Soil Structure under Forest and Pasture Land-Uses Affecting Compressive Behavior and Air Permeability in a Subtropical Soil

: Machinery trafﬁc and animal trampling can deform the soil and, consequently, impair soil pore functioning. This study aimed to evaluate how soil structure affects the compressibility, physical properties and air permeability of a Typic Paleudalf under forest, pasture and eucalyptus. Soil samples with preserved structure were used to determine soil physical (bulk density, porosity, degree of water saturation at 33 kPa-tension, air permeability) and mechanical properties (soil deformation, precompression stress, compressibility index). After these evaluations, each soil sample was fragmented, sieved, and the metal rings ﬁlled with structureless soil, and underwent the same determinations as the samples with preserved structure. For loads greater than the precompression stress (load greater than 200 kPa), soil with non-preserved structure had the largest deformation. An increase in bulk density decreased macropores linearly (R 2 = 0.77 and 0.87, respectively, to preserved and non-preserved soil structure) and air ﬂow exponentially. The soil with preserved structure was less susceptible to further compaction. Air ﬂow was greatest in soils with lower bulk density, microporosity and water saturation degree, and a high volume of macropores. Soil structure (preserved and non-preserved) had more signiﬁcative differences in microporosity, compressibility index, soil deformation, and bulk density at the end of the compression test.


Introduction
Soils are responsible for many processes essential to life [1], serving as a substrate to support plant growth, a reservoir of nutrients [2,3], and the site for many biological processes involving the decomposition and cycling of animal and plant compounds [4][5][6].Soil influences air and water quality through interactions with the atmosphere, and as a system for storing [7] and purifying water flowing through the soil profile [8].
Soil structural quality is essential for proper pore functioning for water and air flow and biological activity, all important for the maintenance of life.In cultivated soil, its structure is affected by machinery traffic, animal trampling and soil tillage, for example.Management practices that alter the classes of pores with larger diameters will directly affect flows of air and water in the soil [9].Soil tillage alters the mechanical strength of soil aggregates, pore continuity, and hydraulic, gas and heat fluxes [10].Soil as a three-phase system (solid-air-water) has limited resilience and sustainability and, when stress-supporting limits are exceeded, soil properties and functions are affected, particularly pore size and distribution, affecting the flux of water, gas and heat [11].Soil precompression stress and the compression index are useful indicators to apprise, respectively, soil load bearing capacity and susceptibility to compaction [12,13].When pressure applied to the soil is lower than the precompression stress, elastic (recoverable) deformation occurs in the soil, and physical properties undergo minor changes.However, with pressure greater than the load-bearing capacity of the soil, plastic deformation (nonrecoverable) occurs, and soil physical properties change considerably [12,13].
A soil with good physical quality allows infiltration, retention and availability of water to plants, streams and subsurface; responds to management and resists degradation; allows exchange of heat and gases with the atmosphere and plant roots; and enables root growth [14].Water, oxygen, temperature and root penetration resistance directly affect plant growth.These properties are affected by bulk density, aggregation, aggregate stability and pore size distribution, all indirectly related to crop growth and yield [15].Precompression stress and compressibility index, soil mechanical parameters related to machinery traffic and to animal trampling, are associated with soil structure [16][17][18] and plants [18], and their knowledge may help maintain soil structure adequate for its functioning [19].The knowledge on the transition from elastic to plastic properties and changes of function of the soil is essential to increase or at least maintain soil functions such as fluxes, root penetrability, filtering and buffering [4].
Thus, for soil physical, chemical and biological processes to contribute to improved environmental quality, soil structure must allow adequate aeration, infiltration and retention of water and exchanges of gases and heat with the atmosphere.Furthermore, field operations that involve soil rupture (e.g., by tillage) and/or machinery traffic and animal trampling can substantially change soil structure (soil loosening or compaction), modifying the conditions that determine root and plant growth and yield, and water and air flows [18,[34][35][36].Remolded soil or non-preserved structure has been used in compressibility tests, especially in engineering tests to demonstrate the soil supportability of buildings [37][38][39][40][41][42], but few studies [43] have focused on soil function.
Considering the importance of soil structure, this study aimed to evaluate the effect of soil structure on its compressibility, physical properties, and air permeability.Our hypothesis was that a loose soil has lower load bearing capacity and is more susceptible to soil compaction than a structured soil; and even with lower bulk density and greater macroporosity, the loose soil has lower air permeability due to decreased pore continuity.Therefore, this study contributes towards a better understanding of the relationship between soil structure, compressibility, and permeability.

Study Site and Treatments
The study area is located in the municipality of Butiá, in the physiographic region of the Southeast Mountain Range (Rio-Grandense Shield) of the Rio Grande do Sul State, southern Brazil, with geographic coordinates of 30 • 06 06 south latitude and 51 • 52 18 west longitude (Figure 1).According to the Köppen system of climatic classification, the climate in the region is "Cfa" type-subtropical, humid, without drought.Based on 30 years of data (1981 to 2020), the minimum and maximum temperature of the hottest (January) and least hot (July) month is, respectively, 19.4 and 30.8 • C and 9.4 and 19.2 • C, and the rainfall varies from 99.6 mm (March) to 149.7 mm (June), with an annual average of 124.37 mm [44].The soil in the area is classified as Typic Paleudalf [45], Umbric Rhodic Acriso or "Argissolo Vermelho Distrófico" by the Brazilian Soil Classification System [47], low-activity clays, moderate A horizon (i.e., not included in other categories of A hori medium texture in the horizon A/clay in the horizon B with gravel, smooth undulated undulated relief, and the soil parent material is granite.The uses or treatments in study were in contiguous areas and were as follows: (1) Anthropized forest: forest composed by tree and shrub species with a heig approximately 4 m, used as shelter for cattle.Due to the possibility of cattle gaining a to this sampling point in the driest periods, this area was called anthropized forest; (2) Pasture: 5-y-old pasture, consisting of brachiaria brizanta (Brachiaria brizanth tercropped with Pensacola (Paspalum lourai) and clover (Trifolium sp.).The pasture installed in an area of 1200 ha under conventional tillage (plowing and harrowing) in Before the pasture, there was natural forest and pasture, and soybean intermittently (3) Eucalyptus 20: a 20-y-old Eucalyptus saligna stand, with conventional tillage to plant the stand in 1986.Before the eucalyptus, the area consisted of pasture; (4) Eucalyptus 4.5: clonal Eucalyptus saligna in a second rotation, with 4.5 years o The original planting occurred in 1993, with soil tillage in strip and a three-stem c The harvesting of eucalyptus in the first cycle, at 8.5 years of age, was performed man with a chainsaw, and the wood extraction was carried out with a Forwarder Valme with a load capacity of 18 Mg, without burning the crop residue.The traffic for the vesting of eucalyptus in the first cycle was at random, with number of passes reachin The soil in the area is classified as Typic Paleudalf [45], Umbric Rhodic Acrisol [46] or "Argissolo Vermelho Distrófico" by the Brazilian Soil Classification System [47], with low-activity clays, moderate A horizon (i.e., not included in other categories of A horizon), medium texture in the horizon A/clay in the horizon B with gravel, smooth undulated and undulated relief, and the soil parent material is granite.The uses or treatments in this study were in contiguous areas and were as follows: (1) Anthropized forest: forest composed by tree and shrub species with a height of approximately 4 m, used as shelter for cattle.Due to the possibility of cattle gaining access to this sampling point in the driest periods, this area was called anthropized forest; (2) Pasture: 5-y-old pasture, consisting of brachiaria brizanta (Brachiaria brizantha) intercropped with Pensacola (Paspalum lourai) and clover (Trifolium sp.).The pasture was installed in an area of 1200 ha under conventional tillage (plowing and harrowing) in 2001.Before the pasture, there was natural forest and pasture, and soybean intermittently; (3) Eucalyptus 20: a 20-y-old Eucalyptus saligna stand, with conventional tillage used to plant the stand in 1986.Before the eucalyptus, the area consisted of pasture; (4) Eucalyptus 4.5: clonal Eucalyptus saligna in a second rotation, with 4.5 years of age.
The original planting occurred in 1993, with soil tillage in strip and a three-stem chisel.
The harvesting of eucalyptus in the first cycle, at 8.5 years of age, was performed manually with a chainsaw, and the wood extraction was carried out with a Forwarder Valmet 890 with a load capacity of 18 Mg, without burning the crop residue.The traffic for the harvesting of eucalyptus in the first cycle was at random, with number of passes reaching up to 16.The second planting of eucalyptus was carried out between the rows in 2002.Before the first planting in 1993, the area was used for soybean and pasture.
Soil particle size distribution and total organic carbon content in soils are presented in Table 1.

Soil Sampling and Analyses
Samples of soil with preserved or undisturbed structure, as it was in the field, were collected in September 2006.For this purpose, three trenches in each use were opened and, in each trench, two samples per soil layer were collected, totaling six replicates per layer.The samples were collected in metal cylinders of 2.5 cm height and 6.1 cm diameter, in the 0.025 to 0.05 m, 0.10 to 0.125 m, and 0.20 to 0.225 m soil layers.These were saturated by capillarity and, later, positioned on a tension table at 0.60 m of water column to determine the microporosity [49].Soil macroporosity was calculated by the difference between the total porosity and the microporosity.The total porosity was calculated by the equation: where Tp is the total porosity (m 3 m −3 ), Bd is the bulk density (Mg m −3 ), and Pd is the particle density (Mg m −3 ).Soil particle density was determined by the method proposed by Gubiani et al. [50], in soil samples with non-preserved structure collected in September 2006 in three trenches within each use, in the 0.00-0.05;0.10-0.20 and 0.20-0.40m soil layers.The soil samples with preserved structure were re-saturated, equilibrated at a tension of 33 kPa using Richards pressure chambers [51] and, then, submitted to the air permeability test using an air permeameter.Permeability was calculated as: where K is the air permeability (m s −1 ), ρ is the air density in the moment of measurement (kg m −3 ), g is the acceleration of gravity (m s −2 ), ∆v is the reading on the flowmeter (m 3 ), L is the height of the cylinder (m), ∆t is the time (minutes), ∆p is the air pressure applied (hPa) and A is the area of the cylinder (m 2 ).We used ρ = 1.169 kg m −3 , g = 9.81 m s −2 , ∆t = 1 min, and ∆p = 1 hPa.Air density was calculated as: where ρ is the air density in the moment of measuring (kg m −3 ), ρ n is the standard air density (kg m −3 ), T n is the standard temperature ( • K), p is the atmospheric pressure in the measurement (mbar), p n is standard atmospheric pressure (mbar), and T is the temperature in the measurement ( • K).We used: ρ n (atmospheric pressure of 1013 mbar and temperature of 273.15After the air permeability test, the soil samples were submitted to a uniaxial compression test in the laboratory, with a five minutes application of successive static loads of 12.5, 25, 50, 100, 200, 400, 800 and 1600 kPa in the Terraload model S−450 (Durham Geo-Enterprises) consolidator, with pressure applied by compressed air.Maximum soil deformation was determined by following the methodology of Silva et al. [52], without considering pore water pressure changes during the test, since our apparatus had no such capability.Although this loading time might be considered a short interval in the multistep loading because of water pressure, as discussed in Rosa et al. [53], with the possibility of saturation and prevention, this loading time allows more than 99% of soil deformation.After the compression test, the soil samples were oven dried at 105 • C.
Before the compression test, soil bulk density (Bd) and degree of water saturation at 33 kPa matric tension (Sd) were calculated.Based on the vertical displacement measured in the laboratory by the consolidometer after the application of each load, the deformation (Def) of the soil at the end of the test was calculated.The compressibility index (Ci) and the precompression stress (Pcs) were calculated using Casagrande's method [54].Soil compression curves were plotted relating the observed bulk density to the applied pressure in the uniaxial compression test.
After performing the determinations of macroporosity, microporosity, total porosity, air permeability, bulk density and compressibility with the soil samples with preserved structure, the samples from each ring were unstructured so that the particles passed through a 2 mm mesh sieve.Then the rings were filled with their respective soil (particles > and <2 mm), suffering a slight compaction so that all the soil filled the ring, maintaining the original bulk density of the soil sample.That soil was named non-preserved structure.The samples went through the same processes (saturated by capillarity, submitted at 0.60 m tension on a tension table and at 33 kPa tension using Richards pressure chambers, and oven dried at 105 • C) and determinations (bulk density, macroporosity, microporosity, total porosity, air permeability, degree of water saturation at 33 kPa matric tension and compressibility) of the samples with preserved structure.
A completely randomized design was used, comparing samples with preserved and non-preserved structure for each soil layer and land use.The analysis of variance and the Tukey test of means were performed considering 5% significance, as well as regression analysis considering the properties evaluated.

Results
Soil macroporosity, total porosity and initial bulk density were not significantly (p > 0.05) influenced by soil structure (Tables 2 and 3), whereas microporosity was significantly (p < 0.05) influenced by soil structure (preserved and non-preserved), with an increase in the soil with non-preserved structure in the eucalyptus areas.However, this increase in microporosity was not reflected in significant differences in total porosity (p > 0.05).The unstructured, sieving and reorganization of soil particles during sample accommodation in the cylinder with non-preserved structure may have contributed to the increase in the microporosity and decrease (not statistically significative) in the macroporosity.Although the initial bulk density was equal for soil with preserved and non-preserved structure, the latter soil reached the highest values at the end of the compression test and, consequently, the largest soil deformation (Table 3, Figures 2-5).As the initial bulk density increased, there was a decrease in soil deformation, and this decrease was more pronounced in soil with preserved structure (Figure 6a).Macropores decreased as bulk density increased (R 2 = 0.77 and 0.87, respectively, for preserved and non-preserved soil structure) (Figure 6b); therefore, the soil became less compressive, i.e., lower deformation occurred (R 2 = 0.88 and 0.32, respectively, for preserved and non-preserved soil structure) (Figure 6c).With an increase in the initial bulk density, there was an increase in the range of the final bulk density between soil with preserved and non-preserved structure (R 2 = 0.46 and 0.74, respectively, for preserved and non-preserved soil structure) (Figure 6d).Although the initial bulk density was equal for soil with preserved and non-preserved structure, the latter soil reached the highest values at the end of the compression test and, consequently, the largest soil deformation (Table 3, Figures 2-5).As the initial bulk density increased, there was a decrease in soil deformation, and this decrease was more pronounced in soil with preserved structure (Figure 6a).Macropores decreased as bulk density increased (R 2 = 0.77 and 0.87, respectively, for preserved and non-preserved soil structure) (Figure 6b); therefore, the soil became less compressive, i.e., lower deformation occurred (R 2 = 0.88 and 0.32, respectively, for preserved and non-preserved soil structure) (Figure 6c).With an increase in the initial bulk density, there was an increase in the range of the final bulk density between soil with preserved and non-preserved structure (R 2 = 0.46 and 0.74, respectively, for preserved and non-preserved soil structure) (Figure 6d).Soil precompression stress was similar for both types of soil structure (p > 0.05) (Table 4), while differences between soil structure occurred at loads greater than the precompression stress, i.e., greater than 200 kPa (Figures 2-4).Soil compressibility index was affected by soil structure (p < 0.05) for forest and pasture uses, where the non-preserved structure presented highest values (Table 4), i.e., the soil was more susceptible to compaction.The   Soil precompression stress was similar for both types of soil structure (p > 0.05) (Table 4), while differences between soil structure occurred at loads greater than the precompression stress, i.e., greater than 200 kPa (Figures 2-4).Soil compressibility index was affected by soil structure (p < 0.05) for forest and pasture uses, where the non-preserved structure presented highest values (Table 4), i.e., the soil was more susceptible to compaction.The Soil precompression stress was similar for both types of soil structure (p > 0.05) (Table 4), while differences between soil structure occurred at loads greater than the precompression stress, i.e., greater than 200 kPa (Figures 2-4).Soil compressibility index was affected by soil structure (p < 0.05) for forest and pasture uses, where the non-preserved structure presented highest values (Table 4), i.e., the soil was more susceptible to compaction.The increase in bulk density (R 2 = 0.79 and 0.76, respectively, for preserved and non-preserved soil structure) (Figure 7a) and the degree of water saturation (R 2 = 0.78 and 0.65, respectively, for preserved and non-preserved soil structure) (Figure 6b) was associated with a decrease in the compressibility index.Soil structure condition (preserved or not) had few influences on bulk density, macroporosity, total porosity, air permeability, and precompression stress (p > 0.05).However, for loads greater than the precompression stress (load greater than 200 kPa), the soil with non-preserved structure had greater deformation.The results show that compaction The degree of water saturation was affected by soil structure type (preserved and non-preserved) (p < 0.05) only in forest (Table 4).By decreasing macroporosity (R 2 = 0.87 and 0.91, respectively, for preserved and non-preserved soil structure) (Figure 8a) and increasing microporosity (R 2 = 0.47 and 0.71, respectively, for preserved and non-preserved soil structure) (Figure 8b), there was an increase in the degree of water saturation.Air permeability did not differ statistically (p > 0.05) between soil structure types (Table 4).Soil structure condition (preserved or not) had few influences on bulk density, macroporosity, total porosity, air permeability, and precompression stress (p > 0.05).However, for loads greater than the precompression stress (load greater than 200 kPa), the soil with non-preserved structure had greater deformation.The results show that compaction reduced macropores and air flow; as a consequence the soil experienced less deformation with further loading and was less susceptible to additional compaction.Increase in air permeability was associated with an exponential decrease in bulk density (Figure 9a), degree of water saturation (Figure 9b), microporosity (Figure 9d), and an increase in macropores (Figure 9c).This behavior shows that the air flow occurred mainly in the macropores.By increasing macroporosity (Figure 9c) and reducing the degree of water saturation (Figure 9b), more pores were available for air flow.

Discussion
We observed that soil structure (preserved and non-preserved) had significant influence, especially on microporosity, compressibility index, soil deformation, and bulk density at the end of the compression test.Increasing bulk density and degree of water saturation decreased air permeability, soil deformation, macropores and compression index Soil structure condition (preserved or not) had few influences on bulk density, macroporosity, total porosity, air permeability, and precompression stress (p > 0.05).However, for loads greater than the precompression stress (load greater than 200 kPa), the soil with non-preserved structure had greater deformation.The results show that compaction reduced macropores and air flow; as a consequence the soil experienced less deformation with further loading and was less susceptible to additional compaction.

Discussion
We observed that soil structure (preserved and non-preserved) had significant influence, especially on microporosity, compressibility index, soil deformation, and bulk density at the end of the compression test.Increasing bulk density and degree of water saturation decreased air permeability, soil deformation, macropores and compression index (Figure 10).

(c) (d)
Figure 9. Regression between physical properties for soil with preserved and non-preserved structure.

Discussion
We observed that soil structure (preserved and non-preserved) had significant influence, especially on microporosity, compressibility index, soil deformation, and bulk density at the end of the compression test.Increasing bulk density and degree of water saturation decreased air permeability, soil deformation, macropores and compression index (Figure 10).Chiseling and/or harrowing disaggregate the soil and modify the relation of mass/volume in the field [19,34,55].In our case, in the laboratory the soil mass was the same for both preserved and non-preserved structure, justifying the similar values of bulk Chiseling and/or harrowing disaggregate the soil and modify the relation of mass/volume in the field [19,34,55].In our case, in the laboratory the soil mass was the same for both preserved and non-preserved structure, justifying the similar values of bulk density and total porosity in both structural conditions.However, when the soil sample with non-preserved structure was disrupted, sieved and rearranged in the metal ring, the relation between microporosity and macroporosity was modified, with an increase in the microporosity and a decrease in the macroporosity.Pores in the soil with preserved structure were more continuous, formed by the decomposition of roots and biological activity [56,57], while pores in the soil with non-preserved structure were randomly distributed in the soil mass.
We expected greater values of precompression stress for undisturbed soil samples because of the history of loads applied by machinery traffic and animal trampling, and differences in the precompression stress values when comparing the soil structure (preserved and non-preserved), but our expectations were not confirmed.For instance, the pressure applied on the soil by forest machines and by horse-hoof can exceed 300 kPa [58].When comparing soil tillage treatments (no-tillage, chisel plow, conventional tillage), Veiga et al. [43] obtained differences using undisturbed soil samples, but less difference in the precompression stress between treatments when using remolded soil samples, and suggested that remolding soil samples eliminates the effect of age hardening and soil aggregation.
Furthermore, in our study, we did not observe significative differences in the porosity and bulk density that could differently influence the precompression stress values when comparing preserved and non-preserved soil structure.For instance, Suzuki et al. [59] demonstrated the negative and positive correlation between precompression stress with, respectively, total porosity and bulk density, and Nunes et al. [18] showed correlation of precompression stress with macroporosity and bulk density.
Our precompression stress values (31.24 to 54.40 kPa) were low, suggesting a possible effect of mineralogy and contents of gravel and sand.The studied soil was derived from granite substrate and may have contained micas in its mineralogical composition, including in the clay fraction 1:1 clay minerals, such as kaolinite dominant in the clay fraction, and feldspars in the sand and silt fractions [60].Horn and Lebert [61] stated that soil compressibility depends on soil strength, particle size distribution, type of clay mineral, content and type of organic substances, root distribution, soil bulk density, soil distribution, pore size and pore continuity in soil and simple aggregates, and water content and/or water potential.The resistance of the soil to decrease its volume when subjected to pressure is less pronounced in sandy soils and less aggregated.The increment in clay content increased the precompression stress, while the compression index decreased in denser soils and increased in clay soils and with higher void index, except in higher soil moisture [62].Sandy soils retain less water on their surfaces [63][64][65][66][67], and present greater friction resistance between soil particles, which makes it difficult for particles to move to close-together positions [68].
Other studies [69,70] have shown higher precompression stress values (77 to 183 kPa for natural forest, annual crop and pasture areas in Oxisols, and values larger than 230 kPa for non-irrigated and irrigated grazing systems in Hapludalf) than those obtained in our study (31.24 to 54.40 kPa).However, Capurro et al. [71] showed similar values (35 to 47 kPa) in Vertissol under grazing cattle to those of our study, while Horn et al. [58] found values lower than 60 kPa in Inceptisol under forest.Suzuki et al. [59] verified precompression stress values ranging from 57.09 to 232.42 kPa depending on the sampling position (wheel line, interline planting, line planting and near the peach plant) and soil depth in a peach orchard, and that values were correlated positively with bulk density and negatively with total porosity.
As shown, there is a wide range of precompression stress values in the literature, in different soil types, use and management, and in our study the lower values may have been associated to mineralogy, gravel and sand influence.Corroborating with this suggestion, in the same site of the present study, Suzuki et al. [7] verified that soil texture (sand, silt and clay) and organic matter presented greater correlation with mean weight-diameter of aggregates than with properties related to soil structure, such as porosity and bulk density.The authors also found that, even in a small amount, gravel decreased the mean weight-diameter of aggregates because its low reactivity and greater diameter hindered the formation of stable aggregates.
Although the initial bulk density was equal for soil with preserved and non-preserved structure, the latter soil reached the highest values of bulk density at the end of the compression test and, consequently, the largest soil deformation.When the internal soil strength is high, the rigidity of the pore system will be more pronounced, and the more elastic the soil will be within the recompression load range [73].As the initial bulk density increased, there was a decrease in soil deformation and macropores, making the soil less compressive.Suzuki et al. [17,59] also observed that soil with larger bulk density had smaller deformation and was less susceptible to soil compaction (larger load bearing capacity) when submitted to external loads.Powers et al. [74] observed that soil bulk density augmented with increased compaction, particularly in soils with low or moderate initial bulk density, while for soils with higher bulk density this increase was small.This behavior was attributed to the difficulty in compressing smaller pores, caused by high bulk density and pores filled with water.Soil deformation occurs when particles are able to separate and move towards each other, having their movements limited by friction and bonds between particles.Therefore, the more compact the soil and the closer the particles, the greater the friction forces, which are responsible for resistance [75].
With increasing bulk density, the soil becomes less compressive and less susceptible to compaction.Additionally, increasing the degree of water saturation increases moisture, firstly in micropores, resulting in the pore-pressure effect.Water in the micropores receives the applied load and, as the drainage of these pores is very slow, decreases soil susceptibility to compaction.Pore-pressure is the pressure exerted by water that occupies the pore space of the soil and corresponds to a force that can delay the consolidation of a cohesive soil [13].
When soil aggregates (from homogeneous via prismatic to subangular blocky and finally crumbly structure) are formed, the accessibility of particle and pore surfaces is better and maintains site productivity and biodiversity.However, soil compaction and deformation result in a platy rigid structure that is difficult for the roots to access water, ions and gas and change flux directions, and this occurs within the virgin compression stress range [4].Mentges et al. [76] mention that the type of soil structure (prismatic, massive, for example) should be considered in studies that relate elastic parameters of soil.
Soil with non-preserved structure presented the highest values of compressibility index, while the increase in bulk density and degree of water saturation decreased the compressibility index.Other authors [17,59,77] also observed that the increase in bulk density decreased the compressibility index, while Reichert et al. showed that by increasing moisture, the compressibility index increased as well [77].
Although the types of soil structure did not show statistical differences for macroporosity and air permeability, a greater permeability was expected in samples with preserved structure due to greater pore continuity associated with the activity and root decomposition, while in samples with non-preserved structure there were possible less-continuous pores due to soil disruption and rearrangement.Mechanical deep-ploughing or soil loosening result in less dense soil layers, but they deprive soils of their internal strength and destroy pore continuity and the increased sensitivity to further soil settlement [78].Even with lower total porosity than in conventional tillage, soil under no-tillage in agricultural areas generally conducts water more efficiently [79], due to bioporosity [57,80].Mando et al. [56] found the efficiency of biological pores in increasing water infiltration.
We observed that increase in air permeability was associated with a decrease in bulk density, microporosity and degree of water saturation, and an increase in macropores, demonstrating that the air flow occurred mainly in the macropores; while increasing macroporosity and reducing degree of water saturation caused more pores to be available for air flow.During compaction, the larger pores responsible for soil aeration decreased and were replaced by smaller pores, mainly pores that retain water.
This decrease in aeration porosity can be 1.5-2 times greater than the decrease in total pore space.The decrease in the oxygen diffusion coefficient, however, will depend on the geometry and stability of aeration pore channels and deformation degree during compaction [81].Horn et al. [58] found that soils with low bulk density generally have high air permeability.Soil compaction caused by a tractor changed the pore orientation that persisted two years after the traffic event in a Typic Argiudoll [82].A long-term no-tillage (around 25 years old), increased soil bulk density and reduced air-filled porosity and macroporosity, but created a continuous and stable pore organization system, which is one of the most important properties for gas transport through soils [83].
With reduction in soil moisture, there was an increase in air permeability because of a greater amount of water-free and continuity of pores available for air flow [84,85].Mentges et al. [84] also found that, in areas under no-tillage for annual crops, the increase in permeability is greater in sandy soils than in clayey ones.In an area with eucalyptus, the variation in soil saturated hydraulic conductivity and in air permeability was related to pore size distribution, especially for the >300 µm diameter pores [8].When a load is applied to the soil surface, the stress is transmitted three-dimensionally through the solid, liquid and gas phases.If air permeability in the soil is high enough to allow the immediate Soil Syst.2022, 6, 98 15 of 19 deformation of the pores filled with air, the air flow can be interrupted by changes in water content or pore-pressure [13].

Conclusions
Our results contribute towards a better understanding of the relationship between soil structure, compressibility and air permeability in a Typic Paleudalf under forest, eucalyptus and pasture, with gravel and clay content ranging from 325 to 595 g kg −1 .Total porosity and initial soil bulk density were not influenced by soil structure (preserved and non-preserved), but the relation between macroporosity and microporosity was influenced; moreover, by increasing bulk density, there was a decrease in macropores and in deformation of soil under loading.
Precompression stress was low (<54.40kPa) and similar between soil structure (condition preserved and non-preserved), refuting one of our hypotheses that preserved structure would have a larger precompression stress due to the history of loads applied by machinery traffic and animal trampling.Structure effect occurred for loads above the precompression stress (load larger than 200 kPa), where non-preserved structure presented a larger deformation.Compressibility index was highest for non-preserved soil under forest and pasture uses.With an increase in bulk density and degree of water saturation, the compressibility index decreased.
Air permeability was not affected by soil structure (preserved and non-preserved) in this soil with presence of gravel, and increase in air permeability was associated with a decrease in bulk density, microporosity and degree of water saturation, and increase of macropores, refuting our hypothesis since we expected lower air permeability in the loose soil due to the absence of pore continuity.
Soil structure (preserved and non-preserved) significantly influenced microporosity, compressibility index, soil deformation and bulk density at the end of the compression test.
In terms of farm management, both soil structures (preserved and non-preserved) require greater care in machinery traffic and animal trampling because of compaction susceptibility, especially for loads larger than the precompression stress, that can overcompact the soil, increasing bulk density and decreasing macroporosity and air flow.However, loose soil (non-preserved soil structure) requires more care, especially for loads greater than 200 kPa, when the soil becomes more compressive (greater deformation) than preserved structure.

Figure 1 .
Figure 1.Map of Brazil, with Rio Grande do Sul State shown as hatched; map of Rio Grande d State with Butiá shown as hatched; and image from Google Earth with the land uses studied.I of Google Earth dated 5 September 2005.

Figure 1 .
Figure 1.Map of Brazil, with Rio Grande do Sul State shown as hatched; map of Rio Grande do Sul State with Butiá shown as hatched; and image from Google Earth with the land uses studied.Image of Google Earth dated 5 September 2005.

Figure 2 .
Figure 2. Soil compression curve for soil with preserved and non-preserved structure in the 0.025-0.05m soil layer for four land uses.Vertical error bars for each pressure indicate the least significance difference, while vertical bars that accompany the superior and inferior axes in the figure indicate the precompression stress value for the preserved and non-preserved soil structures.

Figure 2 .Figure 3 .Figure 4 .Figure 3 .Figure 3 .Figure 4 .Figure 4 .
Figure 2. Soil compression curve for soil with preserved and non-preserved structure in the 0.025-0.05m soil layer for four land uses.Vertical error bars for each pressure indicate the least significance difference, while vertical bars that accompany the superior and inferior axes in the figure indicate the precompression stress value for the preserved and non-preserved soil structures.

Figure 5 .
Figure 5. Soil with preserved and non-preserved structure with the same bulk density before the uniaxial compression test and the differences in soil deformation and pores decrease at the end of the test.

Figure 6 .
Figure 6.Regression between physical properties for soil with preserved (P) and non-preserved (NP) structure.Macro = macroporosity; Bdi and Bdf = bulk density in the beginning and in the final of the uniaxial compression test, respectively; Def = soil deformation in the final of the uniaxial compression test.

Figure 5 .
Figure 5. Soil with preserved and non-preserved structure with the same bulk density before the uniaxial compression test and the differences in soil deformation and pores decrease at the end of the test.

Figure 5 .
Figure 5. Soil with preserved and non-preserved structure with the same bulk density before the uniaxial compression test and the differences in soil deformation and pores decrease at the end of the test.

Figure 6 .
Figure 6.Regression between physical properties for soil with preserved (P) and non-preserved (NP) structure.Macro = macroporosity; Bdi and Bdf = bulk density in the beginning and in the final of the uniaxial compression test, respectively; Def = soil deformation in the final of the uniaxial compression test.

Figure 6 .
Figure 6.Regression between physical properties for soil with preserved (P) and non-preserved (NP) structure.Macro = macroporosity; Bdi and Bdf = bulk density in the beginning and in the final of the uniaxial compression test, respectively; Def = soil deformation in the final of the uniaxial compression test.

Figure 7 .Figure 8 .
Figure 7. Regression between physical properties for soil with preserved and non-preserved structure.Bdi = bulk density in the beginning of the uniaxial compression test; Dws = degree of water saturation; Ci = compressibility index.

Figure 7 .
Figure 7. Regression between physical properties for soil with preserved and non-preserved structure.Bdi = bulk density in the beginning of the uniaxial compression test; Dws = degree of water saturation; Ci = compressibility index.

Figure 7 .Figure 8 .
Figure 7. Regression between physical properties for soil with preserved and non-preserved structure.Bdi = bulk density in the beginning of the uniaxial compression test; Dws = degree of water saturation; Ci = compressibility index.

Figure 8 .
Figure 8. Regression between physical properties for soil with preserved and non-preserved structure.Macro = macroporosity; Micro = microporosity; Dws = degree of water saturation.

Figure 9 .
Figure 9. Regression between physical properties for soil with preserved and non-preserved structure.

Figure 9 .
Figure 9. Regression between physical properties for soil with preserved and non-preserved structure.

Figure 10 .
Figure 10.Scheme showing alterations on bulk density and degree of water saturation, with changes on physical and compressive properties and air permeability.

Figure 10 .
Figure 10.Scheme showing alterations on bulk density and degree of water saturation, with changes on physical and compressive properties and air permeability.

Table 1 .
Mean values of gravel, particle size distribution and total organic carbon for the studied land uses and six soil layers.

Table 2 .
Coefficient of variation (cv) and mean values of macroporosity, microporosity and total porosity, for soil with preserved (Pres) and non-preserved (NPres) structure under different land uses and layers.
Means followed by same letters in a given line, for each physical property, do not differ statistically from each other by Tukey's test at 5% significance.

Table 3 .
Coefficient of variation (cv) and mean values of bulk density in the beginning and in the end of the uniaxial compression test, and deformation, for soil with preserved (Pres) and non-preserved (NPres) structure under different land uses and layers.
Means followed by same letters in a given line, for each physical property, do not differ statistically from each other by Tukey's test at 5% significance.

Table 4 .
Coefficient of variation (cv) and average values of precompression stress, compressibility index, degree of water saturation and air permeability for soil with preserved (Pres) and nonpreserved (NPres) structure under different land uses and layers.
Means followed by same letters in a given line, for each physical property, do not differ statistically from each other by Tukey's test at 5% significance.Soil Syst.2022, 6, x FOR PEER REVIEW 11 of 20