Ground-mount vs roof-mount PV systems: the real trade-offs

Choosing between ground-mount and roof-mount is rarely a free choice - available land, roof condition, and grid connection point usually narrow the options before economics enter the discussion. But where both are genuinely on the table, the decision hinges on yield potential, foundation and structural cost, cabling losses, permitting complexity, and long-term operations access. This article is for developers and asset owners comparing the two at feasibility stage, not for homeowners with a single obvious roof.

Ground-mount and roof-mount systems are not simply the same panels on different structures. They differ in achievable tilt and azimuth, in the civil works required, in how DC cabling is routed and how much of it is needed, in the regulatory pathway, and in exposure to theft, vandalism, and access constraints during operation. Each of these differences moves the levelised cost of electricity (LCOE) by a measurable margin.

We work through yield and tilt freedom, foundations and structural cost, cost per kWp, cabling and losses, permitting, and security and O&M access, with a worked cost-per-kWp comparison at the end.

Yield and tilt freedom

Roof-mount systems are constrained by the existing roof pitch and orientation. A roof pitched at 15 degrees facing 30 degrees east of south is fixed - the array follows it unless costly tilt-up frames are added, which then increase wind loading and shading between rows. Ground-mount systems can be tilted to the latitude-optimised angle and oriented true south (or true north in the southern hemisphere), typically recovering 3-8% more specific yield (kWh/kWp/year) than a poorly oriented roof of the same nominal capacity, depending on how far the roof deviates from optimal.

Ground-mount also permits single-axis or dual-axis tracking, which is rarely practical on a roof. Single-axis horizontal trackers typically add 15-25% annual energy yield over a fixed-tilt ground array in mid-latitude sites with a good direct/diffuse irradiance ratio, at the cost of additional mechanical complexity, maintenance, and roughly 10-20% higher capex per kWp for the tracking structure and drive system.

Roof-mount does have one yield advantage in dense urban or industrial contexts: it uses no additional land, so on a per-hectare basis it is unambiguously more space-efficient, even though its per-kWp yield may be lower.

Foundations and structural considerations

Ground-mount foundations depend heavily on soil conditions. Driven steel piles are the default for granular and firm cohesive soils and are fast to install, but soft or high-water-table ground may require screw piles, ballasted concrete blocks, or in extreme cases pad footings with geotechnical engineering input - all of which raise cost and schedule risk. A geotechnical survey is a prerequisite for any utility-scale ground-mount tender; skipping it is one of the most common causes of cost overruns in the sector.

Roof-mount avoids foundation works entirely but requires a structural assessment of the existing roof: dead load capacity, wind uplift resistance of the mounting system, and (for older industrial roofs) verification of purlin spacing and roof sheet fixation. A structural survey showing inadequate load capacity can eliminate roof-mount as an option outright or force a de-rated, lower-density layout that significantly reduces installable kWp.

Wind and snow loading

Ground-mount tables are typically engineered per EN 1991-1-4 (wind actions) and local snow load codes, with a lower deck height in high-wind zones being a common mitigation. Roof-mount systems inherit the building's existing structural wind and snow design and add local uplift loads at the mounting points, which is why ballasted, non-penetrating roof systems dominate on flat industrial roofs in low-to-moderate wind zones - they reduce risk of roof membrane puncture while relying on ballast mass rather than fixings for uplift resistance.

Cost per kWp

These figures vary substantially by market and scale; residential roof-mount below 10 kWp often runs meaningfully higher per kWp than either category above due to fixed mobilisation costs. The comparison is most meaningful at matched capacity (500 kWp - 2 MWp C&I scale) where both options are genuinely competing.

Cabling and losses

Ground-mount plants typically have longer DC and AC cable runs from array to inverter to point of connection, increasing both material cost and resistive losses. A well-designed ground-mount layout keeps voltage drop under 1-1.5% on DC strings by careful string routing and combiner box placement, per the guidance in cable sizing and voltage drop, but this requires deliberate design attention that is less critical on a compact roof where cable runs may be under 20 m.

Roof-mount benefits from proximity to the building's electrical intake, often eliminating the need for a dedicated substation or long MV cable run that a ground-mount plant of equivalent capacity would require if the grid connection point is distant from the array field.

Permitting

Ground-mount projects above a few hundred kWp frequently trigger full planning permission processes, environmental impact screening (particularly for land-use change, drainage, and ecology), and in agricultural areas, additional scrutiny over loss of farmland. Timelines of 6-18 months for planning consent are common in much of Europe for utility-scale ground-mount, and grid connection queue position is often the binding constraint rather than the planning process itself.

Roof-mount on existing commercial or industrial buildings frequently qualifies for simplified or permitted-development-style processes in many jurisdictions, since it does not change land use, though listed buildings, conservation areas, and roofs requiring structural modification can still require full consent. This permitting asymmetry is one of the strongest practical arguments for roof-mount wherever a structurally suitable roof exists.

Security and O&M access

Ground-mount plants require perimeter fencing, and increasingly, intrusion detection and CCTV, to mitigate theft of cabling and modules and to satisfy insurer requirements - a recurring opex line that has no direct equivalent for most roof-mount systems. O&M access to ground-mount arrays is generally straightforward: technicians walk the rows, vegetation management (mowing or grazing) is a routine cost, and module cleaning or replacement uses standard access equipment.

Roof-mount O&M access is constrained by height, roof condition, and health and safety requirements for working at height, which typically mandate anchor points, guardrails, or mobile elevated work platforms depending on roof type. Module replacement or thermal inspection on a steep or fragile roof can cost several times the equivalent task on a ground-mount array, and roofs with fragile fibre-cement or ageing sheeting need specific fall-arrest planning under most occupational safety regulations.

Worked example

A developer compares two 800 kWp options at the same site: a ground-mount array on adjacent farmland with fixed-tilt racking, and a roof-mount array on a nearby warehouse with adequate structural capacity.

Ground-mount: mounting + foundations = EUR 150/kWp x 800 kWp = EUR 120,000. Add fencing/security EUR 10/kWp x 800 = EUR 8,000. Estimated specific yield at optimal tilt: 1,150 kWh/kWp/year.

Roof-mount: mounting = EUR 60/kWp x 800 kWp = EUR 48,000, no foundation or fencing cost. Roof pitch is suboptimal (10 degrees east-facing), reducing specific yield to an estimated 1,060 kWh/kWp/year (roughly 8% below the ground-mount optimum).

Annual energy: ground-mount = 800 x 1,150 = 920,000 kWh/year. Roof-mount = 800 x 1,060 = 848,000 kWh/year, a difference of 72,000 kWh/year in the ground-mount's favour.

Structure-related capex difference: ground-mount incurs EUR 128,000 extra (foundations + fencing) versus roof-mount's EUR 48,000, a gap of EUR 80,000. At an assumed wholesale/PPA value of EUR 0.09/kWh, the extra 72,000 kWh/year from ground-mount is worth EUR 6,480/year, giving a simple payback on the EUR 80,000 structural cost premium of roughly 12.3 years - longer than many investors' hurdle period, illustrating that higher ground-mount yield does not automatically offset its higher structural cost. Run project-specific numbers through the payback calculator and system sizing calculator before committing to either option.

Common mistakes

• Comparing headline cost-per-kWp figures without normalising for site-specific soil conditions or roof structural condition, both of which swing costs by 20-30%.
• Assuming roof-mount always has lower LCOE - a suboptimally oriented, shaded, or fragile roof can erase the capex advantage entirely.
• Skipping a geotechnical survey for ground-mount, leading to foundation redesign mid-construction.
• Underestimating security and fencing opex for ground-mount sites in areas with a history of cable theft.
• Treating O&M access cost as negligible for pitched or fragile roofs, when working-at-height requirements can make routine maintenance disproportionately expensive.
• Ignoring grid connection distance when comparing options - a ground-mount site far from the point of connection can incur MV cabling costs that dominate the entire cost comparison.

Frequently asked questions

Which option has lower LCOE?

Neither has a universal advantage; ground-mount typically wins on yield and cabling proximity to a nearby substation, while roof-mount wins on avoided foundation cost and simplified permitting. Site-specific modelling is required in each case.

Can tracking be retrofitted to a ground-mount fixed-tilt array?

Generally no in a cost-effective way - tracking requires a different foundation design, spacing, and drive infrastructure from the outset, so the decision must be made at initial design stage.

Does roof-mount always avoid planning permission?

No. Many jurisdictions still require consent for larger commercial roof systems, listed buildings, or where structural modifications are needed, though the process is typically faster than for greenfield ground-mount.

How much does row spacing affect ground-mount yield?

Inter-row spacing controls self-shading losses in winter months; a tighter row pitch increases land-use efficiency (kWp/hectare) but can cost several percent of annual yield, so the trade-off should be optimised against land cost per hectare, not fixed by rule of thumb.

Is module theft a material risk for ground-mount plants?

It varies strongly by region and site remoteness; where it is a known risk, mesh fencing, intrusion detection, and marked or serialised modules meaningfully reduce loss frequency and are usually cheaper than the insurance premium uplift from having no security measures.

Standards and references

• EN 1991-1-4: Eurocode 1 - actions on structures, wind actions (relevant to both ground-mount racking and roof-mount uplift design).
• IEC 61215 / IEC 61730: PV module design qualification and safety, relevant to structural mounting compatibility.
• NREL: reports on utility-scale PV balance-of-system cost benchmarking.
• Fraunhofer ISE: Photovoltaics Report, cost and yield benchmarking data.
• IEA PVPS: national survey reports comparing ground-mount and rooftop deployment trends.
• JRC/PVGIS: irradiance and yield estimation tool for tilt/azimuth sensitivity analysis.

For further design context, see MPPT explained, cable sizing and voltage drop, the inverter sizing calculator, the voltage drop calculator, and the glossary.