A
Peer review
Interpretations are checked by a second qualified reviewer before release [VERIFY: review steps and who signs off].
Methodology
Characterize. Model. Decide.
Three stages take a project from raw measurements to a decision. Each records what was done, so the work can be checked.
01 / The three stages
From measurement to decision.
The stages are the same for a single borehole and for a storage site. What changes is the amount of data and the number of iterations.
-
Characterize
Establish the geology, fluids and physical properties from logs, core, tests and existing records. State how well each is known, and which gaps matter for the decision. -
Model
Connect observations across scales, from core plug to log to formation, and test how the system behaves under plausible scenarios. Identify which inputs the result is most sensitive to. -
Decide
Turn results into options: a go or no-go, a monitoring plan, an investment case or a next measurement. Each comes with its range and the assumptions behind it.
02 / Uncertainty
Ranges, not single numbers.
Subsurface quantities are never known exactly. Reporting one number hides how wrong it could be.
-
Inputs as ranges
Each input (thickness, porosity, area, fluid properties) is given a range based on the data, not a single best guess. -
Carried through
Ranges are combined by repeated sampling, so the result is a distribution rather than a product of best guesses. -
Reported by case
Results are quoted as P90, P50 and P10, with the convention stated in every report. -
Sensitivity
The inputs that drive the spread are ranked, which shows which measurement would narrow it most.
03 / Data and quality control
Evidence someone else can check.
The table lists the checks applied to borehole data before interpretation [VERIFY: written QA procedure].
| Check | What is done | Record kept |
|---|---|---|
| Calibration | Tool response checked before and after each run | Calibration sheets with the data |
| Repeatability | A repeat section logged on every run | Main and repeat passes compared |
| Depth control | All runs matched to one reference log and a fixed datum | Depth shifts by interval |
| Corrections | Hole size and fluid corrections applied | Corrections and parameters listed |
| Core tie | Logs matched to core gamma and plug data where core exists | Calibration cross-plots |
| Audit trail | Interpretation parameters and software versions recorded | Parameter table in the report |
04 / Review and independence
B
Independence
DeepSink is paid for the analysis, not the outcome. A finding that the evidence does not support a project is reported as plainly as one that does.
C
Confidentiality
Client data is held under access control and is not shared or reused without written consent [VERIFY: storage and retention].
05 / Next step
See the method on your data.
Tell us what you hold and what you need to decide. We will set out what the data can and cannot establish.