Scaling of pores in 3D images of Latosols (Oxisols) with contrasting mineralogy under a conservation management system.
Soils present a heterogeneous structure made up of the arrangement of solid and organic particles with a formation of porous space (Klein and Libardi 2002) distributed according to a complex hierarchical organisation (Taina et al. 2010). Assessment of soil structure can be made based on the study of pore diameter distribution, which is primarily influenced by the clay fraction mineralogy, which expresses specific attributes with marked variability in the porous space (Ferreira et al. 1999). In turn, the pore distribution, particularly of structural pores, is closely related to dynamic processes of water and solutes in the soil (Luo et al. 2010), the diffusion of gases (Munkholm et al. 2012), and microbiological processes and decomposition of organic matter (Martin et al. 2012), which arc attributes fundamental to natural and agricultural ecosystems (Gomes et al. 2012), as well as promoting root growth (Tracy et al. 2010).
Knowledge of the porous space with distinct geometry enables the understanding of a wide variety of its key functions (Luo et al. 2010; Munkholm et al. 2012), as well as understanding how agricultural systems alter the spatial distribution of these pores, especially of the structural or inter-aggregate pores (Dexter 2004; Lima et al. 2005; Carducci et al. 2011). A soil-conservation management system for coffee cultivation that primarily aims to improve and/or preserve soil quality has been employed in the Minas Gerais State, in the Cerrado biome, particularly in the Alto Sao Francisco River Valley, Brazil. This system is based on the use of practices that seek to improve the soil physical and chemistry quality at greater depth (0.60 m) along the row, thus enabling a deeper root system (Serafim et al. 2011) and consequent benefits.
The visualisation and quantification of soil pores at a high level of detail, non-destructively obtained through the use of X-ray computed tomography (CT) imaging, allow a detailed analysis of the internal three-dimensional (3D) structures of the components of the soil (colloids, water and air) (Tippkotter et al. 2009), as well as identification of modifications promoted by external agents (Luo et al. 2010; Pires et al. 2011).
The objective of this research was to evaluate the spatial and morphological configuration of the pore space in 3D images of Latosols of different mineralogy under a soil-conservation management system in a coffee crop area.
Materials and methods
Description and characterisation of the study area
The study was conducted in a coffee area under a soil-conservationist management system in the municipality of Sao Roque de Minas, physiographic region of the Alto Sao Francisco River, in Minas Gerais State, Brazil. The cultivated area is 52 ha and has coordinates 20[degrees]15'43"S, 46[degrees]22'17"E, altitude 876 m, at the top of the hill, and 20[degrees]11'35"S, 46[degrees]22'07"E, altitude 841m in the middle third of the hill. The climate is classified as humid temperate with dry winters and rainy summers according to Koppen climate classification (Cwa).
Macromorphological description of soil structure was performed according to Santos et al. (2013) and it is presented in Table 1, along with some factors that limit the soil use and management. The soils were classified according to Brazilian Soil Taxonomy (Embrapa Solos 2013) as Red-Yellow Latosol (RYL) and Red Latosol (RL). Most Latosols in Brazil correspond to Oxisols in the US Soil Taxonomy (Soil Survey Staff 2006) and to Ferralsols in the World Reference Base (IUSS Working Group WRB 2006).
According to the principles of conservation management, coffee (Coffea arabica L.) cv. yellow Catucai was planted at a narrow row spacing of 2.50 by 0.65 m. Soil preparation included one ploughing and two harrowings, with application of ameliorants over the total area (4 Mg [ha.sub.-1] of limestone + 1.92Mg [ha.sub.-1] of gypsum). We used a subsoiler followed by a spade fertiliser for opening of the 60-cm-deep, 50-cm-wide furrow, and applied the formula 08-44-00 nitrogen (N)-phosphorus (P)-potassium (K), enriched with 1.5% Zn and 0.5% B as base fertiliser. Coffee seedlings were planted between the second half of October and the first half of November 2005.
After planting, 7kgm 1 of gypsum was applied along the row, this ameliorant being covered with soil material mixed with the inter-row plant material (mixture piled up at the base of the coffee plant stem). Together with the grass pasture crop, Brachiaria decumhens (syn. Urochloa) was planted between the planting lines, and was periodically cut with a brush cutter, which minimises competition with the main crop and allows the plant residue produced to be distributed along the row as well as between rows. This practice aims to improve soil structure and serves as protection against erosive agents (Lima et al. 2012). Crop operations were carried out using animal traction equipment; only the harvest was done mechanically. Nutritional monitoring and fertilisation management of the coffee crop was conducted based on leaf analysis (Serafim et al. 2011).
Soil sampling and physical, chemical and mineralogical characterisation
For sampling and characterisation of soil samples, three trenches were dug at random positions lengthwise along the row, with dimensions 0.70 m (width) by 1.50 m (length) by 1.50 m (depth), and then samples were collected for laboratory analysis. At sampling (September 2011), the crop had 6 years of cultivation.
Intact soil cores were sampled by hand, using plexiglass tubes (0.065 m diameter by 0.14 m length) equipped with a specially designed aluminium sampling ring at depths of 0.20-0.34, 0.80-0.94 and 1.50-1.64 m in the plant row, between plants (0.65 m) and slightly below the layer that had traces of gypsum, with three replications for each soil class, totalling 18 soil cores.
In the laboratory, the disturbed soil samples collected after the undisturbed samples collection were air-dried and passed through a 2-mm mesh sieve.
Particle-size analyses were conducted via slow shaking using NaOH (1 mol [L.sup.-1]) as a chemical dispersant in contact with the sample for 16 h (Embrapa Solos 2011). Sand was determined (2.00-0.05 mm) by sieving, clay content (<0.002 mm) by the pipette method, and silt (0.05-0.002 mm) by difference (Table 2).
The Si[O.sub.2], [Al.sub.2][O.sub.3] and [Fe.sub.2][O.sub.3] contents were determined by sulfuric acid digestion and used in calculations of the Ki (Si[O.sub.2]/Al203) and Kr [Si[O.sub.2]/([Al.sub.2][O.sub.3] + [Fe.sub.2][O.sub.3])] molecular ratios (Embrapa Solos 2013). Kaolinite, gibbsitc, hematite and goethite contents were derived from X-ray diffraction measurements and stoichiometric ratios derived from their ideal chemical formulae, according to Resende et al. (1987) (Table 3).
The tomographic analysis permits observation of the structural components of the soil in their natural forms, allowing better overall visualisation of the soil structure and porous space.
First, the soil cores in plexiglass tubes were dehydrated in an oven at 40[degrees]C until constant weight, to minimise possible interference of water films on the X-ray attenuation, and then were scanned at 120kV and 170 mA with an integration time of 3500 mS, generating an 2D axial projection of X-ray attenuation imagery, in microCT scan (EVS/GE MS8x-130), third-generation preclinical, cone-beam, equipped with tungsten X-ray tube. Excitation energy of 100kV and 130mA was employed for all samples, in the Soil Image Laboratory at the University of Guelph, Canada.
In order to obtain higher accuracy in the analysis and elimination of possible structural alterations generated during the sampling, a 0.033-m slice from the middle of the core was selected for scanning. As the X-ray source emits polychromatic X-rays (Clausnitzer and Hopmans 2000), we employed a pre-filter with a high-pass copper foil (0.5 mm) in order to reduce beam-hardening artefacts and maximise of the contrast between the different phases of the soil core (solid and air).
The 2D axial projections were acquired and reconstructed with 20-[micro]m spatial resolution (pixel size) and they were saved in 16-bit radiometric resolution. Then 3D sub-volumes of interest were selected in the exact centre of each original image and the reconstruction was done using proprietary filtered back-projection software called 'eXplore Reconstruction Utility' (GE Healthcare 2006). The final isometric volume (666 by 666 by 550 voxels) was reconstructed at 60-[micro]m voxel size, to maximise both region of interest and spatial resolution, within a manageable file size of 500 MB (Fig. 1).
For the purpose of comparing the X-ray attenuation, in CT imagery, we used the values in Hounsfield scale (defined relative to air [-1000 HU] and water [0 HU]) and a calibration procedure through the use of two capillary tubes (one filled with water and other with air inserted between the inner wall of the plexiglass tube and the sample at the time of scanning). Subsequently, the coefficients of water and air were calculated and a Gaussian smoothing filter (radius of 1 voxel) was used to reduce image noise and artefacts with the aid of MicroView (GE Healthcare 2006) before subsequent analyses in NIH ImageJ (Ferreira and Rasband 2011; Rasband 2012).
Digital image processing and analyses
The 3D image processing followed the protocol of the Soil Image Laboratory, University of Guelph, Canada. This is the first step in 3D image processing, which involves the conversion of each greyscale voxel value of the greyscale image (proportionally expressing the locations of the X-ray attenuation coefficients) into a binary image, distinguishing the void and non-void, done in the selected sub-volume images.
The thresholding was done in ImageJ (National Institutes of Health) by a method based on the work of Schliiter et al. (2010), employing both Laplacian edge detection and seeded-region-growing to assign the voxels associated with the zero-crossing, but first enforcing clamping of the grey-scale image (considering the histogram peak positions for air and solid), to minimise unnecessary edges.
With the public domain image analysis software ImageJ (Rasband 2012), all analyses are done in full 3D mode, which allows differentiation of data categories according to the desired micromorphological size and shape classes. To obtain spatial data, the 'Semivariance 3D' plugin of the ImageJ was used, with the X-ray attenuation value of the greyscale image. We obtained [??](h) semivariances for the 3D image, i.e. the orthogonal directions (X, Y, Z), where the semivariance is standardised [0-1]. The construction of the experimental semivariogram identified the spatial dependence amplitude of the variable under study and defined the spatial variability structure (Goovaerts 1999).
The 'Analyze Particle' function of ImageJ was used to calculate the pore dimensions (volume and area). This function detects and measures the objects in the binary image to obtain data relative to pores present in the image volume under study (n = 550). From this result, the equivalent diameter to a sphere (3D images, spatial geometry simulation) and, as the shape factor, the sphericity of the pore (0, less rounded; angular, 1, more rounded and smooth) were calculated by Eqn 1:
[PSI] = [[pi].sup.1/3][(6F).sup.2/3] / A (1)
where [pi] [approximately equal to] 141592 ... 8, V is volume ([mm.sup.3]) and A is area ([mm.sup.2]).
The classification of 3D shapes of the pores was performed according proposals put forward by Bullock et al. (1985). The results were obtained via the plugin Analyze Particle, selecting the 'ellipsoids' option on the binary images. Analyse Particle fits an ellipsoid to each object (void), and the three axis [major axes (a), intermediate (b) and lower (c)] of the ellipsoid are generated, which are then used to categorise the pore (b/a and c/b ratio) as: acicular-planar), acicular, planar, platy, channels, disc and spheroid, all the steps made by ImageJ software.
Spatial analyses consisted of constructing experimental semivariograms and fitting a theoretical semivariogram to explain the structure of the data variance. The semi-variance [??](h) is a function of the distance h, which is estimated in a discrete set of lag distances expressed by a scatterplot that allows the variographic analysis of the spatial dependence amplitude of the variables studied (Faraco et al. 2008), in this case the X and Y orthogonal directions (horizontal direction) and Z (vertical direction), defining then the parameters required for the estimation of characteristics resulting from the spatial variability structure (Avila et al. 2010).
The standardised semivariograms were estimated by the classical method, through the estimator:
[??](h) = 1/2N(h)[[summation].sup.N(h).sub.i-1]] [[([i.sub.i]) - Z([i.sub.i] + h)].sup.2] (2)
in which [??][h) is the semivariance estimator, N(h) is the number of pairs of measured values, Z([i.sub.i]) and Z([i.sub.i] + h), separated by a vector distance h, are realisations of the random variable Z([i.sub.i]) (Journel and Huijbregts 1978; Isaaks and Srivastava 1989). We then proceeded to estimate the experimental semivariance for each replication, and with the average semivariance, it was fitted to the exponential model (Eqn 3), which presented the better fit of data, and then determined model parameters nugget effect ([C.sub.0]), sill ([C.sub.0] + [C.sub.1]) and theoretical and practical range (a), in R language (R Development Core Team 2012), specifically with the 'geoR' package (Ribeiro and Diggle 2001), both having free access and in accordance with the GPL (General Public Licence):
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. (3)
Subsequently, prediction intervals (PI) were constructed for the set of replications, in order to compare the spatial variability, as with the PI, so it can be stated that 95% of the samples were predicted.
The detectable pores were classified based on data mining by the 'randomForest' package (Liaw and Wiener 2012), and other functions developed in R language (R Development Core Team 2012), where it was possible to generate various pore diameter classes and to detect the most contrasting ones.
Results and discussion
The empirical semivariograms indicated a second-order stationarity for the variables evaluated, reflected by clear and well-defined sill, where it was asymptotically reached by the exponential fit of the theoretical model (Figs 2-7). This theoretical model represents an average continuity of spatial variability, i.e. this fit can be explained by the high ease of alteration of the soil-structure-related attributes promoted by human action (Bottega et al. 2013).
Through empirical and theoretical semivariograms, as well as the PIs, the spatial dependence in all combinations (2 soils x 3 depths x 3 orthogonal directions) was detected in the 3D greyscale images (Figs 2-7), since it was not sufficient to verily the anisotropy of soil structure by the image analysis only by visual observation of orthogonal directions of the semivariograms.
Through theoretical semivariograms of the X and Y (horizontal) and Z (vertical) orthogonal directions, it was possible to verify the variability of the spatial dependence of the structural components (mineral portion and pore space) of Latosols on a micrometric scale.
Estimates of adjusted semivariograms that best describe the behaviour of the soil structure spatial variability are presented in Table 4. The higher ranges of spatial dependence for the orthogonal directions occurred in RL, indicating its lower variability and higher spatial continuity (Bottega et al. 2013), i.e. the range indicates the limit distance in which the sample points are correlated among themselves (Vieira 2000).
The high weathering and leaching undergone by RL was reflected in the clay (faction mineralogy with marked presence of oxides of iron (Fe; hematite and goethite) and aluminium (Al; gibbsite) (Table 3). Camargo et al. (2008a) attributed higher ranges of spatial dependence to these minerals, because crystallographically, they are more homogeneous.
The Al and Fe oxides, particularly in RL, contribute to the formation of the granular-type structure, where the distribution of minerals in relation to the colloidal size and relatively soluble material follows an agglutinated pattern (Ferreira et al. 1999), in more spherical micro-aggregates (diameter <0.84 mm) (Ghidin et al. 2006), conferring a high predominance of interaggregate or compound packed pores, which are in turn arranged in interconnected cavities (Ferreira et al. 1999; Vidal-Torrado et al. 1999; Bispo et al. 2011). This can be seen in pore diameter distribution, where there was a predominance of greater number and volume of pores mainly in the diameter range 0.2-0.6 mm, detected in 3D images at the three depths (Figs 5-7).
For the 0.80-0.94 m depth in RL, the lower ranges are due to the transition zone between the soil layers that underwent the action provided by conservation system and the soil layer that preserves the intrinsic attributes of the soil. At this depth, a narrowing of the PI was observed related to the lower variability of the sampling.
The RYL, which was dominated by kaolinite in the clay fraction (Table 3), showed lower ranges of spatial dependence than RL, confirming the results obtained by Camargo el al. (2008fi), where they attributed this aspect to kaolinite, due to it presenting a positive correlation with the formation of aggregates >2 mm. According to those authors, kaolinite favours the predominance of cavity-type pores; that is, the mineral particles are enveloped in a dense colloidal-size and relatively soluble material, with little tendency towards the development of microstructure (Ferreira et al. 1999), reducing the structural organisation (Pereira et al. 2010) and thereby limiting the formation of larger diameter pores along the studied profile (Ghidin et al. 2006) (Fig. 8).
Regarding the Pis, the highest amplitudes were observed at a depth of 0.80-0.94 m for RYL orthogonal directions, which coincided with the transition between the BC and C horizons where is the some more easily weatherable primary minerals are present. The RL at 0.20-0.34 m depth showed the joint effects of soil tilling occurred during planting-row preparation, altering the soil structure (Oliveira et al. 2004). The contribution of organic matter from the high input of organic residue from coffee and Brachiuria sp., their residues being scattered on the coffee line and inter-row, probably aids the formation and stabilisation of aggregates (Costa Junior et al. 2012).
The greater spatial dependence range (high value) generally occurred in the Z direction (vertical) in both soils (granular and blocky structure) preferentially at depths of 0.20-0.34 and 1.50-1.64 m (Table 3) and in the Y direction (horizontal) at 0.80-0.94 m. This means that these directions are more correlated, and therefore have more homogeneous structural organisation, possibly with greater pore continuity and presence of larger diameter pores (Schafffath et al. 2008), unlike the soils of temperate regions with platy structure and under freeze-thaw processes, where the highest spatial variability is concentrated in the vertical direction (Z) with low values of spatial range (Taina et al. 2013).
Those same authors detected a high value of spatial range in all directions, which reflects the structure type; however, in the Latosols, the spatial range value observed was much lower, and that indicated the complex organisation of microgranular structure.
In all situations, the nugget effect ([C.sub.0]) was null; i.e. the good spatial resolution (60 ([micro]m) of images generated a higher level of soil organisation details, so within the limits of the range and sill, most of the variability can be explained by the spatial component. Thus, the strong spatial dependence of the orthogonal directions for the soils evaluated was evident (Schaffrath et al. 2008). The sills for all combinations have very similar values, with lower values, however, for RL.
Through analysis of the generated 3D images, a large number of pores at the boundaries between macro-and mesopores for both soils could be detected. Subsequently, the classification of these pores by diameter was performed by means of data mining (Liaw and Wiener 2012), which allowed reproduction of a larger number of classes in order to maximise the difference between level combinations.
The results of the diameter distribution of detectable pores and sphericity are shown in Fig. 8. With the generation of 26
pore-diameter classes, it was possible to detect the contrasts between the soils within an interval of 0.2-1 mm, which corresponds to the boundary between the thin macropores and large mesopores (Bullock et al. 1985). The distinction between the soils was evident in the intervals from 0.2 to 0.4 mm and from 0.4 to 0.6 mm, where the highest number and volume of visible pores occurred, i.e. those that occupy the major part of the soil matrix.
Both the number and volume of the pores followed similar trends within each soil, being more evident in RL throughout the studied depths. The microgranular structure (Vidal-Torrado et al. 1999; Bispo et al. 2011), by favouring the formation of a larger quantity of voids especially inter-aggregate ones (Ferreira et al. 1999; Volland-Tuduri et al. 2005; Carducci et al. 2011), helps to explain this trend.
With respect to the shape factor, indicating the roundness and surface roughness of the evaluated pores, the sphericity was used (0, less rounded and angular; 1, more rounded and smooth) by simulation of the spatial geometry of the pores. The sphericity increased with decreasing pore diameter (Fig. 8), as observed by Tippkotter et al. (2009).
The predominance of more spherical pores was evident in the smaller diameter classes at 0.20-0.34 m depth, especially in RL. This result is attributed to the structure type of RL favouring the presence of higher gibbsite content than RYL (Table 3), which acts in the formation of very small micro-aggregates and more rounded micropcds (Ferreira et al. 1999; Vidal-Torrado et al. 1999; Bispo et al. 2011).
This higher sphericity can also be related to the effect of the management system adopted, which turns over the soil and promotes the formation of smaller, more rounded aggregates, corroborating observations of Cremom et al. (2009). The larger diameter pores (>0.8 mm) were more angular and rougher. At greater depths, some similarity between the soils was observed.
The study of pore diameter distribution is traditionally based on construction of the water retention curve, which relates the diameter of the capillary pore to the tension applied to the samples, in order to give a diameter value for each point of the curve. Based on this curve and using mathematical modelling, Carducci et al. (2013) detected the bimodal characteristic of RL, marked by the presence of structural and textural pores, evidenced by the two inflection points of the curves along the soil profile, as well as the influence of the gibbsitic mineralogy, resulting in greater water-retention capacity.
This bimodal scaling of RL pores was confirmed by the high number of high-volume pores detected in the CT images (Fig. 8), which seems to indicate the existence of an extensive network of well-connected pores (Volland-Tuduri et al. 2005; Luo et al. 2010), typical of the Cerrado biome Latosols (K.er 1997).
The Latosols have high water infiltration (Mello and Curi 2012) and percolation into the structural pores (Ferreira et al. 1999), and retain large amounts of water with very high energy, especially in the textural pores (matric potential <-1500kPa), quickly releasing the water present in the structural pores, as observed by the abrupt slope of the curve (Carducci et al. 2011, 2013) , which represents the structural pore distribution (Dexter 2004).
On the other hand, the RY L presented lower number and pore volume at the measured depths and its pore scaling had a more homogeneous distribution among the generated diameter classes, especially at 0.20-0.34 m depth (Figs 2-4), which can contribute to the gradual release of water to the coffee roots, thus prolonging the time of available water supply for the plants. Note that the coffee plant needs to absorb large amounts of water to maintain high fructification (Rena and Guimaraes 2000).
According to Gantzer and Anderson (2002), a more accurate assessment of the spatial heterogeneity of the different soil structure components requires the use of tools that adequately quantify the morphology of the pores and their connections. Measures of number, area, perimeter and roundness are useful tools. However, they have a limited capacity to describe the systematic behaviour of a large volume of reconstructed images. These analyses are beneficial in describing the interactions among the soil physical, biological and chemical processes, thus discriminating the effects of tillage systems on soil attributes.
From the classification of the 3D shapes of the pores, we found that high percentages of pores in the classifications acicular (needle-like cylindrical), channels and spheroid occur along the profile for both Latosols (Fig. 9).
Pores of a more spherical shape were detected in the gibbsitic Latosol, confirming the results obtained with the shape-factor variable (Fig. 8), as well as most of the pores of the platy, planar and acicular classes, which may be similar to the fissural pores (Castro et al. 2003), occurring in the kaolinitic Latosol.
The presence of certain types of pores reveals the role of factors external to the organisation of soil particles that can influence their formation or transformation. Therefore, monitoring of the pore-class distribution in agricultural areas is relevant, since these minimum substantive changes can modify the dynamic processes occurring in the soil, which can be reflected in the crop yield (Lima et al. 2005).
The channel pores detected in 3D images usually more closely resemble cylinders and are indicative of biological activity (biopores), i.e. they are formed by the action of roots and soil organisms (Castro et al. 2003; Genro et al. 2004; Lima et al. 2005, 2012). There was a significant percentage of this morphological class (-30%) at the three studied depths. When opening the trenches for morphological description and soil sampling, a strong development of the coffee plant roots with an extensive root branching network and consequent greater soil exploration was visible.
Based on the above, it is noteworthy that the management system adopted here in the first years of implementation probably abruptly altered the soil structure. However, after 6 years of cultivation, the reorganisation of the structure as well as the voids in the Latosols could have occurred, as observed by Lima et al. (2012) in a study of structural recovery of a Red-Yellow Latosol promoted by coast-cross grass (Cynodort sp.), confirmed by morphological similarities between depths of 0.20-0.34 and 1.50-1.64 m, which represent the direct effect of the conservation system and intrinsic attributes of the soil, respectively.
The pore morphology was probably influenced by the joint actions of the soil management practices adopted, which aim to improve the soil physical quality. The greater amount of organic residue on the soil surface from Brachiaria spp. maintained between rows, as well as the renewal and growth of roots along the soil profile (Luo et al. 2010; Lima et al. 2012), may contribute to increasing the organic matter content and its decomposition, which would release soil microbiota-activating substances and thereby act to stabilise the aggregates (Costa Junior et al. 2012; Martin et al. 2012).
Greater spatial variability occurred in the horizontal orthogonal direction (X and Y) of the 3D image. The pores detected were different between the Latosols studied, mainly at the 0.20-0.34 m depth. The largest number and volume of pores was found in the Red Latosol (RL). Soil class and the management system both have impacts on the 3D pore characteristics and need to be considered together in order to better characterise the causes of pore variability. The sphericity shape factor was similar for both soils, but with greater emphasis on pore classes with a diameter <0.4 mm, mainly at the 0.20-0.34 m depth. The highest percentage of spheroid pores occurred in the gibbsitic Latosol, whereas platy pores were more abundant in the kaolinitic Latosol.
The authors thank CNPq for the granting of scholarships, FAPEMIG for financial support, the Consorcio de Pesquisas Cafeeiras Embrapa Cafe for providing vehicles and the Empresa Agropecuaria Piumhi (AP) in the person of Alessandro Oliveira for logistic support and utilisation of the experimental area. The authors acknowledge the UFLA for providing institutional support to accomplish this project.
Received 27 August 2013, accepted 12 November 2013, published online 1 April 2014
Avila LF, Mello CR, Silva AM (2010) Continuidade e distribuifao espacial da umidade do solo em bacia hidrografica da Serra da Mantiqueira. Revista Brasileira de Engenharia Agricola e Ambiental 14, 1257-1266. dot: 10.1590/S1415-43662010001200002
Bispo FHA, Silva AC, Vidal-Torrado P, Souza Junior VS (2011) Highlands of the upper Jcquitinhonha Valley, Brazil. II mineralogy, micromorphology, and landscape evolution. Revista Brasileira de Ciencia do Solo 35, 1081-1091. doi: 10.1590/S0100-06832011000400002
Bottega EL, Queiroz DM, Pinto FAC, Souza CMA (2013) Variabilidade espacial de atributos do solo em sistema de semeadura direta com rotacao de culturas no cerrado brasileira. Revista Ciencia Agronomica 44, 1-9. doi: 10.1590/S1806-66902013000100001
Bullock P, Fedoroff N, Jongerius A, Stoops G, TursinaT (1985) 'Handbook for soil thin section description.' (Waine Research Publications: UK)
Camargo LA, Marques Junior J, Pereira GT, Horvat RA (2008a) Variabilidade espacial de atributos mineralogicos de um Latossolo sob diferentes formas do relevo. I--mineralogia da fragao argila. Revista Brasileira de Ciencia do Solo 32, 2269-2277. doi: 10.1590/0S100-06832008000600006
Camargo LA, Marques Junior J, Pereira GT, Horvat RA (20086) Variabilidade espacial de atributos mineralogicos de um Latossolo sob diferentes formas do relevo. II--correlagao espacial entre mineralogia e agregados. Revista Brasileira de Ciencia do Solo 32, 2279-2288. doi: 10.1590/S0100-06832008000600007
Carducci CE, Oliveira GC, Severiano EC, Zeviani WM (2011) Modelagem da curva de retengao de agua de Latossolos utilizando a equagao duplo van Genuchten. Revista Brasileira de Ciencia do Solo 35, 77-86. doi: 10.1590/S0100-06832011000100007
Carducci CE, Oliveira GC, Zeviani WM, Lima VMP, Serafim ME (2013) Bimodal pore distribution on soils under conservationist management system for coffee crop. Engenharia Agricola 33, 291 302. doi: 10.1590/S0100-69162013000200008
Castro SS, Cooper M, Santos MC, Vidal-Torrado P (2003) Micromorfologia do solo: bases e aplicagoes. In 'Topicos em ciencia do solo', (lids RF Novais, VVH Alvarez, CE Schaefer) pp. 107-164. (Sociedade Brasileira de Ciencia do Solo: Vigosa, Brazil)
Clausnitzer V, Hopmans JW (2000) Pore-scale measurements of solute breakthrough using microfocus X-ray computed tomography. Water Resources Research 36, 2067-2079. doi:10.1029/2000WR900076
Costa Junior C, Piccolo MC, Siqueira Neto M, Camargo PB, Cerri CC, Bemoux M (2012) Carbono em agregados do solo sob vegetagao nativa, pastagem e sistemas agricolas no Bioma Cerrado. Revista Brasileira de Ciencia do Solo 36, 1311 1322. doi: 10.1590/S0100-068320120004 00025
Cremom C, Rosa Junior EJ. Serafim ME, Ono FB (2009) Analise micromorfometrica de agregados de um Latossolo Vermelho Distroferrico em diferentes sistemas de manejo. Acta Scientiarum 31, 139-146. doi: 10.4025/actasciagron.v3111.6640
Dexter AR (2004) Soil physical quality part 1: theory, effects of soil texture, density, and organic matter, and effects on root growth. Geoderma 120, 201-214.
Embrapa Solos (2011) 'Manual de metodos de analise de solos.' 2nd edn (Centro Nacional de Pesquisa de Solos: Rio de Janeiro)
Embrapa Solos (2013) 'Sistema brasileira de classificagao de solos.' 3th edn (Centro Nacional de Pesquisa de Solos: Rio de Janeiro)
Faraco MA, Uribe-Opazo MA, Silva EAA, Johann JA, Borssoi JA (2008) Selegao de modelos de variabilidade espacial para elaboragao de mapas tematicos de atributos fisicos do solo e produtividade da soja. Revista Brasileira de Ciencia do Solo 32, 463-476. doi: 10.1590/S010006832008000200001
Ferreira T, Rasband W (2011) IJ, 1.45m. ImageJ user guide. Available at: http://imagej.nih.gov/ij/docs/guide/index.html
Ferreira MM, Femades B, Curi N (1999) Mineralogia da fracao argila e estrutura de Latossolos da regiao sudeste do Brasil. Revista Brasileira de Ciencia do Solo 23, 507 514.
Gantzer CJ, Anderson SH (2002) Computed tomographic measurement of macroporosity in chisel-disk and no-tillage seedbeds. Soil & Tillage Research 64, 101-111. doi: 10.1016/S0167-1987(01)00248-3
Genro SA, Jr, Reinert DJ, Reichert JM (2004) Variabilidade temporal da resistencia a penetragao de um Latossolo argiloso sob semeadura direta com rotagao de cultures. Revista Brasileira de Ciencia do Solo 28, 477 484. doi: 10.1590/S0100-06832004000300009
Ghidin AA, Melo VF, Lima VC, Lima JMJC (2006) Topossequencias de Latossolos originados de rochas basalticas no Parana. II--relagao entre mineralogia da fragao argila e propriedades fisicas dos solos. Revista Brasileira de Ciencia do Solo 30, 307 319. doi: 10.1590/S0100-06832006000200011
Gomes JBV, Fernandes MF, Barreto AC, Araujo Filho JC, Curi N (2012) Soil attributes under agroecossystcms and forest vegetation in the coastal tablelands of Northeastern Brazil. Ciencia e Agrotecnologia 36, 649-664. doi: 10.1590/S1413-70542012000600007
Goovaerts P (1999) Geostatistics in soil science: state-of-the-art and perspectives. Geoderma 89, 1-15. doi: 10.1016/S0016-7061 (98)00078-0
GE Healthcare (2006) Microview Analysis 2.2. Technical Publication. Direction 2407688. Revision 1. Available at: www.oucom.ohiou.edu/ou-microct
Isaaks EH, Srivastava RM (1989) 'An introduction to applied geostatistics.' (Oxford University Press: New York)
IUSS Working Group WRB (2006) 'World reference base for soil resources 2006.' World Soil Resources Reports No. 103. (FAO: Rome)
Journel A, Huijbregts CJ (1978) 'Mining geostatistics.' (Academic Press: San Diego, CA, USA)
Ker JC (1997) Latossolos do Brasil: uma revisao. Geonomos 5, 17 40. Klein VA, Libardi PL (2002) Dcnsidadce distribuigao dodiametro dos poros de um Latossolo Vermelho sob diferentes sistemas dc uso e manejo. Revista Brasileira de Ciencia do Solo 26, 857-867.
Liaw A, Wiener M (2012) Package 'randomForest'. Breiman and Cutler's Random Forests for classification and regression. Version 4.6-7. Available at: http://stat-www.berkcley.edu/users/breiman/RandomForests
Lima HV, Lima CLR, Leao TP, Cooper M, Silva AP, Romero RE (2005) Trafego de maquinas agricolas e altcragoes de bioporos em area sob pomar de laranja. Revista Brasileira de Ciencia do Solo 29, 677-684. doi: 10.1590/S0100-06832005000500003
Lima VMP, Oliveira GC, Serafim ME, Curi N, Evangelista AR (2012) lntervalo hidrico otimo como indicador de melhoria da qualidade estrutural de Latossolo degradado. Revista Brasileira de Ciencia do Solo 36, 71-78. doi: 10.1590/S0104-07072005000400015
Luo L, Lin H, Li S (2010) Quantification of 3-D soil macropore networks in different soil types and land uses using computed tomography. Journal of Hydrology 393, 53-64. doi: 10.1016/j.jhydrol.2010.03.031
Martin SL, Mooney SJ, Dickinson MJ, West HM (2012) The effects of simultaneous root colonization by three Glomus species on soil pore characteristics. Soil Biology & Biochemistry 49, 167 173. doi: 10.1016/j.soilbio.2012.02.036
Mello CR, Curi N (2012) Hydropedology. Ciencia e Agrotecnologia 36, 137 146. doi: 10.1590/S1413-70542012000200001
Munkholm LJ, Heck RJ. Deen B (2012) Soil pore characteristics assessed from X-ray micro-CT derived images and correlations to soil friability. Geoderma 181-182, 22-29. doi:10.1016/j.geoderma.2012.02.024
Oliveira GC, Dias Junior MS, Resck DVS, Curi N (2004) Caracterizacao quimica e fisico-hidrica de um Latossolo Vermelho apos vinte anos de Manejo e cultivo do solo. Revista Brasileira de Ciencia do Solo 28, 327-336. doi: 10.1590/S0100-06832004000200011
Pereira TTC, Ker JC, Schaefer CEGR, Barros NF, Neves JCL, Almeida CC (2010) Genese de Latossolos e Cambissolos desenvolvidos de rochas peliticas do grupo Bambui--Minas Gerais. Revista Brasileira de Ciencia do Solo 34, 1283-1295. doi:10.1590/S0100-06832010000400026
Pires LF, Cassaro FAM, Saab SC, Brinatti AM (2011) Characterization of changes in soil porous system by gamma-ray tomography. Nuclear Instruments & Methods in Physics Research. Section A. Accelerators. Spectrometers, Detectors and Associated Equipment 644, 68-71. doi: 10.1016/j.nima.2011.04.025
R Development Core Team (2012) R: a language and environment for statistical computing. Institute for Statistics and Mathematics, Vienna, Austria. Available at: www.R-project.org
Rasband W (2012) Image J: Image processing and analysis in Java (Online). Research Services Branch, National Institute of Health. Bethesda, MD, USA. Available at: http://rsb.info.nih.gov/ij/index.html
Rena AB, Guimaraes PTG (2000) Sistema radicular do cafeeiro: estrutura, distribuicao, atividade e fatores que o influenciam. Documentos 37. EPAMIG, Belo Horizonte, MG, Brasil.
Resende M, Bahia Filho AFC, Braga JM (1987) Mineralogia da argila de Latossolos estimada por alocacao a partir do teor total de oxidos do ataque sulfurico. Revista Brasileira de Ciencia do Solo 11, 17-23.
Ribeiro PJ Jr, Diggle PJ (2001) geoR: a package for geoestatistical analysis. R-NEWS 1, 15-18.
Santos RD, Lemos RC, Santos HG, Ker JC, Anjos LHC (2013) 'Manual de descricao e coleta de solo no campo'. (Sociedade Brasileira de Ciencia do Solo: Vicosa, MG, Brasil)
Schaffrath VR, Tormena, Fidalski J, Andrade ACG (2008) Variabilidade e correlacao espacial de propriedades fisicas de solo sob plantio direto e preparo convencional. Revista Brasileira de Ciencia do Solo 32, 1369-1377. doi: 10.1590/S0100-06832008000400001
Schluter S, Weller U, Vogel HJ (2010) Segmentation of X-ray microtomography images of soil using gradient masks. Computers & Geosciences 36, 1246 1251. doi:10.1016/j.cageo.2010.02.007
Serafim ME, Oliveira GC, Oliveira AS, Lima JM, Guimaraes PTG, Costa JC (2011) Sistema conservacionista e de manejo intensivo do solo no cultivo de cafeeiros na regiao do Alto Sao Francisco, MG: um estudo de caso. Bioscience Journal 27, 964-977.
Soil Survey Staff (2006) 'Keys to Soil Taxonomy.' 8th edn (United States Department of Agriculture, Natural Resources Conservation Service: Washington, DC)
Taina IA, Heck RJ, Elliot TR, Scaiff N (2010) Micromorphological and X-ray uCT study of Orthic Humic Glcysols under different management conditions. Geoderma 158, 110-119. doi:10.1016/j.geoderma.2010.02. 009
Taina IA, Heck RJ, Deen W, Ma EYT (2013) Quantification of freeze--thaw related structure in cultivated topsoils using X-ray computer tomography. Canadian Journal of Soil Science 93, 533-553. doi: 10.4141/cjss2012-044
Tippkotter R, Eickhorst T, Taubner H, Gredner B, Rademaker G (2009) Detection of soil water in macropores of undisturbed soil using microfocus X-ray tube computerized tomography (uCT). Soil & Tillage Research 105, 12 20. doi: 10.1016/j.still.2009.05.001
Tracy SR, Roberts JA, Black CR, Mcneill A, Davidson R, Mooney SJ (2010) The X-factor: visualizing undisturbed root architecture in soils using X-ray computed tomography. Journal of Experimental Botany 61, 311-313. doi:10.1093/jxb/erp386
Vidal-Torrado P, Lepsch IF, Castro SS, Cooper M (1999) Pedogenese em uma sequencia Latossolo podzolico na borda de um plato na depressao periferica paulista. Revista Brasileira de Ciencia do Solo 23, 909-921.
Vieira SR (2000) Geoestatistica em estudo de variabilidade espacial do solo. In 'Topicos cm ciencia do solo'. Vol. 1. (Eds RF Novais, VHV Alvarez, CEGR Schaefer) pp. 1-55. (Sociedade Brasileira de Ciencia do Solo Vicosa, MG, Brasil)
Volland-Tuduri N, Bruand A, Brossard M, Balbino LC, Oliveira MIL, Martins ES (2005) Mass proportion of microaggregates and bulk density in a Brazilian clayey Oxisol. Soil Science Society of America Journal 69, 1559-1564. doi:10.2!36/sssaj2003.0344
Carla Eloize Carducci (A,E), Geraldo Cesar de Oliveira (B), Nilton Curi (B), Richard John Heck (C), and Diogo Francisco Rossoni (D)
(A) Federal University of Santa Catarina, Curitibanos Campus (UFSC), Curitibanos, SC 89520-000, Brazil.
(B) Soil Science Department of Federal University of Lavras (UFLA), Lavras, MG 37200-000, Brazil.
(C) School of Environmental Science, University of Guelph, Guelph, ON N1G 2W1, Canada.
(D)Statistical Department, State University of Maringa (UEM), Maringa, PR 87020-900, Brazil.
(E) Corresponding author. Email: email@example.com
Table 1. Soil structure and factors limiting the use and management of the B horizon of a Red-Yellow Latosol (RYL) and Red Latosol (RL) ED, Effective depth; TC, texture class, where c is very clayey; TS, terrain slope; S, specific, in which d is dystrophic. Adapted from Serafim et at. (2011) Soil Structure Degree ED TC class Type Class (m) RYL Subangular blocks Small Weak >2.1 c RL Granular Very Small Strong >2.2 c Limiting factor Soil Drainage TS Erosion S class (%) RYL Well drained 4-8 Not apparent d RL Well drained 9-14 Not apparent d Table 2. Particle size distribution (g [kg.sup.-1]) of the B horizon of a Red-Yellow Latosol (RYL) and Red Latosol (RL) Depth RYL RL (m) Clay Silt Sand Clay Silt Sand 0.20-0.34 716 210 74 869 66 65 0.80-0.94 724 213 63 895 46 59 1.50-1.64 700 238 62 904 39 57 Table 3. Chemical and mincralogical characterisation of the B horizon of a Red-Yellow Latosol (RYL) and Red Latosol (RL) Ki, Molecular ratio Si[O.sub.2]/[Al.sub.2][O.sub.3]; Kr, molecular ratio Si[O.sub.2]/[Al.sub.2][O.sub.3] + [Fe.sub.2][O.sub.3]); Kt, kaolinite; Gb, gibbsite; Hm, hematite; Gt, goethite Soils Si[O.sub.2] [Al.sub.2] [Fe.sub.2] [O.sub.3] [O.sub.3] (sulfuric acid digestion) (g [kg.sup.-1]) RYL 243.3 285.3 122.1 RL 127.2 364.1 158.2 Soils Ki Kr Kt Gb Hm Gt (%) RYL 1.45 114 52.30 40.46 0.69 6.55 RL 0.59 0.46 27.34 54.02 11.92 6.72 Table 4. Theoretical exponential model parameters adjusted for the orthogonal directions (X, Y, Z) of a Red-Yellow Latosol (RYL) and Red Latosol (RL) [C.sub.0], Nugget effect was null; [C.sub.0] + [C.sub.1],sill Soils Depth Orthogonal ([C.sub.0] + Theoretical Practical (m) direction [C.sub.1]) range range (mm) (mm) RYL 0.20-0.34 X 0.97 0.96 2.88 RYL 0.20-0.34 Y 0.95 0.90 2.70 RYL 0.20 0.34 Z 0.94 0.97 2.91 RYL 0.80-0.94 X 0.87 0.86 2.58 RYL 0.80 0.94 Y 0.93 1.00 3.00 RYL 0.80-0.94 Z 0.87 0.94 2.82 RYL 1.50-1.64 X 0.95 0.97 2.91 RYL 1.50 1.64 Y 0.96 0.85 2.55 RYL 1.50 1.64 Z 0.93 1.00 3.00 RL 0.20-0.34 X 0.86 0.99 2.97 RL 0.20 0.34 Y 0.87 1.01 3.03 RL 0.20-0.34 Z 0.87 1.17 3.51 RL 0.80 0.94 X 0.93 0.87 2.61 RL 0.80-0.94 Y 0.94 0.88 2.64 RL 0.80 0.94 Z 0.91 0.88 2.64 RL 1.50 1.64 X 0.95 1.02 3.06 RL 1.50-1.64 Y 0.92 1.01 3.03 RL 1.50 1.64 Z 0.98 1.11 3.33
|Printer friendly Cite/link Email Feedback|
|Author:||Carducci, Carla Eloize; de Oliveira, Geraldo Cesar; Curi, Nilton; Heck, Richard John; Rossoni, Diogo|
|Date:||May 1, 2014|
|Previous Article:||The biochar dilemma.|
|Next Article:||Changes in soil carbon fractions due to incorporating corn residues in organic and conventional vegetable farming systems.|