We present a 3-D lithospheric-scale model covering the area of Germany that images the regional characteristics of the structural configuration and of the thermal field. The structural model resolves major sedimentary, crustal and lithospheric mantle units integrated from previous studies of the Central European Basin System, the Upper Rhine Graben and the Molasse Basin, together with published geological and geophysical data. A combined workflow consisting of 3-D structural, gravity and thermal modelling is applied to derive the 3-D thermal configuration. The modelled temperature distribution is highly variable in response to an imposed heterogeneous distribution of thermal properties assigned to the different units. First order variations in the temperature field are mainly attributed to the thermal blanketing effect from the sedimentary cover, the variability in the amount of radiogenic heat produced within the different crystalline crust compartments and the implemented topology of the thermal Lithosphere-Asthenosphere Boundary.
Being a key topic for the present-day scientific and industrial community, the global climate change leaves us no choice but developing a strategy of provision of renewable energy resources, such as geothermal. Geothermal energy is transported by conduction and convection from deeper parts of the earth towards the surface and can be extracted by geothermal power and heating plants from natural and/or engineered reservoirs. Using such energy requires knowing the temperature distribution in the light of the causative processes, the latter being influenced by the tectonic, geological and hydrogeological setting of the target area. Knowledge about reservoir's temperatures can be obtained directly by costly drilling, that only provides local information on temperature at a specific depth without any knowledge of the relevant physical processes. For regional exploration, complementary workflows use integrated 3-D structural and physics-based models that help to predict the temperature distribution in the subsurface by taking into account the heterogeneous structural configuration as well as the, often non-linear, causative processes (e.g. Cacace et al., 2010; Scheck-Wenderoth et al., 2014).
Though several regional models have focussed on different regions of Germany (Sippel et al., 2013; Scheck-Wenderoth et al., 2014; Przybycin et al., 2015; Freymark et al., 2017), a consistent subsurface structural and thermal model for the whole territory of Germany is still missing. Here, we integrate all available information of these studies into a consistent 3-D model of Germany referred to as 3-D-Deutschland (3-D-D) hereafter, that provides the background to be further used in future regional-to-local investigations. Therefore we integrate three regional structural models based on data comprising borehole measurements, seismic profiles, isopach and geological maps and constrained by gravity and thermal data (Fig. 1): (1) the Central European Basin System (CEBS, Maystrenko and Scheck-Wenderoth, 2013), (2) the Molasse Basin (MOLA, Przybycin et al., 2014) and (3) the Upper Rhine Graben (URG, Freymark et al., 2017).
Geographical map of the regional model boundaries and seismic profiles used to solve discrepancies. The 3-D-D model area is shown in red and spans 1000 km in North-South and 643 km in East-West direction. Map coordinates in this and in all following figures are given in km in UTM Zone 32N coordinate system.
The applied workflow comprises (1) structural modelling to correlate the lithostratigraphic units derived from the different input models, followed by (2) a validation of the derived configuration with 3-D gravity modelling and, lastly, (3) the calculation of the conductive thermal field. The main challenge in joining the three existing models was to identify layers of corresponding lithostratigraphy and to merge these for the whole 3-D-D model area (Fig. 1, Table 1). Therefore, we revisited published seismological data to overcome inconsistencies across the boundaries of the input models.
Dominant lithologies, densities and thermal properties assigned to layers of the 3-D-D model for gravity and conductive thermal steady-state calculations: (1) Scheck-Wenderoth et al. (2014); (2) Przybycin et al. (2015); (3) Freymark et al. (2017); (4) Norden et al. (2008); (5) Vilà et al. (2010). Final validated parameter values are marked bold; (*) stars indicate values adjusted in response to the 3-D effects of the larger volumes of certain units in the 3-D-D model compared to the input models.
The data used for this study include the ETOPO1 Global Relief Model (Amante and Eakins, 2009), the configurations of the three regional models and the seismic profiles of the deep seismic experiments EUGENO-S (EUGENO-S Working Group, 1988), the DEKORP (e.g. Meissner and Bortfeld, 2014), the EGT (Blundell et al., 1992) and the ALP2002 (Brückl et al., 2007). These profiles helped to constrain the structural configurations across the boundaries of the regional models (Fig. 1). For the gravity modelling stage we rely on the global combined reference gravity field model EIGEN-6C4 (Förste et al., 2014; Ince et al., 2019). Geological depth maps (Bayrisches Staatsministerium für Wirtschaft, 2004) and thickness maps of Boigk and Schöneich (1974) were also used to correlate boundaries of the Triassic across the three models (Lechel, 2017). Results of the Eastern Alpine Seismic Investigation (EASI) project (Hetényi et al., 2018) were integrated to refine the Mohorovičić discontinuity (Moho) in the SE part (Fig. 1) of the model. To derive a consistent Lithosphere-Asthenosphere boundary (LAB) we integrated results from the three input models, validating the interpolated boundary with lithospheric thicknesses derived from receiver functions data (Geissler et al., 2010).
The
Thicknesses of the major structural layers of the 3-D-D model:
The
The
Consistent merging of the stratigraphic surfaces of the original input
models proved to be a non-trivial task due to the differing resolution and
related uncertainties of the marginal domains caused by insufficient data
coverage. In order to resolve geometric conflicts among the three models and
to integrate additional available data, all surface grids of the original
models were first correlated according to their lithostratigraphic sequences
and then loaded into Petrel. Discrepancies between the units were removed by
revisiting seismic profiles across the boundaries of the regional models
(see Sect. 2.1.1; Fig. 1). After integration, data for each unit were
interpolated using the convergent interpolation algorithm of Petrel into a
regular grid with 1 km spacing (Lechel, 2017). 3-D gravity modelling was
carried out using IGMAS
Thicknesses of the major structural layers of the 3-D-D model:
To assess the distribution of deep temperatures resulting from the
structural configuration of the 3-D-D model, we calculate the present day
conductive thermal field under steady-state conditions. For that purpose we
use GOLEM (Jacquey and Cacace, 2017; Cacace and Jacquey, 2017) – a 3-D
thermal-hydraulic-mechanical simulator based on a Galerkin finite-element
technique. Consistently with all previous studies (Scheck-Wenderoth et al.,
2014; Przybycin et al., 2015; Freymark et al., 2017), we assign uniform
lithology-dependent thermal properties to each resolved stratigraphic unit
(Table 1). As an upper boundary condition we consider the spatially variable
annual average surface temperature (DWD, 2019) onshore, together with
constant 4
The resulting 3-D-D model comprises 31 lithostratigraphic units (Table 1): seawater, 14 sedimentary units, 14 crystalline crustal units and 2 lithospheric mantle units. The thicknesses of cumulative sedimentary, crustal and mantle layers are shown in Fig. 2a–d.
The top surface of the
The
Modelled temperatures at:
Depths of the isotherms:
The next-deeper major layer, the
The
The
Comparison of thermal field of 3-D-D model at 1 km b.s.l. with
The thickness of the
The
The predicted temperatures vary in response to the heterogeneous
distribution of the thermal properties associated with the different
lithological units (Table 1). The highest shallow temperatures (Fig. 4a–d)
are predicted within the Upper Rhine Graben, though elevated shallow
temperatures are also modelled for the CEBS and the Molasse Basin. The
coldest shallow domains are predicted in areas where the crystalline crust crops out or is close to the surface. Thus the short-wavelength variations of shallow temperatures are mainly influenced by the blanketing effect of a less conductive sediments compared to a more conductive crystalline rocks or rock salt. The modelled surface heat flow varies in the range of 60–100 mW m
Accordingly, temperatures of 100
In order to quantify the consistency of the newly derived 3-D-D model with the
three input models (that were properly validated with measured temperatures)
we analysed the temperature differences at 1 km b.s.l (Fig. 6a–c). The
models are consistent concerning the regional pattern and the range of
temperature variations. The higher misfits (up to 30
Cross-plots of the temperature values in the coinciding grid nodes together with their linear fit (inset plots in Fig. 6a–c) reveal that the most of the discrepancies are found for the MOLA model, mainly in the Alpine area and along the boundary with the URG model (Fig. 6b). In the CEBS part (Fig. 6a) main discrepancies are on the margins of the area and can be explained by a smaller extent of the 3-D-D model (Fig. 1) leading, e.g. to the lack of thermal impact from the unconsidered shallow LAB in the NW part of the larger CEBS model (Maystrenko and Scheck-Wenderoth, 2013). Discrepancies in the URG area form two linear trends in the cross-plot (inset plot in Fig. 6c) corresponding to the areas of negative and positive temperature misfit, although the linear fit is the best of all three models. A proper comparison to the results of the statistical interpolation of measured temperatures by Agemar et al. (2012) would be beneficial, however these measured temperatures are not freely available and the temperature profiles in Agemar et al. (2012) don't display calibration wells. Moreover, there is a high probability that interpolation over salt structures predicts wrong temperature distributions: temperatures from wells penetrating the salt show a far larger chimney effect, which was nicely demonstrated by Fuchs and Balling (2016). Therefore, model validation is possible only in terms of regional pattern of subsurface temperature variations and ranges of temperatures at certain depths, while a more detailed calibration would require local models of higher structural resolution that consider also the effects of coupled heat and fluid transport.
It is clear that the derived model is built on assumptions that introduce
uncertainties. First of all, the assumption of steady state may not be valid
and neglecting the effects of previous glaciations or rifting phases would
over- or underestimate shallow temperatures. Sensitivity studies have shown
that such effects would influence the absolute values of temperatures but
not their regional distribution pattern (Majorowicz and Wybraniec, 2011).
Neglecting the influence of coupled fluid and heat transport may also lead
to erroneous conclusions as the conductive thermal model that best
reproduces the observed heat flow considers an “effective” thermal
conductivity that contains the superposed effects of different heat
transport processes (Cacace et al., 2010). Again, local studies would be
required to better quantify the absolute impact of this simplification. The
quality and quantity of observations decreases with depth. The indirect
methods of deriving the depth to the thermal LAB come with their own
limitations and future improvement would require more densely spaced passive
seismic experiments. Comparing the results for our model with seismically
constrained LAB and a model with the LAB assumed at a constant depth of 120 km, we estimated the impact on temperature variations in the upper 3 km in
the range of
The derived lithospheric-scale 3-D-D model resolves the first order trends in structure, density and temperature configuration of all of Germany and can serve as a data-consistent background for smaller-scale structural, geothermal and stress field studies, both in academy and industry. It demonstrates how first order variations in the structurally controlled distribution of thermal properties influence the regional thermal field. Apart from providing first order deep temperature variations, the model also provides a basis for rheological modelling that will help to relate observed seismicity to strength distribution. It can be a starting point for refinement of local models of a higher spatial resolution if a denser data coverage is available.
The developed 3-D-D structural model has been published with GFZ Data Services (see Anikiev et al., 2019) and is publicly accessible.
AL in collaboration with JB and MSW prepared the initial structural model by compiling and correlating the interfaces of the three input models. JB assisted in revision of the structural and density model in correspondence with gravity simulations. MSW, MC and MLGD contributed to discussion of the thermal modelling results and limitations of the derived model. DA revised the initial structural model, performed all gravity and thermal simulations, made comparison analysis and prepared the figures and the manuscript with contributions from all co-authors.
The authors declare that they have no conflict of interest.
This article is part of the special issue “European Geosciences Union General Assembly 2019, EGU Division Energy, Resources & Environment (ERE)”. It is a result of the EGU General Assembly 2019, Vienna, Austria, 7–12 April 2019.
We are grateful to Hans-Jürgen Götze and Sabine Schmidt for
permission to use the 3-D modelling software IGMAS
The article processing charges for this open-access publication were covered by a Research Centre of the Helmholtz Association.
This paper was edited by Christopher Juhlin and reviewed by Niels Balling and one anonymous referee.