium; V₀ is the initial value of the variable; r is the exponential rate of variation (yr^-1); and t is time under cropping or under pasture (yr). To fit the model depicted by Eq. [1] to the dependent variables under cropping, data from crop-pasture rotations and from continuous cropping rotations were pooled. The half-life (tt_1/2 ) of each variable was calculated as t_1/2 = 0.693r^-1. The values of tm represent the estimated elapsed time between the beginning of a cropping or pasture period and the decrease or increase to half of the total estimated variation. Soil organic C data from Rotations 1,4, and 7 (short-term pastures before cropping, Fig. 1) were fitted to Eq. [1] separately from data from Rotations 2 and 3 (long-term pastures before cropping, Fig. 1) as a function of years under cropping.

For the analyses of variance, years were included as an independent variable, so that they were carried on as if treatment arrangement had been in sub-subplots (Little and Hills, 1978). Statistical analysis and curve fitting were performed with the OLM and NLIN (DUD method) procedures of the Statistical Analysis System (SAS Institute, 1985), respectively.


RESULTS AND DISCUSSION

Year X rotation interaction was highly significant (P < 0.01; Table 1) for all the variables analyzed. This indicates that the stated rotations produced differential changes with time of the analyzed dependent variables.

During periods under pasture, the levels of SOC, LPC, MEN, and ASI increased, whereas during periods under cropping, they decreased. As an example of the effect of cropping and pasture periods, changes in SOC with time for Rotations 1, 2, and 5 (Fig. 1) are shown in Fig. 2. Continuous cropping (Rotation 5) produced a continuous decrease in SOC. On the other hand, periods under pasture after periods under cropping (Rotations 1 and 2) reversed the fall in SOC produced by conventional cropping. Similar results had already been reported by other researchers (Greenland, 1981; Tisdall and Oades, 1982; Diaz Rosello, 1992; Angers et al., 1992; Haynes et al., 1991).

Nitrogen fertilization did not significantly affect (P > 0.05) SOC (Table 1), neither by interacting with rotation and/or year nor by itself. These results do not agree with those reported by other researchers (Doran and Smith, 1987; Havlin et al., 1990; Campbell and Zentner, 1993; Robinson et al., 1996) with regard to either increases in SOC or decreases in its rate of decline, due to the greater amounts of C returned to the soil associated with N fertilization. For model-fitting purposes, SOC of fertilized and nonfertilized subplots were averaged.


Soil Organic Carbon

It can be seen in Table 2 that the values of V0 and r for the case of cropping after a longer term pasture were higher than when cropping after a short-term pasture (Fig. 3^a). Including short-term pastures in the rotation meant shorter periods of recovery of SOC and, consequently, SOC at the beginning of the cropping period was lower (Fig. 2). This confirms that it was adequate to fit SOC observed data according to the duration of the previous period under pasture. However, V_e was high (29.3 g kg^-1, Table 2) and the same for both cases of cropping after pasture. Likewise, total estimated variations of SOC (V₀ — V_e, = 7.9 and 5.2 g kg^-1 for previous long- and short-term pasture, respectively, Table 2) were small. These values indicate that the soil under study contains a highly stable organic fraction that shows small variation even under continuous conventional tillage for long periods. This is due to the great proportion of highly protected mineralassociated humic substances that forms the SOM of these soils. These results contrast with those obtained with other soils in which equilibrium values were lower, and SOM content decreases due to conventional cropping were more drastic (Lamb et al, 1985; Dalai and Mayer, 1986a).

Fig. 2. Soil organic C between 1981 and 1993 for continuous cropping, 75% cropping-25% pasture (period under pasture 1984-1986), and 50% cropping-50% pasture (period under pasture 1984-1989).

Table 2. Results of fitting the Bartholomew and Kirkham (1960) model to soil organic carbon (SOC), light-fraction carbon (LFC), microbial biomass nitrogen (MUN), and aggregate stability index (ASI) as a function of time under cropping or pasture.

*Units apply only to the values of Ve, and Vo *Ve = value of the variable at equilibrium; Vo = initial value of the variable; r = exponential rate of variation. § t1/2 = half-life. ¶ SOC,, = soil organic C under agriculture following a 4- to 5-yr pasture. # SOC, = soil organic C under agriculture following a 2-yr pasture.

Fig. 3. Soil organic C as a function of years under cropping and years under pasture. Curves result from fitting observed data to a model proposed by Bartholomew and Kirkham (1960). SOCL = estimated soil organic C after a long-term pasture; SOCS = estimated soil organic C after a short-term pasture.

Figure 3b shows observed and fitted values of SOC during periods under pasture. The fast recovery of SOC after pasture establishment was defined by a high r (Table 2). This agrees with expected results since morphological characteristics of the species in the pasture mix (especially grasses) make them return great amounts of aboveground and root biomass residues with a very uniform and complete distribution in the soil (Oades, 1984). It can also be seen in Table 2 that the V₀ corresponding to the period under pasture and the V_e corresponding to the period under cropping are very close. This suggests that the model proposed by Bartholomew and Kirkham (1960) (Eq. [1]) was adequate to describe the variation in SOC under two contrasting managements.


Light-Fraction Carbon and Nitrogen in the Microbial Biomass

The trend of change of LFC and MBN as a function of years under cropping and pasture were very similar.

Therefore, only the variation of LFC with time is shown (Fig. 4). An abrupt fall in LFC in the first years under cropping can be seen (Fig. 4a). Light-fraction C became asymptotic to V_e after 6 to 7 yr under conventional cropping. In an analogous manner, there was a rapid recovery of this organic fraction after pasture establishment (Fig. 4b), reaching values close to Ve in 3 to 5 yr under pasture. Values of r for both variables were higher than those reported for SOC (Table 2). Consequently, t\n values (Table 2) were much lower and similar between variables when comparing the period under pasture with the period under cropping.

These variables are much more sensitive than SOC to system management changes and are early indicators of the variations in organic fraction contents provoked by such changes (Richter et al., 1975; Dalai and Mayer, 1986b,c; Angers et al., 1992; Wander et al., 1994; Bremer et al., 1994). Microbial biomass N represents only 2 to 10% of soil N (Yong et al., 1990), but its turnover is approximately fivefold faster than that of other soil organic N fractions (Marumoto et al., 1982) and it accounts for a large portion of available N in soil (Power, 1994).

Fig. 4. Light-fraction C as a function of years under cropping and years under pasture. Curves result from fitting observed data to a model proposed by Bartholomew and Kirkham (1960).

Something similar occurs with other labile organic fractions (Doran and Smith, 1987) such as LPC (Cambardella and Elliott, 1992, 1994; Janzen, 1987; Janzen et al., 1992; Dalai ancHVlayer, 1986b; Bremer et al., 1994). The fall in MEN and LPC during the period under cropping means decreasing N supply to crops and a consequent increase in N fertilizer need (Doran and Smith, 1987; Power, 1994).


Aggregate Stability Index

The behavior of this variable as a function of years under cropping and pasture was almost the same as that reported for LPC and MEN, and the corresponding values of r and tm were quite similar (Table 2). This indicates a close relationship between the labile organic fraction content in soil and its aggregate stability (Cambardella and Elliott, 1993, 1994; Golchin et al., 1994,
1995).

Figure 5 shows the relationship between SOC and ASI. It can be seen that ASI values under cropping were relatively constant across all the explored range of SOC (29.6-37.7 g kg^-1)- On the other hand, ASI increased with increases in SOC (r = 0.64, P < 0.01) when periods under pasture were analyzed (Fig. 5). These results agree with the fact that aggregate stability and SOC are closely related (Greenland, 1981; Oades, 1984). However, it appears that SOC was not enough by itself to explain ASI variations because ASI values were different at the same SOC under cropping or under pasture, respectively.

Fig. 5. Relationship between aggregate stability index and soil organic C.

Labile SOM fractions content and dynamics are more closely related to aggregate stability than SOC (Cambardella and Elliott, 1993; Golchin et al., 1994). Only for a few situations were the values of LPC and ASI available for the same treatment factor combination and sampling time. However, there exists some evidence of a relationship between both variables during the period under pasture. Across part of the explored range of LPC under cropping (0.88-1.15 g kg^-1), ASI values ranged between 35.0 and 21.0, and averaged 30.0 ± 6.2, which is similar to Ve and V0 of the situation under cropping and pasture, respectively (Table 2). On the other hand, an increase of 57.0 in ASI (43.0 to 100.0) under pasture corresponded to an increase of 0.47 g kg^-1 in LPC (1.26 to 1.73 g kg^-1). These results seem to confirm that the relationship between the organic fraction of soil and aggregate stability was different according to whether the system was under pasture or conventional cropping.

Frequent soil tillage and seasonal growth of crops under cropping contrasts with the continuous pasture growth (aboveground and root biomass) without tillage for relatively long periods. The latter favors an abundant and uniform addition of SOC- and LFC-forming materials (Oades, 1984; Haynes and Francis, 1993) that, through the action of an increased microbial biomass, contribute to form larger and more stable aggregates (Haynes and Francis, 1993). Likewise, root system growth of pasture, especially if the pasture is dominated by grasses, facilitates soil particle aggregation and strengthens the resulting aggregates (Oades, 1984).

ACKNOWLEDGMENTS

This study was made possible thanks to the long-term croppasture rotation experiment at Balcarce, Argentina, conducted since 1976 in the Unidad Integrada Balcarce (UIB) with financial support of the Institute Nacional de Tecnologia Agropecuaria. We want to express our gratitude to the scientists who have been in charge of the experiment during different periods, to the field personnel who have helped with most of the field work, and to the personnel of the Soil, Plant, and Water Analysis Laboratory of the UIB, who ran a great part of all soil analyses. We also want to express our appreciation to Dr. John W. Doran for his valuable suggestions to improve the abstract of this paper.

REFERENCES

AUTHORS:

Guillermo A. Studdert, Hernan E. Echeverría, and Elda M. Casanovas
Facultad de Ciencias Agrarias (U.N.M.P.)- Estacion Experimental Agropecuaria Balcarce (I.N.T.A.), Unidad Integrada Balcarce, Buenos Aires, Argentina.

Readers' Rating:

Rate this article

Send to a friend