Wind and snow loads on PV structures under Eurocode

Rooftop and ground-mount PV structures fail from wind far more often than from snow, but the two loads interact in Eurocode design, and getting the combination wrong is a common source of under-designed mounting systems. This article is for installers, structural checkers, and technical due-diligence reviewers who need to verify that a mounting system's load certification actually matches the site conditions, not just a generic datasheet claim.

EN 1991-1-4 (wind actions) and EN 1991-1-3 (snow loads) are the reference documents across most of Europe, each with National Annexes that adjust base values, terrain categories, and combination factors. A module clamp rated for "2400 Pa" means nothing without knowing which load case, which zone, and which partial safety factors produced that number.

The goal here is not to replace a structural engineer but to give enough grounding to read a mounting system's technical approval, cross-check it against a specific roof or field, and spot the assumptions that do not transfer from one site to another.

Wind load fundamentals under EN 1991-1-4

The basic wind velocity v_b is a regional value from the National Annex map, typically given as a 10-minute mean at 10 m height in open terrain (terrain category II), with a 50-year return period. This is converted to a peak velocity pressure q_p(z) that accounts for:

• Terrain category (0 to IV): sea/lake shoreline through dense urban, affecting the roughness length z0 and the mean wind profile.
• Orography factor: hills, ridges, and escarpments accelerate flow locally; flat terrain gives c_o = 1.0.
• Turbulence intensity: higher in rough terrain, which increases peak gust factors even though mean speed is lower.
• Reference height z: for low buildings this is often the building height; for open-field trackers it is the array height above grade.

The peak velocity pressure feeds into the wind force via pressure coefficients c_pe (external) and c_pi (internal, usually zero for open-back PV structures but not for parapet-mounted or building-integrated systems).

Terrain and exposure factors in practice

A module rated by the manufacturer for terrain category 0 (coastal, fully open) will see 30-50% lower peak pressure at the same nominal wind speed than one placed in terrain category III (suburban, scattered obstacles) evaluated at the same base wind speed - the terrain factor works the other way in the exposure factor c_e(z), which is higher for smoother terrain because turbulence is lower but mean speed is higher near the surface. This is a frequent point of confusion: rougher terrain reduces mean wind speed close to the ground but increases gustiness, and the net effect on c_e(z) depends on height. For low-profile ground-mount arrays (1.5-3 m height) in terrain category III or IV, c_e(z) is typically in the 1.8-2.2 range; in terrain category 0-I it can reach 2.5-3.0 at the same height because mean speeds are higher even though turbulence is lower.

Edge, corner, and zone effects

Wind pressure on a roof or open array is never uniform. EN 1991-1-4 divides roof and wall surfaces into zones (F, G, H, I, J for flat roofs; similar schemes for duopitch and monopitch) with pressure coefficients that escalate sharply near edges and corners. For a flat roof with a parapet under 0.3 m, corner zone F can see c_pe values of -1.6 to -2.0 (suction) versus -0.7 for the central zone H at the same wind direction - roughly double to triple the local suction.

This is why mounting system installation manuals specify denser ballast or additional clamping in a 1.5-3 m perimeter band and at building corners. Skipping this detail because "the datasheet says 60 kg/m2 average ballast" is a common and serious error: average ballast across a roof is not the same as zone-specific ballast, and corner modules under-ballasted by even 20% have been documented lifting in gusts well below the design wind speed.

For open-rack ground-mount and tracker rows, the equivalent concept is edge rows and end-of-row tables, which see higher net pressure coefficients than interior rows due to reduced flow blockage from neighboring rows. Tracker manufacturers typically specify a stow (flat or near-flat) position above a threshold wind speed, commonly 15-20 m/s sustained, precisely to reduce the effective c_p by presenting a smaller and more favorable profile to the flow.

Uplift and the net pressure coefficient

For pitched or tilted PV arrays, the critical failure mode is usually uplift (suction) rather than downward pressure, because the array acts as an airfoil-like obstruction and the low-pressure zone above the tilted surface is larger and more variable than the high-pressure zone on the windward underside. The net pressure coefficient c_p,net for rack-mounted PV on a flat roof (per approaches consistent with EN 1991-1-4 Annex and supplementary guidance such as SEAOC PV2-2017, widely referenced even in European practice for array-specific coefficients) can reach -2.5 to -3.5 in corner zones at low tilt angles.

Uplift force per module area is:

F_uplift = c_p,net x q_p(z) x A

Worked example

Take a 2.0 m2 module in a corner zone of a flat commercial roof, tilt 10 degrees, mounting height 0.6 m, in terrain category III, with a National Annex basic wind velocity v_b = 24 m/s and a design return period giving q_p(z) = 0.85 kN/m2 at the relevant reference height (typical for this terrain/height/speed combination once terrain and orography factors are applied).

Using a corner-zone net pressure coefficient c_p,net = -2.8:

F_uplift = 2.8 x 0.85 kN/m2 x 2.0 m2 = 4.76 kN per module

At two clamps per module edge (4 clamps typical), each clamp must resist roughly 4.76 kN / 4 = 1.19 kN of uplift, before applying the partial safety factor for wind actions (gamma_Q = 1.5 in the fundamental combination). The factored design uplift per clamp becomes 1.19 x 1.5 = 1.79 kN. A clamp rated for 1.2 kN pull-out on the specific roofing membrane in question is under-designed for this zone even though it might be adequate for an interior-zone module on the same roof.

This is the arithmetic that should appear, module type by module type and zone by zone, in a proper site-specific structural check - not just a single headline ballast or clamp number.

Snow loads under EN 1991-1-3

Snow load design uses a characteristic ground snow load s_k from the National Annex map (altitude- and zone-dependent), converted to a roof snow load:

s = mu_i x C_e x C_t x s_k

where mu_i is the shape coefficient (typically 0.8 for flat and low-pitch roofs, higher for valleys and drift-prone geometries), C_e is the exposure coefficient (0.8 for windswept sites, 1.0 for normal, 1.2 for sheltered sites where snow is not removed by wind), and C_t is the thermal coefficient (usually 1.0 for PV arrays unless there is a heated envelope reducing snow retention below the modules).

For ground-mount and elevated rooftop PV, two effects matter beyond the base roof snow load:

• Snow drift accumulation at row-to-row gaps in tracker and fixed-tilt arrays, which can locally double the design snow depth compared to open ground.
• Load transfer through mounting rails: on sloped rooftop systems, snow sliding off the array can add a line load at the eave-side rail, distinct from the distributed load on the modules themselves.

Combined wind and snow cases

EN 1990's combination rules require checking wind and snow as leading and accompanying actions in turn - snow rarely governs simultaneously with peak wind in most of Europe (heavy snow events are typically calmer), but low-altitude coastal sites with occasional wet snow and moderate wind need both cases checked explicitly rather than assumed bounding by the wind-only case.

Mounting system certification: what to check

A wind tunnel report is only valid for the tested tilt, height, row spacing, and open-area ratio; extrapolating it to a denser row spacing or a taller structure without re-analysis is a documented cause of field failures.

Roof capacity checks

Before adding a PV array to an existing roof, the structural check must confirm that dead load (array weight, typically 12-20 kg/m2 for rail-mounted rooftop systems, up to 25-40 kg/m2 for ballasted systems) plus live snow load plus wind uplift/downforce stays within the roof's original design envelope, including any reserve already consumed by other rooftop equipment (HVAC units, parapets, previous retrofits). Many roofs designed decades ago carry little to no reserve capacity, particularly older industrial roofs with lightweight steel decking; a capacity assessment by a structural engineer, not just the PV installer, is required before ballasted systems are installed on such roofs. For system-level load budgeting alongside electrical design, see the system sizing calculator.

Common mistakes

• Using a manufacturer's average ballast or clamp spacing across the whole roof instead of zone-specific values for edges and corners.
• Applying a wind tunnel report's pressure coefficients to a different tilt angle, row spacing, or mounting height than what was tested.
• Assuming wind-only governs and skipping the explicit snow-plus-wind combination check for calm, heavy-snowfall coastal or valley sites.
• Neglecting snow drift accumulation between tracker or fixed-tilt rows, which can exceed the open-field ground snow load locally.
• Treating a roof structural capacity assessment as optional on older or lightweight-deck buildings.
• Ignoring stow-position wind thresholds for trackers, or specifying a stow angle that still presents a large effective area to the design wind direction.

Frequently asked questions

Does a higher module wind-load rating (Pa) always mean a safer installation?

Not by itself. The module frame's static pressure rating covers the glass and frame only; the racking, clamps, and roof or foundation attachment have separate and often more limiting capacity, so the weakest link in the full load path determines actual safety margin.

How much does terrain category change the design wind pressure?

Moving from terrain category IV (dense urban) to terrain category 0 (open coastal) at the same basic wind velocity can increase the exposure factor c_e(z) by 40-70% at typical array heights, because turbulence-driven gust amplification is much lower in smooth open terrain even as mean speeds rise.

Is ballast-only mounting adequate in high wind zones?

It can be, but ballast mass required to resist corner-zone uplift scales quickly with wind speed and array height, often becoming impractical above roughly 30-32 m/s basic wind velocity; mechanical anchoring or a hybrid ballast-plus-anchor design is typically preferred beyond that range, subject to roof penetration and waterproofing constraints.

Do trackers need a different wind analysis than fixed-tilt arrays?

Yes. Trackers must be analyzed both in operating tilt positions (where projected area and pressure coefficients vary continuously with tracking angle) and in the stow position, plus the dynamic behavior during stow transition if a wind event develops faster than the stow actuation time.

How does snow drift between rows affect row spacing decisions?

Tight row spacing on fixed-tilt or tracker arrays can trap drifting snow, raising local snow depth well above the open-field ground snow load and creating asymmetric loading on the downwind module row; sites in heavy snowfall regions sometimes widen row spacing beyond the shading-optimal value specifically to reduce this drift accumulation risk, a trade-off discussed further in relation to array layout and row spacing.

Standards and references

• EN 1991-1-4:2005+A1 - Eurocode 1: Actions on structures, Part 1-4: Wind actions
• EN 1991-1-3:2003 - Eurocode 1: Actions on structures, Part 1-3: Snow loads
• EN 1990:2002+A1 - Eurocode: Basis of structural design (load combination rules)
• SEAOC PV2-2017 - Wind Design for Solar Arrays (referenced for array-specific pressure coefficients)
• Fraunhofer ISE - Photovoltaics Report, mounting and BOS reliability sections
• NREL - Structural and mounting system reliability studies for utility-scale PV
• IEA PVPS Task 13 - Reliability and performance of PV systems, failure mode reporting

For related electrical design constraints that interact with structural layout, see cable sizing and voltage drop and the glossary for term definitions used throughout this article.