NS23A-01 INVITED
Remote sensing and scaling of satellite derived hydrological variables
Remote sensing is an important tool to observe spatial and temporal variability in hydrological studies. Given the ever-growing spatial extents of our studies in hydrological sciences and the considerable overlap with near- surface geophysics, the properties of remotely sensed data deserve to be examined very carefully. Remotely sensors integrate over a spatial area the radiance emitted by the surface. Therefore, it is not possible to separate the contributions of the different space elements by simple methods. However, in some cases, spatial disaggregation may be useful to quantify surface and subsurface heterogeneity. In this paper, I will show examples of soil moisture derived from the satellite sensor AMSR (Advanced Microwave Scanning Radiometer) as well aircraft sensors, PALS (Passive Active L and S Band Radiometer and Radar) and AIRSAR (Aircraft Synthetic Aperture Radar) to show an example of spatial disaggregation. I will also show studies using SMMR (Scanning Multichannel Microwave Radiometer) which deal with spatial scale of soil moisture. Simulation studies will also highlight the impact of spatial heterogeneity in land surface conditions of aircraft and satellite microwave observations and retrievals.
NS23A-02 INVITED
Characterizing Watersheds with Geophysical Methods: Some uses of GPR and EMI in Hydropedological Investigations.
The USDA-NRCS and its cooperators use ground-penetrating radar (GPR) and electromagnetic induction (EMI) as rapid, noninvasive tools to support soil surveys at different scales and levels of resolution. The effective use of GPR is site-specific and generally restricted to soils having low electrical conductivity (e.g., soils with low clay and soluble salt contents). In suitable soils, GPR provides high resolution data, which are used to estimate depths to soil horizons and geologic layers that restrict, redirect, and/or concentrate the flow of water through landscapes. In areas of coarse-textured soils, GPR has been used to map spatiotemporal variations in water-table depths and local ground-water flow patterns. Compared with GPR, EMI can be effectively used across a broader spectrum of soils and spatial scales, but provides lower resolution of subsurface features. EMI is used to refine and improve soil maps prepared with traditional soil survey methods. Differences in apparent conductivity (ECa) are associated with different soils and soil properties (e.g., clay, moisture and soluble salt contents). Apparent conductivity maps provide an additional layer of information, which directs soil sampling, aids the identification and delineation of some soil polygons, and enhances the quality of soil maps. More recently, these tools were used to characterize the hydropedological character of a small, steeply sloping, forested watershed. Within the watershed, EMI was used to characterize the principal soil-landscape components, and GPR was used to provide high resolution data on soil depth and layering within colluvial deposits located in swales and depressional areas.
NS23A-03 INVITED
Improving Crosshole GPR Velocity Tomography at Close Borehole Spacings
Over the past decade, crosshole ground-penetrating radar (GPR) velocity tomography has become a popular technique for high-resolution imaging of subsurface moisture content. With this method, when the spacing between the boreholes is large compared to the length of the antennas being used, reliable subsurface images can be readily obtained because small errors in measured travel-times and antenna locations have a small effect on the resulting images. However, when the borehole spacing is on the order of the antenna length in crosshole GPR tomography, significant problems often occur. Specifically, when attempting to include high-angle ray data into the tomographic inversion procedure at close borehole spacings, numerical artifacts often appear that obscure the true velocity structure of the subsurface. To practically deal with this problem, high-angle ray data are often excluded from crosshole GPR inversions. Although such aperture limitation allows for the recovery of reasonable tomograms, discarding the high-angle data theoretically limits the horizontal resolution of the images that can be obtained. We have determined a number of factors that affect crosshole GPR tomography at close borehole spacings, and that we believe cause the problems experienced with high-angle ray data. First, it is always assumed in crosshole GPR inversions that first-arriving energy travels directly between the antenna centers; we have found, however, that tip-to-tip coupling at high transmitter-receiver angles is often possible. Secondly, interference between feed- and end-radiated pulses from the antennas can cause errors in travel-time picks that vary with transmitter-receiver angle. Thirdly, errors in the calculation of the transmitter fire time can occur due to differences in antenna coupling in the air and ground. Finally, propagation dispersion in the GPR pulse can affect travel-time measurements. All of these factors can lead to errors in picked travel times that vary with transmitter-receiver angle. Because of the close borehole spacing, the errors have a significant effect on the resulting tomographic images. To deal with these problems, we have developed a crosshole GPR inversion procedure that estimates, in addition to subsurface EM-wave velocity, a small number of parameters that describe a travel-time "correction curve"½ as a function of angle. Using this new inversion methodology, we show improved inversion of synthetic crosshole GPR data, and also field data collected at the Boise Hydrogeophysical Research Site between closely spaced wells.
NS23A-04 INVITED
The Support Volume of Geophysical Measurement: How and Why to Define It
The concept of support volume provides a useful framework for considering the integration of various forms of data. Support volume can be defined as the volume of material that is sampled by, or influences, a measurement. The support volume of a measurement can have a significant effect on the determined magnitude of a measured property, and on the determined correlation structure of a property. Data acquired with geophysical methods can be used to obtain information about the magnitude and correlation structure of subsurface properties. It is important, therefore, when using geophysical data with other forms of measurement for subsurface characterization, to define the support volume of the geophysical measurement. The support volume of the geophysical measurement is best defined as the volume of the subsurface to which we can confidently assign a single effective geophysical property. The geophysical support volume is therefore determined by both the physics of the measurement and the processing and/or inversion of the data. In this way the "support volume" of a geophysical measurement combines both measurement and model resolution. At present we do not rigorously quantify, or effectively communicate, the support of our geophysical data. Geophysical models are often presented using discretizations that stem from computational convenience rather than a fundamental understanding of the resolving power of the measurement. An added complication is the non-linear nature of the geophysical support volume, the size and shape of which will be influenced by subsurface properties. The support volume of geophysical data plays a critical role in 1) transforming the geophysical property to the subsurface property of interest, and 2) extracting information about the correlation structure of the subsurface property. We need to develop and adopt a quantitative approach that can be used to define the support volume of geophysical measurement.
NS23A-05 INVITED
Effects of scale and spatial process organisation in hydrogeology
One of the fascinating features of hydro-geological processes is their astounding spatial variability and spatial organisation at all scales. Organisation relates to the spatial arrangement of media characteristics and flow properties including continuity, zones with boundaries between them, the presence of preferential pathways, self-similar organisation and extremes or outliers that occur more often than would be expected based on standard statistical distributions. These organised patterns are linked to the processes that drive and modulate them. Conversely, this heterogeneity may introduce measurement biases and make it difficult to interpret and reconcile data collected at different scales. In this paper I will discuss the concepts of process scales, sampling scales and model scales; and the concept of a scale triplet consisting of spacing, extent and support. This will provide a framework of scale effects in sampling and modelling. The effects will be demonstrated quantitatively for the example of spatially correlated random fields, i.e., when spatial continuity dominates the heterogeneity characteristics. Other types of organisation will also be reviewed and how their scale effects differ from the basic case of random fields. The implications of these scale effects for sampling design will be discussed as will be the implications for modelling spatial flow processes in the subsurface. The degree of non-linearity in the flow processes is critically important for the extent to which data collected at different scales can be interpreted and reconciled. Predictability will be limited if threshold processes exist due to the limited level of detail at which data can be collected. I will emphasise the value of observed spatial patterns vis a vis the sampling of point data. Pattern information is particularly important if spatial organisation is present in the subsurface.
NS23A-06
Predicting Spatial Distribution of Soil Texture with Electromagnetic Induction Mapping and Terrain Analysis Models in Small Watersheds
Spatial pattern modeling of catchment hydrological processes is limited by the availability of time-sensitive high resolution maps of subsurface architecture. Electromagnetic induction (EMI) instruments are gaining wider use for this purpose due to their non-destructive nature, rapid response and ease of integration into mobile platforms. Real-time measurements can infer soil spatial heterogeneity at the small watershed scale. From EMI measurements the soil apparent electrical conductivity (ECa) can be calculated and calibrated to a number of soil properties including: soil salinity, moisture and clay content. The objective of the study is to: 1) infer the textural properties of a watershed through EMI mapping, and 2) compare the topography with the textural distribution using terrain analysis models. The DUALEM 1-S ground conductivity meter along with a Trimble ProXT GPS unit were used to make non-invasive geo-referenced EMI measurements of the 36 ha Reynolds Mountain East watershed on the south side of the larger Reynolds Creek Experimental Watershed in southwestern Idaho. The geo-referenced ECa readings were input into a salinity modeling statistical software package (ESAP) in order to generate an optimal soil sampling plan. Based on this plan, 20 soil samples were obtained and analyzed for soil moisture content, electrical conductivity of the saturation paste extract (ECe) and particle size for clay percentage determination. ESAP was used to estimate the theoretical strength of correlation between ECa and ECe, clay percentage and gravimetric soil moisture content. Terrain analysis software (TauDEM and ArcHydro) were used to evaluate digital elevation models (DEMs) in inferring the influence of topography on the observed field-scale patterns. The results indicate a strong link between clay percentage and the major flow paths due to the movement of finer particles into low lying areas. EMI mapping in conjunction with ESAP statistical sampling analysis provides high spatial resolution soil texture parameters that can be used for modeling watershed hydrological processes.
<a href='http://soilphysics.usu.edu/research/EMI-Texture'>http://soilphysics.usu.edu/research/EMI- Texture</a>