A agricultura moderna sofre grande impacto da mecanização devido ao aumento do tráfego e do peso do maquinário. A cana-de-açúcar (Saccharum officinarum L.) é uma cultura altamente mecanizada e, devido ao tempo de sua permanência no campo, o solo fica muito suscetível à compactação, processo dependente de diversos fatores intrínsecos e extrínsecos deste, podendo ser revertido através de técnicas, como subsolagem ou aração profunda. Porém, tais técnicas geram altos gastos de energia e capital, sendo que alguns manejos, como o controle do tráfego agrícola, contribuem para diminuição da carga exercida sobre o solo e dos custos de produção. Diversos aparelhos foram desenvolvidos para a apuração de atributos do solo diretamente no campo, visando averiguar áreas a serem manejadas por causa da compactação. A penetrometria é uma técnica muito utilizada na verificação de áreas com problemas de compactação, entretanto, existem diversos penetrômetros, que possuem modos de atuação diferenciados. Os objetivos deste estudo foram verificar a eficiência de dois penetrômetros com diferentes modos de introdução da haste metálica no solo (automático e manual) e a correlação dos índices obtidos (resistência à penetração) com outros atributos do solo (teor de água). Os resultados comprovaram que o teor de água no solo altera os resultados de resistência a penetração e a utilização de equações de regressão não linear expressam bem a relação entre estes dois atributos do solo, podendo ser utilizado para correção dos dados antes da comparação de diferentes sistemas de manejo. O ajuste das medições de resistência de acordo com o teor de água no solo diminuiu o intervalo de confiança da média na comparação de diferentes manejos, o que melhora a comparação de atributos que apresentam alta variabilidade, como a resistência do solo a penetração. Os dois métodos de utilização do penetrômetro (Manual e Automático) apresentaram diferenças significativas nas medições, de modo que houve uma tendência das medições manuais superestimarem os valores de resistência. / Modern agriculture has suffered a large impact of mechanization due to increased of traffic and the weight of the machinery. The sugarcane (Saccharum officinarum L.) crop is highly mechanized and, since it is a semi perennial crop, the soil is very susceptible to compaction, a process dependent on a number of intrinsic and extrinsic factors that can be reversed through techniques, such as deep tillage or subsoiling. However, these techniques are expensive, and some crop managements, such as traffic control, contribute to reduction of production costs and the load on the soil. Several devices have been developed for the determination of soil attributes in the field, aiming to determine areas to be managed because of compaction. The penetrometer is a widely used equipment in the verification of areas with compaction problems. However, there are several penetrometers, which have different action mode. The objective of this study was to verify the efficiency of two penetrometers with two modes of introduction of the metal rod into the soil (automatic and manual) and the correlation of obtained data (penetration resistance) with soil water content. Results showed that the soil moisture modify the penetration resistance data and the use of non-linear regression equations express well the relationship between these two soil properties and can be used for correction of the data before comparing different management systems. The modeling of soil resistance measurements with soil moisture decreased confidence interval of the mean in the comparison of different management systems, which improves the comparison of attributes that have high variability, as the soil resistance to penetration. The two penetrometers methods (manual and automatic) showed significant differences in data measurements, showing a tendency for manual measurements overestimate the values of soil resistance.

A agricultura moderna sofre grande impacto da mecanização devido ao aumento do tráfego e do peso do maquinário. A cana-de-açúcar (Saccharum officinarum L.) é uma cultura altamente mecanizada e, devido ao tempo de sua permanência no campo, o solo fica muito suscetível à compactação, processo dependente de diversos fatores intrínsecos e extrínsecos deste, podendo ser revertido através de técnicas, como subsolagem ou aração profunda. Porém, tais técnicas geram altos gastos de energia e capital, sendo que alguns manejos, como o controle do tráfego agrícola, contribuem para diminuição da carga exercida sobre o solo e dos custos de produção. Diversos aparelhos foram desenvolvidos para a apuração de atributos do solo diretamente no campo, visando averiguar áreas a serem manejadas por causa da compactação. A penetrometria é uma técnica muito utilizada na verificação de áreas com problemas de compactação, entretanto, existem diversos penetrômetros, que possuem modos de atuação diferenciados. Os objetivos deste estudo foram verificar a eficiência de dois penetrômetros com diferentes modos de introdução da haste metálica no solo (automático e manual) e a correlação dos índices obtidos (resistência à penetração) com outros atributos do solo (teor de água). Os resultados comprovaram que o teor de água no solo altera os resultados de resistência a penetração e a utilização de equações de regressão não linear expressam bem a relação entre estes dois atributos do solo, podendo ser utilizado para correção dos dados antes da comparação de diferentes sistemas de manejo. O ajuste das medições de resistência de acordo com o teor de água no solo diminuiu o intervalo de confiança da média na comparação de diferentes manejos, o que melhora a comparação de atributos que apresentam alta variabilidade, como a resistência do solo a penetração. Os dois métodos de utilização do penetrômetro (Manual e Automático) apresentaram diferenças significativas nas medições, de modo que houve uma tendência das medições manuais superestimarem os valores de resistência. / Modern agriculture has suffered a large impact of mechanization due to increased of traffic and the weight of the machinery. The sugarcane (Saccharum officinarum L.) crop is highly mechanized and, since it is a semi perennial crop, the soil is very susceptible to compaction, a process dependent on a number of intrinsic and extrinsic factors that can be reversed through techniques, such as deep tillage or subsoiling. However, these techniques are expensive, and some crop managements, such as traffic control, contribute to reduction of production costs and the load on the soil. Several devices have been developed for the determination of soil attributes in the field, aiming to determine areas to be managed because of compaction. The penetrometer is a widely used equipment in the verification of areas with compaction problems. However, there are several penetrometers, which have different action mode. The objective of this study was to verify the efficiency of two penetrometers with two modes of introduction of the metal rod into the soil (automatic and manual) and the correlation of obtained data (penetration resistance) with soil water content. Results showed that the soil moisture modify the penetration resistance data and the use of non-linear regression equations express well the relationship between these two soil properties and can be used for correction of the data before comparing different management systems. The modeling of soil resistance measurements with soil moisture decreased confidence interval of the mean in the comparison of different management systems, which improves the comparison of attributes that have high variability, as the soil resistance to penetration. The two penetrometers methods (manual and automatic) showed significant differences in data measurements, showing a tendency for manual measurements overestimate the values of soil resistance.

Vegetation can increase the resistance of slopes against landsliding. The mechanical contribution of roots to the shear strength of the soil is however difficult to measure in situ. Existing methodologies are time-consuming and therefore not suitable to quantify spatial variability on the slope. Furthermore, some existing methods, for example large in situ shear box testing, can be difficult to apply on remote sites with difficult access, e.g. steep slopes. Therefore in this thesis novel, simple and portable methods to quantify mechanical root-reinforcement in the field were developed. The ‘blade penetrometer method’, one of these new methods, was based on standard penetrometer testing but used an adapted tip shape to increase sensitivity to roots. Root depths and diameters could be quantified based on characteristics of the depth–resistance trace, both in the laboratory and in the field. Several new analytical interpretive models were developed to predict the force–displacement behaviour of roots loaded under various conditions: one assuming roots broke in tension and another assuming roots broke in pure bending. Both methods did take root–soil interaction into account. Based on these models, some roots were shown to have broken in bending and others in tension, depending on plant species and root diameter. Two new methods were developed to measure the root-reinforced soil strength directly. The ‘pin vane’ was an adaptation of a standard field shear vane, replacing the cruciform blades of the latter by prongs to minimise the effects of soil disturbance and root breakage during installation. This was one of the main problems encountered when using standard vanes in rooted soil. This ‘pin vane’ method was qualitatively shown to be able to measure the reinforcing effects of both fine and thick roots (or root analogues), both in the laboratory and the field. This method will be most useful when the strength of densely rooted surface layers is to be analysed, e.g. for erosion resistance purposes. Another newly developed shear device was the ‘corkscrew’. Rotational installation of the screw ensured minimal soil and root disturbance. During vertical extraction the root-reinforced shear strength was mobilised along the interface of the soil plug caught within the screw. The measured extraction force could be related to the reinforced soil strength. This method underestimated the strength in surface layers (especially at 0–125 mm and less so at 125–250 mm depth) but functioned well in deeper soil layers important for landsliding. Although laboratory results were promising, during in situ testing in deeper layers ( > 125 mm) local variation in soil stress, gravel content and water content, combined with low root volumes, made it difficult to accurately quantify the effect of the roots. Where the effect of roots was pronounced, e.g. in more heavily rooted surface layers (0–125 mm), significant positive trends between the measured soil strength and root strength and quantity were found. Measured reinforcements were small compared with various root-reinforcement model predictions but comparable to direct shear tests on rooted soil reported by others. These new methods, although still in the early stages of development, showed promising results for practical use in field conditions. The equipment was simple to use and portable, enabling measurements on sites with difficult accessibility. However, more work is required to validate the interpretive models developed and to calibrate these methods for a wider range of soil and root conditions.

The phenomenon called soil aggregate coalescence occurs at contact-points between aggregates and causes soil strength to increase to values that can inhibit plant root exploration and thus potential yield. During natural wetting and drying, soil aggregates appear to ‘weld’ together with little or no increase in dry bulk density. The precise reasons for this phenomenon are not understood, but it has been found to occur even in soils comprised entirely of water stable aggregates. Soil aggregate coalescence has not been widely observed and reported in soil science and yet may pose a significant risk for crops preventing them from achieving their genetic and environmental yield potentials. This project used soil penetrometer resistance and an indirect tensile-strength test to measure the early stages of aggregate coalescence and to evaluate their effects on the early growth of tomato plants. The early stages of aggregate coalescence were thought to be affected by a number of factors including: the matric suction of water during application and subsequent drainage, the overburden pressure on moist soil in the root zone, the initial size of soil aggregates prior to wetting, and the degree of sodicity of the soil aggregates. Seven mainexperiments were conducted to evaluate these factors. The matric suction during wetting of a seedbed affects the degree of aggregate slaking that occurs, and the strength of the wetted aggregates. The matric suction during draining affects the magnitude of ‘effective stresses’ that operate to retain soil structural integrity as the soil drains and dries out. An experiment was conducted to evaluate the influence of matric suction (within a range of suctions experienced in the field) on aggregate coalescence using soils of two different textures. Sieved aggregates (0.5 to 2 mm diameter) from a coarse-textured and two fine-textured (swelling) soils were packed into cylindrical rings (4.77 cm i.d., 5 cm high) and subjected to different suctions on wetting (near-saturation, and 1 kPa), and on draining (10 kPa on sintered-glass funnels, and 100 kPa on ceramic pressure plates). After one-week of drainage, penetrometer resistance was measured as a function of depth to approximately 45 mm (penetrometer had a recessedshaft, cone diameter = 2 mm, advanced at a rate of 0.3 mm/min). Tensile strength of other core-samples was measured after air-drying using an indirect “Brazilian” crushing test. For the coarse-textured soil, penetrometer resistance was significantly greater for samples wet to near-saturation, despite there being no significant increase in dry bulk density; this was not the case for the finer-textured soils, and it was difficult to distinguish the effects of variable bulk density upon drying from those of the imposed wetting treatments. In both coarse- and fine-textured soils, the tensile strength was significantly greater for samples wet to near-saturation. Thus wetting- and draining-suctions were both found to influence the degree of soil aggregate coalescence as measured by penetrometer resistance and tensile strength. Aggregate coalescence in irrigated crops is known to develop as the growing season progresses. It was therefore thought to be linked to the repeated occurrence of matric suctions that enhance the phenomenon during cycles of wetting and draining. An experiment was conducted to determine the extent of aggregate coalescence in a coarsetextured and two fine-textured (swelling clay) soils during 8 successive cycles of wetting and draining. Sieved aggregates (0.5 to 2 mm diameter) from each soil were packed into cylindrical rings (4.77 cm i.d., 5 cm high) and wetted to near saturation for 24 h. They were then drained on ceramic pressure plates to a suction of 100 kPa for one week, after which penetrometer resistance and tensile strength were measured as described above. The degree of expression of aggregate coalescence depended on soil type. For the coarse-textured soil, repeated wetting and draining significantly increased bulk density, penetrometer resistance and tensile strength. For the fine-textured soil, penetrometer resistance and bulk density did not vary significantly with repeated wetting and draining; on the contrary, there was evidence in these swelling clay soils to suggest bulk density and penetrometer resistance decreased. However, there was a progressive increase in tensile strength as cycles of wetting and draining progressed. The expansive nature of the fine-textured soil appears to have masked the development of aggregate coalescence as measured by penetrometer resistance, but its expression was very clear in measurements of tensile strength despite the reduction in bulk density with successive wetting and draining. Field observations have indicated that aggregate coalescence is first expressed at the bottom of the seedbed and that it develops progressively upward to the soil surface during the growing season. This suggests that overburden pressures may enhance the onset of the phenomenon by increasing the degree of inter-aggregate contact. Soils containing large quantities of particulate organic matter were known to resist the onset of aggregate coalescence to some extent. An experiment was conducted to evaluate the effects of soil organic matter and overburden pressures, by placing brass cylinders of various weights (equivalent to static load pressures of 0, 0.49, 1.47 and 2.47 kPa) on the top of dry soil aggregates (0.5 – 2 mm diameter) having widely different soil organic carbon contents placed in steel rings 5 cm high and 5 cm i.d. With the weights in place, the aggregates were wetted to near-saturation for 24 h and then drained on ceramic pressure plates to a suction of 100 kPa for one week. Bulk density, penetrometer resistance and tensile strength were measured when the samples were removed from the pressure plates and they all increased significantly with increasing overburden pressure in the soil with low organic matter content, but not in the soil with high organic matter content. The amount of tillage used to prepare seedbeds influences the size distribution of soil aggregates produced – that is, more tillage produces finer seedbeds. The size distribution of soil aggregates affects the number of inter-aggregate contact points and this was thought to influence the degree of aggregate coalescence that develops in a seedbed. Previous work has shown that soil organic matter reduces aggregate coalescence and so an experiment was conducted to evaluate the effects of aggregate size and organic matter on the phenomenon. For soils with high and low organic matter contents, aggregate size fractions of < 0.5, 0.5 – 2, 2 – 4, and < 4 mm were packed into soil cores (as above) and wetted to near-saturation then drained to 100 kPa suction as described above. Penetrometer resistance and tensile strength were measured and found to increase directly with the amount of fine material present in the soil cores – being greater in the < 0.5 mm and < 4 mm fractions, and being less in the 0.5 – 2 mm and 2 – 4 mm fractions. In all cases, penetrometer resistance and tensile strength were lower in the samples containing more organic matter. The rate at which soil aggregates are wetted in a seedbed affects the degree of slaking and densification that occurs, and the extent to which aggregates are wetted influences the overall strength of a seedbed. Both wetting rate and the extent of wetting were believed to influence the onset of aggregate coalescence and were thought to be affected by soil organic matter and irrigation technique. An experiment was therefore designed to separate these effects so that improvements to management could be evaluated for their greatest efficacy – that is, to determine whether management should focus on improving irrigation technique or increasing soil organic matter content, or both. The rate of wetting was controlled by spraying (or not spraying) soil aggregates (0.5 – 2 mm diameter) with polyvinyl alcohol (PVA). Samples of coarse- and fine-textured soils were packed into steel rings (as above) and subjected to different application rates of water (1, 10 and 100 mm/h) using a dripper system controlled by a peristaltic pump. Samples were brought to either a near-saturated state or to a suction of 10 kPa for 24 h, and then drained on a pressure plate at a suction of 100 kPa for one week. Measurements of penetrometer resistance and tensile strength were then made as described above. As expected, penetrometer resistance was lower in samples treated with PVA before wetting (slower wetting rates) and in samples held at a greater suction (10 kPa) after initial wetting (greater inter-aggregate strength). The effects were more pronounced in the coarse-textured soil. In both coarse- and fine-textured soils, tensile strengths increased with increasing wetting rate (greatest for 100 mm/h) and extent of wetting (greater when held at near-saturated conditions). The rate of wetting was found to be somewhat more important for promoting aggregate coalescence than the extent of wetting. Because aggregate coalescence often occurs with little or no increase in bulk density, an explanation for the increase in penetrometer resistance and tensile strength is unlikely to be explained by a large increase in the number of inter-aggregate contacts. An increase in the strength of existing points of inter-aggregate contact was therefore considered in this work. For inter-aggregate bond strengths to increase, it was hypothesized that small increases in the amount of mechanically (or spontaneously) dispersed clay particles, and subsequent deposition at inter-aggregate contact points could increase aggregate coalescence as measured by penetrometer resistance and tensile strength. An experiment was devised to manipulate the amount of spontaneously dispersed clay in coarse- and fine-textured soils of high and low organic matter content. The degree of sodicity of each soil was manipulated by varying the exchangeable sodium percentage (ESP) of soil aggregates (0.5 – 2mm) above and below a nominal threshold value of 6. Dry aggregates were then packed into steel rings (as above) and subjected to wetting near saturation, then draining to a suction of 100 kPa for one week as described above. Measurements were then taken of penetrometer resistance and tensile strength, both of which were affected by ESP in different ways. In the coarse-textured soil, sodicity enhanced aggregate slaking and dispersion, which increased bulk density. While penetrometer resistance also increased, its effect on aggregate coalescence could not be separated from a simple effect of increased bulk density. Similarly, the effect of sodicity on aggregate coalescence in the fine-textured soil was confounded by the higher water contents produced by greater swelling, which produced lower-than-expected penetrometer resistance. Measurements of tensile strength were conducted on air-dry samples, and so the confounding effects of bulk density and water content were eliminated and it was found that tensile strength increased with sodicity in both coarse- and fine-textured soils. The presence of dispersed clay was therefore implicated in the development of aggregate coalescence in this work. Finally, a preliminary evaluation of how the early stages of aggregate coalescence might affect plant growth was attempted using tomatoes (Gross lisse) as a test plant. Seeds were planted in aggregates (0.5 – 4 mm) of a coarse- or fine-textured soil packed in steel rings. These were wetted at a rate of 1 mm/h to either near-saturation (for maximum coalescence) or to a suction of 10 kPa (for minimum coalescence) and held under these conditions for 24 h. All samples were then transferred to a ceramic pressure plate for drainage to 100 kPa suction for one week. Samples were then placed in a growth-cabinet held at 20C with controlled exposure to 14 h light/day. Germination of the seeds, plant height, and number and length of roots were observed. Germination of the seeds held at near-saturation in both coarse- and fine-textured soils was delayed by 24 h compared with seeds held at 10 kPa suction. Neither the number nor the length of tomato roots differed significantly between the different treatments and soils. In the coarse-textured soil, however, the total root length over a period of 14 days was somewhat greater in the uncoalesced samples than in the coalesced samples, but this difference was not statistically significant. These results suggest that aside from delaying germination, aggregate coalescence may not have a large effect on early growth of tomato plants. However, this is not to say that detrimental effects may not be manifest at later stages of plant growth, and this certainly needs to be evaluated, particularly because aggregate coalescence increase with repeated cycles of wetting and draining. In conclusion, the primary findings of the work undertaken in this thesis were: • Rapid wetting of soil aggregates to near-saturation enhanced the onset of soil aggregate coalescence as measured by (in some cases) penetrometer resistance at a soil water suction of 100 kPa, and (in most cases) tensile strength of soil cores in the air-dry state. The rate of wetting appeared to be more important in bringing on aggregate coalescence than how wet the soil eventually became during wetting. This means reducing the rate at which irrigation water is applied to soils may reduce the onset of aggregate coalescence more effectively than controlling the total amount of water applied – though both are important. The literature reports that aggregate coalescence occurs in the field over periods of up to several months, involving multiple wetting and draining cycles, but the work here demonstrated that this can occur over much shorter time periods depending on conditions imposed. • Aggregate coalescence occurred in coarse-textured soils regardless of whether the bulk density increased during wetting and draining. In finer-textured soils, the response to wetting conditions varied and was complicated by changes in bulk density and water content due to swelling. • Small overburden pressures enhanced the onset of aggregate coalescence, but these effects were diminished in the presence of high soil organic matter contents. • Finer aggregate size distributions (which are often produced in the field by excessive tillage during seedbed preparation) invariably led to greater aggregate coalescence than coarser aggregate size distributions. The effects of aggregate size were mitigated to some extent by higher contents of soil organic matter. • Sodicity enhanced aggregate coalescence as measured by tensile strength, but when penetrometer resistance was measured in the moist state, the effects were masked to some extent by higher water contents generated by swelling and dispersion. This work suggests that tensile strength (in the air dry state) may be a more effective measure of aggregate coalescence than penetrometer resistance. • Early plant response to aggregate coalescence was not large, but the response may become magnified during later stages of growth. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1297583 / Thesis (Ph.D.) -- School of Earth and Environmental Sciences, 2007

Over the past 50 years, several advancements have been made in the research area of the mechanics of unsaturated soils. These advancements can be categorized into two groups; (i) development (or improvement) of testing techniques (or apparatus) to determine the mechanical properties of unsaturated soils and (ii) development of (numerical, empirical or semi-empirical) models to estimate the variation of mechanical properties of unsaturated soils with respect to suction based on the experimental results. Implementation of the mechanics of unsaturated soils in conventional geotechnical engineering practice, however, has been rather limited. The key reasons for the limited practical applications may be attributed to the lack of simple and reliable methods for (i) measuring soil suction in the field quickly and reliably and (ii) estimating the variation of mechanical properties of unsaturated soils with respect to suction. The main objective of this thesis research is to develop simple and reliable techniques, models or approaches that can be used in geotechnical engineering practice to estimate sol suction and the mechanical properties of unsaturated soils. This research can be categorized into three parts. In the First Part, simple techniques are proposed to estimate the suction values of as-compacted unsaturated fine-grained soils using a pocket penetrometer and a conventional tensiometer. The suction values less than 300 kPa can be estimated using a strong relationship between the compressive strength measured using a pocket penetrometer and matric suction value. The high suction values in the range of 1,200 kPa to 60,000 kPa can be estimated using the unique relationship between the initial tangent of conventional tensiometer response versus time behavior and suction value. In the Second Part, approaches or semi-empirical models are proposed to estimate the variation of mechanical properties of unsaturated soils with respect to suction, which include: - Bearing capacity of unsaturated fine-grained soils - Variation of bearing capacity of unsaturated fine-grained soils with respect to matric suction - Variation of initial tangent elastic modulus of unsaturated soils below shallow foundations with respect to matric suction - Variation of maximum shear modulus with respect to matric suction for unsaturated non-plastic sandy soils (i.e. plasticity index, Ip = 0 %) In the Third Part, approaches (or methodologies) are suggested to simulate the vertically applied stress versus surface settlement behavior of shallow foundations in unsaturated coarse-grained soils assuming elastic-perfectly plastic behavior. These methodologies are extended to simulate the stress versus settlement behavior of both model footings and in-situ plates in unsaturated coarse-grained soils. The results show that there is a reasonably good comparison between the measured values (i.e. soil suction, bearing capacity, elastic and shear modulus) and those estimated using the techniques or models proposed in this thesis research. The models (or methodologies) proposed in this thesis research are promising and encouraging for modeling studies and practicing engineers to estimate the variation of mechanical behavior of unsaturated soils with respect to matric suction.

The phenomenon called soil aggregate coalescence occurs at contact-points between aggregates and causes soil strength to increase to values that can inhibit plant root exploration and thus potential yield. During natural wetting and drying, soil aggregates appear to ‘weld’ together with little or no increase in dry bulk density. The precise reasons for this phenomenon are not understood, but it has been found to occur even in soils comprised entirely of water stable aggregates. Soil aggregate coalescence has not been widely observed and reported in soil science and yet may pose a significant risk for crops preventing them from achieving their genetic and environmental yield potentials. This project used soil penetrometer resistance and an indirect tensile-strength test to measure the early stages of aggregate coalescence and to evaluate their effects on the early growth of tomato plants. The early stages of aggregate coalescence were thought to be affected by a number of factors including: the matric suction of water during application and subsequent drainage, the overburden pressure on moist soil in the root zone, the initial size of soil aggregates prior to wetting, and the degree of sodicity of the soil aggregates. Seven mainexperiments were conducted to evaluate these factors. The matric suction during wetting of a seedbed affects the degree of aggregate slaking that occurs, and the strength of the wetted aggregates. The matric suction during draining affects the magnitude of ‘effective stresses’ that operate to retain soil structural integrity as the soil drains and dries out. An experiment was conducted to evaluate the influence of matric suction (within a range of suctions experienced in the field) on aggregate coalescence using soils of two different textures. Sieved aggregates (0.5 to 2 mm diameter) from a coarse-textured and two fine-textured (swelling) soils were packed into cylindrical rings (4.77 cm i.d., 5 cm high) and subjected to different suctions on wetting (near-saturation, and 1 kPa), and on draining (10 kPa on sintered-glass funnels, and 100 kPa on ceramic pressure plates). After one-week of drainage, penetrometer resistance was measured as a function of depth to approximately 45 mm (penetrometer had a recessedshaft, cone diameter = 2 mm, advanced at a rate of 0.3 mm/min). Tensile strength of other core-samples was measured after air-drying using an indirect “Brazilian” crushing test. For the coarse-textured soil, penetrometer resistance was significantly greater for samples wet to near-saturation, despite there being no significant increase in dry bulk density; this was not the case for the finer-textured soils, and it was difficult to distinguish the effects of variable bulk density upon drying from those of the imposed wetting treatments. In both coarse- and fine-textured soils, the tensile strength was significantly greater for samples wet to near-saturation. Thus wetting- and draining-suctions were both found to influence the degree of soil aggregate coalescence as measured by penetrometer resistance and tensile strength. Aggregate coalescence in irrigated crops is known to develop as the growing season progresses. It was therefore thought to be linked to the repeated occurrence of matric suctions that enhance the phenomenon during cycles of wetting and draining. An experiment was conducted to determine the extent of aggregate coalescence in a coarsetextured and two fine-textured (swelling clay) soils during 8 successive cycles of wetting and draining. Sieved aggregates (0.5 to 2 mm diameter) from each soil were packed into cylindrical rings (4.77 cm i.d., 5 cm high) and wetted to near saturation for 24 h. They were then drained on ceramic pressure plates to a suction of 100 kPa for one week, after which penetrometer resistance and tensile strength were measured as described above. The degree of expression of aggregate coalescence depended on soil type. For the coarse-textured soil, repeated wetting and draining significantly increased bulk density, penetrometer resistance and tensile strength. For the fine-textured soil, penetrometer resistance and bulk density did not vary significantly with repeated wetting and draining; on the contrary, there was evidence in these swelling clay soils to suggest bulk density and penetrometer resistance decreased. However, there was a progressive increase in tensile strength as cycles of wetting and draining progressed. The expansive nature of the fine-textured soil appears to have masked the development of aggregate coalescence as measured by penetrometer resistance, but its expression was very clear in measurements of tensile strength despite the reduction in bulk density with successive wetting and draining. Field observations have indicated that aggregate coalescence is first expressed at the bottom of the seedbed and that it develops progressively upward to the soil surface during the growing season. This suggests that overburden pressures may enhance the onset of the phenomenon by increasing the degree of inter-aggregate contact. Soils containing large quantities of particulate organic matter were known to resist the onset of aggregate coalescence to some extent. An experiment was conducted to evaluate the effects of soil organic matter and overburden pressures, by placing brass cylinders of various weights (equivalent to static load pressures of 0, 0.49, 1.47 and 2.47 kPa) on the top of dry soil aggregates (0.5 – 2 mm diameter) having widely different soil organic carbon contents placed in steel rings 5 cm high and 5 cm i.d. With the weights in place, the aggregates were wetted to near-saturation for 24 h and then drained on ceramic pressure plates to a suction of 100 kPa for one week. Bulk density, penetrometer resistance and tensile strength were measured when the samples were removed from the pressure plates and they all increased significantly with increasing overburden pressure in the soil with low organic matter content, but not in the soil with high organic matter content. The amount of tillage used to prepare seedbeds influences the size distribution of soil aggregates produced – that is, more tillage produces finer seedbeds. The size distribution of soil aggregates affects the number of inter-aggregate contact points and this was thought to influence the degree of aggregate coalescence that develops in a seedbed. Previous work has shown that soil organic matter reduces aggregate coalescence and so an experiment was conducted to evaluate the effects of aggregate size and organic matter on the phenomenon. For soils with high and low organic matter contents, aggregate size fractions of < 0.5, 0.5 – 2, 2 – 4, and < 4 mm were packed into soil cores (as above) and wetted to near-saturation then drained to 100 kPa suction as described above. Penetrometer resistance and tensile strength were measured and found to increase directly with the amount of fine material present in the soil cores – being greater in the < 0.5 mm and < 4 mm fractions, and being less in the 0.5 – 2 mm and 2 – 4 mm fractions. In all cases, penetrometer resistance and tensile strength were lower in the samples containing more organic matter. The rate at which soil aggregates are wetted in a seedbed affects the degree of slaking and densification that occurs, and the extent to which aggregates are wetted influences the overall strength of a seedbed. Both wetting rate and the extent of wetting were believed to influence the onset of aggregate coalescence and were thought to be affected by soil organic matter and irrigation technique. An experiment was therefore designed to separate these effects so that improvements to management could be evaluated for their greatest efficacy – that is, to determine whether management should focus on improving irrigation technique or increasing soil organic matter content, or both. The rate of wetting was controlled by spraying (or not spraying) soil aggregates (0.5 – 2 mm diameter) with polyvinyl alcohol (PVA). Samples of coarse- and fine-textured soils were packed into steel rings (as above) and subjected to different application rates of water (1, 10 and 100 mm/h) using a dripper system controlled by a peristaltic pump. Samples were brought to either a near-saturated state or to a suction of 10 kPa for 24 h, and then drained on a pressure plate at a suction of 100 kPa for one week. Measurements of penetrometer resistance and tensile strength were then made as described above. As expected, penetrometer resistance was lower in samples treated with PVA before wetting (slower wetting rates) and in samples held at a greater suction (10 kPa) after initial wetting (greater inter-aggregate strength). The effects were more pronounced in the coarse-textured soil. In both coarse- and fine-textured soils, tensile strengths increased with increasing wetting rate (greatest for 100 mm/h) and extent of wetting (greater when held at near-saturated conditions). The rate of wetting was found to be somewhat more important for promoting aggregate coalescence than the extent of wetting. Because aggregate coalescence often occurs with little or no increase in bulk density, an explanation for the increase in penetrometer resistance and tensile strength is unlikely to be explained by a large increase in the number of inter-aggregate contacts. An increase in the strength of existing points of inter-aggregate contact was therefore considered in this work. For inter-aggregate bond strengths to increase, it was hypothesized that small increases in the amount of mechanically (or spontaneously) dispersed clay particles, and subsequent deposition at inter-aggregate contact points could increase aggregate coalescence as measured by penetrometer resistance and tensile strength. An experiment was devised to manipulate the amount of spontaneously dispersed clay in coarse- and fine-textured soils of high and low organic matter content. The degree of sodicity of each soil was manipulated by varying the exchangeable sodium percentage (ESP) of soil aggregates (0.5 – 2mm) above and below a nominal threshold value of 6. Dry aggregates were then packed into steel rings (as above) and subjected to wetting near saturation, then draining to a suction of 100 kPa for one week as described above. Measurements were then taken of penetrometer resistance and tensile strength, both of which were affected by ESP in different ways. In the coarse-textured soil, sodicity enhanced aggregate slaking and dispersion, which increased bulk density. While penetrometer resistance also increased, its effect on aggregate coalescence could not be separated from a simple effect of increased bulk density. Similarly, the effect of sodicity on aggregate coalescence in the fine-textured soil was confounded by the higher water contents produced by greater swelling, which produced lower-than-expected penetrometer resistance. Measurements of tensile strength were conducted on air-dry samples, and so the confounding effects of bulk density and water content were eliminated and it was found that tensile strength increased with sodicity in both coarse- and fine-textured soils. The presence of dispersed clay was therefore implicated in the development of aggregate coalescence in this work. Finally, a preliminary evaluation of how the early stages of aggregate coalescence might affect plant growth was attempted using tomatoes (Gross lisse) as a test plant. Seeds were planted in aggregates (0.5 – 4 mm) of a coarse- or fine-textured soil packed in steel rings. These were wetted at a rate of 1 mm/h to either near-saturation (for maximum coalescence) or to a suction of 10 kPa (for minimum coalescence) and held under these conditions for 24 h. All samples were then transferred to a ceramic pressure plate for drainage to 100 kPa suction for one week. Samples were then placed in a growth-cabinet held at 20C with controlled exposure to 14 h light/day. Germination of the seeds, plant height, and number and length of roots were observed. Germination of the seeds held at near-saturation in both coarse- and fine-textured soils was delayed by 24 h compared with seeds held at 10 kPa suction. Neither the number nor the length of tomato roots differed significantly between the different treatments and soils. In the coarse-textured soil, however, the total root length over a period of 14 days was somewhat greater in the uncoalesced samples than in the coalesced samples, but this difference was not statistically significant. These results suggest that aside from delaying germination, aggregate coalescence may not have a large effect on early growth of tomato plants. However, this is not to say that detrimental effects may not be manifest at later stages of plant growth, and this certainly needs to be evaluated, particularly because aggregate coalescence increase with repeated cycles of wetting and draining. In conclusion, the primary findings of the work undertaken in this thesis were: • Rapid wetting of soil aggregates to near-saturation enhanced the onset of soil aggregate coalescence as measured by (in some cases) penetrometer resistance at a soil water suction of 100 kPa, and (in most cases) tensile strength of soil cores in the air-dry state. The rate of wetting appeared to be more important in bringing on aggregate coalescence than how wet the soil eventually became during wetting. This means reducing the rate at which irrigation water is applied to soils may reduce the onset of aggregate coalescence more effectively than controlling the total amount of water applied – though both are important. The literature reports that aggregate coalescence occurs in the field over periods of up to several months, involving multiple wetting and draining cycles, but the work here demonstrated that this can occur over much shorter time periods depending on conditions imposed. • Aggregate coalescence occurred in coarse-textured soils regardless of whether the bulk density increased during wetting and draining. In finer-textured soils, the response to wetting conditions varied and was complicated by changes in bulk density and water content due to swelling. • Small overburden pressures enhanced the onset of aggregate coalescence, but these effects were diminished in the presence of high soil organic matter contents. • Finer aggregate size distributions (which are often produced in the field by excessive tillage during seedbed preparation) invariably led to greater aggregate coalescence than coarser aggregate size distributions. The effects of aggregate size were mitigated to some extent by higher contents of soil organic matter. • Sodicity enhanced aggregate coalescence as measured by tensile strength, but when penetrometer resistance was measured in the moist state, the effects were masked to some extent by higher water contents generated by swelling and dispersion. This work suggests that tensile strength (in the air dry state) may be a more effective measure of aggregate coalescence than penetrometer resistance. • Early plant response to aggregate coalescence was not large, but the response may become magnified during later stages of growth. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1297583 / Thesis (Ph.D.) -- School of Earth and Environmental Sciences, 2007

The phenomenon called soil aggregate coalescence occurs at contact-points between aggregates and causes soil strength to increase to values that can inhibit plant root exploration and thus potential yield. During natural wetting and drying, soil aggregates appear to ‘weld’ together with little or no increase in dry bulk density. The precise reasons for this phenomenon are not understood, but it has been found to occur even in soils comprised entirely of water stable aggregates. Soil aggregate coalescence has not been widely observed and reported in soil science and yet may pose a significant risk for crops preventing them from achieving their genetic and environmental yield potentials. This project used soil penetrometer resistance and an indirect tensile-strength test to measure the early stages of aggregate coalescence and to evaluate their effects on the early growth of tomato plants. The early stages of aggregate coalescence were thought to be affected by a number of factors including: the matric suction of water during application and subsequent drainage, the overburden pressure on moist soil in the root zone, the initial size of soil aggregates prior to wetting, and the degree of sodicity of the soil aggregates. Seven mainexperiments were conducted to evaluate these factors. The matric suction during wetting of a seedbed affects the degree of aggregate slaking that occurs, and the strength of the wetted aggregates. The matric suction during draining affects the magnitude of ‘effective stresses’ that operate to retain soil structural integrity as the soil drains and dries out. An experiment was conducted to evaluate the influence of matric suction (within a range of suctions experienced in the field) on aggregate coalescence using soils of two different textures. Sieved aggregates (0.5 to 2 mm diameter) from a coarse-textured and two fine-textured (swelling) soils were packed into cylindrical rings (4.77 cm i.d., 5 cm high) and subjected to different suctions on wetting (near-saturation, and 1 kPa), and on draining (10 kPa on sintered-glass funnels, and 100 kPa on ceramic pressure plates). After one-week of drainage, penetrometer resistance was measured as a function of depth to approximately 45 mm (penetrometer had a recessedshaft, cone diameter = 2 mm, advanced at a rate of 0.3 mm/min). Tensile strength of other core-samples was measured after air-drying using an indirect “Brazilian” crushing test. For the coarse-textured soil, penetrometer resistance was significantly greater for samples wet to near-saturation, despite there being no significant increase in dry bulk density; this was not the case for the finer-textured soils, and it was difficult to distinguish the effects of variable bulk density upon drying from those of the imposed wetting treatments. In both coarse- and fine-textured soils, the tensile strength was significantly greater for samples wet to near-saturation. Thus wetting- and draining-suctions were both found to influence the degree of soil aggregate coalescence as measured by penetrometer resistance and tensile strength. Aggregate coalescence in irrigated crops is known to develop as the growing season progresses. It was therefore thought to be linked to the repeated occurrence of matric suctions that enhance the phenomenon during cycles of wetting and draining. An experiment was conducted to determine the extent of aggregate coalescence in a coarsetextured and two fine-textured (swelling clay) soils during 8 successive cycles of wetting and draining. Sieved aggregates (0.5 to 2 mm diameter) from each soil were packed into cylindrical rings (4.77 cm i.d., 5 cm high) and wetted to near saturation for 24 h. They were then drained on ceramic pressure plates to a suction of 100 kPa for one week, after which penetrometer resistance and tensile strength were measured as described above. The degree of expression of aggregate coalescence depended on soil type. For the coarse-textured soil, repeated wetting and draining significantly increased bulk density, penetrometer resistance and tensile strength. For the fine-textured soil, penetrometer resistance and bulk density did not vary significantly with repeated wetting and draining; on the contrary, there was evidence in these swelling clay soils to suggest bulk density and penetrometer resistance decreased. However, there was a progressive increase in tensile strength as cycles of wetting and draining progressed. The expansive nature of the fine-textured soil appears to have masked the development of aggregate coalescence as measured by penetrometer resistance, but its expression was very clear in measurements of tensile strength despite the reduction in bulk density with successive wetting and draining. Field observations have indicated that aggregate coalescence is first expressed at the bottom of the seedbed and that it develops progressively upward to the soil surface during the growing season. This suggests that overburden pressures may enhance the onset of the phenomenon by increasing the degree of inter-aggregate contact. Soils containing large quantities of particulate organic matter were known to resist the onset of aggregate coalescence to some extent. An experiment was conducted to evaluate the effects of soil organic matter and overburden pressures, by placing brass cylinders of various weights (equivalent to static load pressures of 0, 0.49, 1.47 and 2.47 kPa) on the top of dry soil aggregates (0.5 – 2 mm diameter) having widely different soil organic carbon contents placed in steel rings 5 cm high and 5 cm i.d. With the weights in place, the aggregates were wetted to near-saturation for 24 h and then drained on ceramic pressure plates to a suction of 100 kPa for one week. Bulk density, penetrometer resistance and tensile strength were measured when the samples were removed from the pressure plates and they all increased significantly with increasing overburden pressure in the soil with low organic matter content, but not in the soil with high organic matter content. The amount of tillage used to prepare seedbeds influences the size distribution of soil aggregates produced – that is, more tillage produces finer seedbeds. The size distribution of soil aggregates affects the number of inter-aggregate contact points and this was thought to influence the degree of aggregate coalescence that develops in a seedbed. Previous work has shown that soil organic matter reduces aggregate coalescence and so an experiment was conducted to evaluate the effects of aggregate size and organic matter on the phenomenon. For soils with high and low organic matter contents, aggregate size fractions of < 0.5, 0.5 – 2, 2 – 4, and < 4 mm were packed into soil cores (as above) and wetted to near-saturation then drained to 100 kPa suction as described above. Penetrometer resistance and tensile strength were measured and found to increase directly with the amount of fine material present in the soil cores – being greater in the < 0.5 mm and < 4 mm fractions, and being less in the 0.5 – 2 mm and 2 – 4 mm fractions. In all cases, penetrometer resistance and tensile strength were lower in the samples containing more organic matter. The rate at which soil aggregates are wetted in a seedbed affects the degree of slaking and densification that occurs, and the extent to which aggregates are wetted influences the overall strength of a seedbed. Both wetting rate and the extent of wetting were believed to influence the onset of aggregate coalescence and were thought to be affected by soil organic matter and irrigation technique. An experiment was therefore designed to separate these effects so that improvements to management could be evaluated for their greatest efficacy – that is, to determine whether management should focus on improving irrigation technique or increasing soil organic matter content, or both. The rate of wetting was controlled by spraying (or not spraying) soil aggregates (0.5 – 2 mm diameter) with polyvinyl alcohol (PVA). Samples of coarse- and fine-textured soils were packed into steel rings (as above) and subjected to different application rates of water (1, 10 and 100 mm/h) using a dripper system controlled by a peristaltic pump. Samples were brought to either a near-saturated state or to a suction of 10 kPa for 24 h, and then drained on a pressure plate at a suction of 100 kPa for one week. Measurements of penetrometer resistance and tensile strength were then made as described above. As expected, penetrometer resistance was lower in samples treated with PVA before wetting (slower wetting rates) and in samples held at a greater suction (10 kPa) after initial wetting (greater inter-aggregate strength). The effects were more pronounced in the coarse-textured soil. In both coarse- and fine-textured soils, tensile strengths increased with increasing wetting rate (greatest for 100 mm/h) and extent of wetting (greater when held at near-saturated conditions). The rate of wetting was found to be somewhat more important for promoting aggregate coalescence than the extent of wetting. Because aggregate coalescence often occurs with little or no increase in bulk density, an explanation for the increase in penetrometer resistance and tensile strength is unlikely to be explained by a large increase in the number of inter-aggregate contacts. An increase in the strength of existing points of inter-aggregate contact was therefore considered in this work. For inter-aggregate bond strengths to increase, it was hypothesized that small increases in the amount of mechanically (or spontaneously) dispersed clay particles, and subsequent deposition at inter-aggregate contact points could increase aggregate coalescence as measured by penetrometer resistance and tensile strength. An experiment was devised to manipulate the amount of spontaneously dispersed clay in coarse- and fine-textured soils of high and low organic matter content. The degree of sodicity of each soil was manipulated by varying the exchangeable sodium percentage (ESP) of soil aggregates (0.5 – 2mm) above and below a nominal threshold value of 6. Dry aggregates were then packed into steel rings (as above) and subjected to wetting near saturation, then draining to a suction of 100 kPa for one week as described above. Measurements were then taken of penetrometer resistance and tensile strength, both of which were affected by ESP in different ways. In the coarse-textured soil, sodicity enhanced aggregate slaking and dispersion, which increased bulk density. While penetrometer resistance also increased, its effect on aggregate coalescence could not be separated from a simple effect of increased bulk density. Similarly, the effect of sodicity on aggregate coalescence in the fine-textured soil was confounded by the higher water contents produced by greater swelling, which produced lower-than-expected penetrometer resistance. Measurements of tensile strength were conducted on air-dry samples, and so the confounding effects of bulk density and water content were eliminated and it was found that tensile strength increased with sodicity in both coarse- and fine-textured soils. The presence of dispersed clay was therefore implicated in the development of aggregate coalescence in this work. Finally, a preliminary evaluation of how the early stages of aggregate coalescence might affect plant growth was attempted using tomatoes (Gross lisse) as a test plant. Seeds were planted in aggregates (0.5 – 4 mm) of a coarse- or fine-textured soil packed in steel rings. These were wetted at a rate of 1 mm/h to either near-saturation (for maximum coalescence) or to a suction of 10 kPa (for minimum coalescence) and held under these conditions for 24 h. All samples were then transferred to a ceramic pressure plate for drainage to 100 kPa suction for one week. Samples were then placed in a growth-cabinet held at 20C with controlled exposure to 14 h light/day. Germination of the seeds, plant height, and number and length of roots were observed. Germination of the seeds held at near-saturation in both coarse- and fine-textured soils was delayed by 24 h compared with seeds held at 10 kPa suction. Neither the number nor the length of tomato roots differed significantly between the different treatments and soils. In the coarse-textured soil, however, the total root length over a period of 14 days was somewhat greater in the uncoalesced samples than in the coalesced samples, but this difference was not statistically significant. These results suggest that aside from delaying germination, aggregate coalescence may not have a large effect on early growth of tomato plants. However, this is not to say that detrimental effects may not be manifest at later stages of plant growth, and this certainly needs to be evaluated, particularly because aggregate coalescence increase with repeated cycles of wetting and draining. In conclusion, the primary findings of the work undertaken in this thesis were: • Rapid wetting of soil aggregates to near-saturation enhanced the onset of soil aggregate coalescence as measured by (in some cases) penetrometer resistance at a soil water suction of 100 kPa, and (in most cases) tensile strength of soil cores in the air-dry state. The rate of wetting appeared to be more important in bringing on aggregate coalescence than how wet the soil eventually became during wetting. This means reducing the rate at which irrigation water is applied to soils may reduce the onset of aggregate coalescence more effectively than controlling the total amount of water applied – though both are important. The literature reports that aggregate coalescence occurs in the field over periods of up to several months, involving multiple wetting and draining cycles, but the work here demonstrated that this can occur over much shorter time periods depending on conditions imposed. • Aggregate coalescence occurred in coarse-textured soils regardless of whether the bulk density increased during wetting and draining. In finer-textured soils, the response to wetting conditions varied and was complicated by changes in bulk density and water content due to swelling. • Small overburden pressures enhanced the onset of aggregate coalescence, but these effects were diminished in the presence of high soil organic matter contents. • Finer aggregate size distributions (which are often produced in the field by excessive tillage during seedbed preparation) invariably led to greater aggregate coalescence than coarser aggregate size distributions. The effects of aggregate size were mitigated to some extent by higher contents of soil organic matter. • Sodicity enhanced aggregate coalescence as measured by tensile strength, but when penetrometer resistance was measured in the moist state, the effects were masked to some extent by higher water contents generated by swelling and dispersion. This work suggests that tensile strength (in the air dry state) may be a more effective measure of aggregate coalescence than penetrometer resistance. • Early plant response to aggregate coalescence was not large, but the response may become magnified during later stages of growth. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1297583 / Thesis (Ph.D.) -- School of Earth and Environmental Sciences, 2007

The phenomenon called soil aggregate coalescence occurs at contact-points between aggregates and causes soil strength to increase to values that can inhibit plant root exploration and thus potential yield. During natural wetting and drying, soil aggregates appear to ‘weld’ together with little or no increase in dry bulk density. The precise reasons for this phenomenon are not understood, but it has been found to occur even in soils comprised entirely of water stable aggregates. Soil aggregate coalescence has not been widely observed and reported in soil science and yet may pose a significant risk for crops preventing them from achieving their genetic and environmental yield potentials. This project used soil penetrometer resistance and an indirect tensile-strength test to measure the early stages of aggregate coalescence and to evaluate their effects on the early growth of tomato plants. The early stages of aggregate coalescence were thought to be affected by a number of factors including: the matric suction of water during application and subsequent drainage, the overburden pressure on moist soil in the root zone, the initial size of soil aggregates prior to wetting, and the degree of sodicity of the soil aggregates. Seven mainexperiments were conducted to evaluate these factors. The matric suction during wetting of a seedbed affects the degree of aggregate slaking that occurs, and the strength of the wetted aggregates. The matric suction during draining affects the magnitude of ‘effective stresses’ that operate to retain soil structural integrity as the soil drains and dries out. An experiment was conducted to evaluate the influence of matric suction (within a range of suctions experienced in the field) on aggregate coalescence using soils of two different textures. Sieved aggregates (0.5 to 2 mm diameter) from a coarse-textured and two fine-textured (swelling) soils were packed into cylindrical rings (4.77 cm i.d., 5 cm high) and subjected to different suctions on wetting (near-saturation, and 1 kPa), and on draining (10 kPa on sintered-glass funnels, and 100 kPa on ceramic pressure plates). After one-week of drainage, penetrometer resistance was measured as a function of depth to approximately 45 mm (penetrometer had a recessedshaft, cone diameter = 2 mm, advanced at a rate of 0.3 mm/min). Tensile strength of other core-samples was measured after air-drying using an indirect “Brazilian” crushing test. For the coarse-textured soil, penetrometer resistance was significantly greater for samples wet to near-saturation, despite there being no significant increase in dry bulk density; this was not the case for the finer-textured soils, and it was difficult to distinguish the effects of variable bulk density upon drying from those of the imposed wetting treatments. In both coarse- and fine-textured soils, the tensile strength was significantly greater for samples wet to near-saturation. Thus wetting- and draining-suctions were both found to influence the degree of soil aggregate coalescence as measured by penetrometer resistance and tensile strength. Aggregate coalescence in irrigated crops is known to develop as the growing season progresses. It was therefore thought to be linked to the repeated occurrence of matric suctions that enhance the phenomenon during cycles of wetting and draining. An experiment was conducted to determine the extent of aggregate coalescence in a coarsetextured and two fine-textured (swelling clay) soils during 8 successive cycles of wetting and draining. Sieved aggregates (0.5 to 2 mm diameter) from each soil were packed into cylindrical rings (4.77 cm i.d., 5 cm high) and wetted to near saturation for 24 h. They were then drained on ceramic pressure plates to a suction of 100 kPa for one week, after which penetrometer resistance and tensile strength were measured as described above. The degree of expression of aggregate coalescence depended on soil type. For the coarse-textured soil, repeated wetting and draining significantly increased bulk density, penetrometer resistance and tensile strength. For the fine-textured soil, penetrometer resistance and bulk density did not vary significantly with repeated wetting and draining; on the contrary, there was evidence in these swelling clay soils to suggest bulk density and penetrometer resistance decreased. However, there was a progressive increase in tensile strength as cycles of wetting and draining progressed. The expansive nature of the fine-textured soil appears to have masked the development of aggregate coalescence as measured by penetrometer resistance, but its expression was very clear in measurements of tensile strength despite the reduction in bulk density with successive wetting and draining. Field observations have indicated that aggregate coalescence is first expressed at the bottom of the seedbed and that it develops progressively upward to the soil surface during the growing season. This suggests that overburden pressures may enhance the onset of the phenomenon by increasing the degree of inter-aggregate contact. Soils containing large quantities of particulate organic matter were known to resist the onset of aggregate coalescence to some extent. An experiment was conducted to evaluate the effects of soil organic matter and overburden pressures, by placing brass cylinders of various weights (equivalent to static load pressures of 0, 0.49, 1.47 and 2.47 kPa) on the top of dry soil aggregates (0.5 – 2 mm diameter) having widely different soil organic carbon contents placed in steel rings 5 cm high and 5 cm i.d. With the weights in place, the aggregates were wetted to near-saturation for 24 h and then drained on ceramic pressure plates to a suction of 100 kPa for one week. Bulk density, penetrometer resistance and tensile strength were measured when the samples were removed from the pressure plates and they all increased significantly with increasing overburden pressure in the soil with low organic matter content, but not in the soil with high organic matter content. The amount of tillage used to prepare seedbeds influences the size distribution of soil aggregates produced – that is, more tillage produces finer seedbeds. The size distribution of soil aggregates affects the number of inter-aggregate contact points and this was thought to influence the degree of aggregate coalescence that develops in a seedbed. Previous work has shown that soil organic matter reduces aggregate coalescence and so an experiment was conducted to evaluate the effects of aggregate size and organic matter on the phenomenon. For soils with high and low organic matter contents, aggregate size fractions of < 0.5, 0.5 – 2, 2 – 4, and < 4 mm were packed into soil cores (as above) and wetted to near-saturation then drained to 100 kPa suction as described above. Penetrometer resistance and tensile strength were measured and found to increase directly with the amount of fine material present in the soil cores – being greater in the < 0.5 mm and < 4 mm fractions, and being less in the 0.5 – 2 mm and 2 – 4 mm fractions. In all cases, penetrometer resistance and tensile strength were lower in the samples containing more organic matter. The rate at which soil aggregates are wetted in a seedbed affects the degree of slaking and densification that occurs, and the extent to which aggregates are wetted influences the overall strength of a seedbed. Both wetting rate and the extent of wetting were believed to influence the onset of aggregate coalescence and were thought to be affected by soil organic matter and irrigation technique. An experiment was therefore designed to separate these effects so that improvements to management could be evaluated for their greatest efficacy – that is, to determine whether management should focus on improving irrigation technique or increasing soil organic matter content, or both. The rate of wetting was controlled by spraying (or not spraying) soil aggregates (0.5 – 2 mm diameter) with polyvinyl alcohol (PVA). Samples of coarse- and fine-textured soils were packed into steel rings (as above) and subjected to different application rates of water (1, 10 and 100 mm/h) using a dripper system controlled by a peristaltic pump. Samples were brought to either a near-saturated state or to a suction of 10 kPa for 24 h, and then drained on a pressure plate at a suction of 100 kPa for one week. Measurements of penetrometer resistance and tensile strength were then made as described above. As expected, penetrometer resistance was lower in samples treated with PVA before wetting (slower wetting rates) and in samples held at a greater suction (10 kPa) after initial wetting (greater inter-aggregate strength). The effects were more pronounced in the coarse-textured soil. In both coarse- and fine-textured soils, tensile strengths increased with increasing wetting rate (greatest for 100 mm/h) and extent of wetting (greater when held at near-saturated conditions). The rate of wetting was found to be somewhat more important for promoting aggregate coalescence than the extent of wetting. Because aggregate coalescence often occurs with little or no increase in bulk density, an explanation for the increase in penetrometer resistance and tensile strength is unlikely to be explained by a large increase in the number of inter-aggregate contacts. An increase in the strength of existing points of inter-aggregate contact was therefore considered in this work. For inter-aggregate bond strengths to increase, it was hypothesized that small increases in the amount of mechanically (or spontaneously) dispersed clay particles, and subsequent deposition at inter-aggregate contact points could increase aggregate coalescence as measured by penetrometer resistance and tensile strength. An experiment was devised to manipulate the amount of spontaneously dispersed clay in coarse- and fine-textured soils of high and low organic matter content. The degree of sodicity of each soil was manipulated by varying the exchangeable sodium percentage (ESP) of soil aggregates (0.5 – 2mm) above and below a nominal threshold value of 6. Dry aggregates were then packed into steel rings (as above) and subjected to wetting near saturation, then draining to a suction of 100 kPa for one week as described above. Measurements were then taken of penetrometer resistance and tensile strength, both of which were affected by ESP in different ways. In the coarse-textured soil, sodicity enhanced aggregate slaking and dispersion, which increased bulk density. While penetrometer resistance also increased, its effect on aggregate coalescence could not be separated from a simple effect of increased bulk density. Similarly, the effect of sodicity on aggregate coalescence in the fine-textured soil was confounded by the higher water contents produced by greater swelling, which produced lower-than-expected penetrometer resistance. Measurements of tensile strength were conducted on air-dry samples, and so the confounding effects of bulk density and water content were eliminated and it was found that tensile strength increased with sodicity in both coarse- and fine-textured soils. The presence of dispersed clay was therefore implicated in the development of aggregate coalescence in this work. Finally, a preliminary evaluation of how the early stages of aggregate coalescence might affect plant growth was attempted using tomatoes (Gross lisse) as a test plant. Seeds were planted in aggregates (0.5 – 4 mm) of a coarse- or fine-textured soil packed in steel rings. These were wetted at a rate of 1 mm/h to either near-saturation (for maximum coalescence) or to a suction of 10 kPa (for minimum coalescence) and held under these conditions for 24 h. All samples were then transferred to a ceramic pressure plate for drainage to 100 kPa suction for one week. Samples were then placed in a growth-cabinet held at 20C with controlled exposure to 14 h light/day. Germination of the seeds, plant height, and number and length of roots were observed. Germination of the seeds held at near-saturation in both coarse- and fine-textured soils was delayed by 24 h compared with seeds held at 10 kPa suction. Neither the number nor the length of tomato roots differed significantly between the different treatments and soils. In the coarse-textured soil, however, the total root length over a period of 14 days was somewhat greater in the uncoalesced samples than in the coalesced samples, but this difference was not statistically significant. These results suggest that aside from delaying germination, aggregate coalescence may not have a large effect on early growth of tomato plants. However, this is not to say that detrimental effects may not be manifest at later stages of plant growth, and this certainly needs to be evaluated, particularly because aggregate coalescence increase with repeated cycles of wetting and draining. In conclusion, the primary findings of the work undertaken in this thesis were: • Rapid wetting of soil aggregates to near-saturation enhanced the onset of soil aggregate coalescence as measured by (in some cases) penetrometer resistance at a soil water suction of 100 kPa, and (in most cases) tensile strength of soil cores in the air-dry state. The rate of wetting appeared to be more important in bringing on aggregate coalescence than how wet the soil eventually became during wetting. This means reducing the rate at which irrigation water is applied to soils may reduce the onset of aggregate coalescence more effectively than controlling the total amount of water applied – though both are important. The literature reports that aggregate coalescence occurs in the field over periods of up to several months, involving multiple wetting and draining cycles, but the work here demonstrated that this can occur over much shorter time periods depending on conditions imposed. • Aggregate coalescence occurred in coarse-textured soils regardless of whether the bulk density increased during wetting and draining. In finer-textured soils, the response to wetting conditions varied and was complicated by changes in bulk density and water content due to swelling. • Small overburden pressures enhanced the onset of aggregate coalescence, but these effects were diminished in the presence of high soil organic matter contents. • Finer aggregate size distributions (which are often produced in the field by excessive tillage during seedbed preparation) invariably led to greater aggregate coalescence than coarser aggregate size distributions. The effects of aggregate size were mitigated to some extent by higher contents of soil organic matter. • Sodicity enhanced aggregate coalescence as measured by tensile strength, but when penetrometer resistance was measured in the moist state, the effects were masked to some extent by higher water contents generated by swelling and dispersion. This work suggests that tensile strength (in the air dry state) may be a more effective measure of aggregate coalescence than penetrometer resistance. • Early plant response to aggregate coalescence was not large, but the response may become magnified during later stages of growth. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1297583 / Thesis (Ph.D.) -- School of Earth and Environmental Sciences, 2007

The phenomenon called soil aggregate coalescence occurs at contact-points between aggregates and causes soil strength to increase to values that can inhibit plant root exploration and thus potential yield. During natural wetting and drying, soil aggregates appear to ‘weld’ together with little or no increase in dry bulk density. The precise reasons for this phenomenon are not understood, but it has been found to occur even in soils comprised entirely of water stable aggregates. Soil aggregate coalescence has not been widely observed and reported in soil science and yet may pose a significant risk for crops preventing them from achieving their genetic and environmental yield potentials. This project used soil penetrometer resistance and an indirect tensile-strength test to measure the early stages of aggregate coalescence and to evaluate their effects on the early growth of tomato plants. The early stages of aggregate coalescence were thought to be affected by a number of factors including: the matric suction of water during application and subsequent drainage, the overburden pressure on moist soil in the root zone, the initial size of soil aggregates prior to wetting, and the degree of sodicity of the soil aggregates. Seven mainexperiments were conducted to evaluate these factors. The matric suction during wetting of a seedbed affects the degree of aggregate slaking that occurs, and the strength of the wetted aggregates. The matric suction during draining affects the magnitude of ‘effective stresses’ that operate to retain soil structural integrity as the soil drains and dries out. An experiment was conducted to evaluate the influence of matric suction (within a range of suctions experienced in the field) on aggregate coalescence using soils of two different textures. Sieved aggregates (0.5 to 2 mm diameter) from a coarse-textured and two fine-textured (swelling) soils were packed into cylindrical rings (4.77 cm i.d., 5 cm high) and subjected to different suctions on wetting (near-saturation, and 1 kPa), and on draining (10 kPa on sintered-glass funnels, and 100 kPa on ceramic pressure plates). After one-week of drainage, penetrometer resistance was measured as a function of depth to approximately 45 mm (penetrometer had a recessedshaft, cone diameter = 2 mm, advanced at a rate of 0.3 mm/min). Tensile strength of other core-samples was measured after air-drying using an indirect “Brazilian” crushing test. For the coarse-textured soil, penetrometer resistance was significantly greater for samples wet to near-saturation, despite there being no significant increase in dry bulk density; this was not the case for the finer-textured soils, and it was difficult to distinguish the effects of variable bulk density upon drying from those of the imposed wetting treatments. In both coarse- and fine-textured soils, the tensile strength was significantly greater for samples wet to near-saturation. Thus wetting- and draining-suctions were both found to influence the degree of soil aggregate coalescence as measured by penetrometer resistance and tensile strength. Aggregate coalescence in irrigated crops is known to develop as the growing season progresses. It was therefore thought to be linked to the repeated occurrence of matric suctions that enhance the phenomenon during cycles of wetting and draining. An experiment was conducted to determine the extent of aggregate coalescence in a coarsetextured and two fine-textured (swelling clay) soils during 8 successive cycles of wetting and draining. Sieved aggregates (0.5 to 2 mm diameter) from each soil were packed into cylindrical rings (4.77 cm i.d., 5 cm high) and wetted to near saturation for 24 h. They were then drained on ceramic pressure plates to a suction of 100 kPa for one week, after which penetrometer resistance and tensile strength were measured as described above. The degree of expression of aggregate coalescence depended on soil type. For the coarse-textured soil, repeated wetting and draining significantly increased bulk density, penetrometer resistance and tensile strength. For the fine-textured soil, penetrometer resistance and bulk density did not vary significantly with repeated wetting and draining; on the contrary, there was evidence in these swelling clay soils to suggest bulk density and penetrometer resistance decreased. However, there was a progressive increase in tensile strength as cycles of wetting and draining progressed. The expansive nature of the fine-textured soil appears to have masked the development of aggregate coalescence as measured by penetrometer resistance, but its expression was very clear in measurements of tensile strength despite the reduction in bulk density with successive wetting and draining. Field observations have indicated that aggregate coalescence is first expressed at the bottom of the seedbed and that it develops progressively upward to the soil surface during the growing season. This suggests that overburden pressures may enhance the onset of the phenomenon by increasing the degree of inter-aggregate contact. Soils containing large quantities of particulate organic matter were known to resist the onset of aggregate coalescence to some extent. An experiment was conducted to evaluate the effects of soil organic matter and overburden pressures, by placing brass cylinders of various weights (equivalent to static load pressures of 0, 0.49, 1.47 and 2.47 kPa) on the top of dry soil aggregates (0.5 – 2 mm diameter) having widely different soil organic carbon contents placed in steel rings 5 cm high and 5 cm i.d. With the weights in place, the aggregates were wetted to near-saturation for 24 h and then drained on ceramic pressure plates to a suction of 100 kPa for one week. Bulk density, penetrometer resistance and tensile strength were measured when the samples were removed from the pressure plates and they all increased significantly with increasing overburden pressure in the soil with low organic matter content, but not in the soil with high organic matter content. The amount of tillage used to prepare seedbeds influences the size distribution of soil aggregates produced – that is, more tillage produces finer seedbeds. The size distribution of soil aggregates affects the number of inter-aggregate contact points and this was thought to influence the degree of aggregate coalescence that develops in a seedbed. Previous work has shown that soil organic matter reduces aggregate coalescence and so an experiment was conducted to evaluate the effects of aggregate size and organic matter on the phenomenon. For soils with high and low organic matter contents, aggregate size fractions of < 0.5, 0.5 – 2, 2 – 4, and < 4 mm were packed into soil cores (as above) and wetted to near-saturation then drained to 100 kPa suction as described above. Penetrometer resistance and tensile strength were measured and found to increase directly with the amount of fine material present in the soil cores – being greater in the < 0.5 mm and < 4 mm fractions, and being less in the 0.5 – 2 mm and 2 – 4 mm fractions. In all cases, penetrometer resistance and tensile strength were lower in the samples containing more organic matter. The rate at which soil aggregates are wetted in a seedbed affects the degree of slaking and densification that occurs, and the extent to which aggregates are wetted influences the overall strength of a seedbed. Both wetting rate and the extent of wetting were believed to influence the onset of aggregate coalescence and were thought to be affected by soil organic matter and irrigation technique. An experiment was therefore designed to separate these effects so that improvements to management could be evaluated for their greatest efficacy – that is, to determine whether management should focus on improving irrigation technique or increasing soil organic matter content, or both. The rate of wetting was controlled by spraying (or not spraying) soil aggregates (0.5 – 2 mm diameter) with polyvinyl alcohol (PVA). Samples of coarse- and fine-textured soils were packed into steel rings (as above) and subjected to different application rates of water (1, 10 and 100 mm/h) using a dripper system controlled by a peristaltic pump. Samples were brought to either a near-saturated state or to a suction of 10 kPa for 24 h, and then drained on a pressure plate at a suction of 100 kPa for one week. Measurements of penetrometer resistance and tensile strength were then made as described above. As expected, penetrometer resistance was lower in samples treated with PVA before wetting (slower wetting rates) and in samples held at a greater suction (10 kPa) after initial wetting (greater inter-aggregate strength). The effects were more pronounced in the coarse-textured soil. In both coarse- and fine-textured soils, tensile strengths increased with increasing wetting rate (greatest for 100 mm/h) and extent of wetting (greater when held at near-saturated conditions). The rate of wetting was found to be somewhat more important for promoting aggregate coalescence than the extent of wetting. Because aggregate coalescence often occurs with little or no increase in bulk density, an explanation for the increase in penetrometer resistance and tensile strength is unlikely to be explained by a large increase in the number of inter-aggregate contacts. An increase in the strength of existing points of inter-aggregate contact was therefore considered in this work. For inter-aggregate bond strengths to increase, it was hypothesized that small increases in the amount of mechanically (or spontaneously) dispersed clay particles, and subsequent deposition at inter-aggregate contact points could increase aggregate coalescence as measured by penetrometer resistance and tensile strength. An experiment was devised to manipulate the amount of spontaneously dispersed clay in coarse- and fine-textured soils of high and low organic matter content. The degree of sodicity of each soil was manipulated by varying the exchangeable sodium percentage (ESP) of soil aggregates (0.5 – 2mm) above and below a nominal threshold value of 6. Dry aggregates were then packed into steel rings (as above) and subjected to wetting near saturation, then draining to a suction of 100 kPa for one week as described above. Measurements were then taken of penetrometer resistance and tensile strength, both of which were affected by ESP in different ways. In the coarse-textured soil, sodicity enhanced aggregate slaking and dispersion, which increased bulk density. While penetrometer resistance also increased, its effect on aggregate coalescence could not be separated from a simple effect of increased bulk density. Similarly, the effect of sodicity on aggregate coalescence in the fine-textured soil was confounded by the higher water contents produced by greater swelling, which produced lower-than-expected penetrometer resistance. Measurements of tensile strength were conducted on air-dry samples, and so the confounding effects of bulk density and water content were eliminated and it was found that tensile strength increased with sodicity in both coarse- and fine-textured soils. The presence of dispersed clay was therefore implicated in the development of aggregate coalescence in this work. Finally, a preliminary evaluation of how the early stages of aggregate coalescence might affect plant growth was attempted using tomatoes (Gross lisse) as a test plant. Seeds were planted in aggregates (0.5 – 4 mm) of a coarse- or fine-textured soil packed in steel rings. These were wetted at a rate of 1 mm/h to either near-saturation (for maximum coalescence) or to a suction of 10 kPa (for minimum coalescence) and held under these conditions for 24 h. All samples were then transferred to a ceramic pressure plate for drainage to 100 kPa suction for one week. Samples were then placed in a growth-cabinet held at 20C with controlled exposure to 14 h light/day. Germination of the seeds, plant height, and number and length of roots were observed. Germination of the seeds held at near-saturation in both coarse- and fine-textured soils was delayed by 24 h compared with seeds held at 10 kPa suction. Neither the number nor the length of tomato roots differed significantly between the different treatments and soils. In the coarse-textured soil, however, the total root length over a period of 14 days was somewhat greater in the uncoalesced samples than in the coalesced samples, but this difference was not statistically significant. These results suggest that aside from delaying germination, aggregate coalescence may not have a large effect on early growth of tomato plants. However, this is not to say that detrimental effects may not be manifest at later stages of plant growth, and this certainly needs to be evaluated, particularly because aggregate coalescence increase with repeated cycles of wetting and draining. In conclusion, the primary findings of the work undertaken in this thesis were: • Rapid wetting of soil aggregates to near-saturation enhanced the onset of soil aggregate coalescence as measured by (in some cases) penetrometer resistance at a soil water suction of 100 kPa, and (in most cases) tensile strength of soil cores in the air-dry state. The rate of wetting appeared to be more important in bringing on aggregate coalescence than how wet the soil eventually became during wetting. This means reducing the rate at which irrigation water is applied to soils may reduce the onset of aggregate coalescence more effectively than controlling the total amount of water applied – though both are important. The literature reports that aggregate coalescence occurs in the field over periods of up to several months, involving multiple wetting and draining cycles, but the work here demonstrated that this can occur over much shorter time periods depending on conditions imposed. • Aggregate coalescence occurred in coarse-textured soils regardless of whether the bulk density increased during wetting and draining. In finer-textured soils, the response to wetting conditions varied and was complicated by changes in bulk density and water content due to swelling. • Small overburden pressures enhanced the onset of aggregate coalescence, but these effects were diminished in the presence of high soil organic matter contents. • Finer aggregate size distributions (which are often produced in the field by excessive tillage during seedbed preparation) invariably led to greater aggregate coalescence than coarser aggregate size distributions. The effects of aggregate size were mitigated to some extent by higher contents of soil organic matter. • Sodicity enhanced aggregate coalescence as measured by tensile strength, but when penetrometer resistance was measured in the moist state, the effects were masked to some extent by higher water contents generated by swelling and dispersion. This work suggests that tensile strength (in the air dry state) may be a more effective measure of aggregate coalescence than penetrometer resistance. • Early plant response to aggregate coalescence was not large, but the response may become magnified during later stages of growth. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1297583 / Thesis (Ph.D.) -- School of Earth and Environmental Sciences, 2007

A thesis submitted to the Faculty of Engineering, Cape Peninsula University of Technology, Cape Town, in partial fulfilment of the requirements for the M-Tech Degree in Civil Engineering” Cape Town 16 MAY 2011 / Dwellings in impoverished and upper class urban areas of the Western Cape have undergone serious structural failure due to problematic underlying sand deposits, generally known as Cape Flats windblown found in areas such as Delft, Blue Downs and the Atlantic Beach Golf Estate. The problem is compounded further when moisture penetrates below the footings and reaches saturation in the winter months. When poorly graded sand with a high fines content is coupled with vibration (through earth tremors), liquefaction may occur and without proper precautions, this can lead to inadequate foundation design, more than expected settlement and eventual structural failure. Some sand deposits are highly contaminated with organic debris, leading to compressible and collapsible conditions. Little knowledge is locally available of liquefiable conditions in windblown dune sand and what long-term settlement can be expected. The Western Cape is not known as an area exposed to serious seismic activity, but an updated (and more locally applicable set of data) is needed to eliminate possible erroneous foundation design. Samples have been extracted from typical sandy sites in the Western Cape where windblown dune sand is evident. Laboratory tests have been carried out on representative samples for closer examination and have been placed inside a purpose built calibration chamber that facilitates easy densification and probe testing. A DMT (flat blade dilatometer) was used to hydraulically penetrate the chamber sand sample to varying depths (up to 800 mm). The device can measure (with reduction formula) horizontal stress, angle of friction, bearing capacity and settlement. An accelerometer was attached to the chamber wall and vibration measured with the sand in varying states of moisture. The DMT is an unexplored in-situ soil testing device in South Africa and so far the outcome indicates favourably compared to other devices such as the Dynamic Cone Penetrometer (DCP. The DMT has the ability to measure the in-situ stiffness, strength and stress history parameters of soil for better site characterisation. Settlement within the chamber is easily measured. The DMT has, for example, indicated that sand from the Philippi area are a problematic founding soil and should be treated with special care at shallow founding levels. The horizontal stress index is low and according to the available knowledge on soil stress history, these sand, coupled with low densities, can liquefy easily and result in structural damage. The West Coast dune sand, being coarser and easily consolidated, poses less of a problem under liquefiable conditions. A suitable terrain device for easy on-site manoeuvrability is required to assist the DMT in further testing.