Hydrogeophysics and remote sensing for the design of hydrogeological conceptual models in hard rocks – Sardón catchment (Spain)
Introduction
Although groundwater resources in hard rock aquifers are generally limited in term of productivity, they are strategically important in many regions of the world because they constitute a unique source of water supply for population and agriculture (Cook, 2003, Singhal and Gupta, 2010). Hard rock aquifers are characterized by high heterogeneity, which leads to difficulties in groundwater prospecting, boreholes implementation and water resources management. This heterogeneity exhibit a complex pattern that results from the interaction of factors such as mineralogy and texture of lithologies, regional and local tectonics, and paleoclimate. An overall layout of the general conceptual model of hard rock aquifers, both from horizontal extent and depth-wise structure, was described by e.g. Lloyd (1999), Dewandel et al. (2006) and Lachassagne et al. (2011). Its description includes from top to bottom: (i) an upper weathered layer, so-called saprolite, that has typically a storage function; (ii) an underlying fissured layer that has a transmissive function; and (iii) a fresh basement composed of massive, unaltered rocks with low primary and secondary porosity. The mapping of such structures and retrieval of their hydraulic properties is essential to design a hydrogeological conceptual model of hard rock aquifers.
A hydrogeological conceptual model is a pictorial representation of a groundwater flow system that summarizes available geological and hydrogeological information of a study area (Anderson and Woessner, 1992). Its purpose is to help hydrogeologists to understand the behavior of a hydrogeological system and to support quantitative modelling. The more complex the geological setting, the more important is a hydrogeological conceptual model. Therefore, particularly in hard rock aquifers, the design of a reliable hydrogeological conceptual model is a critical step in quantitative hydrogeological system assessment typically carried out by groundwater modelling, as it strongly conditions the reliability of such models.
Besides horizontal and depth-wise heterogeneities, another important challenge in hydrogeological studies of hard rocks is the scale dependence of aquifer parameters. Based on worldwide dataset of different fractured rocks, Illman (2006) observed an asymptotic increase of permeability from laboratory to regional scale. Sánchez-Vila et al. (1996), Neuman and Federico (2003) and de Marsily et al. (2005) presented a theoretical framework to explain this observation in different rock types. They pointed out the relevance of spatial organization and connectivity of low and high hydraulic conductivity (Κ) zones to explain the scale effect. Dewandel et al. (2012) studied the spatial distribution of aquifer parameters at the catchment scale in a deeply weathered crystalline aquifer in India (Maheswaram catchment). They observed that Κ was relatively homogenous in the fissured zone at the scale of few hundreds of meters to around one kilometer. This observation is remarkable because this range is similar to the cell-size of numerical groundwater models at the catchment scale.
The classical, physically based approach aims to analyze hydrological processes at the local scale and to scale-up the results to the catchment scale (Sivapalan et al., 2003a). However, large scale preferential flow paths in the subsurface may not be observed at the local scale, and thus measurement of properties at the local scale may be not adequate to model the catchment scale hydrodynamics. To account for heterogeneities and scale dependence in hard rock aquifers, the concept of top-down approach (Sivapalan et al., 2003b), also known as downward approach, is more adequate. The downward approach consists of analyzing the hydrology of a catchment by interpreting data obtained at the catchment scale to find patterns in the observed data. It is inherently cross-scale and multi-methods approach. Robinson et al. (2008) presented a review of electrical and magnetic geophysical methods to study the hydrology of watersheds using the downward approach. As a general methodology, they proposed to use airborne electromagnetic system to identify the large-scale dominant structures and associated hydrological processes. The zones of interest were afterwards surveyed using ground-based, local geophysical methods. They also presented study cases in different geological settings where geophysics was used to identify the dominant hydrological processes and to quantify hydrological parameters or variables.
In hard rock aquifers, remote sensing (RS) techniques are frequently used to detect the main hydrogeological features at the catchment scale. Lineament detection based on DTM processing allows to identify the main fault zones, while geomorphological classification supports the mapping of pediments (sub-horizontal hard rock erosion front), inselbergs and weathered areas (Meijerink et al., 2007, Srinivasa Rao et al., 2000). Lachassagne et al. (2001) and Vouillamoz (2003) indicated RS and photo interpretation techniques as particularly suitable to obtain a first characterization of a study area to be subsequently complemented at the local scale by geophysical methods. Additionally, Lachassagne et al. (2001) developed a downscaling methodology based on GIS and multi-criteria analysis to map high-yield zones in the hard rock aquifer of Massif Central (France). The applied methodology integrated terrain parameters, such as lithology type, slope map, thickness of weathered and fissured zones, fracture network information, obtained by RS and DTM analysis, conventional field work and geophysical surveys.
Subsurface data are generally scarce because invasive methods such as borehole drilling and associated aquifer tests are expensive and time-consuming. Hydrogeophysics provides non-invasive, efficient methods of subsurface data acquisition to identify subsurface rock heterogeneities and potential high water yield zones in hard rock aquifers (Krishnamurthy et al., 2008, Lloyd, 1999). Each hydrogeophysical method has its own characteristics and capability with respect to aquifer characterization, so the selection of the appropriate one must be done as a function of the objectives of a survey and geological settings. Hydrogeophysical methods such as geoelectric and electromagnetic have been widely used to retrieve hard rock hydrogeological structures (Dutta et al., 2006, Ramalho et al., 2012) and aquifer parameters using empirical, area-specific relationships (Chandra et al., 2008, Kirsch and Yaramanci, 2009b). Among the hydrogeophysical methods, the magnetic resonance soundings (MRS) has definite advantage for quantitative groundwater assessment because of its most direct relation to in-situ subsurface water (Legchenko et al., 2004, Lubczynski and Roy, 2007). Such direct relation is a result of selective excitation of the water molecule's hydrogen nuclei (1H+) and detection of its corresponding precession signal through nuclear magnetic resonance (NMR). The 1D inversion of the measured MRS data allows defining subsurface layers characterized by thickness, MRS water content and decay time constant. The application of the MRS output to surface-based groundwater evaluation is detailed in several articles, such as Plata and Rubio (2007), Roy and Lubczynski (2003) and Yaramanci (2009). In particular, the aquifer flow parameters, hydraulic conductivity (Κ) and transmissivity (T), are derived from both MRS water content and decay time constant. The storage parameters, i.e. specific yield (Sy) and elastic storativity (Se), are related with the MRS water content (Legchenko et al., 2004, Lubczynski and Roy, 2007, Vouillamoz et al., 2005, Vouillamoz et al., 2007), although recent studies also included the decay time constant (Vouillamoz et al., 2012, Vouillamoz et al., 2014). The MRS sensitivity of water detection was illustrated with a synthetic study of Legchenko et al. (2006) who showed that a 20 m thick layer with water content of 2 % can be detected down to ~ 50 m under low noise conditions (~ 5 nV stacked noise). However, the MRS signal of a deep layer with low water content overlain by a surficial layer with high water content will be attenuated and might not be detected (Legchenko et al., 2006, Vouillamoz et al., 2005). For these reasons, in hard rocks, MRS is suitable to detect groundwater in the saprolite reservoir but the fissured reservoir characterized by low water content is usually hardly detectable (Baltassat et al., 2005, Legchenko et al., 2006, Vouillamoz et al., 2005, Wyns et al., 2004).
In this study, we propose a multi-technique methodology to define the geometry and the hydrogeological parameters of hard rock aquifers as major contribution to the design of hydrogeological conceptual models. The proposed methodology is based on the downward approach and focuses on the integration of RS techniques and hydrogeophysical methods with hydrogeological field data acquisition methods. Our specific objective was to identify the main hydrogeological features such as high and low hydraulic conductivity zones, their spatial distribution and connectivity and characterize them in the context of groundwater flow at the catchment scale. The proposed method is particularly suitable for areas with borehole data scarcity, such as the granitic Sardón Catchment (~ 80 km2, Fig. 1) where this study was realized. That catchment was selected as pilot research area due to hard-rock hydrogeology, semi-arid climate and scarcity of groundwater resources. By applying the proposed method, we revised the former hydrogeological conceptual model of the Sardón catchment (Lubczynski and Gurwin, 2005) with the aim to upgrade the existing numerical groundwater model in a follow up study (not part of this paper), quantifying typically underestimated subsurface fluxes such as groundwater evaporation and groundwater transpiration (Lubczynski, 2011).
Section snippets
Study area
The Sardón catchment (∼ 80 km2, Fig. 1) is located west of Salamanca, in the Castilla y León province (Spain). The terrain elevation ranges from 730 to 860 m a.s.l. (Fig. 2). Geologically, the study area belongs to the Central Iberian Zone of the Iberian Meseta. The Sardón catchment is predominantly covered by anatexic granites of the Hercynian mega-structure known as the Tormes Gneiss Dome (Instituto Geológico and Minero de España, 1991a, Instituto Geológico and Minero de España, 1991b, López, 2004
Methodology
The proposed methodology is based on 3 steps: (1) detection of the main hydrogeological features of the catchment based on remote sensing (RS) techniques of digital image processing; (2) characterization of the main hydrogeological features at the local scale using qualitative and quantitative hydrogeophysics; and (3) drilling of 5 boreholes (25–48 m deep) and performing of slug tests to verify the hydrogeophysical interpretation and to calibrate the MRS parameters. The results from RS and
Lineament mapping
The images obtained by the application of high-pass filters on the high resolution digital terrain model (DTM) from the Centro Nacional de Informacíon Geográfica are presented in Fig. 2. The images of Fig. 2b to g highlight clearly the lineaments that were manually digitized and interpreted as presented in Fig. 2a. By using high-pass filters over DTM, the linear contrasts in terrain elevation of adjacent cells were enhanced. The obtained lineaments corresponded explicitly to geological and
Conclusion
We proposed a multi-technique methodology to contribute to the design of hydrogeological conceptual model in hard rock aquifers. The method is based on a downward approach that combines: (i) remote sensing techniques such as digital image processing on satellite images and digital terrain models; (ii) hydrogeophysic; and (iii) hydrogeological field data acquisition. We applied that method to the hard rock Sardón catchment, contributing to the improvement of the hydrogeological conceptual model
Acknowledgement
This study was funded by the Fundação para a Ciência e a Tecnologia through the Programa Operacional Potencial Humano of the QREN Portugal 2007–2013 (PhD scholarship SFRH / BD / 27425 / 2006) and by ITC Faculty (University of Twente). The authors would like to thank Anatoly Legchenko (LTHE) for the NUMISLite loan during the MRS survey of 2009 and José Martínez Fernández (CIALE) for the support in the field logistic. Finally, we thank the land owners of the Sardón area, in particular Lucidio
References (79)
- et al.
Multi-resolution, object-oriented fuzzy analysis of remote sensing data for GIS-ready information
ISPRS J. Photogramm. Remote Sens.
(2004) - et al.
Contribution of geophysical surveys to groundwater modelling of a porous aquifer in semiarid Niger: an overview
Compt. Rendus Geosci.
(2009) - et al.
Estimation of hard rock aquifers hydraulic conductivity from geoelectrical measurements: A theoretical development with field application
J. Hydrol.
(2008) - et al.
Constraining complex aquifer geometry with geophysics (2-D ERT and MRS measurements) for stochastic modelling of groundwater flow
J. Appl. Geophys.
(2013) - et al.
A generalized 3-D geological and hydrogeological conceptual model of granite aquifers controlled by single or multiphase weathering
J. Hydrol.
(2006) - et al.
A conceptual hydrodynamic model of a geological discontinuity in hard rock aquifers: Example of a quartz reef in granitic terrain in South India
J. Hydrol.
(2011) - et al.
Upscaling and regionalizing hydraulic conductivity and effective porosity at watershed scale in deeply weathered crystalline aquifers
J. Hydrol.
(2012) - et al.
Topsoil thickness prediction at the catchment scale by integration of invasive sampling, surface geophysics, remote sensing and statistical modeling
J. Hydrol.
(2011) Imaging of groundwater with nuclear magnetic resonance
Prog. Nucl. Magn. Reson. Spectrosc.
(2008)Strong field evidence of directional permeability scale effect in fractured rock
J. Hydrol.
(2006)