Background station - minimum requirements (EU framework and EEA Air Quality Index)

Airbox V3 project. This document covers background stations only: urban, suburban and rural
background. Traffic and industrial stations have their own documents in this folder.

Sources used (abbreviations used in the text):

1. Purpose and scope

This document lists the minimum a background station must meet to:

1. be a compliant sampling point under the AAQD: correct siting, reference or equivalent methods,
   data quality objectives, QA/QC and documentation; and
2. be shown in full colour on the EEA European Air Quality Index map.

What it covers:

• Station types background with area types urban, suburban and rural (including rural sub-types).
• Index pollutants PM2.5, PM10, NO2, O3 and SO2. CO is mentioned only for public information.

What it does not cover:

• Traffic and industrial stations.
• Monitoring supersites (AAQD Art. 10, Annex VII). These are background sites too, but they have extra
  pollutant lists (UFP, BC, PM2.5 chemical composition, and so on). They are out of scope here.
• Network design, meaning how many stations a zone needs. That is the competent authority's job (AAQD Annex III).
  It is mentioned only where it affects a single station.

Important framing:

• The EEA index is a short-term health communication tool. It is not a compliance tool (EEA index page).
• The index uses whatever E2a data arrive. Compliance is a separate, stricter matter.
• A station can show in full colour and still fail the AAQD data quality objectives. The aim here is to meet both.
• Data reach the EEA only through the national competent authority's E2a feed (IPR). For Romania
  that is ANPM, which runs the national network RNMCA.
• A private station therefore appears on the index only if ANPM accepts and reports it as a registered
  sampling point.

2. What makes a station a "background" station

2.1 Legal definitions (AAQD Art. 4)

The AAQD itself has only two background categories, urban and rural. A suburban background site
legally counts as an urban background location, because Art. 4(23) covers "urban and suburban areas".
The separate "suburban" label comes from the Eionet area classification (2.3). The ozone siting table in
Annex IV B.4 also treats suburban separately (section 5.1).

2.2 Station classification by predominant source (Eionet vocabulary stationclassification)

Background is the residual class. Concept URI: http://dd.eionet.europa.eu/vocabulary/aq/stationclassification/background

The classification is reported per pollutant / measurement configuration (IPR Annex II D(ii) item 22).
In principle a site can be background for O3 and something else for another pollutant. For an Airbox
background station, every configuration should be classified background.

2.3 Area classification (Eionet vocabulary areaclassification, IPR Guidance Table 14)

Notes from IPR Guidance:

• The rural sub-type distances are "only indicative".
• "Border" means the actual edge of the built-up area, not the administrative boundary.
• Area types are mutually exclusive.

The exact vocabulary concept IDs rural-nearcity, rural-regional and rural-remote match the
Eionet code list as referenced in the brief. The guidance text names them "Rural – near city",
"Rural – regional" and "Rural – remote". Check the IDs against the live code list before filling Dataset D.

2.4 Representativeness area (AAQD Annex IV B)

Rules for determining the representativeness area (AAQD Annex IV B.5):

• It may consist of non-contiguous parts, but must stay within the air quality zone.
• The annual mean of the observed concentration is the metric for a given year.
• If modelling is used to determine it, the model must be fit for purpose.

3. Mandatory and recommended pollutants

3.1 Minimum set for full colour on the EEA index (ETC HE section 2)

"For background and industrial stations: Either PM2.5 or PM10 needs to be measured as well as NO2 and O3."

If any of these is missing, the station is drawn semi-transparent (section 8).

3.2 Recommended set and reasoning

Legal context for NO2 at O3 sites:

• NO2 must be measured at a minimum of 50 % of the O3 sampling points required by Annex III Table 2 (AAQD Art. 9(4a)).
• That NO2 measurement must be continuous except at rural background stations, where other methods are allowed (AAQD Art. 9(4a)).
• For the EEA index, however, NO2 must arrive as hourly values. In practice this means a continuous NO2 analyser even at a rural background site.

4. Measurement requirements

4.1 Time resolution, units, reference conditions

4.2 Reference methods and equivalence (AAQD Annex VI)

Equivalence rules:

• Any other method may be used if equivalence is demonstrated (AAQD Annex VI B.1).
• For PM, a method that shows a consistent relationship to the reference method is also acceptable,
  for example an automatic monitor meeting EN 16450:2017. Its results must be corrected to be
  reference-equivalent (AAQD Annex VI B.1).
• The Commission's guidance on the demonstration of equivalence applies (Annex VI B.3). Corrections
  should be applied retroactively where appropriate (Annex VI B.4).
• Type-approval test reports from accredited laboratories in other Member States must be accepted (Annex VI D).
• Low-cost sensors can at best reach the indicative class (CEN/TS 17660-1/-2 Class 1). Whether ANPM
  accepts them into E2a is ANPM's decision (../EU_AQI_requirements.md section 6.5).

4.3 Maximum measurement uncertainty (AAQD Annex V A, 95 % confidence)

Short-term means (Annex V A, Table 2):

Long-term (annual) means (Annex V A, Table 1):

How to apply these values (Annex V A):

• Before 2030 the relative values apply.
• From 2030, uncertainty must not exceed the absolute or the relative value, whichever is higher.
• The values apply in the region of the limit or target values. They do not apply to information or alert thresholds.
• Calculate uncertainty per the relevant EN standard. For non-standard methods use JCGM 100:2008 and ISO 5725-5.
  For indicative measurements without a CEN standard, use the equivalence guidance.
• The Directive gives no hourly uncertainty objective for O3, only the 8 h mean.
  The design goal for hourly O3 should therefore follow EN 14625 performance criteria. Those criteria are not in the local sources.

4.4 Minimum data coverage (AAQD Annex V B)

Data coverage = share of the calendar year with valid data.

O3 seasonal rule (Annex V B footnotes 1 and 2):

• Fixed O3: the coverage must be met for the full calendar year and separately for April–September
  and for October–March. For AOT40, it must also be met during the AOT40 calculation period.
• Indicative O3: coverage applies to April–September only. There is no winter criterion.

Other coverage rules (Annex V B):

• Fixed measurements of SO2, NO2, CO, O3, PM10 and PM2.5 must run continuously all year.
• Coverage must be met for specific sub-periods (quarter, month, weekday) so that data losses do not bias results.
• Normal maintenance must not take place during pollution peak periods. For a background station this
  means mainly summer O3 episodes and winter PM episodes.

4.5 Calibration and QA/QC (AAQD Annex V E)

Not verified: the specific zero/span check intervals, calibration frequencies and performance criteria
are set inside EN 14211, EN 14625, EN 12341/EN 16450 and EN 14212. These are paid standards and are not
in the local sources. Get them from the standards themselves or from the national reference laboratory.

5. Siting requirements for background stations

5.1 Macro-scale (AAQD Annex IV B)

General (all background sites):

Additional O3 criteria (Annex IV B.4). These apply to fixed and indicative measurements:

The Annex IV B.4 table distinguishes "rural" from "rural background" for O3. A rural background O3 site
has the strictest distance rule (≥ 20 km), stricter than the general 5 km rural background rule of B.2(ca).
Rural and rural background O3 sites should, where appropriate, be coordinated with forest monitoring
under Reg. (EC) 1737/2006 (Annex IV B.4).

Network context (for discussion with ANPM, not a per-station obligation):

• Each zone's minimum fixed sampling points must include at least one background location point (Annex III A.1).
• Urban background and hotspot point counts for NO2 and PM must not differ by more than a factor of 2 (Annex III A.1).
• In agglomerations, at least 50 % of O3 sampling points must be suburban (Annex III A, Tables 2 and 4, footnote).
• Rural background O3: at least 1 sampling point per 50 000 km² as a national average. For complex terrain, 1 per 25 000 km² is recommended (Annex III C.2).

5.2 Micro-scale (AAQD Annex IV C)

These apply "in so far as is practicable":

The traffic rule (≥ 25 m from major junctions, ≤ 10 m from the kerb, Annex IV C(e)) does not apply
to background sites. A background site must not look like a traffic site: if the inlet is within about
10 m of a busy kerb, the station would normally be classified traffic.

For the O3 inlet, the > 10 m from the nearest road rule applies to every road, and the distance
should increase with traffic. For PM and NO2 at a background site, the Directive gives no fixed road
distance. The no-dominant-single-source rule (B.2(c)) and the representativeness requirement are what
keep the site away from roads in practice.

5.3 What to document (AAQD Annex IV D)

The competent authority must document the site selection. For a station offered to ANPM, prepare at least:

6. Station and sampling point metadata (Dataset D)

The national authority registers the station in Eionet Dataset D. Incoming E2a data are QA-checked
against the latest Dataset D, so it must be up to date before any data flow (IPR Guidance, data flow overview).

The complete Dataset D field list is in IPR Annex II D(ii). This document lists only the fields with
background-specific content. The full list is in ../EU_AQI_requirements.md section 4.3.

7. Data delivery to the EEA index

How quickly after the hour data must reach the EEA to show for that hour is not specified in the
local sources. The only statement is "transmitted hourly". Agree a latency target with ANPM, for example
the hourly value available to ANPM within a few minutes of the end of the hour. That figure is a
suggestion, not a requirement.

8. How the station appears on the EEA index

8.1 Display rules

The gap-filling and missing-hour rules come from the EEA index page via ../EU_AQI_requirements.md.
They were not re-verified against a local copy of that page for this document.

Practical consequence for a background station:

• If the O3 analyser goes down for several days, the O3 sub-index becomes "not determined" after the 3-of-4-days rule.
• It is unclear from the sources whether the station then drops to semi-transparent. This is likely, because
  O3 is part of the minimum set, but it is not explicitly stated.
• Either way, O3 uptime is critical for keeping full colour.

8.2 Bands (µg/m³, 1 h means), in force since 4 July 2025 (ETC HE Table 5.2)

Notes:

• The "very poor" upper limits for O3 (180), NO2 (150) and SO2 (275) equal the AAQD information thresholds (ETC HE section 5).
• Health messages are unchanged from the previous index (ETC HE Table 6.1; reproduced in ../EU_AQI_requirements.md section 2.4).
• Category colours: good = turquoise, fair = green, moderate = yellow, poor = red, very poor = dark red,
  extremely poor = purple. Approximate hex values are in ../EU_AQI_requirements.md section 2.3 and should be checked against the live site.

9. Checklist (background station)

Sensors / instruments

• NO2 analyser installed (EN 14211 chemiluminescence, or demonstrated equivalent), also giving NO/NOx
• O3 analyser installed (EN 14625 UV photometry, or demonstrated equivalent)
• PM2.5 monitor installed (EN 12341, or EN 16450-type automatic monitor with correction)
• PM10 monitor installed (recommended in addition to PM2.5)
• SO2 (EN 14212) and CO (EN 14626) analysers considered and decision documented (optional)
• Equivalence demonstration / type-approval report available for every non-reference instrument (AAQD Annex VI B)
• Expected uncertainty meets Annex V A (fixed or indicative class declared per pollutant)

Data format

• Hourly means produced for every pollutant (half-hour raw data aggregated to 1 h)
• Gas concentrations standardised to 293 K / 101.3 kPa; PM at ambient conditions (Annex VI C)
• Units µg/m³ (CO mg/m³)
• PM values corrected to EN 12341 equivalence before reporting
• Timestamps in ISO 8601 extended format, end of period, fixed offset without summer time (+02:00 or Z)
• Validity flag (1/2/3/−99/−1) attached to every hourly value
• Verification flag (3 for live data) attached to every hourly value
• Values below −DL discarded; values between −DL and DL flagged 2

Siting

• Location qualifies as background: not traffic, not industrial, no single dominant source (IPR Table 12; Annex IV B.2(c))
• Representativeness of several km² (urban background) documented
• Rural background: no urban area, major road or industrial site within 5 km (Annex IV B.2(ca))
• Rural background O3: ≥ 20 km from urban and industrial areas, not in an inversion-prone or coastal local-wind site (Annex IV B.4)
• Suburban O3 site placed downwind of the city under ozone-forming conditions (Annex IV B.4)
• Inlet airflow free over ≥ 270° (≥ 180° at a building line) (Annex IV C(a))
• Inlet ≥ 1.5 m from buildings, balconies, trees and obstacles (Annex IV C(a))
• Inlet height 0.5–4 m, or a higher inlet justified and documented (Annex IV C(b))
• O3 inlet > 10 m from the nearest road and away from furnaces/flues (Annex IV C(g))
• Inlet not next to local sources; exhaust does not recirculate (Annex IV C(c)(d))

Metadata

• EoI station code agreed with ANPM (ROxxxxA format); new code planned if the station moves
• Station name follows IPR Table 8 conventions
• Coordinates with ≥ 5 decimals, ETRS89/WGS84, plus altitude
• Station classification background set for every measurement configuration
• Area classification set (urban / suburban / rural / rural-nearcity / rural-regional / rural-remote), with distance evidence
• Inlet height, method, equipment, detection limit and equivalence reference filled in for each sampling point
• Maps, compass-point photos, representativeness and micro-scale deviations documented (Annex IV D)

Delivery

• Agreement with ANPM / local APM to integrate the station into RNMCA and the E2a feed
• Dataset D record registered before the first E2a data are sent
• Hourly data transfer to ANPM in their ingest format, with local buffering and back-fill
• Arrival verified in the EEA Download Service (UTD/E2a) and on the index map in full colour
• Process for validated (E1a) data and Dataset D updates by 30 September of the following year

QA

• QA/QC system in place: maintenance, zero/span and calibration schedule per EN standards (Annex V E.1(b))
• Calibration traceable to an accredited (EN ISO/IEC 17025) laboratory (Annex V E.1(a))
• PM correction factor established by co-location and reviewed periodically
• Data coverage target ≥ 85 % per year for NO2 and O3 and ≥ 85 %/75 % for PM (fixed)
• O3 coverage ≥ 85 % in April–September and in October–March separately (Annex V B, footnote 1)
• Maintenance scheduled outside pollution peaks (summer O3 episodes, winter PM episodes) (Annex V B)
• Downtime alarms for O3, NO2 and PM, so no pollutant is missing at the same hour on ≥ 3 of the 4 previous days
• Site review at least every 5 years (Annex IV D.9)