Standardization of Techniques for X-ray Computed Tomography of Soil
The adaptation of X-ray computed tomography (X-ray CT) to the study of intact soil represents a considerable advancement in our ability to understand pedogenesis, the physical, chemical and biological processes occurring in soils, as well as soil quality. A survey of the literature documenting related research initiatives over the past couple of decades reveals a plethora of approaches to the acquisition, processing and interpretation of X-ray CT imagery of soil. While this is natural, and represents the exploratory phase of this field, the lack of standardization of techniques fundamentally limits the comparability of results. This contribution examines issues associated with the development and adoption of standards for X-ray CT of soil. Particular attention is given to imaging protocols and image segmentation procedures.
Quantification of Soil Pore Structure Based on Minkowski-Functions
The porous structure in soils and other geologic media is typically a complex 3-dimensional object. Most of the physical material properties including mechanical and hydraulic characteristics are immediately linked to this structure which can be directly observed using non-invasive techniques as e.g. X-ray tomography. It is an old dream and still a formidable challenge to related structural features of porous media to their physical properties. In this contribution we present a scale-invariant concept to quantify pore structure based on a limited set of meaningful morphological functions. They are based on d+1 Minkowski functionals as defined for d-dimensional bodies. These basic quantities are determined as a function of pore size obtained by filter procedures using mathematical morphology. The resulting Minkowski functions provide valuable information on pore size, pore surface area and pore topology having the potential to be linked to physical properties. The theoretical background and the related algorithms are presented and the approach is demonstrated for the structure of an arable topsoil obtained by X-ray micro tomography. We also discuss the fundamental problem of limited resolution which is critical for any attempt to quantify structural features at any scale.
Time-resolved thermography for prediction of localized anisotropy in earth materials
Investigation of earth materials in a non-destructive manner is gaining impetus in environmentally sensitive areas and novel techniques are required for large-scale applications that account for variability in local composition. One such method, utilized in industrial and built-environment analysis, is non-temporal thermal imaging for detection of deformation, insular weakness, or structural discrepancies. The primary characteristics that resolve thermal response of an earth material are thermal diffusivity (á), volumetric heat capacity (C), and distribution of various mineral phases that compose a sample. While a non-temporal thermal image is resultant from super-imposing spatial distribution of the á and C, a time-resolved thermal image depicts the distribution of mineral phases and the correlated á and C. Established time-resolved thermography is prevalent in industrial quality control and for isotropic analysis of homogeneous specially- manufactured materials. Technical investigation using time-resolved thermography has resulted in characterization of sub-surface structures in homogeneous materials. In order to adapt this technique for application in the earth material sciences, the anisotropy of á, C, and spatial distribution must be accounted for. The purpose of this study is to relate a novel time-resolved thermal tomography approach to intact sample micromorphology of a multi-mineral earth material sample. The primary objective is a correlated prediction of micromorphology at depth via time-domain thermography. Secondary objectives include a standard approach to thermal tomography for intact earth material analysis, as well as a comparison of local and effective thermal properties as a function of variability in composition. Three earth material samples (one cemented by gypsum crystal growth, one cemented by carbonate crystals, and one is a crystalline sandstone) were subjected to a step-heating protocol with a duration that ensured penetration of the sample thickness. The resulting surface thermal emission (surface radiative temperature, T) was recorded at 3 second intervals until a steady state was reached. The variability in the slope of time resolved surface radiative temperature (T/t0.5) was compared to compositional indices extracted from computed tomographic imagery of the identical region contributing to the T(t). Variation around the generalized slope of time resolved surface radiative temperature for each sample (gypsum = 3.84; carbonate = 3.82 ; crystalline sandstone = 3.32) was described as a function of change in micromorphology at depth. Comparison of variation in T/t0.5 and micromorphometric variation encouraged the authors to pursue further development of time-resolved thermography application to earth materials.
Combining Interannual Airborne Lidar and Diurnal Oblique Thermal Imagery to Investigate Glacial Ice-Cored Moraine Dynamics
As glacier extents in the Canadian Rockies diminish, the influence of groundwater and other baseflow inputs to headwater river systems will gradually take on a more important role in terms of their contribution to the available water resource. It is believed that commensurate with decreasing exposed ice extents there is a volumetric increase in the reservoir of debris covered and therefore insulated ice cored moraines adjacent to and surrounding glacier margins. There is plenty of geomorphic evidence that clearly illustrates these storages of buried ice are increasing during the current period of glacier recession but to date little effort has focussed on investigating the relative proportions of buried ice melt vs. exposed ice melt in the Rockies, and the impact this might have on future water resources. In this paper, we present the results of multi-temporal lidar collections over the Peyto Glacier (2000, 2002, 2006, 2007) that accurately quantify rates of volumetric moraine downwasting. The spatial variability of lateral moraine downwasting is high but the average across the 7 year period was approximately 1.5 m p.a. While some of the downwasting can be attributed to slope creep, slumping and general mass wasting processes, the lack of debris volumes at the foot of slope clearly indicate that some other process is in operation and in fact dominates. By comparing the ratios of moraine to exposed ice downwasting, we find that up to 5% of the mass loss from the Peyto basin is from these marginal moraine areas, and we further hypothesize that most of this volume is lost in the form of melt water from ice cored moraines. This hypothesis was evaluated by collecting oblique diurnal thermal imagery to map the rise and decay in temperature signal over a large area of moraine adjacent to the Peyto Glacier. While no exposed ice was visible anywhere on the moraine slope, the presence of shallow ice core beneath the debris covered surface could be inferred in areas of slow temperature rise during daylight solar heating and rapid thermal decay after sunset. It is presumed that this apparent increased loss of heat in certain areas of the moraine is being used to drive internal melt processes. With some calibration and energy balance modeling, this technique will be further developed to map debris cover depth and assist in the parameterisation of glacial hydrological melt models.
Quantitative geomorphic analysis with LiDAR DEMs: Case-studies from Boreal landscapes
Many common digital terrain analysis techniques are problematic when applied to very high-resolution digital elevation models (DEMs), such as those derived from airborne Light Detection and Ranging (LiDAR) surveys. As many researchers are acquiring LiDAR DEMs to facilitate research in Canada's Boreal regions, new image analysis techniques are required that can exploit the full level of detail observed in these datasets. This talk will present several new quantitative methods for landscape segmentation and analysis using LiDAR digital elevation imagery, with a particular emphasis on the spatial characterization of northern peatlands and forested wetlands. A major focus of the talk will be on the new technique of hydrogeomorphic edge-detection, which provides a flexible and scale-adaptive approach for segmenting landscapes into hydrologic response units. Potential applications in hydrological modelling and ecohydrological prediction will also be discussed.
High-Resolution Digital Mapping of Soil Surface Water Content at the Field Scale Using Ground Penetrating Radar
Measuring soil surface water content spatial variability is essential for many environmental and agricultural researches and engineering applications, as this variable controls important key processes of the hydrological cycle such as infiltration, runoff, evaporation, and energy exchanges between the earth and the atmosphere. In particular, the characterization of spatial patterns and heterogeneities over a continuous range of scales is presently subject to intensive research for developing, calibrating and testing distributed hydrological models, with, e.g., the installation of field- to watershed-scale observatories. In that respect, ground penetrating radar (GPR) appears to be a promising tool for real-time, high resolution digital soil mapping at the field scale. Yet existing GPR techniques for quantitative soil characterization still suffer from a series of limitations, mainly arising from the strong simplifying assumptions that are commonly made with respect to electromagnetic wave propagation phenomena. We have developed a new GPR methodology based on full-waveform forward and inverse modelling, that inherently maximizes radar information retrieval capabilities thanks to an accurate electromagnetic model and system calibration procedure. The radar system consists of a vector network analyzer combined with an off- ground, zero-offset, ultra-wideband horn antenna, thereby setting up a stepped-frequency continuous-wave (SFCW) GPR. A full-waveform model describes accurately the radar signal by accounting for (1) all antenna effects and antenna-soil interactions through a linear system of frequency dependent, complex transfer functions, and (2) wave propagation in three-dimensional multilayered media through a Green's function as exact solution of Maxwell's equations. A fast procedure was developed to evaluate the involved spatial Green's function from its spectral counterpart, whose integral is singular. The soil electromagnetic properties and their vertical distribution are estimated by inverse modeling using various iterative optimization strategies, depending on the model complexity. The method presents especially considerable advantages compared to the current surface characterization techniques using GPR, namely, the ground wave and common reflection methods. The proposed methodology was successfully validated for a series of model configurations of increasing complexity. For the particular case of soil surface water content retrieval, we especially addressed the impact of shallow soil layering on the inverse estimates in case it is or not accounted for in the inverse model configuration. The results show that thin layers should not be neglected, especially when high contrasts between soil layers are encountered. The method is now routinely used for real-time, automated mapping of soil surface water content in the field. GPR-derived maps are compared to ground-truth measurements and satellite radar data products. Stochastic approaches are used for assessing the uncertainty on the inverse estimates. The proposed method constitutes in particular a robust alternative to other GPR approaches for shallow soil characterization.
Calibration of Time Domain Reflectometry Using Undisturbed Peat Samples
At present, freeze-thaw experiments on 60 cm diameter by 90 cm deep peat/permafrost cores from our central Mackenzie River basin long-term field site are being conducted to better elucidate coupled heat and moisture transport processes occurring in the active layer. Time domain reflectometry (TDR) along with 15 cm and 30 cm probes is being used to monitor the depth-specific volumetric soil moisture (VSM) in these cores. Organic soils are characterized by high porosities and soil density increases many folds at successive depths. Therefore, we calibrated the TDR for different depths and for peat with two different origins (sphagnum moss and lichen). A simple apparatus was developed to ensure a more homogenous soil moisture distribution in the undisturbed samples used for calibration. Methodology and results for the calibration as well as comparisons with mixing models at different densities are presented and the effects of different soil phases on calibration are discussed.