H23H-01 INVITED
Assessing the Reliability of Geoelectric Imaging Results for Permafrost Investigations
The effects of global climate change on mountain permafrost are of increasing concern; warming thaws permafrost, thereby increasing the risk of slope instabilities. Consequently, knowledge of the extent and location of permafrost are important for construction and other geotechnical and land-management activities in mountainous areas. Geoelectric imaging is a useful tool for mapping and characterizing permafrost occurrences. To overcome the generally poor electrical contacts in the active layer, geoelectric surveys usually involve coupling the electrodes to the ground via sponges soaked in salt water. The data are processed and inverted in terms of resistivity models of the subsurface. To monitor the evolution of mountain permafrost, time-lapse geoelectric imaging may be employed. A challenging aspect in geoelectric imaging of permafrost is the very large resistivity contrast between frozen and unfrozen material. Such a contrast makes inversion and interpretation difficult. To assess whether features at depth are required by the data or are artifacts of the inversion process, the reliability of models needs to be evaluated. We use two different approaches to assess the reliability of resistivity images in permafrost investigations: (i) depth of investigation (DOI) and (ii) resolution matrix maps. To compute the DOI, two inversions of the same data set using quite different reference resistivity models are carried out. At locations where the resistivity is well constrained by the data, the inversions yield the same results. At other locations, the inversions yield different values that are controlled by the reference models. The resolution matrix, which is based on the sensitivity matrix calculated during the inversion, quantifies the degree to which each resistivity cell in the model can be resolved by the data. Application of these two approaches to field data acquired in the Swiss Alps and Jura Mountains suggests that it is very difficult to obtain dependable bedrock resistivity information beneath occurrences of massive ice. The reliability tests, which tell us to what depth the resistivity images are trustworthy, help us explain (i) erratic and non-geologic features in the inversion models and (ii) suspicious changes in geoelectric time-lapse imaging. The DOI and resolution matrix techniques do not provide exactly the same information about model reliability. Instead, a combination of the two approaches prevents over-interpretations or misinterpretations of inversion results.
H23H-02 INVITED
Limitations and Perspectives of Electrical Resistivity and Refraction Seismic Tomography in Frozen Ground
Determining the subsurface ice and unfrozen water content in cold regions are important tasks in all kind of cryospheric studies, but especially on perennial or seasonal frozen ground, where little insights can be gained from direct observations at the surface. In the absence of boreholes, geophysical methods are often the only possibility for visualising the subsurface characteristics, and their successful application in recent years included 2D/3D monitoring and even quantifying the ice and unfrozen water content evolution within the subsurface. Due to the strong sensitivity of electrical resistivity, permittivity and seismic velocity to the phase change from unfrozen water to ice, the application of electrical, electromagnetic and seismic techniques has been especially successful. Within these methods, Electrical Resistivity Tomography (ERT) is often favoured due to its comparatively easy and fast data processing, its robustness against ambient noise and its good performance even in cold and irregular environments. Numerous studies have now shown that ERT is principally suitable to map ground ice, differentiate between ice-poor and ice-rich occurrences, monitor freezing, thawing and infiltration processes, and determine the origin of the ice, i.e. a differentiation between buried glacier ice and segregation ice. However, in practice, a number of uncertainties often prohibit a reliable determination of the material composition from ERT surveys alone. Sources of uncertainty are based on the necessity to choose a set of inversion parameters for calculating the specific resistivity distribution from the measured apparent resistivity data set. In addition, high contact resistances at the surface, large topographic gradients, measurement geometry and the presence of fine material or saline pore water can influence the obtained specific resistivity values. Consequently, it is generally advisable to apply a combination of different geophysical methods at each field site. In this contribution we will analyse the reliability of ERT and refraction seismic results regarding the above uncertainties by using multiple inversions and forward modelling techniques for data sets from different permafrost occurrences in the European Alps and Antarctica. Addressing the inherent uncertainty of ERT an inversion ensemble approach is introduced using multiple inversions of the same data set and clustering techniques to obtain the dominant, and therefore most reliable, subsurface features. In addition, combined electrical resistivity and seismic surveys are presented for a detailed quantification of the material composition in frozen ground.
H23H-03 INVITED
Combination of Geophysical Methods for Measuring the Structure of Rock Glaciers
The internal structure of talus-derived rock glaciers was surveyed by a combination of electrical resistivity and refraction seismic measurements to study effects of lithological and thermal situations on the development of rock glaciers. P-wave velocity, direct current (DC) resistivity and/or year-round ground surface temperature were measured on 32 rock glaciers in the Alps and 6 rock glaciers in the Brooks Range, Alaska. In the Alps, two lithological situations were compared: one resulted in bouldery rock glaciers covered with matrix-free boulders and the other pebbly rock glaciers consisting of matrix-supported pebbles and cobbles. Different structure accompanying permafrost thawing was studied through the comparison of vegetated rock glaciers with (non- vegetated) bouldery/pebbly rock glaciers, because such vegetated rock glaciers have been regarded as relict in the Alps. In the Brooks Range located in the continuous permafrost area, bouldery rock glaciers are mostly vegetated but probably contain permafrost. The structural difference of rock glaciers between the Alps and much colder and drier Brooks Range is also discussed. In the Alps, subsurface P-wave velocity differed significantly between the bouldery/pebbly rock glaciers (>2 km/s) and the vegetated rock glaciers (<1 km/s). In addition, the bouldery/pebbly rock glaciers had lower surface temperatures than the vegetated rock glaciers. These results indicate the presence of permafrost in the former and the absence in the latter. DC resistivity differed between the bouldery (10-5000 kohmm) and pebbly (0.5-20 kohmm) rock glaciers, whereas the subsurface resistivity had similar values between the pebbly and vegetated types. The bouldery rock glaciers in the Brooks Range lacked high resistivity core (>100 kohmm). Thus, the difference in the resistivity does not reflect the ground temperature (e.g. frozen or unfrozen) but the structural difference (e.g. bouldery or pebbly). In particular, high resistivity values (>100 kohmm) in rock glaciers probably indicate highly ice-supersaturated structure. Such ice-supersaturation in bouldery rock glaciers in the Alps corresponds to the highest potential of snow burial for this type of rock glaciers, having the late-lying thick snow and large rockwalls occasionally providing rock avalanches.
H23H-04 INVITED
Estimating snow depth, snow water equivalent, and stratigraphy at high resolution using microwave radar
Snow water equivalent (SWE) estimates are a critical component of the hydrological system in cold terrestrial environments, yet our ability to estimate SWE accurately over large areas remains limited. Point measurements of SWE at automatic weather stations (e.g. SNOTEL sites) and manual snowpit measurements are costly, time consuming and difficult to interpolate between, primarily due to the large variability that often exists in snow properties over short distances (< 10 m). Microwave remote sensing offers a promising alternative as large spatial coverage and high temporal resolution can be achieved, however, interpretation of SWE from microwave sensors remains problematic. This is largely due to the difficulty in obtaining ground-truth measurements at a sufficient resolution and scale within air- and space-borne instrument footprints for testing and improving SWE retrieval algorithms. Ground-based microwave radar measurements were used to estimate snow depth, SWE and stratigraphy during recent, intensive, remote sensing ground-truth campaigns in Colorado (NASA Cold Lands Processes Experiment (CLPX) I (2002/03) and II (2006-07)) and Alaska (NASA AMSR-Ice06). These sampling campaigns were coincident with active and passive, air- and space-borne microwave measurements, and covered a wide range of snow conditions. Ground-based radar, due to its relatively small antenna footprint, allows ground-truth measurements of snow depth, SWE, and stratigraphy within the footprint to be well-characterized using traditional manual in-situ methods. These manual measurements, covering a wide range of snow conditions throughout 3 major campaigns are used to quantify the accuracy of estimating snow depth and SWE from the radar measurement. Due to the very high sample rate, these radar-derived snow depth and SWE estimates can be made several orders of magnitude faster than with traditional techniques. These high-resolution data, covering length scales from centimeters to kilometers, provide key information for understanding the spatial distribution and variation of snow depth, SWE and stratigraphy, in a wide range of environments.
H23H-05
Exploring Permafrost with Multi-Channel GPR
The active layer of permafrost soils is typically characterized by a spatially varying depth that is related to surface features, soil thermal properties, and water flow. Generally, a considerable variation of the liquid water content is associated with this spatial structure. The concomitant variation of both, depth of reflector and water content, prevents the quantitative exploration of such environments with traditional single-channel GPR. Indeed, we demonstrate that simple-minded interpretations of single-channel GPR-measurements may lead to qualitatively wrong results. Multi-channel GPR is an attractive solution to this problem since it allows to monitor depth and water content independently and at roughly the same operating speed as single-channel instruments. We demonstrate this for sites in the Aksai Chin region of Western Tibet, China, where we explored the surface topography of the ice table in an alluvial fan and relate it to the subsurface hydrology. As a second example the ice table beneath a road is determined as a prerequisite to assess its impact and stability.
H23H-06
Geophysical imaging of a temperate glacier's hydrologic system in 1, 2, and 3 dimensions
Between 1999 and 2007, we conducted a series of geophysical experiments on Bench Glacier in the Chugach Range of southeast Alaska. Objectives of the experiments include measuring 1) the distribution of water in the snow pack, 2) water storage and routing within the glacier, and 3) the geometry of flow paths at the bed of the glacier. To accomplish these objectives, we have deployed georadar at frequencies ranging from 5 MHz to 1 GHz in 1D, 2D, and 3D single and multi-offset configurations. We have conducted time lapse georadar monitoring of the glacier at the annual, seasonal, daily, and minute time scales. We have employed state of the art data processing and analysis tools such as reflection tomography to produce accurate radar velocity profiles, and 3D coherence cube imaging to identify the 3D distribution of voids and fractures within the glacier. We have conducted passive seismic monitoring, 3D seismic reflection imaging, and shear wave seismic reflection surveying designed to both to image hydrologic structures and to provide a direct measure of the glaciers elastic properties. Here we provide an overview of the project and to present notable findings to date. These findings include 1) that there is lateral heterogeneity in the liquid water content of the snow and significant lateral flow within the snow pack, 2) that Bench Glacier is stratified with an upper layer containing few voids and little water, and a lower layer containing significantly greater water content and many voids, and 3) the void spaces in the lower layer appear to be comprised of both fractures and randomly distributed irregular void space. Video borehole measurements coupled with azimuthal anisotropy in the geophysical measurements indicate that the voids are primarily subvertically oriented, have a preferred azimuthal orientation that is oblique to the glacier flow direction, and are present at scales ranging from mm to 10s of m.