Distribution of water-stable aggregates and aggregating agents in Cerrado Oxisols

Ž .Geoderma 93 1999 85–99

Distribution of water-stable aggregates andaggregating agents in Cerrado Oxisols

Henry Neufeldt a,), Miguel A. Ayarza b, Dimas V.S. Resck c,Wolfgang Zech a

a Institute for Soil Science and Soil Geography, UniÕersity of Bayreuth, 95440 Bayreuth, Germanyb ( )Centro Internacional de Agricultura Tropical CIAT , A.A. 6713, Cali, Colombia

c ( )Centro de Pesquisa Agropecuaria dos Cerrados EMBRAPA-CPAC , C.P. 08223, 73301-970´Planaltina, DF, Brazil

Received 19 March 1998; received in revised form 25 February 1999; accepted 11 May 1999

Abstract

The effects of structural changes on Cerrado Oxisols in Brazil after land-use change are stillinsufficiently studied. Therefore, topsoil samples of loamy and clayey Cerrado Oxisols under crop,pasture and reforestation were fractionated using a wet-sieving procedure to obtain the distributionof water-stable aggregates and compared to samples from natural savanna. The results werecorrelated to organic and inorganic soil compounds to identify the main aggregating agents anddiscussed in relation to changes in the pore-size distribution of the soils.

Clayey soils showed a significantly higher macro-aggregation than loamy soils. Compared tonatural savanna, macro-aggregation was clearly reduced at the crop sites, whereas aggregation ofsoils under pasture and reforestation was only slightly affected.

In both the clayey and the loamy soils, bonding of polysaccharides was the main aggregatingagent. In the clayey soils, liming was also very important for disaggregation by weakening theelectrostatic forces between positively and negatively charged soil compounds, whereas in theloamy soils, binding of macro-aggregates by roots was significant. The introduction of crop–pas-ture rotations was therefore proposed to take advantage of the strong rooting and polysaccharideproduction of pastures.

Management-induced disaggregation strongly affected the pore-size distribution by compactingthe soils. Thereby, macro-porosity was reduced and the amount of meso-pores was increased,while micro-porosity was unaffected from management and only differed between the twosubstrates. Considering the low pore space at plant-available matrix potentials typical of Cerrado

) Corresponding author. Fax: q49-921-55-2246; E-mail: [email protected]

0016-7061r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.Ž .PII: S0016-7061 99 00046-4

( )H. Neufeldt et al.rGeoderma 93 1999 85–9986

Oxisols, the increase might be important for annual crops during the frequent dry spells in therainy season. q 1999 Elsevier Science B.V. All rights reserved.

Keywords: aggregate fractionation; aggregating agents; water retention; land-use change; oxisols;Brazilian Cerrado

1. Introduction

Increasing mechanisation and expansion of the cultivated area during the pastthree decades have strongly increased the agricultural production in CentralBrazil’s Cerrado region and productivity is believed to increase considerably in

Ž .the near future Goedert, 1989 . The sustainability of management on theŽ .Cerrado soils, most of which are Oxisols Goedert, 1983 , is therefore of high

ecological and socioeconomical significance.Oxisols are known for their stable microstructure which is caused by the

electrostatic forces between positively charged oxyhydroxides and negativelyŽ .charged kaolinite and organic matter El-Swaify, 1981 . The structural stability

leads to a unique water retention behaviour with an extremely high drainage anda rather low amount of ecologically important meso-pores, which is why Oxisolsare less prone to erosion than many other tropical soils, while plants may

Ž .quickly suffer from draught El-Swaify, 1981; Bartoli et al., 1992 .Changes in the structure of Brazilian Oxisols after land-use change have been

Ž . Ž .described by Roth and Pavan 1991 , Roth et al. 1991 , Castro Filho and LoganŽ . Ž . Ž .1991 , de Campos et al. 1995 and Westerhof et al. 1999 . Their resultsindicated that conventional tillage practices physically broke macro-aggregatesinto smaller units, leading to new surfaces and subsequently the loss of labileorganic matter due to a mineralisation flush. Liming had effects on aggregationbecause the pH rise altered the electrostatic forces between soil organic matterŽ .SOM , polyvalent cations and pedogenic oxyhydroxides. The establishment ofno-tillage systems or pastures frequently reduced the degradation or led to a

Ž .structural improvement Roth et al., 1991; Westerhof et al., 1999 .Chemical, physical and biological effects on soil structure are closely related

to each other and depend on management practices, SOM contents and dose oflime application as recently shown for clayey Cerrado Oxisols by Westerhof et

Ž .al. 1999 , but it is unclear how these factors interact and which contribute themost. Likewise, it is known that liming has an effect on bulk density and

Žinfiltration of Brazilian Oxisols Castro Filho and Logan, 1991; Roth and Pavan,.1991 , but how changes in soil structure act on the pore-size distribution and

thus influence drainage or the plant available water content has rarely beenstudied.

We attempted to address these questions for differently managed CerradoŽ .Oxisols of contrasting texture and land-use practices by i determining the

Ž .aggregate distributions, ii identifying the major aggregating agents by simple

( )H. Neufeldt et al.rGeoderma 93 1999 85–99 87

and multiple regression of a series of possibly aggregating organic and inorganicŽ .compounds and iii comparing the results with the water retention character-

istics of the soils.

2. Material and methods

2.1. Study area and site history

The study area is located at 48.18W 19.18S, near Uberlandia, Minas GeraisˆState, Brazil, at 900–950 m altitude on the Tertiary tableland. Mean annualtemperature is 228C and average precipitation is 1650 mm, 90% of which occursbetween October and April. A coarse-loamy, mixed, isohyperthermic Typic

ŽHaplustox medium textured, dystrophic, red–yellow Latosol according to the.Brazilian classification system and a very fine, allitic, isohyperthermic Anionic

Ž .Acrustox very clayey, allitic, red–yellow Latosol were selected for sampling.On both soil types, a conventional maizersoybean crop rotation, a degraded

Žforage grass pasture and a forest plantation site pine on the clayey and.eucalyptus on the loamy substrate were selected. The sites were chosen because

of their comparatively long management histories and the short distance betweenthem. A native Cerrado savanna nearby was selected for control. An overviewof the management histories is presented in Table 1.

Table 1Management history of the studied sites

Ž .Treatment Beginning Fertilisation application year Yieldy1 y1 y1Ž . Ž .ha ha yr

Very fine Anionic AcrustoxŽ .Conventional 1985 soybean: 7 N, 70 P, 70 K kg ; 2400 kg soybean;Ž .maizersoybean maize: 60 N, 80 P, 60 K kg ; 2700 kg maize

Ž . Ž .crop rotation dolomitic lime annually intercropBrachiaria decumbens 1986 300 kg partially acidulated rock 0.5 cattle

Ž .pasture phosphate 34 kg P andŽ .2000 kg dolomitic lime 1986

3 ŽPinus Caribaea 1975 20 g monocalcium 11.7 m pulpŽ . .reforestation superphosphaterseedling 1975 production

Coarse-loamy Typic HaplustoxŽ .Conventional 1986 soybean: 8 N, 80 P, 80 K kg ; 2400 kg soybean;Ž .maizersoybean crop maize: 110 N, 80 P, 80 K kg ; 6300 kg maize

Ž .rotation dolomitic lime annuallyŽ .Brachiaria decumbens 1987 rice 1986 : 8 N, 28 P, 16 K, 0.5 cattle

pasture 1500 kg lime, 700 kg CaSO4Ž .Eucalyptus citriodora 1982 Ca-apatite 1982 unknown

reforestation


2.2. Soil sampling and pretreatment

Soil sampling was done in March 1995, at the end of the rainy season. Oneach treatment, three plots of 50=50 m2 were selected which lay about 50 m

Ž .apart from each other. Five undisturbed topsoil samples 0–12 cm were takenper plot with an Uhland corer and kept refrigerated at 58C after carefullybreaking large peds along fissures to pass an 8-mm screen. For chemicalanalyses, an aliquot of the five samples per plot was dried at 408C, sieved -2mm and ground for complete homogenisation after removing visible roots. Forthe determination of the water retention curve, three undisturbed soil cores perplot were taken from the 0–10 cm and the 10–20 cm increments with 100 cm3

cylinders.

2.3. Aggregate fractionation

Ž .Field-moist samples 40–50% gravimetric water content were used forŽ .aggregate fractionation according to Beare and Bruce 1993 . The samples were

capillary wetted on the uppermost sieve until field capacity and then fractionatedinto aggregates of 8–2 mm, 2–1 mm, 1–0.5 mm, 0.5–0.25 mm and 0.25–0.05

Ž .mm diameter for 30 min following EMBRAPA 1979 . Soil passing the -0.05mm sieve was collected after allowing to settle for 12 h and aspirating thesupernatant. All fractions were dried at 408C and weighed. Fractionation recov-ery lay between 96% and 100%. For analytical analyses of the aggregatefractions, all replicates per plot were combined, visible roots discarded and thesoil ground.

( )2.4. Determination of particulate organic carbon POC

An aliquot of each aggregate fraction was sieved after dispersion withhexametaphosphate to retain particulate organic matter together with the )0.05

Ž .mm sand Elliott et al. 1991 . Because the low carbon content of the sandfractions did not permit a direct measurement of POC, soil passing the sieve wasdried, ground and used for its indirect determination. POC was calculated as the

Ž .difference between soil organic carbon SOC in the whole soil and in thesand-free fraction.

2.5. Analytical methods

Ž .SOC and total N were determined by dry combustion Elementar Vario EL .The pH was determined in water at a soil:solution ratio of 1:10. Exchangeable

Ž .cations were extracted according to EMBRAPA 1979 . Polysaccharides wereŽ .extracted with a two-step acid hydrolysis Amelung et al., 1997 . Noncellulosic

Ž .polysaccharides NCP were hydrolysed with 1 M HCl at 1008C for 5 h and


Ž .cellulosic polysaccharides CP were digested by treating the residue with 12 MH SO . The polysaccharides were determined colorimetrically with the MBTH2 4

Ž .reagent as described by Johnson and Sieburth 1977 . Glucose equivalents weredivided by 2.5 to obtain the polysaccharide C content. Pedogenic Fe and Al

Ž . Ž .compounds Fe , Al were extracted with dithionite–citrate–bicarbonate DCBd dŽ .solution Mehra and Jackson, 1960 . The extraction of X-ray amorphous Fe and

Ž .Al oxyhydroxides Fe , Al was carried out with acid oxalate reagent followingo oŽ .Blume and Schwertmann 1969 . Cations were determined by atomic absorption

Ž .spectroscopy Shimadzu AA-660 . All analyses were executed in duplicate. Thegrain-size distribution was determined with the sieve–pipette method afterdispersion in 0.1 M NaOH, and the pore-size distribution was obtained gravi-metrically by centrifugation of the saturated soil cores at consecutively higherspeeds and final drying at 1058C for the determination of bulk density and pore

Ž .volume EMBRAPA, 1979 .

2.6. Statistical analysis

Ž .Statistical analyses were executed with Statistica software Statsoft . SoilŽ .effects on the distribution of water-stable aggregates WSA were verified with

MANOVA using Tukey’s HSD test at p-0.05, whereas management effectsŽ .were explained by the mean"95% confidence interval ns3 only. Single and

multivariate regressions were calculated to determine functional relationshipsbetween aggregation and independent variables. The Durbin–Watson test forserial correlation was applied to the regressions and d compared with bench-

Ž .mark values calculated by Savin and White 1977 prior to accepting anyequation.

3. Results and discussion

3.1. Physical and chemical characterisation of the soils

Under natural conditions, the chemical characteristics of the soils were typicalŽ .of Cerrado Oxisols Goedert, 1983 as indicated by very low nutrient and P

Ž .contents and a low pH leading to an Al-dominated exchange complex Table 2 .Differences in nutrient contents between the substrates were rather small. Due tothe limiting conditions, the plants possibly lowered the nutrient concentrations tothe minimum threshold value of the soils. Liming and fertilisation raised plantavailable P and the pH, exchanging Al for Ca and Mg, and led to an increase ofCEC . The greater CEC rise in the clayey soils is probably related to the highere e

clay content because fertilisation application was comparable.The considerably lower bulk density and higher pore volume of the clayey

soils was mainly provided by fine pores due to their higher intra-aggregate pore


Table 2Ž .Selected physical and chemical surface soil 0–12 cm properties of differently managed Cerrado

Oxisols

Ž .Treatment Texture % pH P Exchange complexH O Mehlich2y1Ž .mg kg Ž .Clay Silt Sand Ca Mg K Al CEC BS %e

y1Ž .cmol kgc

Very fine Anionic AcrustoxCerrado 67 7 26 4.7 3 0.05 0.05 0.22 0.93 1.25 25Crop 66 12 22 5.6 19 2.33 0.34 0.34 0.11 3.12 96Pasture 66 13 21 5.6 11 1.69 0.36 0.25 0.13 2.43 95Pinus 66 7 27 4.5 3 0.01 0.02 0.11 0.81 0.95 16

Coarse-loamy Typic HaplustoxCerrado 16 0 84 5.0 3 0.06 0.07 0.19 0.44 0.76 42Crop 16 0 84 6.0 9 1.13 0.21 0.18 0.08 1.60 95Pasture 17 0 83 5.4 2 0.56 0.05 0.08 0.22 0.91 76Eucalyptus 15 0 85 4.8 4 0.05 0.04 0.14 0.70 0.93 25

Ž .space Bui et al., 1989 , whereas macro-and meso-pores were not significantlyŽ .different between the substrates Table 3 . Management effects resulted in an

increase of bulk density and thus a reduction of pore volume due to compactionby heavy machinery, liming or both, and were most accentuated at the croppedsites followed by the reforestations and the pastures. The compaction led to areduction of macro-pores, but also increased the amount of meso-pores undercrop, whereas micro-pores were entirely unaffected.

SOC, NCP and CP were two to four times higher in the clayey than in theloamy substrate, whereas POC showed no significant differences between the

Ž .soil types Table 4 . POC consisted of fresh to strongly decomposed plantfragments and fine roots which might have an entangling effect on aggregatesŽ .Dorioz et al., 1993 . Since POC contents were comparable between the twosoils, they represented a much higher proportion of SOC in the loamy than inthe clayey soils. NCP are mainly of microbial origin, while CP are mostly

Ž .plant-derived Guggenberger et al., 1994 and were determined because of theirŽ .aggregate bonding properties Dorioz et al., 1993 . NCP corresponded to

81–86% of total polysaccharides and both NCP and CP were highly correlatedŽ UUU.to SOC r)0.92 . Total polysaccharide contents ranged from 15% to 22%

of SOC, a figure which is comparable to topsoil polysaccharide contents of theŽ .Great Plains Amelung et al., 1997 , indicating that there are no pronounced

differences in the amount of polysaccharides between tropical Oxisols andtemperate Mollisols. Management effects on SOC, NCP and CP were restrictedto the pine reforestation on the clayey and the crop treatment on the loamysubstrate, while POC was lowest in the two crop treatments.

Pedogenic oxyhydroxides were at least twice as high in the clayey substrateŽ .Table 4 , reflecting the close relationship between organic compounds and clay

()

H.N

eufeldtetal.r

Geoderm

a93

199985

–99

91

Table 3Ž . Ž .Distribution "95% confidence intervals of macro- )0.25 mm and micro-aggregates -0.25 mm , bulk density and pore volume at specific matrix

potentials in differently managed clayey and loamy Cerrado OxisolsaTreatment Macro-aggregates Micro-aggregates Bulk density Pore volume Pore volume at hPa %

y3Ž .Mg m %Ž .% -63 63–15 000 )15 000

Very fine Anionic AcrustoxCerrado 91.2"4.8 8.6"4.9 0.84"0.02 76"2 39"1 11"1 26"1Crop 82.3"0.9 17.6"0.9 0.93"0.04 68"2 23"3 17"3 28"0Pasture 89.4"0.7 10.5"0.6 0.89"0.06 73"4 30"2 14"3 29"0Pinus 87.6"1.2 12.0"1.3 0.92"0.04 68"3 28"2 13"2 27"1

aAverage 87.6 b 12.2 a 0.90 a 71 b 30 a 14 a 28 b

Coarse-loamy Typic HaplustoxCerrado 85.2"0.6 14.6"0.8 1.15"0.05 58"3 35"3 11"2 12"0Crop 70.3"0.3 29.6"0.3 1.38"0.06 48"2 19"2 17"2 12"1Pasture 79.8"1.7 20.0"1.7 1.23"0.03 54"2 26"2 16"2 12"0Eucalyptus 77.8"0.5 22.0"0.5 1.27"0.05 52"2 27"1 14"3 11"1Average 78.3 a 21.6 b 1.26 b 53 a 27 a 15 a 12 a

aBetween the two soil types, values followed by the same lower case letter are not significantly different at the p-0.05 level of Tukey’s HSD test.

()

H.N

eufeldtetal.r

Geoderm

a93

199985

–99

92

Table 4SOC, organic compounds and pedogenic oxides "95% confidence intervals in differently managed clayey and loamy Cerrado Oxisols. The values in

Ž .brackets represent the proportion % of organic compounds to SOC and of oxalate to DCB extractable oxyhydroxidesy1Ž .Treatment SOC g kg CrN POC NCP CP Fe Fe Al Ald o d o

y1Ž .g kg

Very fine Anionic AcrustoxŽ . Ž . Ž . Ž . Ž .Cerrado 23.5"0.3 17.8"0.8 2.3 9.8 3.8 16.2 0.76 3.2 46.6"3.6 2.1 4.5 10.1"3.6 1.8 17.8Ž . Ž . Ž . Ž . Ž .Crop 22.9"0.4 16.6"0.2 1.4 6.1 3.6 15.7 0.60 2.6 50.8"1.3 1.6 3.1 17.0"1.8 2.0 11.8Ž . Ž . Ž . Ž . Ž .Pasture 24.6"0.9 17.8"0.2 3.1 12.6 4.5 18.3 0.84 3.4 53.5"2.2 1.9 3.5 16.2"3.5 1.9 11.7Ž . Ž . Ž . Ž . Ž .Pinus 21.5"1.2 20.0"0.6 2.9 13.5 2.5 11.6 0.53 2.5 48.0"2.2 1.6 3.3 12.1"2.9 1.9 15.7

Coarse-loamy Typic HaplustoxŽ . Ž . Ž . Ž . Ž .Cerrado 9.8"0.4 15.9"0.6 3.8 38.8 1.6 16.3 0.38 3.9 16.0"0.5 0.8 5.0 3.9"0.6 0.5 12.8Ž . Ž . Ž . Ž . Ž .Crop 7.1"0.3 16.8"0.3 1.5 21.1 1.1 15.5 0.22 3.1 15.6"1.0 0.7 4.5 4.8"0.7 0.5 10.4Ž . Ž . Ž . Ž . Ž .Pasture 9.3"0.7 17.5"0.8 2.6 28.0 1.5 16.1 0.32 3.4 19.3"0.7 0.8 4.1 3.9"0.3 0.4 10.3Ž . Ž . Ž . Ž . Ž .Eucalyptus 10.2"0.2 18.9"0.1 2.5 24.5 1.6 15.7 0.35 3.4 17.2"0.6 0.8 4.7 6.0"0.5 0.5 8.3


Ž .content Feller and Beare, 1997 . Over 95% of the Fe oxyhydroxides were incrystalline form, suggesting an advanced weathering degree. The proportion ofAl was probably higher because DCB only extracts a small fraction of totalo

Ž .pedogenic Al Wada, 1989 . Differences in oxyhydroxide contents between thetreatments presented no clear trends, possibly reflecting natural variation ratherthan management effects.

3.2. Distribution of water-stable aggregates

Ž .In general, macro-aggregates )0.25 mm were slightly, but significantlylower in the loamy than in the clayey soils due to the higher amounts of soil

Ž .compounds per unit volume in the clayey substrate Table 3 . A more detailedŽview of the aggregate distribution for which the 8–2 mm and the 2–1 mm as

well as the 1–0.5 mm and the 0.5–0.25 mm classes were combined because of.similar distribution patterns revealed that the higher amount of macro-aggre-

Ž .gates in the clayey soils mainly resulted from large macro-aggregates 8–1 mm ,Ž .while small macro-aggregates 1–0.25 mm showed comparable proportions in

Ž .both soils Fig. 1 . The 0.25–0.05 mm fraction was always clearly higher,whereas the -0.05 mm fraction was generally lower in the loamy soils than inthe clayey substrate.

Management effects indicated that in comparison to natural Cerrado, allŽ .treatments tended to loose macro-aggregates Table 3 . However, the reduction

was only expressive at the cropped sites. Concomitantly, a land-use-specific

Fig. 1. Management effects on the distribution of water-stable aggregates "95% confidenceŽ . Ž .interval in clayey left and loamy right Cerrado Oxisols. The same lowercase letter above an

aggregate class indicates that differences between the soil types are not significant with Tukey’sHSD test at p-0.05.


Ž .aggregate distribution formed with time Fig. 1 . Small macro-aggregates in-creased at the expense of large macro-aggregates under pine reforestation and atthe cropped sites, regular tillage resulted in a strong enrichment of micro-ag-gregates.

The loss of macro-aggregation at the cropped sites can be attributed to tillage,Žliming and the subsequent loss of organic matter Castro Filho and Logan, 1991;.de Campos et al., 1995; Westerhof et al., 1998 . Differently, the macro-aggre-

gate distribution under pine is probably related to a strongly reduced rootingintensity in the topsoil which could otherwise hold large macro-aggregatestogether. Under pine, the fine-root maximum shifted from the first centimetersof the mineral soil into the thick moder horizon, a typical feature of pinereforestations in the Cerrado region, probably because nutrient availability washigher in the organic layer than in the mineral soil and because the organic

Ž .matter may complex toxic Al Puhe, 1994 .

3.3. Binding agents of aggregation

The influence of several soil compounds on aggregation was studied withregression analyses. The results indicated significant correlations in the loamysubstrate with all organic compounds, Ca, Mg, Al and pH, whereas in the clayey

Ž .substrate, only POC, Fe and Ca were significantly correlated Table 5 . In botho

soil types together, all correlations except for Ca, Mg and K were significant.However, SOC, Fe , Al and Al were only well correlated because of thed d o

Table 5Ž . aPearson correlation coefficients r and significance levels of bulk soil constituents vs. fraction

of aggregates )0.25 mm in clayey and loamy Cerrado Oxisols

Predictors Correlation coefficients

Clayey Acrustox Loamy Haplustox Both OxisolsŽ . Ž . Ž .ns12 ns12 ns24

UU UUUSOC 0.291 0.768 0.779UUU UUUNCPqCP 0.327 0.887 0.768UUU UUUNCP 0.275 0.884 0.755UUU UUUCP 0.553 0.868 0.830

U UUU UPOC 0.615 0.827 0.448UAl y0.571 0.419 0.500dUUUAl y0.424 0.117 0.669oUUUFe y0.280 0.226 0.706d

U UUUFe 0.684 0.380 0.790oUU UUpH y0.520 y0.786 y0.605

U UUCa y0.666 y0.758 y0.021Mg y0.520 y0.096 0.180

UAl 0.655 0.477 0.462

aUp-0.05, UUp-0.01, UUUp-0.001.


strongly contrasting contents between the two substrates. Due to these discrep-ancies, it seemed necessary to calculate multiple regressions for each substrateand for both soil types together to evaluate the most important aggregating

Ž .agents. The best fits of macro-aggregation MA vs. independent variableswhich presented an acceptable serial correlation according to Savin and WhiteŽ .1977 are presented below:

MA s11.33 NCPqCP q2.43 POCq52.08 R2 s0.88Ž .loam

MA s3.07 NCPqCP y0.02 Caq77.77 R2 s0.93Ž .clay

MA s3.30 NCPqCP y5.77 pHq2.14 POCq97.60 R2 s0.93Ž .both

These equations indicate that polysaccharides played an important role forŽ .aggregation in both soils. According to Dorioz et al. 1993 , extracellular

polysaccharides are especially effective in gluing soil particles together at theŽ .micro-aggregate 5–200 mm scale, although packing effects may also influence

macro-aggregation up to 1000 mm. However, since in the studied Oxisols moreŽ .than 70% of the polysaccharides are bound to the clay fraction Neufeldt, 1998 ,

they probably mainly act on micro-aggregation and are at least temporarilyprotected from decomposition. A minor contribution to aggregation might also

Žarise from ion exchange and by complexation of polysaccharides Cheshire,.1979 .

POC was important for aggregation in the loamy substrate only, possiblybecause of its higher proportion when compared to the clayey soils. However,POC corresponded to only 10% of R2, indicating that its contribution was rather

Ž .small. Similar conclusions were drawn by Gijsman 1996 for ColombianOxisols. Differently, for the clayey substrate, Ca corresponded to nearly half of

Ž .the variation. Since Ca was strongly correlated to the pH rs0.89 and AlŽ .rsy0.90 after log-transformation and correction for valence, respectively,the pH rise and subsequent cation exchange after liming seems to dominatedisaggregation in the clayey substrate next to bonding by polysaccharides. Forboth soils together, the best fit included polysaccharides, pH and POC which

2 Žcontributed with 59%, 29% and merely 6% of the R change, respectively Fig..2 .

These results could have strong implications for management practices asthey suggest that aggregation in the loamy soils depends to a greater extent onthe enmeshing of aggregates by roots so that no-tillage systems or crop–pasturerotations might be able to invert tillage-induced disaggregation as suggested by

Ž . Ž .de Campos et al. 1995 and Gijsman 1996 . Chemical disaggregation in theclayey substrate is probably inevitable because liming is essential to successfulagriculture under the low pH and nutrient conditions of natural Cerrado Oxisols.

Ž . Ž .However, according to Roth et al. 1991 and Castro Filho and Logan 1991 ,liming may also stabilise Oxisol aggregates if application doses are sufficiently

Ž .high, and Westerhof et al. 1998 found that liming only had disaggregating


Fig. 2. Predicted vs. observed fraction of aggregates )0.25 mm"95% confidence intervals ofŽ . Ž .clayey v and loamy ` Cerrado Oxisols.

effects when SOM contents were low. This could explain the small influence ofliming on the pastures, especially due to their high polysaccharide contents. Thestructural improvement of Oxisols after the introduction of pastures to conven-tional crops might therefore be principally related to the increase of polysaccha-rides, and supports crop–pasture rotations to reduce soil slaking.

3.4. Effects of disaggregation on pore-size distribution

Ž . Ž .Roth and Pavan 1991 and Castro Filho and Logan 1991 explained thereduction of pore volume after land-use change in Brazilian Oxisols with liming,but physical compaction by heavy machinery or cattle trampling could alsocontribute in the treatments under study. Since macro-aggregation was highly

Ž UUU.correlated to the pore volume of the soils rs0.89 , it appears thatmanagement-induced disaggregation led to soil compaction.

ŽThe loss of pore volume mainly occurred among the macro-pores B)50.mm as indicated by the close correlation between macro-aggregates and

Ž UUU.macro-porosity rs0.80 . This corresponds to the results of Roth et al.Ž .1991 who detected a reduction of macro-porosity after the introduction ofconventional and no-tillage practices on an Oxisol from South Brazil. Regres-sion analyses of the aggregate classes showed that the increase of the 0.25–0.05mm fraction was mainly responsible for the loss of macro-porosity, probably by

Ždetachment and subsequent transport of aggregates during heavy rains Suther-. Ž .land et al., 1996 . However, Sutherland et al. 1996 found that mainly the


0.25–2 mm aggregates were responsible for soil sealing in an Oxisol fromHawaii.

Ž .The amount of meso-pores B 50–0.2 mm was significantly correlated withŽ UU.the -0.05 mm fraction rs0.61 , suggesting that after land-use change, the

diameter of some macro-pores was reduced by small micro-aggregates. Roth etŽ .al. 1991 found a comparable increase in the amount of meso-pores after

Ž .land-use change and an experiment by Curmi et al. 1994 confirmed thatmechanical compaction of an Oxisol derived from basalt reduced inter-aggregatepores of 1–100 mm diameter. Considering the low pore space at plant-available

Ž .matrix potentials in Oxisols Bartoli et al., 1992 , the increase of meso-porescould prevent annual crops from suffering draught during dry spells in the rainy

Ž .season veranicos which have a probability of over 30% in the study regionŽ . Ž .Assad et al., 1994 . However, Moraes et al. 1995 determined that thedevelopment of soybean roots in an Oxisol was strongly reduced due tomechanical compaction.

Ž .Micro-pores B-0.2 mm were not functionally correlated with any aggre-gate class because they were not affected by management practices. Likewise,

Ž .Curmi et al. 1994 found that compaction had no effects on intra-aggregatepores of -1 mm diameter because their amount is only determined by the

Ž .mineralogy of the soils Bui et al., 1989 .These results indicate strong effects of disaggregation on water retention after

land-use change. However, whether the compaction has net positive effects dueto the increase of plant-available water or whether the loss of pore space must beconsidered negative for cropping systems needs to be studied in greater detail.

4. Conclusions

Cerrado Oxisols of both clayey and loamy texture are characterised by a highmacro-aggregate stability under natural conditions. Land-use change leads to theenrichment of micro-aggregates under cropping systems, whereas pastures andeucalyptus are not clearly affected. Under pine, there is a management-specificincrease of small macro-aggregates at expense of large macro-aggregates due tothe shift of the fine root maximum from the mineral surface soil into the thickmoder horizon.

Aggregation in the clayey soils is preferentially caused by bonding of mineralparticles with polysaccharides and electrostatic forces between polyvalent metalcations and negatively charged organic matter and kaolinite. The reduction ofpolysaccharides and liming are therefore the most important disaggregatingagents. In loamy Cerrado Oxisols, the binding of larger aggregates by roots is ofgreater significance next to bonding with polysaccharides. To invert disaggrega-tion, crop–pasture rotations are proposed, especially for the loamy soils which


would benefit from the strong rooting and polysaccharide production of pas-tures.

Compaction after land-use change is induced by disaggregation and leads to areduction of macro-pores and the enrichment of meso-pores, whereas micro-poresare only controlled by the texture of the soils. However, whether the increase ofthe generally low pore space at plant-available matrix potentials is positive forannual crops or not requires further research.

Acknowledgements

This BMZ-financed study was carried out within GTZ project 94.7860.3-01.100 in collaboration with CIAT, EMBRAPA-CPAC, and the University ofBayreuth. We are greatly indebted to A. Farias for the determination ofpolysaccharides and oxyhydroxides.

References

Amelung, W., Flach, K.W., Zech, W., 1997. Climatic effects on soil organic matter compositionin the Great Plains. Soil Sci. Soc. Am. J. 61, 115–123.

Assad, E.D., Sano, E.E., Araujo, A.G., 1994. In: Precipitacao pluviometrica e veranico nos´ ˜ ´cerrados. EMBRAPA-CPAC, Brasılia, pp. 35–44, Technical report.´

Bartoli, F., Burtin, G., Guerif, J., 1992. Influence of organic matter on aggregation in Oxisols richin gibbsite or in goethite: II. Clay dispersion, aggregate strength and water-stability. Geoderma54, 259–274.

Beare, M.H., Bruce, R.R., 1993. A comparison of methods for measuring water-stable aggregates:implications for determining environmental effects on soil structure. Geoderma 56, 87–104.

Blume, H.P., Schwertmann, U., 1969. Genetic evaluation of profile distribution of aluminum, ironand manganese oxides. Soil Sci. Soc. Am. Proc. 33, 438–444.

Bui, E.N., Mermut, A.R., Santos, M.C.D., 1989. Microscopic and ultramicroscopic porosity of anOxisol as determined by image analysis and water retention. Soil Sci. Soc. Am. J. 53,661–665.

Castro Filho, C., Logan, T.J., 1991. Liming effects on the stability and erodibility of someBrazilian Oxisols. Soil Sci. Soc. Am. J. 55, 1407–1413.

Cheshire, M.V., 1979. Nature and Origin of Carbohydrates in Soils. Academic Press, London.Curmi, P., Kertzman, F.F., Queiroz Neto, J.P., 1994. Degradation of structure and hydraulic

Ž .properties in an Oxisol under cultivation Brazil . In: Ringrose-Voase, A.J., Humphreys, G.S.Ž .Eds. , Soil Micromorphology. Elsevier, Amsterdam, pp. 569–579.

de Campos, B.C., Reinert, D.J., Nicolodi, R., Ruedell, J., Petrere, C., 1995. Estabilidade estruturalde um latossolo vermelho-escuro distrofico apos sete anos de rotacao de culturas e sistemas de´ ´ ˜manejo de solo. R. Bras. Ci. Solo 19, 121–126.

Dorioz, J.M., Robert, M., Chenu, C., 1993. The role of roots, fungi and bacteria on clay particleorganization. An experimental approach. Geoderma 56, 179–194.

Elliott, E.T., Palm, C.A., Reuss, D.E., Monz, C.A., 1991. Organic matter contained in soilaggregates from a tropical chronosequence: correction for sand and light fraction. Agric.Ecosystems Environ. 34, 443–451.

Ž .El-Swaify, S.A., 1981. Physical and mechanical properties of Oxisols. In: Theng, B.K.G. Ed. ,Soils with Variable Charge. New Zealand Society of Soil Science, Lower Hutt, pp. 303–324.


EMBRAPA, 1979. Manual de metodos de analise de solos. SNLCS, Rio de Janeiro.´ ´Feller, C., Beare, M.H., 1997. Physical control of soil organic matter dynamics in the tropics.

Geoderma 79, 69–116.Gijsman, A.J., 1996. Soil aggregate stability and soil organic matter fractions under agropastoral

systems established in native savanna. Aust. J. Soil. Res. 34, 891–907.Goedert, W.J., 1983. Management of the Cerrado soils of Brazil: a review. J. Soil Sci. 34,

405–428.Goedert, W.J., 1989. Regiao dos cerrados: potencial agrıcola e polıtica para seu desenvolvimento.˜ ´ ´

Pesq. Agropec. Bras. 24, 1–17.Guggenberger, G., Christensen, B.T., Zech, W., 1994. Land-use effects on the composition of

organic matter in particle-size separates of soil: I. Lignin and carbohydrate signature. Eur. J.Soil Sci. 45, 449–458.

Johnson, K.M., Sieburth, J.McN., 1977. Dissolved carbohydrates in seawater: I. A precisespectrophotometric analysis for monosaccharides. Mar. Chem. 5, 1–13.

Mehra, O.P., Jackson, M.L., 1960. Iron oxide removal from soils and clays by a dithionite–citratesystem buffered with sodium bicarbonate. Clays Clay Miner. 7, 317–327.

Moraes, M.H., Benez, S.H., Libardi, P.L., 1995. Efeitos da compactacao em algumas propriedades˜fısicas do solo e seu reflexo no desenvolvimento das raızes de plantas de soja. Bragantia 54,´ ´393–403.

Neufeldt, H., 1998. Land-use effects on chemical and physical properties of Cerrado Oxisols.Ž .Bayreuther Bodenk. Ber. Bayreuth 59.

Ž w x .Puhe, J., 1994. Die Wurzelentwicklung der Fichte Picea abies L. Karst. bei unterschiedlichenŽ .chemischen Bodenbedingungen. Ber. Forschungszentrum Waldokosysteme Gottingen 108.¨ ¨

Roth, C.H., Pavan, M.A., 1991. Effects of lime and gypsum on clay dispersion and infiltration insamples of a Brazilian Oxisol. Geoderma 48, 351–361.

Roth, C.H., de Castro Filho, C., de Medeiros, G.B., 1991. Analise de fatores fısicos e quımicos´ ´ ´relacionados com a agregacao de um latossolo roxo distrofico. R. Bras. Ci. Solo 15, 241–248.˜ ´

Savin, N.E., White, K.J., 1977. The Durbin–Watson test for serial correlation with extremesample sizes or many regressors. Econometrica 45, 1989–1996.

Sutherland, R.A., Watung, R.L., El-Swaify, S.A., 1996. Splash transport of organic carbon andassociated concentrations and mass enrichment ratios for an Oxisol, Hawaii. Earth SurfaceProcesses Landforms 21, 1145–1162.

Ž .Wada, K., 1989. Allophane and imogolite. In: Dixon, J.B., Weed, S.B. Eds. , Minerals in SoilEnvironments. SSSA, Madison, pp. 1051–1087, Book Series No. 1.

Westerhof, R., Buurman, P., van Griethuysen, P., Ayarza, M., Vilela, L., Zech, W., 1999.Aggregation studied by laser diffraction in relation to plowing and liming in the Cerrado regionin Brazil. Geoderma 90, 277–290.

Documents

Distribution of water-stable aggregates and aggregating agents in Cerrado Oxisols