The Geological Survey of the Netherlands (TNO-GSN) provides public information on the subsurface (http://www.dinoloket.nl). Among this information are (hydro)geological models with hydraulic parametrization. GeoTOP is a voxel model of the upper 50 meters with voxels of 100m x 100m x 0.5m. The basis of the hydraulic parametrisation is a large collection of laboratory measurements of the hydraulic conductivity determined for samples from undisturbed cores, which TNO-GSN collects systematically throughout the Netherlands. The values are upscaled from labortatory to voxel scale for each combination of geological unit and lithoclass.
In order to further test and improve the upscaling, TNO-GSN is preparing a field test focussing on the resistance of a basal Holocene clay layer. The field test will consist of pumping from the aquifer below the clay layer combined with infiltration in various infiltration wells, high frequency registration of heads in 23 piezometers, and determination of (the variation of) flow velocities over the thickness of the pumped aquifer. Nine months before the start of the field test, existing piezometers around the site have been equipped with automatic pressure transducers. These record the groundwater heads above and below the clay layer with high frequency in order to determine background variations of the head due to the tide, precipitation and evaporation as well as surface water levels. The measurements will be continued at least until the field test is fully completed.
Exploratory model simulations have shown that 1) a reliable estimate of the resistance requires the measurement of a drawdown above the clay layer, and 2) only a small drawdown will develop above the clay layer even though the pumping will run for several months. This means that a small drawdown due to the pumping has to be separated from the background influences. This is already important during the execution of the field test in order to be able to optimize the pumping. For this purpose, detailed time series modelling of the groundwater heads from the existing piezometers is carried out. Challenges in this modelling are 1) the different time scales of the tide, precipitation, evaporation, and surface water level changes, 2) the inhomogeneity of the time series with years of bi-monthly manual measurements and 9 months of pressure transducer data with 15 minute intervals, and 3) non-linearity in the precipitation and evaporation response due to drainage and the unsaturated zone.
Although, time series modelling is not the appropriate tool for the interpretation of the aquitard resistance from the field test, it is valuable because the time series models can determine the background variations of the groundwater head and allow to check the performance of the test and to tune the pumping.