P50, P75, and P90 are probability-based estimates that describe the likelihood of achieving specific energy production levels over the operational lifetime of a wind project.
What Do These Numbers Mean?
P50 — The Median Estimate
P50 represents the median or 50th percentile estimate of annual energy production (AEP). There is a 50% probability that actual production will exceed the P50 value. P50 is used as the base case for project economics and equity return calculations.
P75 — Conservative Estimate
P75 represents a 75% probability of exceedance. Lenders commonly use P75 for financial modelling when assessing equity returns under conservative scenarios. The P75 is determined by the combined uncertainty budget of the energy yield assessment.
P90 — Very Conservative Estimate
P90 represents a 90% probability of exceedance. This is the most risk-averse estimate — used for debt sizing, Debt Service Coverage Ratio (DSCR) calculations, stress testing, and project valuation in acquisitions. The P90 is the single most scrutinised output in lender due diligence.
How Are They Calculated?
The spread between P50 and P90 is determined by the combined uncertainties in the energy yield prediction, propagated in quadrature across all major categories:
- Wind measurement uncertainty (anemometer calibration, data recovery)
- Long-term wind climate variability (inter-annual variability, reference dataset quality)
- Wind flow modelling uncertainty (WAsP, WindPRO, CFD for complex terrain)
- Wake loss uncertainty (Jensen, Larsen, or RANS-based wake models)
- Turbine performance uncertainty (power curve, availability)
- Loss factor assumptions (electrical, environmental, curtailment)
A one-sigma combined uncertainty of approximately 8–10% AEP translates to a P50-to-P90 spread of roughly 10–13%, which is typical for well-measured onshore sites.
Practical Example
For a 50 MW wind farm:
- P50: 150 GWh/year — median expectation, base case for equity
- P75: 138 GWh/year — conservative equity scenario
- P90: 127 GWh/year — used for debt sizing
The 15% spread between P50 and P90 reflects the combined uncertainty in the assessment. A lender's technical advisor will challenge any assumption that artificially compresses this spread.
What Drives the P90 Down
Short onsite measurement campaigns, weak long-term reference correlation, aggressive MCP methods, or under-characterised inter-annual variability all widen the uncertainty range and push the P90 lower. Investing in a longer, higher-quality measurement campaign with proper calibration directly improves the debt case.