Email or Username
Password

Please contact our Customer Service Team if you are unable to log in at clientservices@accessintel.com or 1-888-707-5814.
Expected annual energy yield (PVout) is a fundamental number for every utility-scale photovoltaic (PV) project. It informs the design, shapes the budget, feeds the financial model, and influences what investors and lenders are willing to accept.
Behind every expected yield estimate, however, is a range of uncertainty. Part of it comes from the solar resource itself. Part comes from the quality of the input data, the modeling approach, assumptions about losses, and the way site-specific conditions are represented. Snow, soiling, clipping, terrain, shading, thermal behavior, degradation, bifacial effects, component parameters—all of these factors can influence the final result.
Uncertainty has typically been treated as a technical reporting item. In today’s PV market, though, uncertainty affects how projects are designed, valued, financed, and approved. It is not only a technical issue. It is a commercial variable.
The same yield uncertainty can mean different things to different project stakeholders. For engineers, it affects how confidently they can optimize the design. For investors, it changes the strength of the return case. For lenders, it influences how much debt the project can support.
This is why the industry needs to go beyond just reporting uncertainty. The more important question is how much of that uncertainty can be reduced before it starts shaping project decisions in costly ways.
Engineers use energy yield estimates to make practical design decisions. These include tracker configuration, row spacing, DC/AC ratio (the ratio of direct current to alternating current capacity), inverter loading, string design, cable sizing, terrain adaptation, clipping strategy, and loss assumptions.

When uncertainty is low and well understood, design options can be compared with greater confidence. Engineers can better judge whether a higher DC/AC ratio is justified, whether tighter spacing improves project economics, or whether additional equipment will deliver enough extra energy to pay back.
When uncertainty is high or poorly defined, the design process becomes more cautious. Conservative decisions begin to feel safer, even if they are not always optimal.
This can create two types of inefficiency. A project may be overdesigned, with extra capacity, larger margins, or more conservative layouts added to protect against unknowns. Or it may be under-optimized, with energy left on the table because the model does not properly capture site-specific behavior such as seasonal soiling, complex shading, clipping, or bifacial albedo.
For engineers, uncertainty is therefore not an abstract probability range. It affects the confidence behind every design trade-off.
Investors do not invest in a single production figure. They invest in a range of possible outcomes.
The P50 yield—the annual energy production estimate that has a 50% probability of being exceeded—is typically used as the expected production case. But investment committees also look closely at downside scenarios. They need to know whether the project still works if production is lower than expected, capital expenditure (CAPEX) increases, financing becomes more expensive, or merchant prices weaken.
This is where the relationship between P50 and P90 (the production level that has a 90% probability of being exceeded) becomes important. A project may show an attractive return under the P50 case. But if uncertainty is high, the P90 case may be significantly weaker. The wider the gap between expected and conservative production, the more fragile the investment case becomes.
Return on equity (ROE) is often presented as one figure, but in reality, it is more useful to see it as a range. The P50 ROE shows what the project may deliver under expected production. The P90 ROE shows what the investor might face under a downside production case, assuming other variables remain equal.
Reducing uncertainty may not change the expected yield. However, it can improve the conservative yield case and narrow the gap between P50 and P90 (Figure 1). That can make the downside return more resilient—often the case that matters most when investment decisions are being approved.

For investors, the question is not only how much the project can earn. It is also how much returns can deteriorate before the investment case becomes difficult to defend.
Lenders approach yield uncertainty through the lens of debt repayment. Their main concern is whether the project can generate enough cash flow to service debt under conservative assumptions.
This is usually assessed through metrics such as debt service coverage ratio, or DSCR. In simple terms, DSCR measures whether project income is sufficient to cover debt payments.
Banks often assess projects using conservative production assumptions, such as P90 energy. However, it is a mistake to assume that lenders simply apply an annual uncertainty discount across the full project life. In real project finance, that approach can be too crude.
If production is mechanically reduced every year over a 20- or 25-year period, it can materially weaken DSCR, loan life coverage ratio, and equity returns. A project may start to look less bankable on paper, even when the risk could be managed in a more precise way.
Lenders usually deal with uncertainty through financing structure. This may include debt sizing, DSCR thresholds, reserve accounts, dividend restrictions, covenants, guarantees, or sponsor support. The goal is to make sure the project remains robust under conservative assumptions. For lenders, uncertainty is real, but it is usually managed through structure rather than a simple annual cut to production.
Quantifying uncertainty is necessary. It improves transparency and gives stakeholders a clearer view of project risk. But reporting uncertainty does not automatically improve the project.
If uncertainty remains high, each stakeholder reacts defensively. Engineers add buffers. Investors focus more heavily on downside returns. Lenders reduce leverage or tighten financing terms. This defensive behavior can affect the project even if the expected yield remains attractive.
That is why uncertainty reduction matters. It can move the discussion from “how do we protect ourselves against this risk?” to “how much confidence do we have in the project’s real performance?” This is a different conversation. And it can have real financial consequences. For large utility-scale projects, the financial benefit of reducing uncertainty can justify the additional effort and cost.
Imagine a utility-scale PV project with a defined expected P50 yield. Under a standard approach, the project uses acceptable but limited inputs, simplified assumptions, and a conventional modeling process. The P50 yield may look strong, but the uncertainty range is relatively wide. As a result, the P90 yield sits noticeably lower.
The project may still be financeable, but only within tight limits. The lender sizes debt conservatively to protect DSCR. The investor sees a weaker downside return. The engineer has less room to justify more optimized design choices.
Now imagine the same project with better solar resource data, longer historical time series, more realistic modeling, higher temporal resolution where relevant, and stronger validation of site-specific losses. The P50 yield may remain the same. But uncertainty falls, and the P90 yield improves.
Nothing physical has changed. The site is the same. The equipment may be the same. The expected production has not increased.
What has changed is confidence. That confidence can create more headroom in the financial model. It can strengthen the downside return case. It can support more efficient debt sizing. It can also give engineers a stronger basis for design optimization. In other words, reducing uncertainty can improve the project without increasing the expected yield (Figure 2).

Not all uncertainty can be removed. Interannual variability, for example, reflects natural year-to-year weather variation. It can be understood and quantified, but not eliminated. However, other sources of uncertainty can often be reduced.
A good starting point is the quality of solar resource data. Long-term, validated solar radiation datasets help project teams better understand expected conditions and variability. Where possible, long historical time series should be used instead of relying only on typical meteorological year data.
Temporal resolution also matters. Sub-hourly data can be valuable when short-term effects influence project performance, including irradiance peaks, clipping, inverter behavior, and temperature dynamics.
Modeling assumptions should also be improved. Fixed “rules of thumb” for losses can be replaced with physics-based models where possible, especially for soiling, albedo, temperature, snow, and other site-specific effects.
In more complex layouts, optical losses may require advanced methods such as ray tracing. In challenging regions, ground measurements and local validation can further improve confidence.
Component data should not be overlooked either. Datasheets need to be checked, and model parameters should reflect the equipment that will actually be installed.
The industry often talks about better data and better modeling as technical improvements. They are, but their impact goes further.
For engineers, they support better design decisions. For investors, they make downside returns more defensible. For lenders, they improve confidence in conservative production assumptions.
This is why PV yield uncertainty should not be treated as a footnote in an energy yield report. It is a project-level issue that influences design quality, investment resilience, and financing efficiency.
At the same time, uncertainty reduction should be proportionate to the project and market context. While the financial impact can justify deeper data, modeling, and validation work on utility-scale projects, the same investment may not always be worthwhile for smaller assets or in markets where energy prices, curtailment, or interconnection risks dominate the business case.
The goal is not to eliminate uncertainty completely. That is impossible. The goal is to reduce what can be reduced, quantify what remains, and avoid letting avoidable uncertainty make good projects look riskier than they are.
—Pablo Caballero is an industrial engineer and technical writer at Solargis. He has extensive experience in the renewable energy and software development sectors. He specializes in technical writing and content marketing, and is driven by a passion for bridging gaps between audiences, technology, and business.
Modern process industries are experiencing fluctuating market conditions and tight operational margins, leading chemical engineers to rely on real-time data to boost efficiency and reduce costs. Yet, many organizations are at different stages in their digital transformation journey. Some are just starting, while others are looking to optimize existing solutions. This webinar explores practical ways […]
Sponsored by dataPARC
Sponsored by RENTECH
Sponsored by technosylva
The inaugural Data Center POWER eXchange brought the data center power ecosystem together in Denver for a focused, one-day summit on the infrastructure shaping the future of digital growth. From site selection and development to operations and forward planning, DPX connected leaders across data center ownership, utilities, engineering, power generation, and technology. Watch to see […]
archives + podcasts + webinars + e-letters + video + exclusive content + events + much more!
For over 142 years, POWER has been the unwavering pillar of the power and energy industry, offering unparalleled insights on the latest trends, best practices, and news through our robust channels: print, digital media, and in-person events.
POWER is at the forefront of the global power market, providing in-depth news and insight on the end-to-end electricity system and the ongoing energy transition. We strive to be the “go-to” resource for power professionals, offering a wealth of information on innovative business practices, sound safety measures, useful productivity enhancements, and much more.
Sept. 28—30, 2026
Omni Shoreham
Washington, D.C.
Learn More
Trends, best practices, and news through: print, digital media, and in-person events. Subscribe
Exclusive E-letters to POWER up your knowledge. Sign up
© 2026 Access Intelligence, LLC – All Rights Reserved

source