The utility-scale solar sector has grown at extraordinary pace, and the flood risk assessments submitted to lenders have not always kept pace with the sophistication that international project finance now demands. Lenders' technical advisors reviewing FRAs are increasingly rigorous — the days of a desktop study passing unchallenged are over.
After reviewing a significant number of FRAs submitted for solar projects across multiple regions, five errors appear consistently. None of them are obscure technical failures. Most stem from one of three root causes: applying generic methodology without site-specific calibration, structuring the report for the wrong audience, or underestimating what IFC Performance Standards actually require.
Key finding
In projects where the FRA was challenged by lender TAs, the average additional time to financial close was approximately four months — during which the project was fully capitalised but unable to draw down. The cost of a properly structured FRA is a small fraction of that carrying cost.
Mistake 1 — Using a single rainfall source without cross-validation
01
Relying solely on one gridded rainfall dataset without cross-validation
Gridded rainfall products at ~0.25° resolution are the standard starting point for flood risk assessment — and they are the correct starting point. But a grid cell of that size covers a substantial area, missing significant sub-grid variability in rainfall intensity, particularly for short-duration extreme events.
FRAs that base their entire frequency analysis on one gridded dataset alone — without cross-validating against available gauge records or satellite rainfall — present a significant methodological vulnerability. Lender TAs with experience will identify this immediately and request justification for why no cross-validation was performed. In several cases we have reviewed, the TA has requested a revised hydrological study, adding 6–8 weeks to the timeline.
The fix
Perform rainfall frequency analysis using the gridded dataset as the primary source, cross-validated against any available gauge records and regional IDF curves. Document all data sources, their limitations, and the justification for the final frequency estimates in a dedicated data section — this takes additional time at scoping but prevents a full revision later.
Mistake 2 — Flood extent based on DEM analysis alone
02
Deriving flood extent from DEM analysis, not a calibrated hydraulic model
DEM-based flood boundary delineation is faster and cheaper than building a calibrated hydraulic model. It is also, in most cases, not sufficient for a bankable solar FRA. DEM-based methods cannot account for hydraulic effects — backwater, flow constrictions, floodplain storage, or channel-floodplain interaction — and frequently underestimate flood extent in the flat terrain typical of utility-scale solar sites.
The specific risk for solar sites: the areas most at risk of being under-assessed are the areas most likely to flood.
The fix
Build a calibrated 1D or 2D hydraulic model using the peak discharges from the hydrological model as boundary conditions. Calibrate against any available observed flood data. Document and justify the model choice. Deliver model files alongside the report.
"A flood hazard map produced without a calibrated hydraulic model is based on assumptions, not physics."
— HYDRisk Engineering TeamMistake 3 — Assessing the site boundary, not the full catchment
03
Limiting analysis to the site boundary rather than the full contributing catchment
A solar farm's contributing upstream catchment can be many times larger than the site itself. An FRA that characterises only rainfall and runoff within the site boundary will significantly underestimate discharge arriving from upstream. This is particularly common where sites sit at the downstream end of large dryland catchments — the site can look low-risk in isolation, while concentrated runoff from tens of thousands of hectares upstream arrives within hours of an extreme event.
The fix
Delineate the full upstream catchment using DEM-based watershed analysis. Sub-divide by land use and soil type. Calculate runoff from the entire contributing area — not just the site footprint — and present the catchment boundary map as a standalone deliverable.
| Region type | Typical catchment : site ratio | Risk if ignored |
|---|
| Arid dryland corridor | 20:1 – 50:1 | Very high |
| Semi-arid plain | 10:1 – 30:1 | High |
| Coastal / low-relief | 5:1 – 15:1 | Moderate |
Commissioning a solar FRA for your project?
Every HYDRisk solar FRA is built from the full upstream catchment — with a calibrated hydraulic model, IFC-structured report, and 7–14 day delivery.
Mistake 4 — No substation or electrical infrastructure assessment
04
Not separately assessing substation, inverter, and cable infrastructure flood risk
Most FRAs assess flood risk to the site as a whole and conclude that risk is acceptable — without separately assessing the specific vulnerability of the substation, inverter stations, transformer bays, and underground cable routes. This matters because the substation is the most expensive and longest-lead component; replacement of a damaged transformer can take 18–24 months. A lender's TA will specifically look for explicit substation flood risk assessment, and its absence is one of the most consistent sources of technical queries.
The IFC PS4 angle: IFC Performance Standard 4 requires assessment of infrastructure failure consequences. A substation flood event that disrupts grid supply is a PS4-relevant impact — FRAs that do not address this are structurally incomplete for IFC compliance.
The fix
Include a dedicated electrical infrastructure flood risk section. Map substation, inverter, transformer, and cable routes against inundation extent and depth per return period, and provide component-specific recommendations rather than generic mitigation language.
Mistake 5 — Reporting structured for a planning authority, not a lender's TA
05
Writing the report for a planning officer rather than for lender TA review under IFC standards
A planning FRA and a bankable lender FRA require different structures, different technical detail, and different reference frameworks. An FRA written for a planning submission — however technically sound — will frequently generate a significant list of queries from a lender TA simply because the structure and reference framework are wrong.
The fix
Before writing a word, identify the primary audience. For lender FRAs, structure explicitly around IFC PS1 and PS4, include a compliance matrix appendix, write the executive summary for the investment committee, and deliver model files alongside the report.
The common thread — and what to do about it
All five mistakes share a common root: applying a standard approach without first asking what the specific study needs to achieve and for whom. The most effective way to avoid these errors is to define the audience, regulatory framework, and technical requirements before the study begins — not to write a report and then retrofit it after the first query letter arrives.
HYDRisk approach
Every HYDRisk solar FRA begins with a scoping consultation that identifies the lender, framework, and climate requirements before modelling begins — which is what gives a report its best chance of clearing review first time.
Topics:Solar PVIFC PS4Project financeBankable FRA
HYD
HYDRisk Engineering Team
Flood Risk & Hydrology Specialists
HYDRisk delivers bankable flood risk assessments across the US, Europe, the Middle East and Asia.