Aggregate Size, Particulate and Total Organic Carbon in Different Land Uses on a Sandy Loam Soil Exposed to Wind Erosion

Advances in Agricultural Science 06 (2018), 03: 95-111

Aggregate Size, Particulate and Total Organic Carbon in Different Land Uses on a Sandy Loam Soil Exposed to Wind Erosion

Eduardo A. Rienzi 1*, Antonio Marchi 2 and Gabriel Rodriguez 1

1University of Buenos Aires, Faculty of Agronomy, Soil and Water Conservation; Av San Martin 4452 Ciudad Autonoma de Buenos Aires, Argentina.
2Ministerio de Medio Ambiente Campoy Producción de San Luis, Argentina.


After 16 years under crops and pasture, an assessment was performed to determine the aggregation status and soil quality in a sandy loam soil exposed to wind erosion in a semiarid region of Argentina. The goal was to test the effectivity of pastures to improve soil qualities that increase the resistance to the erosive process and the degree of soil degradation produced under moldboard tillage. Soil samples from natural forest of Prosopis caldenia (calden) and bare soils were used to compare the effect of Secale cereale (rye) under plowing and cross strips with Eragrostis curvula (weeping love grass). The selected properties measured were aggregate size distribution (ASD), mean weight diameter (MWD), macro to microaggregate ratio, Chepil index, and total organic carbon content (OC). The OC and coarse (>0.250 mm) and fine (>0.05 mm) particulate organic matter, CPOM and FPOM, respectively, in four classes of aggregates (0.250, 0.500, 2.5 and 4.8 mm in size) were recorded. The results indicated that rye and Eragrostis included 80% of the ASD with aggregates smaller than 0.250 mm, 70% in the bare soil and only 60% in the natural forest. The lowest value of Chepil index, i.e., the minimum wind erosion risk, was observed in Prosopis and Eragrostis. Additionally, under natural forest the MWD was the largest, intermediate in Eragrostis and the lowest was measured in the bare soil and rye. The OC content varied with aggregate sizes and land uses, but those values were not consistent with the land use. Only Eragrostis presented a consistent positive relationship between OC and aggregate size. In this sandy loam soil the aggregation seems to depend on CPOM and FPOM. Specifically, the large aggregates in Eragrostis depend on CPOM and under natural forest, on FPOM. However, conventional plowing removed all the influence of particulate organic matter. Our finding is confirming the superior soil quality developed under natural forest and the need to replace the aggressive moldboard tillage system for no tillage due to the physical and biological degradation caused on this soil after 20 years of cropping.

Keywords: Aggregate size distribution, Mean weight diameter, Prosopis caldenia, Eragrostis curvula, Rye Secale cereal, Erosion

How to Cite: Rienzi, E., Marchi, A., & Rodriguez, G. (2018). Aggregate Size, Particulate and Total Organic Carbon in Different Land Uses on a Sandy Loam Soil Exposed to Wind Erosion. Advances in Agricultural Science6(3), 95-111.

1. Introduction

In agroecosystems of semiarid regions, the condition of soil surface plays a very important role in soil sustainability and the control of erosive processes. In the semiarid region of Argentina, near of one-third of the land under cropland present an estimated soil loss from 10 to 50 Mg ha-1 year-1 due to intensive tillage or improper cropping system. The replacement of natural coverture by annual crops exposes the aggregates in soil surface to the erosive agents and the inappropriate management decreases the possibility of developing a strong aggregation status.  At difference with humid zones the critical soil depth affected for wind erosion is in fact the first centimeters in depth as confirmed in wind tunnel experiments (Xing and Guo, 2009; Kohake et al., 2010; Asensio et al., 2016). The soil state, characterized for the aggregation status, the size distribution and the soil organic matter are critical in those environments to prevent soil degradation. In general, sandy loam soils are highly erodible because they dry quickly, present weak aggregates and contain a great part of particles in the most erodible fraction (aggregates from 0.080 to 0.200 mm in diameter) (Buschiazzo and Funk, 2015).The aggregate size distribution and the dry aggregates stability in soil surface are the most important soil characteristics to explain the susceptibility to wind erosion and represent a reliable way to assess the degree of physical degradation in soils under tillage. Hence, to maintain or increase the amount of macro aggregates in soil surface is a key to reduce soil erodibility and enhance soil quality (Zobeck and Popham, 1990). Soil management could increase soil macro aggregates formation (Elliott, 1986; Beare et at, 199­4; Dexter, 2008); this is why quantifying surface conditions by determining the mean weight diameter index (MWD) could be a convenient way to assess cropping systems in this environment (Pikul et al., 2006; Zobeck et al, 2003).The selection of tillage practices, crop sequences and residue management are extremely important to keep the goals of soil sustainability in semiarid regions. Knowing the relationships between organic residues and soil ag­gre­gates would also allow understanding how different soil managements contribute to accumulate soil orga­nic carbon and promote aggregation in sandy loam soils. Golchin et al. (1995) commented that stable aggregates are more dependable on particulate organic matter (POM) than the total organic carbon (OC) content. They observed that the amount of substrates and the turnover are the key to generate stable aggregates. In three over five soils, the occluded POM and clay-associated organic materials were found to be a relatively constant percentage of soil OC regardless of tillage practices, supporting the fact that the residue input is one of the most important aggregation factors in agroecosystems (Golchin et al., 1995; Asensio et al., 2016; Zhao et al., 2018). The amount of substrates, the residues quality and the turnover are the key to generate stable aggregates (Golchin et al., 1995). In three over five soils, the occluded POM and clay-associated organic materials were found to be a relatively constant percentage of soil C regardless of tillage practices, supporting the fact that the C input is one of the most important aggregation factors in soils. In this sense, Pugget et al. (2000) found differences in soil aggregates that were associated to the residues quality due to the presence of C3-C4 crops. The C3 plants are cool-season whereas the C4 plants are warm-season, with differences manifested in the leaf anatomy and enzymes used to carry out photosynthesis. Briefly, these characteristics are important with respect to their nitrogen and water-use efficiency, forage quality, and seasonal production profile, which could determine several differences in aggregation. Hence, assessing aggregate size distribution, MWD, POM and OC appears to be suitable as soil quality indexes for quantifying differences between management systems in soils susceptible to wind erosion, which are highly dependable to a stable soil structure.

 In the study area  where natural fo­rest of Prosopis caldenia (L), the farmers  usually install cross wind trap strips (NRCS, 2007) using Eragrostis curvula (Schrad Nees) to serve like protective barrier against wind erosion on commercial crops like rye Secale cereale (L). After several periods with crops, the position with cross strips are switched, i.e, pastures are replaced by crops and vice versa. This management system was adopted for farmers in order not only to increase their incomes and protect the resource but also as forages in cattle farms. However, the sustainability of such management must be analyzed and the state of soil must be assessed in order to accomplish soil conservation objectives (Carter, 2002). The goal of this work was to assess the effect of Eragrostis curvula when used as a cross strip and rye crops on selected soil physics and biological variables compared to a pristine forest environment. Patches of bare soil were used to illustrate the consequences of maintain uncovered sites in sandy loam soils, under the assumption that this condition represent the worst scenario in this environment. In order to achieve the objectives, we test the assumption of a similar presence of large aggregates under pasture compared with a natural forest and the influence of C3 or C4 plants on the aggregation status. The aggregate size distribution associated with particulate organic matter and soil organic carbon was also analyzed to identify possible ways for improving management systems, under the assumption that the soil resilience could allow continuing the cropping activities.

2. Methods

2.1. Details of the area in study and the selected plots.

The site of study was localized in the research farm of Experimental Field of  EEA INTA, at 8 km from Villa Mercedes, San Luis province, Argentina (33°40’11.34″S, 65°22’13.32″W) (Figure 1) (Google maps, 2018; redesigned with QGIS 3.0.2). The climate of the area is considered arid to semiarid with 550 mm annual mean precipitation and 8 °C mean temperature in July and 23 °C in January (INTA, 2000).  The wind direction is NE predominantly with velocities from 45 to 56 km h-1 measured at 10 m at the end of the winter (South Hemisphere) (INTA, 2000).  The experimental field poses plots with commonly farmer-adopted land uses. We selected 4 contrasting plots (30 ha each, proximately, Figure 1): a) natural forest of Prosopis caldenia (Burk.)(C3 plant), b) natural patches within forest, which present less than 10 % of soil covered with native grass, basically Piptochaetium napostaense (Speg.) Hackel, Setaria leucopila (Lam-Scrib.) K. Schuman and Setaria leiantha Hackel, labeled as “bare soil” c) cross stripes of Eragrostis curvula (Schrad Nees)(C4 plant), and d) plots with rye Secale cereale (L.) (C3 plant),  planting with moldboard plowing and disc.  Plots c) and d) were under the same condition for the last sixteen years, when the usual in the region is to rotate it every 8-10 years depending on economic conditions. The plots have no previous record of cropping activities. No grazing was allowed in the cross strips but they were mechanical harvested for hay production. In the Secale cereale plot, the harvest combine equipment proceeds to distribute the residue through a spreading mechanism. Normally, the amount of residue left in the field reach from 4 to 5 Mg ha-1yr-1 and the plot is under fallow for six month until the new seedbed preparation.  No grazing was allowed in this plot. The site presents a Cramer sandy loam Typic Ustipsament soil (Soil Survey Staff, 1999) or Eutric Cambisol (FAO WRB 2014, Gardi et al, 2014 ) with 67.9 g kg-1 clay, 208.9 g kg-1 silt and 723.2 g kg-1 sand content. The Cramer sandy loam soil described originally in the 80’s in this region presented 5 g kg-1of soil organic carbon content (OC) at 10 cm of soil depth with a C: N ratio of 10, a pH of 7.4 and 0.16% Ca C03 (INTA, 2000).


2.2. Field sampling and soil analysis.

The soil samples were taken from the first 8 cm of soil depth for the four treatments in 3 replicated plots (12 plots total) (Figure 1) to evaluate the condition of aggregation, wind erosion risk, OC content and particulate organic matter. To evaluate the aggregate size distribution and the aggregate stability, the Chepil rotary mechanical sieve was used. Briefly, the device consists of a concentric cylindrical set of sieves in a 4° degree inclined axis (Chepil, 1952). When rotating at 7 rpm, the device exposes the aggregates to a rupture-collision force, thus obtaining the size distribution and the mechanical stability of the aggregates (Chepil, 1952). The abrasion effect is considered more representative than the wet sieving of the forces suffered for aggregates exposed to wind in arid environment (Kohake et al., 2010; Hong et al., 2014). Triplicated

Figure 1. Map of the area used for the study. The rectangle in the map indicates the site where the fields are located. In the detailed sector, the labels indicate A) Fields with Secale cereale and cross strips with Eragrostis curvula. Numbers 1 to 3 identify the plots sampled. B) Fields with Prosopis caldenia and the bare soil patches. Numbers 4 to 6 identify the plots sampled.


soil samples (2 kg each) were used, totalizing 6 kg of soil processed for each plot. Aggregates size distribution was analyzed through a battery of sieves from 8, 4.8, 3, 2, 1, 0.500 and 0.250 up to 0 mm that correspond to the standard procedure for the Agriculture Research Service (ARS- USDA) (Elliot et al., 1989). The soil remaining on each sieve was collected and weighed. The mean weight diameter (MWD) was used to express dry aggregate size distribution (Youker and McGuiness, 1957).

Where: xi is the mean diameter of the size fraction, and wi is the proportion of total sample weight retained on each sieve. Because the mechanical energy involved in the procedure, the remained amount of aggregates on each sieve was considered stable enough to resist wind erosion. Macro to micro aggregate ratio was calculated from the original data by mass and expressed as percentage. The Chepil index, evaluated as the amount of aggregates smaller than 0.840 mm, (Colazo and Buschiazzo, 2010; Kohake et al., 2010), was calculated from the same procedure used to measured ASD, adding the proper sieve. In accordance to Chepil and Woodruff (1963), the turbulence force that drags and lifts particles producing saltation is substantially resisted by particle sizes larger than 0.840 mm. Following the assumption that this particle size defines the boundary of soil erodibility, the larger the value of Chepil index, the greater the risk of wind erosion.

The OC through Walkley Black procedure was recorded for the whole soil samples. Additionally, OC was measured in four diameter size classes of aggregates: a) 0.250, b) 0.500, c) 2.5 and d) 4.8 mm. To determine particulate organic matter (POM), triplicate sub samples (100 g) for the four classes of aggregates previously selected were dispersed overnight with 0.1 M Na4P2O7 diluted to 1 L with distilled water. The material dispersed was washed over sieves in a gently flow of water. Two sieves (0.250 and 0.05 mm in size) were used to separate the organic resi­dues released from the soil aggregates, dried at 55°C and weighted. The material thus collected was placed in porcelain cups and burned in a muffle at 500 °C and weighed. The difference by mass from the material retained on the 0.250 mm sieve was considered a coarse particulate organic matter (CPOM). The difference calculated from the material retained on the 0.050 mm sieve was labeled as fine particulate organic matter (FPOM) (Cambardella and Elliott, 1992; Franzluebbers and Stuedemann, 2002; Carter, 2003).


2.3. Statistical analysis

The PROC MIXED procedure (SAS Institute Inc, Cary, NC) was used for an ANOVA with interactions in an array of 4 land uses x 4 classes of aggregates size x 3 repetitions to assess significant main effects and interactions and test of means (Neter et al., 1996). The MIXED procedure perform mixed model analysis of variance with repeated measures via covariance structure modeling, allowing to explain the interactions from different source of variations. Canonical discriminant analysis followed by a cluster analysis (PROC CANDISC, ACECLUS and PROC CLUSTER, SAS Institute Inc, Cary, NC) was performed to identify associations with different sources of variability. Canonical discriminant analysis (CANDISC) is a dimension-reduction technique similar to canonical correlation that derives linear combinations of the variables (named canonical variables) to summarize between-class variation. The maximum multiple correlations are found in the first canonical correlation, the second includes the part of the variability not included in the first one and so on. The canonical variables can show differences among the classes, by maximizing the correlation within groups (SAS, 2010). The approximate covariance estimation for clustering procedure (ACECLUS) is used for preprocessing data when neither cluster membership nor the number of clusters is known. It is useful to detect natural clusters regardless of whether some variables have more influence than others. Basically, the procedure obtains estimates of the pooled within-cluster covariance matrix using pairwise differences between observations, not between observations and means (Art et al., 1982). The preprocessing data obtained from ACECLUS was subsequently analyzed with PROC CLUSTER using the Ward’s method (Ward, 1963).


3. Results and Discussion

3.1. Effect on physical properties: Aggregate size distribution (ASD), mean weight diameter (MWD), Chepil index and macro/micro aggregate ratio

The shapes of the ASD curves are evidencing that the land uses determine differences in aggregate formation (Figure 2). Notice that the soil under rye and bare soil patches present more than 80% of the ASD composes by aggregates smaller than 0.250 mm, which is considered the limit of the micro aggregates domain (Six et al., 1999).The soil under Eragrostis present around of 70 % of the ASD composes by aggregates smaller than 0.250 mm in size. On the contrary, the natural forest shows only 60% of the ASD integrated by those aggregates.

It is also noticeable that the natural forest presented a large variety of aggregate sizes compared to the other situations. For example, to reach 80 % of ASD, it was necessary to involve four classes of aggregates compared to only one class under rye or bare soil (Figure 2). This means that under undisturbed forest the aggregation process could be more elaborated, determining a variety of sizes compared to other treatments.

Figure 2. Aggregate size distribution measured (ASD) from different land uses in the Cramer sandy loam soil.

Table 1. Selected soil physical variables measured on different plots in the Cramer sandy loam soil.

Land use Chepil index Macroaggregates from 2 to 4 mm in size (%) Macroaggregates from 4.8 to 8 mm in size (%) Macro to microaggregates ratio (%) MWD


Eragrostis curvula 74.1a 3.2 16.1 43 1.37a
Rye Secale cereal 83.5b 2.4 9.7 25 0.91b
Prosopis caldenia 69.8a 4.7 15.1 69 1.46a
Patches of bare soil 87.6b 1.9 3.9 23 0.57c
Abbreviations: MWD=mean weight diameter; a,b=different letters in same column mean significant difference with p<0.05 according to the LSD test.



Table 2. Probability results from an ANOVA analysis for land use and aggregate sizes as main factors and their interactions in the Cramer sandy loam soil.

Source df CPOM FPOM OC
A Land Use 3 **** **** ****
B Agregate size 3 **** **** ****
A*B 9 **** **** ****
Abbreviation: df= degree of freedom; CPOM = coarse particulate organic matter; FPOM = fine particulate organic matter; OC = organic carbon content; ****= significant with α ≤ 0.001.


In sandy loams Typic Torriorthents and Typic Torrifluvents soils under native shrubs, Canton et al. (2009) measured 7 to 10.2 % of aggregates between 4 to 8 mm in diameter and 19 to 21% between 2 to 4 mm.  In our soils, the same range from 4 to 8 mm reaches a percentage of 16.1% in the Eragrostis plot and 15.1% in the Prosopis plot. In the rye plot was 9.7% and 3.9 % in the bare soil (Table 1). However, the range of aggregates between 2 to 4 mm diameter sizes resulted very low in our plots with values between 1.9 to 4.7% (Table 1). Hence, our conditions produce a bias to large aggregates. Canton et al. (2009) observed that the soil samples close to perennials plants presented the largest proportion of bigger aggregates. Despite the soil condition and the different environment our data are confirming their findings. We notice that the soils in the Canton et al. (2009) study presented 24 % silt+clay compared to our soils with 27.68 % silt + clay, but this difference cannot explain our results.

As a consequence of the ASD composition, under natural forest the MWD was the largest while the lowest was measured in the bare soil patches. The MWD measured under rye was also low and in the Eragrostis plot the value was intermediate (Table 1). Zhao et al. (2017) comparing different land uses found that the forest presented larger MWD than cropland (1.9 vs. 0.9 mm) The soils used in this study were Inceptisols  and Ultisols with silty clay or  silty clay loam  textures  and  Fe oxides as agents of aggregation, very different to our soils. However, must be noted that the presence of the natural forest always determine large aggregates even in different environment and soil textures.

The aggregation status in rye plot under moldboard plowing contrary to expected was low but not negligible, and likely depend on the residue management (5 Mg h-1).  In according to several authors, residue management seems to have decisive influence on the MWD. For example, Colazo and Buschiazzo (2010) found in sandy soils of 21.5% of silt +clay that the large clods created under tillage in wheat/corn/sunflower rotations were sufficient to control wind erosion.

Baumhard et al. (2012) and Abiven et al. (2007) found more relevant the crop rotation on MWD than the tillage effect, due to the greater amount of organic matter added at harvesting. However, Alvaro Fuentes et al. (2008a) in sandy soils attributed the lack of differences observed in MWD in tillage systems to a distribution effect of clay content provoked by the mixing effect of the plough layer. In our case, no difference of clay content was found in the plowing depth in the Cramer series that can explain our results, thus the MWD reflect the real consequences of the soil use and residue management.

Plowing these sandy soils could likely contribute to create clods by increasing the formation of organo-mineral microstructures (Carter et al., 2004; Bossuyt et al., 2002; Colazo y Buschiazzo, 2010). Consequently, our data could support the statement of Waters and Oades (1991) that showed the encrustation of plant debris by mineral particles as an important mechanism in the formation of aggregates. Similarly, Cates et al. (2016) found evidence that macro aggregates are correlated with fresh biomass inputs, which could explain our observations.

To preserve residue in the field seems to have a strong effect on macro aggregates even in dry environment, as was commented by Blankinship et al. (2016). In the same sense, Blanco Canqui and Lal (2009) demonstrated how the MWD was reduced by 50% when the crop residue was systematically removed. Thus, as long as the residue is preserved after harvesting, conventional plowing could maintain some degree of aggregation in this sandy loam soil.

On the contrary, the high MWD values observed under Eragrostis are expected in accordance to the roots distribution and perennial condition of pasture. The slow decomposition rate of the roots provided short-term benefits to the improvement of soil structure (Ball et al., 2005;Six et al., 1999; Tisdall and Oades, 1982) therefore it is known that the effect could be smaller than in a native environment.

Moreover, Bach et al. (2010) mentioned the strong influence of soil texture when analyzing the effect of pasture on MWD. They observed that in restored grasslands this effect was lower than in the native prairies, but the MWD was greater in the silty clay loam than in the loamy fine sand soil. Cates et al. (2016), states that under pastures a soil develops very stable microaggregates within macroaggregates and our data under Eragrostis is supporting their statement. The ratio of macro to micro aggregates in our study follows the order: natural forest>Eragrostis> patches with bare soil, rye (Table 1).

As expected the Chepil index is showing that the soil erodibility was high in the plots under tillage and bare soil, but it was low under natural forest and Eragrostis. The low value of Chepil index implies low risk of wind erosion, confirming the beneficial contribution of both, the natural forest and the pasture in this environment. Pikul et al. (2008) mentioned that systems with a greater fraction of material erodible by wind are denoted by a small MWD value. Our data are showing that the soil qualities under rye and under bare soil patches were compromised, judged with respect to wind erosion.

Conventional (moldboard) plowing in these sandy loam soils should be discontinued and it would be recommended to introduce NT system for cropping. Several researchers (Huang et al., 2015;Bono et al., 2008; Alvaro Fuentes et al., 2008a and b and Shukla et al.,2003) found that NT trend to generate more stable aggregates in the first centimeter of the soils, which is particularly important to control wind erosion. In addition, the soil surface is covered permanently under NT, increasing the protection in this scenario.


3.2. Total organic carbon in aggregates of different sizes.

It was observed that all the plots but rye present larger OC content than the first original description of Cramer sandy loam (5 g kg-1) as recorded in the ‘80s (INTA, 2000) (Figure 3a). The main sources of variation and their interactions resulted significant (Table 2), demonstrating that the land uses produce important consequences on the variable selected.

The highly significant interaction also demonstrates that the OC content varies in the 4 classes of aggregates. For example, the natural forest present the highest OC content with micro aggregates (p<0.05), but not at the other sizes (Figure 3b). In addition, we notice a large variability in the trend of OC content among different sizes of aggregates. In the natural forest, the OC values trend to increase until aggregates of 0.5 mm, reaching the maximum value but decrease when the aggregate size increases. In patches of bare soil, increases exponentially from 0.25 to 0.5, then the OC values decrease and present no differences between 2 and 4.8 mm size (p<0.05).

On the contrary, the plots under rye showed no trend and under Eragrostis it was found a slight lineal tendency to increase the OC (r2= 49%) when the aggregate size increases. The minimum OC was measured in micro aggregates (< 0.250 mm in size) in patches with bare soil, rye and Eragrostis, but in the natural forest the minimum OC was observed with large macro aggregates (4.8 mm in size, Figure 3b).

Madari et al. (2006) working with different tillage and forest on Typic and Rhodic Haplustox soils commented that under natural forest, the OC is largely accumulated in microaggregates. Our data indicate that the large accumulation in a forest of Prosopis occurred not only in micro aggregates but also in macro aggregates (2.5 mm) (Figure 3b).

Blanco Canqui and Lal (2007) in a humid environment on a silty loam soil under tillage observed that to accumulate residue in soil surface modify the distribution of OC in aggregates. The OC concentration in their environment significantly decreased with decrease the aggregate size following a power function (r 2> 90%). The amount of residue apparently controlled the accumulation and begins with 8 Mg ha-1 of mulch rate.

Figure 3. Total organic carbon content (OC) from different land uses in the Cramer sandy loam soil in: a) whole soil samples and b) in four classes of aggregate diameter size.



In our region, Secale cereale is expected to produce between 4 to 5 Mg ha-1 of residue rate, but a power function cannot represent the OC concentration. The reason could be the low residue input, the arid climate or the different soil texture, as was mentioned in Zhao et al. (2018). However, the behavior described by Blanco Canqui and Lal (2007) was observed in the pasture of Eragrostis (Figure 3b), where a power function could explain more than 90% of the data (model not shown).

This finding is raising a question: Does the OC concentration depend on residue or living roots in arid environments? In accordance to the results observed under Eragrostis and natural forest of Prosopis, the living roots seems to determine the difference in our study, but because only one soil texture was assessed, more evidence is necessary to elaborate a final conclusion.


3.3. The role of CPOM and FPOM on the aggregates size.

As occurred with OC, the main sources of variation and their interactions resulted significant (Table 2), demonstrating that land uses generate important consequences on the variables selected (CPOM and FPOM). In addition, the interaction between land use and aggregates indicates that not only land use but also the aggregates size shows completely different values of CPOM and FPOM. The meaning of this interaction is well represented in the Figure 4 (a and b). The amount of CPOM shows no trend when the aggregates size increases. For example, under natural forest the amount of CPOM are not consistent with the aggregates size. Our data reveal that a high content was observed in macro aggregates of 0.500 mm, but not with other large aggregates (2.5 and 4.8 mm) (Figure 4 a).

The patches with bare soil showed a slight trend to increase CPOM with the aggregate size. Also, the plot under rye is showing the same pattern as the forest, although this plot presents very low values. Interestingly, under Eragrostis was found the highest CPOM in micro aggregates of 0.250 mm in size (Figure 4a). Puget et al. (2000) in cultivated soils with 150 to 250 g kg-1of clay content observed also that the amount of CPOM was not consistent with the aggregates size, and they concluded that depends on the characteristics of the plant (C3 or C4 metabolisms).  They assume that the differences in forage quality, and seasonal production profile could produce several effects in soil aggregation. In our study, Eragrostis is representative of a C4 plant, while native forest and rye are C3 plants. Even though are both C3 plants, they determined very different CPOM content. Hence, either the plant cycle (perennial or annual) play a role in the inconsistencies between CPOM and aggregates size or could be attributed to the tillage effect on the rye plot.

The analysis of FPOM is showing a complete different pattern.  Eragrostis plot presented the highest values at two over four classes of aggregates sizes (2.5 and 4.8 mm in sizes) (Figure 4b). The rye was the only plot with FPOM content in aggregates of 0.5 mm (Figure 4b) but showing no trend with the aggregate sizes. The natural forest present no differences between 0.25 and 2 mm size (p<0.05), and increase to reach the maximum with aggregates of 4.8 mm in size. The patches with bare soil showed no FPOM in aggregates from 0.25 to 0.5 mm and the value also increases from aggregates from 2 to 4.8 mm in size, as was observed in the natural forest.


3.4. Relationships among aggregates size, OC, CPOM and FPOM.

The canonical discriminant analysis was used to establish a relationship among the different variables measured with the aggregates size of different land use. The different tests used during the multiple analyses of variance (Wilk’s Lambda, Pillai’s Trace and Hotelling-Lawley Trace) (data not shown) were significant, fulfilling the condition that the groups were different and no homoscedasticity was observed. The proportion of the variability included in each eigenvalue and the raw canonical coefficients found are displayed in Table 3. We can observe that the first three eigenvalues almost account for the total variability found in the aggregates (99%). The results show that more than 66% of the variability was included in the first eigenvalue and 31 % in the second (Table 3). Basically, the first canonical component was a linear combination of FPOM, CPOM and OC, but only FPOM have a substantial and positive association with the aggregates as denoted by its value (Table 3). The second canonical coefficient was composed mainly for CPOM, with positive but limited association to FPOM and OC (Table 3). Notice that the third canonical coefficient, which retains only

Figure 4. Particulate organic matter measured in different land uses in the Cramer sandy loam soil: a) Coarse (CPOM, > 0.250 mm) and b) fine (FPOM, >0.050 mm).


2% of the total variation, is biased almost completely to the OC (Table 3). Due to the majority of the variability were captured in the first two canonical coefficients, it follows that the aggregation in those soils is strongly related to CPOM and FPOM. In other words, fresh and partially decomposed organic matter was responsible for the aggregation in the studied treatments. This condition is common in environment like semiarid zones and sandy soils were the fragility of the aggregation is noticeable (Fuentes et al., 2008a; and Moure et al., 2016).

Additionally, we can derive more information from this analysis. To elucidate the fraction of organic matter responsible to develop aggregation size in the

Figure 5. Clusters analyses applied after canonical discriminant analysis from the biological variables associated with the aggregate size in the Cramer sandy loam soil. Several aggregates from different treatments were labelled for comparison purposes.



Table 3. Canonical discriminant analysis. Eigenvalues variance proportion and canonical coefficients that shows the linear combinations of variables that produce the maximum correlations.

Eigenvalue Proportion Df Pr>F
168.5045 0.66 45 0.0001
 79.7538 0.31 28 0.0001
   6.1932 0.02 13 0.0001
Variable Can1 Can2 Can3
CPOM -0.851 7.592 -2.384
FPOM 8.232 2.312 0.392
OC -0.721 2.821 5.610
Abbreviations: CPOM= Coarse particulate organic matter; FPOM= Fine particulate organic matter; OC= Organic carbon; Can1=first canonical coefficient.


different treatments, the results from clusters analysis was displayed in the Figure 5. Four clusters were isolated as defined by an expected root square (ERsqr) value of 0.85, a root square (Rsqr) of 0.89 and a value for cubic clustering criterion (CCC) of 3.4 (data not shown). Ward’s procedure define pertinence based on analysis of variance, thus we observe two large clusters (clusters 2 and 4) where aggregates of different size and origin share similar relationship with FPOM, CPOM and OC. In other words, the aggregates size did not depend on specific kind of organic matter (Figure 5).

In brief, we observed that the cluster 1 is isolated in function of first canonical coefficient, which is biased to FPOM, and composed exclusively for large aggregates from the Eragrostis plot (2.5 and 4.8 mm). The cluster 2 is negatively associated to both canonical coefficients, while the cluster 3 is composed for aggregates that have null to very low positive association with both canonical coefficients.

At the extreme positive top of this cluster are located the largest aggregates (4.8 mm) from the Prosopis plot, identified in the Figure 5 for comparison purposes. The cluster 4 is defined for high positive value to the first canonical coefficient, integrated only by large aggregates (4.8 mm) from the Prosopis plot.

Moreover, notice that even though clusters 1 and 4 are both positively associated with the 2nd canonical (CPOM biased), the relationship with 1st canonical is negative in Eragrostis plot and positive in Prosopis plot. In other words, under Eragrostis, the large aggregates are defined by CPOM and FPOM, while under Prosopis caldenia the macro aggregates (0.5 mm) are directly related to the presence of FPOM.

If this is a consequence of different C3 –C4 plants behavior is not definitive but when perennial plants are involved, this finding sustains the conclusion of Puget et al (2000), i.e., C4 or C3 plants determine differences in soil aggregation and must be considered in aggregation studies.

The Secale cereale plot due to the periodic tillage and harvesting operation presents the opportunity to observe the consequences of plowing a sandy loam soil. The canonical and cluster analysis reveal that as a consequence of tillage, it was not possible to observe a pattern in their aggregates (Figure 5), i.e., the macro aggregates under rye crop lost the strong relationship with the particulate organic matter, as detected under pasture or forest.

Notice the difference position occupied for the large aggregates from Eragrostis and Prosopis plots comparing to the rye plot (Figure 5, labeled points). Given the natural forest represents the pristine condition of this sandy loam soil, the consequence of tillage was to eliminate FPOM and CPOM as a macro aggregation factor. Kravchenko et al. (2015) in a fine loamy soil found that CT reduced the POM compared to pastures, as occurred on our sandy loam soil. They attributed this finding to the presence of connected pores to the atmosphere that accelerate the POM decomposition, which is a condition generated by the tillage system.

Similar conclusion was obtained by Blanco Moure et al. (2016); CT provokes reduction of POM in their fields, with variation determined by the cropping system. Thus, the continuous use of CT in a fragile environment as we used in our study must be considered unsustainable.


4. Conclusions

The aggregate size distribution showed an important presence of aggregates smaller than 0.250 mm in plowing system (rye, Secale cereale) and cross strips (Eragrostis curvula), which means a low efficacy to promote macro aggregation and reduce soil erodibility in this sandy loam soil. The natural forest of Prosopis caldenia developed a variety of aggregate sizes confirming the high complexity created under undisturbed environment. The values of MWD showed the largest value in Prosopis caldenia plots and intermediate values under Eragrostis and Secale cereale. The behavior of cross strips with Eragrostis curvula on this soil in promoting large aggregates was lower than expected according to its potential. After 16 years of human activity it was observed that the OC varies inconsistently with the aggregate size, and depending on the land use. The minimum OC was measured in micro aggregates of Eragrostis curvula, rye and patches with bare soil, but not under the natural forest of Prosopis caldenia. In general, the OC content present no relationship with aggregate sizes and only under Eragrostis a power function could represent this relationship. Both, the CPOM and FPOM vary among land uses and aggregate sizes, but without a definite trend. Our analysis showed that the aggregation in this type of soil is related to CPOM and FPOM with low dependence on total OC. Interestingly, we found evidence that the largest aggregates under Eragrostis curvula were directly related to CPOM and FPOM. In the natural forest of Prosopis caldenia, two over four classes of aggregates (0.5 and 4.8 mm in size) seem to be associated to FPOM. Yet not conclusive effect on CPOM was observed by comparing the effect of C4 or C3 plants. Our assumption was that the plant cycle plays a role in the CPOM-aggregate size relationship, while tillage affects this behavior. It was clear that to plow this sandy loam soil eliminates the action of both POM fractions as aggregation promotor. The conventional plowing system provokes low MWD, a high wind erosion risk as measured for Chepil index and a low macro to microaggregate ratio, configuring the worst scenario in this environment. This finding strongly supports the need to incorporate a less aggressive tillage system as NT to accomplish the minimum requirement for soil quality. The pasture of Eragrostis curvula used as cross strips demonstrates an acceptable performance by being responsible for an intermediate MWD value, a high macro to micro aggregate ratio and a low Chepil index. The relative high percentage of particles smaller than 0.250 measured in the ASD cannot affect the soil erodibility because the surface is permanently under cover. However, in order to integrate this cross strips in a cropping system, must be necessary not discontinue the soil cover by using chemical fallow or a roller-crimper before started a crop sequence under NT.



The authors want to express their gratitude for the help and support received from the Ing Felix Fernandez, Director of the Management and Soil Conservation laboratory, Faculty of Agronomy UBA. The excellent suggestions and comments made for the anonymous reviewers must be recognized and highly appreciated.



Abiven S, Menasseri S, Angers DA & Leterme P. (2007). Dynamics of aggregate stability and biological binding agents during decomposition of organic materials. European Journal of Soil Science. 58(1): 239-247.

Alvaro-Fuentes J, Arrue JL, Cantero-Martinez C & Lopez MV. (2008a). Aggregate breakdown during tillage in a Mediterranean loamy soil. Soil & Tillage Research. 101(1-2): 62-68.

Alvaro-Fuentes J, Arrue JL, Gracia R & Lopez MV. (2008b). Tillage and cropping intensification effects on soil aggregation: Temporal dynamics and controlling factors under semiarid conditions. Geoderma. 145(3-4): 390-396.

Alvaro-Fuentes J, Lopez MV, Cantero-Martinez C & Arrue JL. (2008c). Tillage effects on in Mediterranean soil organic carbon fractions dryland agroecosystems. Soil Science Society of America Journal. 72(2): 541-547.

Asensio C, Lozano FJ, Gallardo P & Gimenez A. (2016). Soil wind erosion in ecological olive trees in the Tabernas desert (southeastern Spain): a wind tunnel experiment. Solid Earth. 7(4): 1233-1242.

Bach EM, Baer SG, Meyer CK & Six J. (2010). Soil texture affects soil microbial and structural recovery during grassland restoration. Soil Biology & Biochemistry. 42(12): 2182-2191.

Ball BC, Bingham I, Rees RM, Watson CA & Litterick A. (2005). The role of crop rotations in determining soil structure and crop growth conditions. Canadian Journal of Soil Science. 85(5): 557-577.

Baumhardt RL, Johnson GL & Schwartz RC. (2012). Residue and Long-Term Tillage and Crop Rotation Effects on Simulated Rain Infiltration and Sediment Transport. Soil Science Society of America Journal. 76(4): 1370-1378.

Blanco-Canqui H & Lal R. (2007). Soil structure and organic carbon relationships following 10 years of wheat straw management in no-till. Soil & Tillage Research. 95(1-2): 240-254.

Blanco-Canqui H & Lal R. (2009). Crop Residue Removal Impacts on Soil Productivity and Environmental Quality. Critical Reviews in Plant Sciences. 28(3): 139-163.

Blanco-Moure N, Gracia R, Bielsa AC & Lopez MV. (2016). Soil organic matter fractions as affected by tillage and soil texture under semiarid Mediterranean conditions. Soil & Tillage Research. 155: 381-389.

Blankinship JC, Fonte SJ, Six J & Schimel JP. (2016). Plant versus microbial controls on soil aggregate stability in a seasonally dry ecosystem. Geoderma. 272: 39-50.

Bono A, Alvarez R, Buschiazzo DE & Cantet RJC. (2008). Tillage effects on soil carbon balance in a semiarid agroecosystem. Soil Science Society of America Journal. 72(4): 1140-1149.

Bossuyt H, Six J & Hendrix PF. (2002). Aggregate-protected carbon in no-tillage and conventional tillage agroecosystems using carbon-14 labeled plant residue. Soil Science Society of America Journal. 66(6): 1965-1973.

Buschiazzo DE & Funk R. (2015). Wind Erosion of Agricultural Soils and the Carbon Cycle. In: Banwart SA, Noellemeyer, E. and Milne, E., editor. Soil Carbon: Science, Management and Policy forMultiple Benefits CAB International. p. 161-168.

Cambardella CA & Elliott ET. (1992). Particulate soil organic-matter changes across a grassland cultivation sequence. Soil Science Society of America Journal. 56(3): 777-783.

Canton Y, Sole-Benet A, Asensio C, Chamizo S & Puigdefabregas J. (2009). Aggregate stability in range sandy loam soils Relationships with runoff and erosion. Catena. 77(3): 192-199.

Carter MR. (2002) Published. Soil quality for sustainable land management: Organic matter and aggregation interactions that maintain soil functions. Amer Soc Agronomy. p. 38-47.

Carter MR. (2004). Researching structural complexity in agricultural soils. Soil & Tillage Research. 79(1): 1-6.

Carter MR, Angers DA, Gregorich EG & Bolinder MA. (2003). Characterizing organic matter retention for surface soils in eastern Canada using density and particle size fractions. Canadian Journal of Soil Science. 83(1): 11-23.

Cates AM, Ruark MD, Hedtcke JL & Posner JL. (2016). Long-term tillage, rotation and perennialization effects on particulate and aggregate soil organic matter. Soil & Tillage Research. 155: 371-380.

Chepil WS. (1952). Improved rotary sieve for measuring state and stability of dry soil structure. Soil Science Society Proceedings 16 (2), 113-117.

Chepil WS & Woodruff NP. (1963).The physics of wind erosion and its control. Advance in Agronomy 15, 211-302.

Colazo JC & Buschiazzo DE. (2010). Soil dry aggregate stability and wind erodible fraction in a semiarid environment of Argentina. Geoderma. 159(1-2): 228-236.

Dexter AR. (2008). Complexed organic matter controls soil physical properties. Geoderma. 144: 620-627.

Elliot ET. (1986). Aggregate structure and carbon, nitrogen and phosphorus in native and cultivated soils. . Soil Science Society of America Journal. 50: 627-633.

Elliot, W.J., Liebenow, A.A., Laflen, J.M., Kohl, K.D., 1989. A compendium of soil erodibil-

ity data from WEPP cropland soil field erodibility experiments 1987 and 88. NSERL

Report 3, 316.

Franzluebbers AJ & Stuedemann JA. (Year) Published. Particulate and non-particulate fractions of soil organic carbon under pastures in the Southern Piedmont USA. 2002: Elsevier Sci Ltd. p. S53-S62.

Golchin A, Clarke P, Oades JM & Skjemstad JO. (1995). The effects of cultivation on the composition of organic matter and structural stability of soils. Australian Journal of Soil Research. 33(6): 975-993.

Google Maps. (2018). Royal Roads University. Retrieved from

Hong SW, Lee IB, Seo IH, Kwon KS, Kim TW, Son YH & Kim M. (2014). Measurement and prediction of soil erosion in dry field using portable wind erosion tunnel. Biosystems Engineering. 118: 68-82.

Huang MX, Liang T, Wang LQ & Zhou CH. (2015). Effects of no-tillage systems on soil physical properties and carbon sequestration under long-term wheat-maize double cropping system. Catena. 128: 195-202.

INTA Instituto Nacional de Tecnología Agropecuaria Hoja Villa Mercedes Provincia de San Luis. 2000. Estación Experimental Agropecuaria San Luis Gobierno de la Provincia de San Luis Subsecretaria de Estado de Producción y Desarrollo Agropecuario.

Kohake DJ, Hagen LJ & Skidmore EL. (2010). Wind Erodibility of Organic Soils. Soil Science Society of America Journal. 74(1): 250-257.

Kravchenko AN, Negassa WC, Guber AK & Rivers ML. (2015). Protection of soil carbon within macro-aggregates depends on intra-aggregate pore characteristics. Scientific Reports 5.

Madari BE, Reeves JB, Machado P, Guimaraes CM, Torres E & McCarty GW. (2006). Mid- and near-infrared spectroscopic assessment of soil compositional parameters and structural indices in two Ferralsols. Geoderma. 136(1-2): 245-259.

Neter J, Kutner MH, Nachtsheim CJ & Wasserman W. (1996). Applied linear regression models. 3rd ed. McGraw-Hill Chicago (Ill) USA.

NRCS Natural Resources Conservation Services. (2007). CP 24 Cross wind trap strips. USDA Washington DC.

Pikul JL, Schwartz RC, Benjamin JG, Baumhardt RL & Merrill S. (2006). Cropping system influences on soil physical properties in the Great Plains. Renewable Agriculture and Food Systems. 21(1): 15-25.

Puget P, Chenu C & Balesdent J. (2000). Dynamics of soil organic matter associated with particle-size fractions of water-stable aggregates. European Journal of Soil Science. 51(4): 595-605.

QGIS 3.0.2 (2018) Quantum Geographic System

SAS (2010). SAS/STAT®9.3 User’s Guide. SAS Institute Inc., Cary, NC, US.

Six J, Elliott ET & Paustian K. (1999). Aggregate and soil organic matter dynamics under conventional and no-tillage systems. Soil Science Society of America Journal. 63(5): 1350-1358.

Skukla MK, Lal R, Owens LB & Urikefer P. (2003). Land use and management impacts on structure and infiltration characteristics of soils in the North Appalachian region of Ohio. Soil Science. 168(3): 167-177.

Soil Survey Staff (1999). Soil taxonomy. A Basic System of Soil Classification for Making and Interpreting Soil Surveys. 2nd edition. Natural Resources Conservation Service. U.S. Department of Agriculture Handbook 436.

Tisdall JM & Oades JM. (1982). Organic matter and water stable aggregates in soil. . Journal of Soil Science. 33: 141-163.

Ward JH. (1963). Hierarchical grouping to optimize an objective function. Journal of the American Statistical Association.58, 236–244.

Waters AG & Oades JM. (1991). Organic matter in water-stable aggregates. Cambridge: Royal Society of Chemistry. p. 163-74. .

Xing M & Guo LJ. (2009). The dust emission law in the wind erosion process on soil surface. Science in China Series G-Physics Mechanics & Astronomy. 52(2): 258-269.

Zhao HL, Shar AG, Li S, Chen YL, Shi JL, Zhang XY & Tian XH. (2018). Effect of straw return mode on soil aggregation and aggregate carbon content in an annual maize-wheat double cropping system. Soil & Tillage Research. 175: 178-186.

Zhao JS, Chen S, Hu RG & Li YY. (2017). Aggregate stability and size distribution of red soils under different land uses integrally regulated by soil organic matter, and iron and aluminum oxides. Soil & Tillage Research. 167: 73-79.

Zibilske LM & Bradford JM. (2007). Soil aggregation, aggregate carbon and nitrogen, and moisture retention induced by conservation tillage. Soil Science Society of America Journal. 71(3): 793-802.

Zobeck TM & Popham TW. (1990). Dry aggregate size distribution of sandy soils as influenced by tillage and precipitation. Soil Sci. Soc. Am. J. 54 (1), 198–204.

Zobeck TM, Popham TW, Skidmore EL, Lamb JA, Merrill SD, Lindstrom MJ, Mokma DL & Yoder RE. (2003). Aggregate-mean diameter and wind-erodible soil predictions using dry aggregate-size distributions. Soil Science Society of America Journal. 67(2): 425-436.