DPR 142/04 limits — road noise

Regulatory framing

DPR 30/3/2004 no. 142 is the Italian implementing regulation for art. 11 of Law 447/95, specific to road-traffic noise. It defines the noise-pertinence bands and their immission limits, so as to decouple the regime for the noise of a specific infrastructure from the general municipal zoning regime (DPCM 14/11/97). This Open Lab tool checks whether the L_Aeq levels at the receptor comply with those limits; the tables have been verified against the text of Annex 1 of the decree.

The logic of the bands

Near a road, traffic noise dominates. Imposing the DPCM 14/11/97 limits — 50 dB(A) daytime for Class I, “particularly protected areas” — within 250 m of a motorway would be unrealistic: no new road could ever be built. The decree introduces pertinence bands, strips of land measured from the road boundary, within which higher limits apply and for which the infrastructure operator is responsible.

Outside the band the DPCM regime resumes: someone living 300 m from a motorway is covered by the municipal zoning limits, not by the decree’s limits. The tool handles this transition explicitly: beyond the band width it asks for the municipality’s acoustic class and applies Table C of the DPCM.

New and existing roads

The decree distinguishes two regimes, with two tables in Annex 1. For newly built roads (Table 1) mitigation enters the design: the decree requires the identification of design corridors able to guarantee the best protection of receptors and compliance with the table limits. The limits are stricter and the band is single (A and B 250 m, C1 250 m, C2 150 m, D 100 m), with 65 dB(A) daytime / 55 night-time for generic receptors.

For existing and assimilated roads (Table 2) the built reality is a given. The category does not carry the everyday sense of “old”: art. 1 anchors it to a date. Existing means an infrastructure in service, under construction, or with a final design approved when the decree entered into force — 16 June 2004; newly built means one that was in design at that date and does not fall under the previous definition. Table 2 then also covers widenings on the same alignment, parallel additions and re-routings; for parallel additions the band is measured from the boundary of the pre-existing infrastructure, not the new one. It is the distinction that decides which table applies, so check it against the decree before picking the mode in the tool. For A and B the decree allows a more tolerant Band A (0–100 m, 70/60 dB(A)) and a stricter Band B (100–250 m, 65/55). The extra tolerance applies to generic receptors only: for sensitive ones the limit stays 50/40 in both bands. Secondary roads also have two bands, of different width by subtype: Ca (separated carriageways or type IV CNR 1980) has Band A 100 m and Band B 150 m, 250 m in total like motorways; Cb (all others) has the same 100 m Band A but a Band B of only 50 m, 150 m in total. In both cases Band A is 70/60 and Band B 65/55. Urban through-roads stay on a single 100 m band and are separated by configuration, not by limit: Da are those with separated carriageways and inter-district roads, Db all the others. The limit — 70/60 for Da, 65/55 for Db — follows from the classification, it is not the criterion for picking the row. The same holds for C1/C2 and Ca/Cb: the row follows from the road’s classification under the technical criteria the tables refer to, never from the band width or the limit that would suit. Bringing them back within the limits is entrusted to the multi-year remediation activity set out in DM 29/11/2000, which also fixes the order of priority: for schools, hospitals, nursing and care homes the whole pertinence band, for other receptors the band closest to the infrastructure. It does not hold in every case, though: for parallel additions and re-routings art. 5 c. 2 provides that the limits apply from the decree’s entry into force, without waiting for the multi-year remediation.

The limits in a table

New roads — Table 1 (generic receptors):

Existing and assimilated roads — Table 2 (generic receptors):

The two middle columns say different things and must be kept apart: the width is the width of the sub-band as the decree states it (Ca’s Band B is 150 m wide, Cb’s is 50 m); the distance from the boundary is the cumulative interval you are in. Confusing the two is the commonest way to get an implementation wrong. On boundary values the tool closes the intervals on the right: at exactly 100 m the receptor is in Band A, at 100.1 m in Band B.

For sensitive receptors — schools, hospitals, nursing and care homes — the limit is 50/40 dB(A) in every row from A to D, in both bands and in both tables, with the daytime limit only for schools. For E and F that cell too defers to the municipality.

Where the band starts

A detail that in practice shifts results more than it appears to: the bands are measured from the road boundary, not from the carriageway axis and not from the edge of the asphalt. Art. 1 of the decree defines it hierarchically, following the Italian Highway Code: it is the limit of the road property as it results from the acquisition deeds or from the expropriation strips of the approved design. Only in the absence of those does one fall back on the residual criteria — outer edge of the drainage ditch or gutter, foot of the embankment for a road on fill, top edge of the slope for a road in cut. Hard shoulders and pavements may lie inside the road property, but are not in themselves the criterion.

On a motorway with two carriageways, central reserve, shoulders and embankments, the distance between the axis and the road boundary can be twenty metres or more per side. A receptor 110 m from the axis can therefore sit 90 m from the boundary, that is in Band A and not Band B: five decibels of extra limit, decided by an acquisition deed or an expropriation strip. It is why automatic band determination from the real geometry is on the roadmap (v0.2) and not improvised: it needs the boundary alignment, not the centreline.

The limit falls faster than the noise

The stepped regime has a counter-intuitive consequence. Crossing the bands, the limit drops in 5 dB steps, while road noise decays continuously and slowly: for a line source, about 3 dB per doubling of distance. The two do not run parallel, and the most critical point of the cross-section is not necessarily the one closest to the road.

A worked example on an existing motorway, calibrated at 70 dB(A) daytime 25 m from the boundary:

Note that there are two points at zero margin, at 25 m and at 251 m: the one closest to the road is not less critical than the others, it is exactly as critical. What changes on crossing a boundary is the abrupt loss of the margin that had built up.

The 5 dB step at the Band A / Band B boundary is worth, for a line source, moving the receptor from 100 to 316 m: the decree asks for it in one metre. From 100 to 250 m, by contrast, the level falls by 3.98 dB in total.

The step on leaving the band, though, is not a constant: it depends on the municipal acoustic class. From a Band B at 65 dB(A) daytime you go to −15 dB in Class I, −10 in Class II, −5 in Class III, 0 in Class IV, and even +5 dB in Class V or VI, where leaving the band loosens the limit instead of tightening it. The example in the table uses Class III; change the class and the conclusion changes.

Hence an indication for screening, not for priority: the points just past a band boundary deserve a look, because that is where the margin can thin out abruptly. It is not a regulatory criterion, though, and must not be confused with one. Remediation priority is set by the decree: art. 5 c. 3 assigns it to the whole pertinence band for schools, hospitals, nursing and care homes, and to the band closest to the infrastructure for other receptors; DM 29/11/2000 then orders the interventions by a priority index weighing the size of the exceedance and the number and type of people exposed, with higher coefficients for schools and healthcare facilities.

Two caveats on the example, which is purely illustrative.

The first concerns the distances, and follows from the previous section. The band is measured from the road boundary; propagation starts at the equivalent source, that is at the lanes. These are two different coordinates, and the table makes them coincide for readability, as if the source line sat on the boundary. In a real case the source is inside the road property, sometimes twenty metres inside: the level column should be read as an order of magnitude, not as a prediction at that distance. A serious check keeps the two distances apart — the one from the boundary to fix the band, the one from the source to compute propagation.

The second concerns the model: 3 dB per doubling, with no ground effect, atmospheric absorption, reflections, screening or meteorological conditions. A full ISO 9613-2 propagation (2024 edition) adds terms that normally attenuate — ground and atmospheric absorption — but also reflections, screening and propagation-favourable conditions, which can act the other way and non-uniformly across receptors. There is no guarantee that real margins are better than those in the table, nor that the ordering of critical points stays as it is: a barrier can change which receptor is the most exposed. The decree does not, in any case, mandate a calculation method; DM 29/11/2000 sets the functional requirements a prediction model must meet. For the real calculation you need acmap or a model meeting those requirements.

Sensitive receptors

Schools, hospitals, care and rest homes have a privileged regime: 50 dB(A) daytime / 40 night-time throughout the pertinence band, regardless of the subtype — but only for roads A to D. For E and F the whole limits cell, sensitive receptors included, defers to the municipality. The rationale is the specific vulnerability of the activities carried out there. A note in Annex 1 specifies that for schools only the daytime limit applies: the tool does not evaluate the night-time period when the receptor is a school.

Those 50 dB(A), translated into distance: On the same motorway as above — 70 dB(A) at 25 m — reaching 50 dB(A) by geometric divergence alone takes 2500 m. The 40 dB(A) night-time limit has to be computed on the road’s night level, not on the daytime one: starting from 62 dB(A) at 25 m — eight decibels below daytime, a plausible order of magnitude for a motorway — it takes about 4 km; were the night level equal to the daytime one it would take 25 km. No pertinence band is that wide: the sensitive-receptor limit is not attainable by distance inside the band, and is not meant to be. Its function is to force mitigation — barriers, earth bunds, speed reduction, low-noise pavements — not to define a geometric threshold.

The decree also covers the case where mitigation is not enough. Where the outdoor values are not technically attainable — or where technical, economic or environmental considerations show it is preferable to act directly on the receptor — art. 6 moves the check indoors, at the centre of the room, with windows closed and 1.5 m above the floor: 35 dB(A) at night for hospitals, nursing and care homes, 45 dB(A) daytime for schools, 40 dB(A) at night for other residential receptors.

It is the step from mitigation along the propagation path to intervention on the receptor itself, typically the glazing. On who bears the cost, care is needed: it is not automatically the operator. Art. 8 places it on the holder of the building permit where that permit post-dates the decree’s entry into force (for existing infrastructure) or the approval of the infrastructure’s final design (for new works, widenings, parallel additions and re-routings), for the part exceeding the mitigation already provided by the road design. Whoever builds next to a road that was already there carries their own burden. Outside these cases the cost stays with the infrastructure.

The complete tables

The tool fully implements both Annex 1 tables. E roads (urban district roads) and F roads (local) need reading carefully: they too have a pertinence band, of 30 m, and are not “roads without a band”. The decree sets no limits of its own for them — inside those 30 m the values are determined by the municipality through acoustic zoning, following Table C of the DPCM 14/11/97.

The distinction matters because the two cases are not equivalent. Inside the pertinence band — of any road type — the limit refers to the noise of the infrastructure alone. Outside the band, municipal zoning applies to the total noise of all sources, with the infrastructure contributing alongside the others. Same number in the table, different object of the check.

The Table C values the tool uses are: Class I 50/40, II 55/45, III 60/50, IV 65/55, V 70/60, VI 70/70 dB(A). One point on method: the decree says the values for E and F are set by the municipalities in compliance with Table C and with the zoning. The tool applies Table C directly from the class, which is the reasonable approximation for a national pre-screening, but the relevant municipal rules still have to be checked.

Interaction with other regulations

DPR 142/04 coordinates with the DPCM 14/11/97 (zoning, valid outside the band), with DM 29/11/2000 (criteria for noise-containment and abatement plans) and with Legislative Decree 194/05, the implementation of EU Directive 2002/49/EC. On the latter a distinction is due: 194/05 concerns noise mapping, strategic maps, action plans and public information — it is a planning level, not a point compliance check. It neither replaces nor modifies the values of DPR 142/04 and DPCM 14/11/97.

The decree’s limits refer to the noise of one road infrastructure. Where several bands overlap the assessment becomes concurrent: DM 29/11/2000 provides that in the overlap area the total noise must not exceed the greater of the limits applicable to the individual infrastructures. Coexistence with sources of a different nature (rail, industry) requires a multi-source analysis under the respective regulations. The tool handles neither case.

Tool status and roadmap

• v0.1 — Single-receptor compliance check, new and existing modes, tables verified against the decree text.
• v0.2 — Automatic band calculation from real geometry (offset from the road axis with graphics).
• v0.3 — Integration with acmap results via querystring (pre-populated levels).
• v1.0 — Extension to DPR 459/98 for railway noise, with an analogous pattern.

Limitations

The figure to be compared with the limit is not just any L_Aeq, and this is the most serious limitation. For road-noise monitoring, DM 16/03/1998 requires a measurement time of no less than one week, acquisition of the equivalent level hour by hour and the computation of the weekly daytime and night-time mean levels, with the microphone 1 m from the façade, 4 m above the ground, at the most exposed points. And where the infrastructure’s own limits apply, the value must be stripped of the other sources.

The tool does none of this: it takes a number and compares it with the table. An hourly, daily or instantaneous L_Aeq, or one including the noise of other sources, is not interchangeable with the one the regulation asks for, and the verdict is worth what the input is worth. It also checks a single receptor, and handles neither the overlap of several infrastructures’ bands nor the regime of derogations and remediation plans in case of exceedance.

It is therefore a demonstrative pre-screening, not a compliance check and not an expert report. The asymmetry runs one way: an exceedance computed on the road contribution alone is solid evidence, since adding sources can only make it worse; compliance computed this way is not compliance, least of all outside the band, where the limit is compared with total noise. Real impact assessments and checks need a competent acoustic technician on the national ENTECA register (art. 21 of Legislative Decree 42/2017) and a prediction model meeting the functional requirements of DM 29/11/2000 — a requirement on the model, not on the software’s brand name.

References

• Italian DPR 30/3/2004 no. 142 — containment and prevention of road-traffic noise pollution.
• Italian Law 26/10/1995 no. 447 — framework law on noise pollution, art. 11.
• DPCM 14/11/1997 — limit values of noise sources (Table C, outside the band).
• DM 29/11/2000 — noise-containment and abatement plans.
• DM 16/3/1998 — techniques for detecting and measuring noise pollution (referred to by art. 2 of DPR 142/04 for checking the limits).
• Legislative Decree 285/1992 — Italian Highway Code: road classification and definition of the road boundary.
• DM 5/11/2001 — functional and geometric standards for road construction (criteria for the C1/C2 subtypes).
• Legislative Decree 42/2017 — art. 21: national register of competent acoustic technicians (ENTECA).
• Italian Legislative Decree 19/8/2005 no. 194 — EU Directive 2002/49/EC on environmental noise.

The tool is available as an executable tool at /en/tools/italian-road-noise-limits/.