Depth Conversion of the 3D Seismic Interpretation for the Area of Interest in the Jura Ost, Nördlich Lägern and Zürich Nordost Siting Regions
Nagra (2024): Depth Conversion of the 3D Seismic Interpretation for the Area of Interest in the Jura Ost, Nördlich Lägern and Zürich Nordost Siting Regions. Nagra Arbeitsbericht NAB 24-49.
In the course of the repository site selection process, 3D seismic evaluation for the Jura Ost (JO), Nördlich Lägern (NL) and Zürich Nordost (ZNO) siting regions was carried out on the selected pre-stack time migration zero-phase reference seismic datasets (PreSTM-ref, Nagra 2024a, 2024b, 2024c, 2024d).
The interpretation results represent important input for the 3D geological models (Nagra 2024f) and directly and indirectly for the repository performance assessment (Nagra 2024g) and for radiological consequence analysis (Nagra 2024h). This implies a conversion of the interpreted horizons from time to depth domain.
Depth conversion was therefore performed together with an evaluation of depth uncertainties. The latter was carried out to explain the non-uniqueness of models built with impact on the top and base surfaces of the host rock (Opalinus Clays) and on thickness of the host rock, and upper and lower confining units[1].
Within the areas of interest (potential repository zones), an infill stage of interpretation was carried out. The number of stratigraphic horizons were increased from the regional (full survey) horizons to multiple local horizons, in order to generate a higher granularity in the stratigraphic framework and to identify and map velocity change interfaces.
Since critical anomalies were observed in the velocity model associated with the PreSDM seismic vintage, which impaired the confidence on the resulted depth surfaces, and to enable comparability in the approach and analysis of the three siting regions, a layered form of velocity model, i.e. Vo/K, approach, was used.
Several realisations were tested and evaluated exhausting all the available velocity inputs. Depth residuals, i.e. the difference between the computed depth of an event and the actual depth encountered in the well, were used as first level diagnostic for the validity of the velocity model for time depth conversion.
Among the several realisations, a reference case (solution of choice), generally referred to as base case, has to be selected.
There are no globally established criteria to define the base case. Considering the comparability scope of the site investigation, the base case was selected following the three established criteria below:
- Use of comparable seismic products between sites: PreSTM-ref is the most comparable seismic processing product among the three surveys, thus the resulting surfaces from the PreSTM-ref interpretation had to be used
- Conservative approach: selection of the interpretation of the base host rock (Top Lias- 174.7 Toa-100), which is by tendency the shallowest (smaller thickness of host rock)
- Technically acceptable velocity model: model is consistent with data and captures relevant heterogeneity within the data.
Once the base cases for the top and base host rock (170.9 Aal 300 - Top Opalinus and 174.7 Toa 100 -Top Lias, respectively) were defined, depth uncertainties were evaluated by a three-step approach: (i) the 2-standard deviation depth residuals accounting for the minimum value of depth uncertainty and (ii) the depth differences between the selected depth conversion scenarios, providing depth uncertainty lateral distribution and (iii) tapering the uncertainty at well locations down to 2 m (i.e. almost no uncertainty at wells).
In the NL siting region, the depth differences between the Quick Look Processing (QLP) and the PreSTM-ref is relatively small, and it is generally below the 25 m standard deviation of two (2Stdev), except from a small area to the north, where uncertainties are around 45 m. In larger areas of the JO and ZNO siting regions, values are instead up to 45 m and < 60 m, respectively, even after tapering to wells.
Since the uncertainty maps were generated by the difference of PreSTM-ref and QLP depth converted surfaces, the uncertainty values appear to strongly depend on the different static solutions implemented in the two seismic vintages and should be interpreted as a conservative estimation.
The different statics applied in the two time migration data sets and the near-surface velocity model in the depth migration data sets are the largest contributor to depth uncertainty.
In addition, thickness uncertainties were assessed for the host rock and, for the upper and lower confining units[1]. The thickness uncertainty is related to the uncertainty associated with the velocity of the interval.
In the NL siting region, the upper confining unit has a high velocity variability related to the presence of the fast Herrenwis carbonate platform penetrated by two of the four wells. Consequently, the thickness uncertainty value was derived by the average of the uncertainty values calculated separately for the wells penetrating the clastic sequence and for those penetrating the carbonate interval.
In the ZNO siting region, the thickness uncertainty for the upper confining unit might underestimate the uncertainty away from the well-controlled areas, especially when compared with the observed (large) depth uncertainty ranges.
[1] The term lower confining unit used in this report includes the whole Klettgau Formation and the Bänkerjoch Formation and thus a thicker section compared to the term confining units used in other reports (e.g. NTB 24-17, NAB 24-10).