The specific objective of this study is to further understand and quantify the contribution of lateral heterogeneities (in terms of laterally varying lithologies & elastic properties) within gas reservoirs and in the overburden on small-wavelength fluctuations of the subsidence bowl above depleting gas reservoirs. Here we define small-wavelength fluctuations as all fluctuations with a length-scale smaller than the reservoir-scale.
The results will be beneficial to understand at what depth and scale subsurface heterogeneities are expected to contribute to irregularities in subsidence at surface level. The results will give a scientific base for discriminating between the likely role of deep heterogeneities and processes and shallow heterogeneities and processes to the occurrence of small scale fluctuations in the subsidence profile.
The research was commissioned to TNO and started in 2024. The project was finished and evaluated September 2025.
More ...
In general, in subsidence models the subsurface is schematized as a sequence of laterally continuous, more or less horizontal layers with relatively small variations in thickness. Data from seismic surveys and boreholes show that heterogeneities, such as lateral variations in lithologies, layer thickness and material properties, can exist both at reservoir level and at various depths in the overburden of a gas reservoir, whereas these heterogeneities can have varying length scales. Depending on depth, lateral and vertical dimensions and contrasts in material properties, these heterogeneities may have an impact on the shape of the subsidence bowl at surface level, and may potentially be a (partial) source of small-wavelength fluctuations in subsidence, such as observed at surface level.
Subsidence modelling. (Semi-)analytical models such as AESUBS have been frequently used to model the evolution of subsidence above depleting gas reservoirs. These (semi-)analytical approaches are generally computationally efficient, which enables their use in data-assimilation and model calibration against observations of subsidence, and they have been successfully used for modelling the large scale evolution of the subsidence bowl in time. However, (semi-)analytical approaches can only account for vertical variations in lithologies and properties, but cannot account for lateral variations in lithologies and elastic properties. To model the impact of these lateral subsurface heterogeneities on the occurrence of small / intermediate wave-lengths fluctuations in the subsidence profile, numerical modelling techniques are required.
Subsurface data and subsidence measurements. Information on the various scales of lateral heterogeneities in the subsurface can be obtained from seismic surveys, geological maps and cross sections and borehole data available at www.nlog.nl as well as from the GeoTop model of the shallow subsurface. In general, data on subsurface heterogeneities will be more abundant for the shallow subsurface. Data on characteristic small-wavelengths fluctuations in the subsidence profile above Dutch gas reservoirs, such as the Groningen gasfield can be derived from observations, such as levelling-, GPS- and InSar data.
In the study the following issues have to be addressed:
1. Main research question: What is the effect of lateral heterogeneities in the subsurface on the shape of the subsidence bowl above a depleting, compacting gas reservoir? Can they explain the occurrence of small wavelength fluctuations in subsidence such as observed above Dutch depleting gas fields?
2. What characteristic length-scales of subsurface heterogeneities can we expect at reservoir level and in the overburden of producing gas fields?
3. What is the magnitude of typical small wavelength fluctuations in subsidence as observed above depleting gas fields?
4. How does the magnitude of the fluctuations observed compare to the measurement error? How does the magnitude of the small wavelength fluctuations compare to the uncertainty of the predictions with common layer-cake analytical solutions?
5. What is the impact of a major lithological heterogeneity in the subsurface, such as 1) a visco elastic salt-dome on top of an elastic reservoir, 2) deep subglacial valley deposits, on subsidence above a depleting gas field?
6. What is the impact of intra-reservoir heterogeneity and spatial variation in elastic material parameters on fluctuations in subsidence at surface level? What is the effect of heterogeneities in the overburden on fluctuations in subsidence at surface level? In what way do depth and dimension of the heterogeneity (in terms of height and lateral extent) affect the subsidence profile?
7. What is the first order estimate of expected fluctuations in subsidence caused by groundwater lowering? How do wavelengths of fluctuations in subsidence due to the process of reservoir compaction compare to differential surface movements caused by shallow processes like groundwater extraction?
8. How do numerical modelling results compare to layer-cake analytical solutions, in case of subsurface heterogeneity?
9. How do the results impact policy on the Dutch mining activities?
10. How do the results impact TNO’s technical assessment of production plans?
Numerical models, calibrated with realistic parameters, show that shallow heterogeneities can locally amplify vertical and horizontal ground displacements by 1–9%, consistent with short-wavelength fluctuations observed in InSAR data. Although the displacement amplifications are modest, they can result in increases in horizontal strain up to 30–200%, which is significant in assessing potential damage to buildings and infrastructure in regions where horizontal strain not considering small-scale heterogeneities is already relatively high.
All specific research questions were addressed in the report.
The report suggests a methodology for including lateral elastic heterogeneities in predictions of induced strains. Areas with high initial strain, such as gas field edges and regions above shallow small gas fields and salt caverns, are particularly susceptible and should be prioritized in risk evaluations. Developing maps of horizontal strain amplification could support more accurate predictions.
The KEM-panel has evaluated the project:
The research was conducted by a skilled team of researchers. The investigations were carried out in a rigorous and systematic way based on integrating inSAR observations with numerical 2D and 3D finite difference and finite element simulations subsidence modelling, considering subsidence at Groningen in the period from 2016 to mid-2020. The types of heterogeneities that were investigated were carefully chosen so as to be realistic.
Computer simulations, based on realistic conditions, show that shallow differences in constitution of the subsurface can locally increase how much the surface moves due to gas production, both horizontally and vertically, by about 1 to 9%. This matches the patterns seen in satellite measurements.
Even though these local changes in movement are relatively small, they can have impact on the risk of damage to buildings and infrastructure in some regions like the edges of gas fields or areas above shallow gas fields and salt caverns. Such regions should be a focus in risk assessments. The report outlines a way to include this information in future predictions.