Automatic identification of fresh–saline groundwater interfaces from airborne electromagnetic data in Zeeland, the Netherlands

Abstract

ABSTRACTIn a setting of predominantly saline surface waters in Zeeland, the Netherlands, the only available shallow fresh groundwater resource is present in the form of freshwater lenses floating on top of saline groundwater. This fresh water is vital for agricultural, industrial, ecological, water conservation and drinking water functions. An essential first step for managing the usable water properly is to know the location of the fresh–saline groundwater interface. Traditional salinity mapping with ground‐based vertical electrical soundings, electrical cone penetration tests or chloride measurements from groundwater samples is time‐consuming and, therefore, expensive to cover large areas. Airborne electromagnetics, which is fast and can cover large areas in short time, is an efficient alternative. Therefore, a consortium of BGR, Deltares and TNO conducted the project FRESHEM Zeeland during 2014–2017. An area of more than 2000 km² was surveyed using BGR's helicopter‐borne geophysical system totalling 9640 line‐km. The helicopter‐borne electromagnetic data, after inversion to resistivity–depth models, served as baseline information for further interpretation. Without information on lithology, however, an accurate discrimination between fresh and saline groundwater applying fixed resistivity thresholds would fail if clayey sediments exist. Therefore, a probabilistic Monte Carlo approach was developed within the FRESHEM project. This approach combines helicopter‐borne electromagnetic resistivities, 3D geological model (GeoTOP), laboratory results (formation factor and surface conductivity) and local in situ groundwater measurements for the translation of resistivity data to chloride concentration. As such detailed information is generally not available, another approach, which uses only helicopter‐borne electromagnetic results to derive the thicknesses of the freshwater lenses from smooth inversion models, is presented in this paper. The corresponding fresh–saline groundwater interfaces are derived from steepest resistivity–depth (log–log) gradients appearing within a certain resistivity range. The bounds of this range are defined by resistivity values, which predominantly correlate with fresh or saline water, nearly independent of the lithology type. The results of this approach are checked using both synthetic and field data. The latter are compared with electrical cone penetration test measurements, the common threshold approach and the 3D chloride distribution of the FRESHEM results, particularly in an evaluation area where the transition of fresh to saline groundwater is relatively sharp. The fresh–saline groundwater interfaces derived by all methods are quite similar on average with deviations in the order of a metre. Locally, however, greater deviations occur, particularly close to the coast and along creek ridges or dunes, where the elevation of the fresh–saline groundwater interface varies notably.