Rethinking how we assess solar energy

Published March 2025

Insightful reflections on the robustness of solar energy yield assessments from Everoze Partner Jessica McMahon. Solar resource assessment is a key source of uncertainty that can be overlooked. How have you considered uncertainties on your projects?

I’ve been curious for some time about why there’s such a difference in the quality benchmarks for solar and wind energy assessments. In wind energy assessments, loss and uncertainty assumptions are subject to constant refinement and scrutiny. Yet in solar, it’s not uncommon to find assessments (even those approved by lenders) that omit resource assessments, poorly define loss factors or neglect key categories of uncertainty.

The P90/P50 ratio is a key metric used to represent a realistic worst-case scenario for energy production over a given period in the life of a renewable energy project. While a 1-year P90/P50 ratio of 77-85% is typical for wind energy production, solar often falls in the range of 88-93% (i.e. carrying about half the uncertainty of wind assessments). This difference largely stems from the difficulty in quantifying long-term wind conditions across a site for wind assessments, compared to the more stable parameters of irradiance and temperature for solar assessments.

P90 performance is often used to understand a realistic worst-case scenario for energy production.
P90 performance is often used to understand a realistic worst-case scenario for energy production

While it is true that the lower inherent uncertainty of solar energy assessments makes the results more… certain, mistakes can stack up. For example, inaccurate loss assumptions can significantly bias the mean energy estimate, often in the direction of overestimating energy production. Combine that with underestimated uncertainty and the real-world versus modelled performance of solar farms can and does throw up some unpleasant surprises. The accessibility of basic solar energy assessment tools has made it easier to generate quick, cost-effective energy estimates, though the quality can vary. This has created confusion around appropriate industry benchmarks for methodologies and assumptions.

Photovoltaic (PV) technology is complex and constantly evolving. Losses like Light Induced Degradation (LID) are highly specific to the technology type (e.g. poly vs. mono-Si, n-type vs. p-type, doping elements, surface passivation techniques etc) and other losses like degradation are hard to measure and require reliance on aggregated datasets. A robust energy assessment demands careful consideration of all modelling inputs, rather than reliance on software defaults and (possibly biased) manufacturer recommendations. Additionally, the foundation of any accurate energy estimate is a solid resource assessment, whether via ground-based data collection (highly recommended) or the best available satellite-derived datasets.

A great article that further delves into quality challenges in solar energy assessments was recently published in PV Tech, titled “Tackling uncertainty in energy yield forecasts” by Keith McIntosh (thanks Nicolas Chouleur).

For tailored advice on enhancing your wind and solar energy assessments and maximising project insights, don’t hesitate to contact myself or another Everoze expert.