Browsing by Author "Parker, Jack C."

Now showing 1 - 13 of 13
  • Thumbnail Image
    A comparison of uncertainty analysis methods for the modeling of land disposal of petroleum exploration and production waste
    Choong, Peng Kee (Virginia Tech, 1991-05-05)
    Three uncertainty analysis methods, namely : (1) Monte Carlo (MC) simulation, (2) First-Order Second Moment (FOSM) method, and (3) Modified Point Estimate (MPE) method, were compared in terms of computational efficiency and accuracy for evaluation of contaminant concentration at a down-gradient receptor. Benzene and chloride, two contaminants most commonly found in petroleum exploration and production waste, were analyzed for two waste disposal scenarios, i.e. (1) closed waste pit, and (2) landspreading. Using MC method as the basis for comparison, results of the analysis suggested the use of FOSM method as the most attractive alternative to the more tedious MC method. Based on 2000 MC simulations for analysis, the FOSM method required 1.053% of the computational effort required for the MC analysis. Considering a regulatory guideline of 85% probability level of exceedance, results of the analysis indicated that the MPE method consistently underestimated the benzene concentration and overestimated the chloride concentration at a receptor, whereas the FOSM analyses indicate good agreement with MC method. However, in the event where a more precise and complete analysis is mandatory, the MC method remains the most appropriate methodology available.
  • Thumbnail Image
    Determining transport parameters from laboratory and field tracer experiments
    Parker, Jack C.; Van Genuchten, Martinus Th. (Virginia Agricultural Experiment Station, 1984)
    This bulletin describes a nonlinear least-squares inversion method that can be used to identify several parameters in a number of theoretical one-dimensional solute transport models. One of the models discussed is the usual convection-dispersion transport equation that includes terms accounting for linear equilibrium adsorption, zero-order production and/or first-order decay. In addition, a two-site/two-region model is described that can be applied to various non-equilibrium transport problems. Also included is a stochastic model that considers the effects of areal variations in hydraulic fluxes on field-scale solute transport. This last model also has provisions for zero- or first-order production and/or decay. The least-squares inversion method can be used to analyze both spatial and temporal distributions of flux or resident concentrations. A detailed description of the computer program, called CXTFIT, is given in one of the appendices of this bulletin. Several example problems illustrating practical applications of the program are discussed in detail.
  • Thumbnail Image
    Development of phenomenologically-based distribution fitting procedures and spatial processes for mixed population soil properties
    Cooke, Richard (Virginia Tech, 1993-11-08)
    In the literature, two distinct flow phenomena, namely, flow through the main body of the soil, and flow through preferential flow paths, have been identified. Models which try to incorporate the effects of these two phenomena require either an explicit or an implicit knowledge of the probability distribution functions associated with the soil properties affecting flow. In keeping with the fact that these properties are influenced by two distinct phenomena, it is postulated that they should be represented by heterogeneous distribution functions. These distribution functions are, by design, suitable for representing mixed population data. Procedures were developed for fitting heterogeneous distribution functions to data. These procedures are encoded in Microsoft QUICKBASIC with some additional FORTRAN routines. The fitting procedures do not utilize any moment above the second order, and are markedly different from the use of regression methods for fitting multiple parameter distributions. Procedures were developed for two types of mixtures. One type is suitable for instances where a measured quantity is the sum of values from two populations, while the other is applicable when a measured quantity may be from one population or from another, but not from both at the same time or location. The procedures were applied to several data sets, including flow data, infiltrability data, and pH data. In many instances, the use of heterogeneous distributions resulted in an improvement in fit quality as compared to the fit quality for homogeneous distributions. The most dramatic improvement are observed in the fit to extreme data values. Procedures were also developed to incorporate heterogeneous distribution functions into three common processes in Soil and Water Engineering, namely, Monte Carlo simulation, stochastic field generation, and interpolation. In these procedures, data which are best represented by heterogeneous distributions are transformed to Gaussian space and existing Gaussian-based procedures are applied. In several validation efforts the modified processes were found to as good as, or better than, conventional procedures. In the process of developing the modified spatial processes mentioned above, a robust trend surface procedure and a new matrix decomposition procedure were developed. These ancillary procedures were shown to be useful in other engineering applications.
  • Thumbnail Image
    Engineering properties of selected soils in the Virginia Piedmont
    Parker, Jack C.; Amos, Dan F.; Baker, James C. (Virginia Agricultural Experiment Station, 1983-10)
    The Piedmont Province i Virginia, running in a north-south direction, is approximately 50 miles wide along the Maryland border I ard broadens to the south until it encompasses approximately 150 miles along the North Carolina border. It cor prises at least one-third of the land area of the stat , with approximately 60% occupied by woodland and 40% by agriculture, primarily beef or dairy enterprises. Located in the Piedmont j:e the cities of Leesburg, Fairfax, Manassas, Warrenton, Culpeper, Charlottesville, Lynchburg , Bedford, Farmville, Martinsville, Danville and South Boston. I an McHargue, in his book Design With Nature, studied the Potomac River watershed and concluded that "the Piedmont is primaril suitable for urbanization with attendant agriculture and undifferentiated recreation" (McHargue 1969). Because of the development potential of the Piedmont and the intense pressures for future urbanization west of Washington, D.C., in the counties of Fairfax, Prince William and Loudoun; around Richmond in the counties of Hanover, Hen ico, Goochland, Powhatan, Amelia and Chesterfield; an in Albemarle, Amherst, Bedford, Campbell, Franklin, Henry and Pittsylvania counties, the need to intensively study the most widely distributed and potentially important Piedmont soils became apparent....
  • Thumbnail Image
    Mass transfer from non-aqueous phase liquids to the aqueous phase in groundwater systems
    Jayaraman, Krithika M. (Virginia Tech, 1992-07-05)
    As immiscible organic contaminants migrate through the subsurface environment, a significant portion of non-aqueous phase liquids (NAPL's) is trapped by capillary forces and remains in the subsurface as immobile blobs or ganglia. Residual saturations of NAPL on the order of 5-30 percent have been observed in saturated subsurface systems. The NAPL can partition into the aqueous phase and serve as a long-term source of groundwater contamination. NAPL-aqueous mass transfer rates impact the distribution and the rate of movement of the contaminant within the subsurface. The mass transfer coefficient is a function of many variables including aqueous phase velocity and NAPL-aqueous interfacial area. A one-dimensional column apparatus and experimental procedure was developed to study the nature of mass transfer between the aqueous phase (water), and a non-aqueous phase liquid (NAPL) in porous media. The NAPLs used in the experiments were composed of soluble and insoluble (inert) compounds. These experiments were designed to investigate the aqueous mass transfer of benzene, toluene, ethylbenzene, and xylene from and inert compound (soltrol or hexadecane). The results of this research indicate that the rate of NAPL-aqueous interphase mass transfer increases as a function of increasing aqueous phase velocity and percent NAPL saturation. Miller's equation was used to compare the model performance. Comparison of the experimental K values obtained by the literature was performed. The results indicated a satisfactory model performance.
  • Thumbnail Image
    A model of nitrate leaching from agricultural systems in the northern neck of Virginia
    Johnson, Terrence G. (Virginia Tech, 1992-10-01)
    A model (VTCROPS) was developed to simulate the long-term effects of nitrogen (N) leaching to groundwater in the Northern Neck region of Virginia, and ultimately to the Chesapeake Bay. VTCROPS simulates N fate and transport in a soil-plant-atmosphere continuum in a vertical slice between two crop rows, enabling consideration of nonuniform fertilizer placement and root growth patterns. VTCROPS models atmospheric, soil and crop subsystems. Atmospheric conditions (rainfall, temperature, solar radiation) may be directly input by the user or generated using a stochastic climatic generator. The soil subsystem simulates runoff, infiltration, drainage and soil-water redistribution, N immobilization, nitrification, mineralization, denitrification and advective N transport. The crop subsystem simulates plant N uptake, and vegetative and reproductive growth in response to soil and climatic factors, explicitly for maize or wheat. VTCROPS simulates soybean in a crop rotation empirically accounting for leaf area and root growth. The model is capable of simulating long-term cropping sequences under minimum and conventional tillage practices for continuous maize or for rotations involving maize, wheat, soybean and fallow. Critical internal model parameters were calibrated through comparison of output to field data. The sensitivity of output to input variables was determined. Model output is most sensitive to the climatic variables. Model predicted crop performance variables — grain and total dry matter yields and N content — and soil N content were compared with available field data from two sites over a three year period for maize. Data from six sites over a one year period were tested for wheat. Predictions for maize and total N content were fairly accurate, with a tendency to greater error in dry years. Predictions for wheat were somewhat less accurate, but incomplete field data precluded determining the source of discrepancies. Long-term model predictions, for two year crop rotations with minimum and conventional tillage, were evaluated by comparing performance variables with literature values. Appropriate responses were obtained for N transformation processes. Mass conservation for soil-water and N were good. Maize performance variables were within the range of literature values, and were higher under minimum till. Wheat yields and N contents were somewhat higher than values reported in the literature. Nitrogen load is correlated to drainage and water use over the short run, and to rainfall and drainage over longer periods. Minimum tillage did not increase N load to groundwater. Over a year nitrogen load was strongly periodic, with most leaching taking place from January through April. More than 50% of the N load over a rotation was lost during an extended fallow period that followed soybeans. Nitrogen load increased with fertilizer rates, however, N leaching fraction was optimal around rates of 150 — 200 kg/ha. The model was applied to the Virginia counties of Richmond, Westmoreland, Lancaster, King George and Northumberland to assess the potential for long term N leaching to groundwater. Soil surveys indicated that 34 soil map units occurred within 123,000 hectares of cultivated land. In order to reduce the number of simulations, principal component analysis and cluster analysis were employed to subdivide the cultivated area into 10 land units based on different soil properties. Historical climatic data from the area was used to calibrate the stochastic climatic generator. Analyses were performed to determine long-term crop performance and N loads to ground and surface waters in the study area over a 26 year period (13 rotations). Two management systems were applied to the land units. The first management system consisted of a rotation of minimum tilled maize, conventionally tilled wheat, minimum tilled soybeans and a fallow period. The second management system had a similar cropping sequence, but all crops were conventionally tilled. In both cases, fertilizer was applied at a rate of 150 kg-N/ha/crop. With the exception of two land units, mean yield, water use and N uptake over the simulation was fairly uniform among the land units. Runoff and drainage were highly variable between land units and over time within units. Mineralization, denitrification and N load were highly variable both between land units and over time. Nitrogen load ranged from 66 to 131 kg/ha/rotation between land units. Long-term average N loads and N concentrations from the cultivated area and from the total area of the study region were estimated. For this analysis it was assumed that 80% of the cultivated area was under minimum till and 20% under conventional tillage. An area-weighted average of 5.4 million kg-N/ha/year or 29% of total N applied is discharged to groundwater with an average drainage concentration of 9.9 mg/L. The average N concentration from the study area (including uncultivated areas) to groundwater is estimated at 5.1 mg/L. Average N concentration to the Chesapeake Bay from all sources, after dilution with runoff is 4.5 mg/L which is lower than the drinking water standard for nitrate N of 10 mg/L. The possibility of using sewage sludge as a replacement for, or in consort with N fertilizer was investigated for a typical land unit, under a conventionally tilled maize-wheat- soybean-fallow rotation. Simulations were conducted with 100%, 50% and 0% sludge (C—N ratio of 12). With fertilizer N augmenting the sludge, the total N input (250 kg/ha) was the same for each treatment. Mean yields were similar for 50% and zero sludge, but lowered by 10% and 16%, respectively, for maize and wheat with 100% sludge. Discrepancies in yields were attributed to the fact that mineralization rates of sludge are not high enough to supply the crop N requirement during periods of peak uptake. Nitrate leaching was reduced by 41 and 25% with 100 and 50% sludge applications, respectively.
  • Thumbnail Image
    A model of nitrate leaching from agricultural systems in Virginia's Northern Neck
    Johnson, Terrence G.; Parker, Jack C. (Virginia Water Resources Research Center, Virginia Polytechnic Institute and State University, 1993-06)
    A model (VT-CROPS) was developed to simulate the long-term effects of nitrogen (N) leaching to groundwater in the Northern Neck region of Virginia and, ultimately, to the Chesapeake Bay. VT-CROPS simulates N fate and transport in a soil-plant-atmosphere continuum in a vertical slice between two crop rows, enabling consideration of nonuniform fertilizer placement and root growth patterns. VT-CROPS models atmospheric, soil and crop subsystems. Atmospheric conditions (rainfall, temperature, solar radiation) may be entered directly by the user or generated by using a stochastic climatic generator. The soil subsystem simulates runoff, infiltration, drainage and soil-water redistribution, N immobilization, nitrification, mineralization, denitrification, and advective N transport. The crop subsystem simulates plant N uptake, and vegetative and reproductive growth in response to soil and climatic factors, explicitly for maize or wheat. VT-CROPS simulates soybean in a crop rotation, empirically accounting for leaf area and root growth. The model is capable of simulating long-term cropping sequences under minimum and conventional tillage . practices for continuous maize or for rotations involving maize, wheat, soybean, and fallow. Critical internal model parameters were calibrated through comparison of output to field data. The sensitivity of output to input variables was determined. Model output is most sensitive to the climatic variables. Model-predicted crop performance variables-grain and total dry matter yields and N content-and soil N content were compared with available field data from two sites over a three-year period for maize. Data from six sites over a one-year period were tested for wheat. Predictions for maize and total N content were fairly accurate, with a tendency to greater error in dry years. Predictions for wheat were somewhat less accurate, but incomplete field data precluded determining the source of discrepancies. Long-term model predictions, for two-year crop rotations with minimum and conventional tillage, were evaluated by comparing performance variables with literature values. Appropriate responses were obtained for N transformation processes. Mass conservation for soil water and N were good. Maize performance variables were within the range of literature values, and were higher under minimum till. Wheat yields and N contents were somewhat higher than values reported in the literature. Nitrogen load is correlated to drainage and water use over the short run, and to rainfall and drainage over longer periods. Minimum tillage did not increase N load to groundwater. Over a year, nitrogen load was periodic, with most leaching taking place from January through April. More than 50% of the N load over a rotation was lost during an extended fallow period that followed soybeans. Nitrogen load increased with fertilizer rates; however, the N leaching fraction was optimal at rates of 150-200 kg/ha. The model was applied to the Virginia counties of Richmond, Westmoreland, Lancaster, King George, and Northumberland to assess the potential for long-term N leaching to groundwater. Soil surveys indicated that 34 soil map units occurred within 123,000 hectares of cultivated land. To reduce the number of simulations, principal component analysis and cluster analysis were used to subdivide the cultivated area into 10 land units based on different soil properties. Historical climatic data from the area were used to calibrate the stochastic climatic generator. Analyses were performed to determine long-term crop performance and N loads to groundwater and surface waters in the study area over a 26-year period (13 rotations). Two management systems were applied to the land units. The first management system consisted of a rotation of minimum-tilled maize, conventionally tilled wheat, minimum-tilled soybeans, and a fallow period. The second management system had a similar cropping sequence, but all crops were conventionally tilled. In both cases, fertilizer was applied at a rate of 150 kg-N/ha/crop. With the exception of two land units, mean yield, water use, and N uptake over the simulation were fairly uniform among the land units. Runoff and drainage were highly variable between land units and over time within units. Mineralization, denitrification, and N load were highly variable both between land units and over time. Nitrogen load ranged from 66 to 131 kg/ha/rotation between land units. Long-term average N loads and N concentrations from the cultivated area and from the total area of the study region were estimated. For this analysis, it was assumed that 80% of the cultivated area was under minimum till and 20% under conventional tillage. An area-weighted average of 5.4 million kg-N/ha/year, or 29% of total N applied, is discharged to groundwater, with an average drainage concentration of 9.9 mg/I. The average N concentration from the study area (including uncultivated areas) to groundwater is estimated at 5.1 mg/I. Average N concentration to the Chesapeake Bay from all sources, after dilution with runoff, is 4.5 mg/I, which is lower than the drinking water standard for nitrate N of 10 mg/I. The possibility of using sewage sludge as a replacement for, or in consort with, N fertilizer was investigated for a typical land unit, under a conventionally tilled maize-wheat-soybean-fallow rotation. Simulations were conducted with 100%, 50%, and 0% sludge (CN ratio of 12). With fertilizer N augmenting the sludge, the total N input (250 kg/ha) was the same for each treatment. Mean yields were similar for 50% and 0% sludge, but lowered by 10% and 16%, respectively, for maize and wheat with 100% sludge. Discrepancies in yields were attributed to the fact that mineralization rates of sludge are not high enough to supply the crop N requirement during periods of peak uptake. Nitrate leaching was reduced by 41 % and 25% with 100% and 50% sludge applications, respectively.
  • Thumbnail Image
    Observation scale effects on fluid transport behavior of soil
    Albrecht, Karen A. (Virginia Tech, 1985-01-15)
    Variabilities of hydraulic and solute transport properties of soil are examined at three scales: pore-scale, sample volume-scale, and field-scale. Undisturbed soil cores were taken at 19 subsites spaced logarithmically along a 150 m line transect in a Groseclose mapping unit near Blacksburg; Virginia. Three core sizes were taken at each subsite at the soil surface and 0.5 m depth. 'Small' cores were-40x54 mm; 'medium' cores were 60X100 mm; and 'large' cores were 100x150 mm. Macropore effects on solute transport were evaluated using monocontinuum and bicontinuum models. Bicontinuum-predicted solute breakthrough curves (BTC) closely agreed with observed BTC data with mean errors of reduced concentrations
  • Thumbnail Image
    ONESTEP : a nonlinear parameter estimation program for evaluating soil hydraulic properties from one-step outflow experiments
    Kool, J. B.; Parker, Jack C.; Van Genuchten, Martinus Th. (Virginia Agricultural Experiment Station, 1985-03)
  • Thumbnail Image
    Physical and chemical characterization of the Groseclose soil mapping unit
    Kool, J. B.; Albrecht, Karen A.; Parker, Jack C.; Baker, James C. (Virginia Agricultural Experiment Station, 1986)
  • Thumbnail Image
    Preferential movement of solutes through soils
    Bruggeman, Adriana C. Jr. (Virginia Tech, 1997-09-18)
    Detection of unexpectedly high concentrations of agricultural pollutants in ground water have inspired investigations of the role of preferential movement of chemicals through agricultural soils. This research focuses on preferential flow and solute transport processes and the effects of agricultural management practices on these processes. Experimental methods for monitoring preferential flow and solute transport in the field as well as a stochastic, physically-based model for predicting water flow and transport of non-reactive chemicals in heterogeneous soils with naturally occurring macropores were developed and evaluated. Field experiments, aimed at monitoring the occurrence of preferential flow and solute transport, were conducted in a conventionally-tilled and a no-till soybean field in the Coastal Plain of Virginia. A rainfall simulator was used to apply a one-hour storm at rates of 5.0, 6.5 and 7.5 cm/hr to six 1.83 by 1.83 m plots. Chloride was added to the water to serve as a non-reactive tracer. Electrical conductivity equipment provided a useful method for monitoring solute transport. The moisture and solute conditions, observed during a 28-hour period after the start of the rainfall event, clearly indicated the occurrence of preferential flow and solute movement in the field plots. The variability of the solute concentrations in the field plots was generally higher in the no-till plots than in the conventionally-tilled plots. The plots that received rain at 6.5 and 7.5 cm/hr showed more variability than the plots that received rain at 5 cm/hr. The observed solute concentrations indicated that if the solute transport would have taken place by advection only, 61% of the solute transport in the conventionally-tilled plots and 50% of the solute transport in the no-till plots could be attributed to preferential flow. A physically-based, finite element model for simulating flow and solute transport in variably-saturated soils with macropores (MICMAC) was developed. Flow and solute transport are described by the Richards' equation and the convection-dispersion equation. Flow in the macropores is described by the Hagen-Poiseuille equation. An axisymmetric coordinate system is used to simulate the flow and solute transport from the macropore into the surrounding soil matrix, assuming a vertically oriented, surface-vented, cylindrical macropore. Flow and solute transport between the macropore and the soil matrix are driven by the pressure head at the macropore-matrix boundary. To assess the natural heterogeneity of the soil properties a stochastic component was added to the model. Flow and solute transport at the field scale were simulated by regarding the field as a collection of statistically independent, non-interacting vertical soil columns, using Monte Carlo simulation. The sensitivity analysis of the model indicated that, for a soil with macropores, the model is most sensitive to the saturated water content of the soil matrix, the initial moisture content, and the rainfall rate. The model is not very sensitive to the macropore dimensions. Examination of the stochastic approach indicated that the representation of a heterogeneous field as a collection of non-interacting stream columns may substantially underestimate water and solute leaching. A change of 5% in the soil properties of the neighboring soil columns may underpredict the solute leaching, 24 hours after a rainstorm, by 157% for a soil column with a macropore, and by 58% for a soil column without a macropore. These differences decreased to 47% and 8%, respectively, 168 hours after the rainfall. Field application of the model suggested that the model underestimates the leaching of water and solutes from the root zone. However, the computed results were substantially better than the results obtained when no preferential flow component was included in the model. The model performed best under conditions that favored preferential flow, i.e., a high rainfall rate and high initial moisture conditions. The simulated and observed solute concentrations in the root zone agreed reasonably well, although the maxima of the observed data were generally higher than those of the simulated data.
  • Thumbnail Image
    Three-Dimensional Modeling of Solute Transport with In Situ Bioremediation Based on Sequential Electron Acceptors
    Waddill, Dan Wilson (Virginia Tech, 1998-01-29)
    A numerical model for subsurface solute transport is developed and applied to a contaminated field site. The model is capable of depicting multiple species transport in a three-dimensional, anisotropic, heterogeneous domain as influenced by advection, dispersion, adsorption, and biodegradation. Various hydrocarbon contaminants are simulated as electron donors for microbial growth, with electron acceptors utilized in the following sequence: oxygen, nitrate, Mn(IV), Fe(III), sulfate, and CO₂. In addition, the model accounts for products of biodegradation such as Mn (II), Fe(II), H₂S, and CH₄. Biodegradation of each hydrocarbon substrate follows Monod kinetics, modified to include the effects of electron acceptor and nutrient availability. Inhibition functions permit any electron acceptor to inhibit utilization of all other electron acceptors that provide less Gibbs free energy to the microbes. The model assumes that Fe(III) and Mn(IV) occur as solid phase ions, while the other electron acceptors are dissolved in the aqueous phase. Microbial biomass is simulated as independent groups of heterotrophic bacteria that exist as scattered microcolonies attached to the porous medium. Diffusional limitations to microbial growth are assumed to be negligible. In order to verify the accuracy of the computer code, the model was applied to simple, hypothetical test cases, and the results were compared to analytical solutions. In addition, a sensitivity analysis showed that variations in model inputs caused logical changes in output. Finally, the capabilities of the model were tested by comparing model output to observed concentrations of hydrocarbons, electron acceptors, and endproducts at a leaking UST site. The model was calibrated using historical site data, and predictive capabilities of the model were tested against subsequent sets of field data. The model was used to examine the effect of porous media heterogeneities on contaminant transport and biodegradation. The turning bands method was used to produce hypothetical, yet realistic heterogeneous fields describing hydraulic conductivity, initial biomass concentration, and the maximum rate of substrate utilization. When the available electron acceptor concentrations were small compared to the hydrocarbon concentration, the overall rate of hydrocarbon mass loss increased with time, even as hydrocarbon concentrations decreased. This trend is the opposite of what would be predicted by a first order decay model.
  • Thumbnail Image
    Water adsorption, microstructure, and volume change behavior of clay minerals and soil
    Parker, Jack C. (Virginia Polytechnic Institute and State University, 1980)
    Swelling attributable to intracrystalline water adsorption by montmorillonite (MT) and vermiculite (VR) saturated with cations which limit interlayer expansion were calculated as the product of one-half the change in interlayer spacing determined by x-ray diffraction and the difference between total surface area taken as 800 m²/g and external crystal surface area measured by N₂ gas adsorption. Swelling directly attributable to osmotic adsorption was calculated as the product of external surface area and theoretical double layer thickness. In homoionic, monomineralic systems of Ca- and Al-MT and Na-, Ca- and Al-VR, intracrystalline expansion accounted for 60, 70, 21, 56 and 44%, respectively, of total expansion during saturated vapor-wetting and 15, 50, 4, 12 and 13% during submersion in 0.01 N electrolyte solutions. Osmotic adsorption of these systems accounted for < 10% of the expansion on submersion in all cases. Dry, static compaction of Ca-MT resulted in an increase in expansion with increasing initial density which was not explained by changes in predicted osmotic or intracrystalline expansion. This was attributed to expansion caused by gaseous pressures developing ahead of advancing wetting fronts. An inverse relationship was found between pore size and swelling caused by entrapped air pressures. Slow wetting ameliorated this expansion by allowing dissipation of entrapped air. Relaxation of crystal strains during wetting was also implicated as contributing to expansion. Electron microscopy revealed crystal strain decreased in the order: Na-MT > Ca-HT > Al-MT > Na-VR ≃ Ca-VR ≃ Al-VR. Greater expansion by freeze-dried than oven-dried Ca-MT was explained by greater crystal strain observed in the former clay. Greater strain broadening of the (060) x-ray diffraction peak was observed for the freeze-dried clay and its mean b-dimension was slightly smaller than the oven-dried clay. Crystal strain relaxation is suggested to explain the relationship between b-dimension and swelling reported in the literature. Predicted osmotic expansion for two soils was greater than in the pure clay systems and intracrystalline expansion smaller, due to a greater ratio of external to internal surface area in the soils. Osmotic adsorption accounted for 30-70% of the soil swelling during submersion in 0.0lM electrolyte solutions, while intracrystalline expansion accounted for < 10%. The magnitude of swelling due to entrapped air pressures was evaluated from the difference between expansion of atmospheric pressure-wet and vacuum-wet clods. This component accounted for 10-60% of the expansion of undisturbed and compacted samples during the second cycle of wetting from the air-dry state, but did not appear to be directly related to pore size. Changes in structure accompanying air-drying appeared to enhance expansion, especially for the high MT Iredell soil. It was suggested that this may be the result of increases in crystal strain during drying.