QC first
Raw logger data is checked for continuity, sensor validity, shadow, drift, icing, and physically implausible values before any modelling begins.
Methodology
QC first. MCP second. Flow modelling third. Uncertainty last.
The order matters because lender review starts with traceability and ends with the P90 case. The methodology is intentionally narrow so the reasoning chain stays legible.
Long-term correction second
MCP is applied against ERA5, MERRA-2, and where useful NIWE or IMD cross-checks. The goal is a defensible climate, not a flattering fit.
Uncertainty last
Measurement, long-term, spatial, wake, and future-loss uncertainty are combined into the P50 / P75 / P90 case so debt sizing can be defended line by line.
01
Data ingest and QC
Timestamp continuity, logger health, sensor drift, icing, boom shadow, and redundant sensor checks are applied before anything else.
02
Long-term correction
Short onsite records are extended via MCP against long-term reference data chosen for the site, terrain, and project stage.
03
Flow modelling and uncertainty
WAsP, WindPRO, or CFD is selected based on terrain. The uncertainty budget then reflects the same assumptions the lender will see.
What the output includes
- Cleaned time series with exclusions marked.
- Long-term correction memo with reference choice and residuals.
- Flow model selection by terrain, not by habit.
- Loss build-up and uncertainty budget with a clear audit trail.
What the method is trying to prevent
Flattering numbers
Optimistic assumptions are easy to produce and expensive to defend. The workflow is designed to expose them early.
Hidden uncertainty
The report should make the risk visible instead of burying it inside a headline P50 or a polished presentation.
Use the methodology to decide whether the work fits.
If the project needs a more conservative model, a second reference, or a full due diligence review, that gets said directly during the scoping call.