Wind Resource Assessment: A Complete Guide for Engineers

A wind site can look excellent on a map and still disappoint on a balance sheet. The windthat actually arrives over twenty years rarely matches a single good year of data — and thegap between “looks windy” and “is financeable” is where projects quietly die.

Wind resource assessment is the discipline that closes that gap. Done well, it turns apromising location into a number a lender will underwrite. Done poorly, it either oversells amarginal site or buries a good one in conservatism.

This guide covers what wind resource assessment is, how a bankable one is built step bystep, which data sources to trust, and how to read the P50 and P90 figures that decidewhether a project gets funded.

_{Run a WRA in the Vind platform →}

Key takeaways

• Wind resource assessment (WRA) estimates lifetime energy production from a site,expressed as AEP at P50/P90 confidence.
• A bankable result needs at least 12 months of on-site measurement, extended to the longterm with measure-correlate-predict.
• ERA5 and MERRA-2 are the standard long-term reference datasets; the choice can move yourlong-term mean by a few percent.
• P90, not P50, is what lenders size debt against — so lowering uncertainty raises thefinanceable number.
• Vind generates wind roses, AEP and P50/P90 from your site data in one workflow.

Wind resource assessment is the process of estimating the energy a wind project willgenerate across its operating life. The output is an Annual Energy Production (AEP)estimate, expressed with confidence levels so developers, investors and lenders can judgethe risk before committing capital.

It sits between site prospecting and financial close. For the people signing off a project, thevalue shows up in a few concrete ways:

• A financeable number: a defensible P90 a lender will underwrite, not an optimistic guess.
• Comparable sites: a consistent method lets you rank a portfolio of sites on the same basis.
• Lower surprise risk: quantified uncertainty means fewer post-construction shortfalls.
• A shared language: AEP, P50 and P90 mean the same thing to engineering, finance andinvestors.

How a wind resource assessment works

A bankable assessment follows a repeatable sequence. Each step removes a specificsource of uncertainty.

1. Measure on site

A met mast, LiDAR or SoDAR records wind speed and direction at hub height. A full yearcaptures the seasonal cycle — the common minimum for a bankable result.

2. Extend to the long term

One year of on-site data is correlated against decades of reference data using a measure-correlate-predict (MCP) method, so the estimate reflects the long-run climate, not one anomalous year.

3. Model the flow across the site

The wind climate at the mast is extrapolated across the layout — accounting for terrain,roughness and wake losses — to estimate production at every turbine position.

Data sources compared

Long-term reference data is what turns a short on-site record into a lifetime estimate. Thetwo most-used reanalysis datasets behave differently:

“The reanalysis dataset you anchor to can move your long-term mean by a fewpercent — and a few percent of AEP is the difference between a financeableproject and a shelved one.”

Reading P50 and P90

AEP is never a single number. It is a distribution, summarised by two figures:

• P50: the central estimate, expected to be met or exceeded half the time.
• P90: the conservative estimate, expected to be met or exceeded 90% of the time. Debt issized against this.

The gap between them is driven by total uncertainty. Lowering uncertainty — through longer measurement, better reference data and tighter modelling — raises P90 without touching the physical site. That is where most of the financial value of a good assessment is created.