North Dakota CSEM Survey

Case Study: CSEM survey North Dakota

This study evaluates whether land CSEM, complemented by MT to reliably detect the small (<1%) resistivity changes caused by CO₂ injection into a deep saline reservoir (~1.6 km). The work was performed at a CarbonSAFE site in North Dakota, integrating well logs, seismic data, anisotropic resistivity modeling, and a full baseline CSEM/MT survey.

Target interval: stratified reservoir at ~1.6 km depth
Environment: cold-weather operation (−25 °C), surface noise conditions proximity to power plant, and overland HVAC (350,000 V)
Signal expectation due to injection: <5% change in electric/magnetic field 

Vector measurements are required due to acoustic and electrical anisotropy

Objective: demonstrate a field-validated EM monitoring workflow/dataset within 1% repeatability that serves as a foundation for long-term CCUS surveillance. Establish a verifiable vector-electromagnetic dataset.

Workflow Summary

The first part of the monitoring workflow consists of a 3D feasibility (including petrophysical analysis) with onsite noise measurements. The feasibility study results in the optimum sensor selection, survey design, and data acquisition. A fluid substitution model was performed for the reservoir zones, resulting in different time-lapse scenarios. The measured noise from the survey area is integrated to optimize acquisition hardware and survey parameters. When the data is acquired, we use the 3D anisotropic model for Cloud-based near-real-time quality assurance. Once multiple measurements take place, 3D models representing the difference in time-lapse data can be generated, providing a measurable change in the reservoir fluid. This approach reduces errors that otherwise could be introduced with other interpretation techniques, such as 3D inversion.

We are looking for an anomalous CSEM response with time from CCS reservoir after CO₂ injection. The survey setup is shown at the top of the model in the next figure. It shows receiver lines (orange dots) and transmitter (black line at the top right). The colored rectangles represent flat layers of geology with a potential CO₂ reservoir (blue rectangle). Different color curves represent the percentage variations of the EM measurements between brine-saturated and CO₂-flooded reservoir. Colors can be attributed to the receiver line as indicated by the arrows and the corresponding color dots at the receiver locations at the top of the figure.  The anomalous part of the curves outlines the carbon storage reservoir (red curve: right side of reservoir, green and pink curves outline the reservoir).

Geoelectric model (3D anisotropic) and survey plan for the CO2 monitoring survey. The anisotropic model resistivities (blocky curves) were derived from layer equivalencing of the resistivity logs on the right. Also, the survey plan is superimposed on the model depicting transmitter (yellow lines) and receiver (dots) locations. Further on the right are the survey acquisition statistics.

Results of a CSEM monitoring baseline survey. The 1D inversion results for a short section of the most northern line, below. The inversions were automated and unconstraint after the processing parameters were tested with full 3D anisotropic modeling. From the inversion statistics we derive sensitivities throughout the entire depth interval. Superimposed on each inversion result is the borehole model derived from the composite log. The shaded zones mark the two potential CO2 injection intervals.

Key technical results

1. CO₂ anomalies are detectable

The “CSEM reservoir case study” image shows distinct percentage anomalies for different receiver lines. These anomalies outline the reservoir geometry and correlate with the modeled CO₂ plume, demonstrating that time-lapse CSEM sees the reservoir at depth.

2. Integration with logs is essential

The “Well logs are key to CO₂ plume monitoring” panel highlights the necessity of log calibration for both baseline accuracy and time-lapse reliability.

3. Survey design enables deep (>1.5 km) imaging

Using CSEM and broadband MT, the study achieves high-fidelity detection at >1.5 km depth while avoiding false leakage signatures.

4. Large-scale acquisition is operationally robust

• 100% equipment functionality at −25 °C
• 100% Continuous 24/7, for larger than two months including near real-time QC
• 4 transmitters and multicomponent receivers
• CSEM and MT methods deployed

5. Reservoir signature aligns with seismic

Unsupervised 1D inversions plotted against seismic horizons confirm that EM responses track major lithologic boundaries, validating both the model and inversion process.