Local snow effects

Drift at projections, snow overhanging the eaves and the force on snowguards — EN 1991-1-3, section 6.

Snow load

s_k — ground snow load [kN/m²] The characteristic ground value. It enters only μ₂ for the projection. s — snow load on the roof [kN/m²] The load for the most unfavourable UNDRIFTED case, from the snow page. It enters §6.3 and §6.4.

Projection — §6.2

h — projection height [m] The height of the obstruction creating the aerodynamic shelter. §6.2 is written for roofs of very low pitch, close to horizontal.

Pitched roof — §6.3 and §6.4

α — roof pitch [degrees] b — distance to the ridge [m] The horizontal distance from the snowguard row to the next row or to the ridge. Two rows at half the distance each carry half. d — layer thickness [m] The thickness of the snow layer on the roof, for the coefficient k. 0 = derived from s/γ, with γ = 3 kN/m³. Site altitude [m] The NOTE to §6.3(1): the provision is recommended for sites above 800 m. Below that, it is left to the National Annex.

▸ Fill in the data on the left and press Calculate

Tool information

What this page computes

The three local forces that section 5 does not give, to EN 1991-1-3:2005, section 6:

§6.2 — drifting at projections and obstructions;
§6.3 — snow overhanging the edge of the roof;
§6.4 — the force on snowguards.

Why this is not the same as Annex B

This is the distinction that matters, and it is easily missed because the leading term looks identical.

§6.1(2) is explicit: the design situations here are persistent and transient. Annex B is accidental. The consequences follow in a chain:

	§6.2 (here)	Annex B
expression	(5.1): \(s = \mu C_e C_t s_k\)	(5.3): \(s = \mu s_k\)
\(C_e\), \(C_t\)	apply	do not apply
combination	fundamental	accidental
cap on \(\mu\)	2.0	8
\(l_s\)	\(2h\), between 5 and 15 m	\(\min(5h,\ b_1,\ 15)\)

The leading term really is the same: §6.2 gives \(\mu_2 = \gamma h/s_k\) with \(\gamma = 2\) kN/m³, that is exactly \(2h/s_k\) — the Annex B form. But the cap differs by a factor of four, and the drift length is computed differently.

Taking a value from one chapter and using it in the other goes wrong in both directions. The page also shows the uncapped value, so that how much the code cut stays visible.

Two different unit weights, in adjacent clauses

\(\gamma = 2\ \text{kN/m}^3 \quad \text{in §6.2} \qquad\qquad \gamma = 3\ \text{kN/m}^3 \quad \text{in §6.3}\)

This is not a transcription error. They describe two different states of snow: freshly drifted behind an obstruction, against compacted at the eaves, where it has melted and refrozen several times. They sit two pages apart in the code and it is easy to use one for the other.

§6.3 — the overhang, where the formula is subtler than it looks

\(S_e = \frac{k s^2}{\gamma} \qquad \text{(6.4)}\)

with \(k\) given by the NOTE to §6.3(2): \(k = 3/d\), subject to \(k \le d\gamma\), where \(d\) is the layer thickness.

The two branches meet exactly at \(d = 1\) m: there \(3/d = 3\) and \(d\gamma = 3\). Below a metre \(d\gamma\) governs; above it, \(3/d\) does.

The consequence for the result is unexpected. If the thickness is derived consistently from the load, \(d = s/\gamma\), then:

for \(s < 3\) kN/m²: \(k = s\), so \(S_e = s^3/3\) — it grows with the cube of the load;
for \(s \ge 3\) kN/m²: \(k = 9/s\), so \(S_e = 3s\) — it becomes linear.

At \(s = 1\) kN/m² the result is 0.33 kN/m; at \(s = 2\) it is 2.67 kN/m — eight times more for twice the load. The join is continuous at \(s = 3\), where both branches give 9 kN/m.

Applicability has a threshold: the NOTE to §6.3(1) recommends the provision for sites above 800 m altitude. Below that, the conditions of use may be specified by the National Annex. The page shows the value anyway and flags when you are below the threshold.

§6.4 — snowguards

\(F_s = s\, b \sin\alpha \qquad \text{(6.5)}\)

Simple, with one remark in the text worth reading: the friction coefficient between snow and roof may be neglected. That is why the force is directly proportional to the sine of the pitch, with no term reducing it.

\(b\) is measured horizontally from the snowguard row to the next row or to the ridge — not to the eaves. From this follows directly the reason for using several rows: two rows at half the spacing each carry half.

What goes in as \(s\)

§6.3 and §6.4 take the undrifted load, not the maximum on the roof. Using the drifted value would overestimate both results.

A neighbouring check

§5.3.2(2): where the roof edge has a parapet, snow fences or other obstacles, the ordinary roof shape coefficient must not be taken below 0.8 — snow can no longer slide off, so it is not allowed to reduce. That is a section 5 requirement, separate from the forces here.

What it does not cover

Designing the snowguards themselves and their fixings — the page gives the force, not the member.
Checking the eaves in bending under \(S_e\) — the load is given per linear metre, but the cantilever is verified separately.
§6.2 on pitched roofs — the code writes the clause for roofs of very low pitch, close to horizontal.

Expressions (6.1)–(6.5) are transcribed from EN 1991-1-3:2005, pages 21–23, and each has a test that reproduces it by hand.

Snow — the source of \(s\), the undrifted roof load.
Snow drift — Annex B, the accidental case, compared against here.
EN 1990 combinations — the fundamental combination, which the results here feed.