Geological Survey of Canada

Airborne time-domain electromagnetic methods (AEM) are useful for hydrogeological mapping due to their rapid and extensive spatial coverage and high correlation between measured magnetic fields, electrical conductivity, and relevant hydrogeological parameters. However, AEM data, preprocessing and modelling procedures can suffer from inaccuracies that may dramatically affect the final interpretation. We demonstrate the importance and the benefits of advanced data processing for two AEM datasets (AeroTEM III and VTEM) collected over the Spiritwood buried valley aquifer in southern Manitoba, Canada. Early-time data gates are identified as having significant flightdependent signal bias that reflects survey flights and flight lines. These data are removed from inversions along with late time data gates contaminated by apparently random noise. In conjunction with supporting information, the less-extensive, but broader-band VTEM data are used to construct an electrical reference model. The reference model is subsequently used to calibrate the AeroTEM dataset via forward modelling for coincident soundings. The procedure produces calibration factors that we apply to AeroTEM data over the entire survey domain. Inversion of the calibrated data results in improved data fits, particularly at early times, but some flight-line artefacts remain. Residual striping between adjacent flights is corrected by including a mean empirical amplitude correction factor within the spatially constrained inversion scheme. Finally, the AeroTEM and VTEM data are combined in a joint inversion. Results confirm consistency between the two different AEM datasets and the recovered models. On the contrary, joint inversion of unprocessed or uncalibrated AEM datasets results in erroneous resistivity models which, in turn, can result in an inappropriate hydrogeological interpretation of the study area.

A new time-domain design implementation for the VTEM helicopter time-domain EM system, known as “Full-waveform” VTEM, addresses the early time issues that have been limiting its shallow mapping capability for near-surface applications. The Full-waveform design implementation consists of a combination of a) streamed half-cycle recording of transmitter and receiver waveform data, as well as b) continuous system calibration corrections, b) parasitic-noise and transmitter-drift corrections, and d) ideal-waveform-deconvolution corrections that are applied in a separate post-processing step. This leads to an improvement in usable early time data from ~100usec in standard VTEM systems to ~20usec for Full waveform VTEM. This results in a vastly improved near-surface hydrogeologic characterization. The Full-waveform system and theory were previously described by Legault et al. (2012). This paper presents VTEM case-study examples with emphasis on near-surface applications. Introduction

We develop a 3D geological modelling procedure supported by the combination of helicopter time-domain electromagnetic data, seismic reflection data, and water well records for the Spiritwood buried valley aquifer system in Manitoba, Canada. Our procedure is an innovative hybrid of knowledge-driven and data-driven schemes that provides a clear protocol for incorporating different types of geophysical data into a 3D stratigraphic model framework. The limited spatial density of water well bedrock observations precludes detection of the buried valley bedrock topography and renders the water well records alone inadequate for accurate hydrogeological model building. The expert interpretation of the geophysical data allows for leveraging of a spatially extensive dataset with rich information content that would be otherwise difficult to utilize for lithostratigraphic classification.

Buried valleys are important hydrogeological structures in Canada and other glaciated terrains, providing sources of groundwater for drinking, agriculture and industrial applications. Hydrgeological exploration methods such as pumping tests, boreholes coring or ground-based geophysical methods (seismic and electrical resistivity tomography) provide limited spatial information and are inadequate to efficiently predict the sustainability of these aquifers at the regional scale. Airborne geophysics can be used to significantly improve geological and hydrogeological knowledge on a regional scale. There has been demonstrated success at using airborne electromagnetics for mapping and characterization of buried valleys in different geological contexts (Auken et al., 2008; Jørgensen et al., 2003; Jørgensen et al., 2009; Steuer et al., 2009). Despite the fact that both electromagnetic surveys and reflection seismic profiling are used extensively in hydrogeological mapping, integration of the methods is a relatively unexplored discipline (Høyer et al., 2011). The Spiritwood Valley is a Canada-USA trans-border buried valley aquifer that runs approximately NW – SE and extends 500 km from Manitoba, across North Dakota and into South Dakota (Winter et al., 1984). The Spiritwood aquifer system consists of glacially deposited silt and clay with sand and gravel bodies, infilling a broad north-south trending valley that has been identified primarily based on water wells information (Wiecek, 2009). The valley is incised into bedrock consisting of fractured siliceous shale. As part of its Groundwater Geoscience Program, the Geological Survey of Canada (GSC) has been investigating buried valley aquifers in Canada using airborne and ground-based geophysical techniques. To obtain a regional three-dimensional assessment of complex aquifer geometries for the Spiritwood, both geophysical and geological investigations were performed with the aim to develop an integrated conceptual model for a quantitative description of the aquifer system. In 2010, the Geological Survey of Canada conducted an airborne electromagnetic (AeroTEM III) survey over a 1062 km2 area along the Spiritwood Valley, north of the US border (Oldenborger 2010a, 2010b). AEM inversion results show multiple resistive valley features inside a wider, more conductive valley structure within the conductive bedrock (Fig. 1). Furthermore, the complexity of the geometries, spatial distribution and size of the channels is evident. Other ground based data collected in the survey area make it possible to provide some constraints on the AEM resistivity model. Downhole resistivity logs were collected that provide information on the electrical model relative to the geological layers (Crow et al., 2012). In addition, over 10 line-km of electrical resistivity data and 42 km of high resolution landstreamer seismic reflection data (Figs. 2a, 2b) were collected at selected sites (Oldenborger et al., 2012). In this short paper we present results obtained from the data inversion and an example of integration of ancillary seismic data into the AEM inversion. In particular, the elevation to a layer (shale bedrock elevation) as interpreted from seismic is added to the inversion to constrain the resistivity model.

Buried valleys are important hydrogeologic features of glaciated terrains. They often contain valuable groundwater resources; however, they can remain undetected by borehole-based hydrogeologic mapping or prospecting campaigns. Airborne electromagnetic (AEM) surveys provide high-density information that can allow detailed features of buried valleys to be efficiently mapped over large geographic areas. Using AEM data for the Spiritwood Valley Aquifer system in Manitoba, Canada, we developed a 3D electric property model and a geologic model of the buried valley network. The 3D models were derived from voxel-based segmentation of electric resistivity obtained via spatially constrained inversion of two separate helicopter time-domain electromagnetic data sets (AeroTEM and versatile time-domain electromagnetic [VTEM]) collected over the survey area. Because the electric resistivity do not provide unequivocal information on subsurface lithology, we have used a cognitive procedure to interpret the electric property models of the aquifer complex, while simultaneously incorporating supporting information for the assignment of lithology in the 3D geologic model. For the Spiritwood model, supporting information included seismic reflection data and borehole records. These data constrained valley geometry and provided lithologic benchmarks at specific borehole sites and along seismic transects. The large-scale AeroTEM survey provided the basis for modeling the regional extent and connectivity of the Spiritwood Valley Aquifer system, whereas the local-scale VTEM survey provided higher near-surface resolution and insight into a detailed shallow architecture of individual buried valleys and their fill.

Early time or high frequency airborne electromagnetic data (AEM) are desirable for shallow sounding or mapping of resistive areas but this poses difficulties due to a variety of issues, such as system bandwidth, system calibration and parasitic loop capacitance. In an effort to address this issue, a continued system design strategy, aimed at improving its early-channel VTEM data, has achieved fully calibrated, quantitative measurements closer to the transmitter current turn-off, while maintaining reasonably optimal deep penetration characteristics. The new design implementation, known as “Full Waveform” VTEM was previously described by Legault et al. (2012). This paper presents some case-study examples of a Full Waveform helicopter time-domain EM system for near-surface applications

Helicopter time-domain electromagnetic (HTEM) surveys often suffer fromsignificant inaccuracies in the early-time or near-surface data—a problem that can lead to errors in the inversemodel or limited near-surface resolution in the event that early time gates are removed. We present an example illustrating the use of seismic data to constrain the model recovered from an HTEM survey over the Spiritwood buried valley aquifer in Manitoba, Canada. The incorporation of seismic reflection surfaces results in improved near-surface resistivity in addition to a more continuous bedrock interface with a sharper contact. The seismic constraints reduce uncertainty in the resistivity values of the overlying layers, although no a priori information is added directly to those layers. Subsequently, we use electrical resistivity tomography (ERT) and borehole data to verify the constrained HTEM models. Treating the ERT and borehole logs as reference information, we perform an iterative time-shift calibration of the HTEM soundings to achieve regional-scale consistency between the recovered HTEM models and the reference information. Given the relatively small time-shifts employed, this calibration procedure most significantly affects the early-time data and brings the first useable time gate to a time earlier than the nominal first gate after ramp off. Although time shifts are small, changes in the model are observed from the near-surface to depths of 100 m. Calibration is combined with seismic constraints to achieve amodel with the greatest level of consistency between data sets and, thus, the greatest degree of confidence. For the Spiritwood buried valley, calibrated and constrained models reveal more structure in the valley-fill sediments and increased continuity of the bedrock contact.

In glaciated terrain, buried valleys often host aquifers that are significant groundwater resources. However, given the range of scales, spatial complexity and depth of burial, buried valleys often remain undetected or insufficiently mapped. Accurate and thorough mapping of bedrock topography is a crucial step in detecting and delineating buried valleys and understanding formative valley processes. We develop a bedrock mapping procedure supported by the combination of seismic reflection data and helicopter time-domain electromagnetic data with water well records for the Spiritwood buried valley aquifer system in Manitoba, Canada. The limited spatial density of water well bedrock observations precludes complete depiction of the buried valley bedrock topography and renders the water well records alone inadequate for accurate hydrogeological model building. Instead, we leverage the complementary strengths of seismic reflection and airborne electromagnetic data for accurate local detection of the sediment-bedrock interface and for spatially extensive coverage, respectively. Seismic reflection data are used to define buried valley morphology in cross-section beneath survey lines distributed over a regional area. A 3D model of electrical conductivity is derived from inversion of the airborne electromagnetic data and used to extrapolate buried valley morphology over the entire survey area. A spatially variable assignment of the electrical conductivity at the bedrock surface is applied to different features of the buried valley morphology identified in the seismic cross-sections. Electrical conductivity is then used to guide construction of buried valley shapes between seismic sections. The 3D locus of points defining each morphological valley feature is constructed using a path optimization routine that utilizes deviation from the assigned electrical conductivities as the cost function. Our resulting map represents a bedrock surface of unprecedented detailwith more complexity than has been suggested by previous investigations. Our procedure is largely data-driven with an adaptable degree of expert user input that provides a clear protocol for incorporating different types of geophysical data into the bedrock mapping procedure.

In the past three decades, airborne electromagnetic (AEM) systems have been used for many groundwater exploration purposes. This contribution of airborne geophysics for both groundwater resource mapping and water quality evaluations and management has increased dramatically over the past ten years, proving how these systems are appropriate for large-scale and efficient groundwater surveying. One of the major reasons for its popularity is the time and cost efficiency in producing spatially extensive datasets that can be applied to multiple purposes. In this paper, we carry out a simple, yet rigorous, simulation showing the impact of an AEM dataset towards hydrogeological mapping, comparing it to having only a ground-based transient electromagnetic (TEM) dataset (even if large and dense), and to having only boreholes. We start from an AEM survey and then simulate two different ground TEM datasets: a high resolution survey and a reconnaissance survey. The electrical resistivity model, which is the final geophysical product after data processing and inversion, changes with different levels of data density. We then extend the study to describe the impact on the geological and hydrogeological output models, which can be derived from these different geophysical results, and the potential consequences for groundwater management. Different data density results in significant differences not only in the spatial resolution of the output resistivity model, but also in the model uncertainty, the accuracy of geological interpretations and, in turn, the appropriateness of groundwater management decisions. The AEM dataset provides high resolution results and well-connected geological interpretations, which result in a more detailed and confident description of all of the existing geological structures. In contrast, a low density dataset from a ground-based TEM survey yields low resolution resistivity models, and an uncertain description of the geological setting.