the experience of piedmont in air quality

1. Workshop on the preparation of air quality plans organised in co-operation with the Ministry of Environment and Spatial Planning Experiences in establishment of zones and agglomerationsthe Case of Piedmont region (Italy)

2. Foreword

Italian Regions are the Institutions committed to perform air quality assessment required by EU directives

Tasks:

Subdivide territory in zones defined depending on the risk level of exceeding air quality limits

Realise yearly air quality assessment elaborate air quality information to be transmitted to the European Commission

Periodically revise zones definition

Define mitigation actions and pollution reduction plans

Experiences in establishment of zones and agglomerations: Piedmont 3. Piedmont Region:acomplex territory

Area: 25.426 Km

Climate:

Winters : cold, dry andbanks of fog

Summers : cool in the hills and quite hot in the plains

Temperature range during the year: - 5C / +30C

Hydrography : more than4000km

Land cover : 52% agricultural, 40,2% forest and semi natural,6,3% artificial, 1,5% water bodies-wetlands

Protected areas :2 national parks ,67 regional parks,8,3% of the total surface

Nature 2000 network :123SCI,51SPA,15,67% of the total surface

Monitoring stations

Air:85 Water : 129 Meteorological : 322

Experiences in establishment of zones and agglomerations: Piedmont 4.

Population: 4.290.000

Density:169 pop/Km

8 Provinces- 1.206Municipalities

Chief Town:Turin ( 901.000)

Agglomerate : Turin + 11( 1.297.000)

Road net: 22.630 Km( first in Italy !!! )

Vehicles:3.481.736

GDP 2006: 119 Bil (10% GDP Italy)

Businesses:407.137

Piedmont Region: acomplex territoryExperiences in establishment of zones and agglomerations: Piedmont 5.

Several subjects provide the source information: Region, Provinces, Environmental Agency, Municipalities, Private organizations.

Environmental information creation process is transversal to the public bodies with specific competence

PROVINCES Permitting procedures and local planning REGIONAL ENVIRONMENT PROTECTION AGENCY (ARPA) Technical control and monitoring REGIONPlanning and coordinating Environmental Information SystemEnvironmental information in Piedmont Experiences in establishment of zones and agglomerations: Piedmont 6. Regional Air Quality Network

74 public fixed stations

6 public mobile stations

11 private fixed stations

In the fixed network there are: 65 nitrogen oxide analyzers (NO x ) 47 carbon monoxide analyzers (CO) 54 PM 10samplers and analyzers 2PM 2.5samplers 25 sulphur dioxide analyzers (SO 2 ) 33 ozone analyzers (O 3 )16 aromatic hydrocarbon analysers (BTX) Experiences in establishment of zones and agglomerations: Piedmont 7.

Air Quality Management

air quality models have an important place in air quality management. They are essential tools in the development of action plans for improving air quality, which is the ultimate goal of the Member States and local authorities in order to fulfil their obligations under the directives.

From: Guidance on Assessment under the EU Air Quality Directives, EEA 2002 Experiences in establishment of zones and agglomerations: Piedmont 8.

Zoning for Air Quality Management

Starting from the preliminary assessment

DIR 1996/62/CE

(art. 5)

Member States which do not have representative measurements of the levels of pollutants for all zones and agglomerations shall undertake series of representative measurements, surveys or assessments in order to have the data available in time for implementation of the legislation referred to in Article 4 (1).

Experiences in establishment of zones and agglomerations: Piedmont 9. Following the EU Framework Directive (96/62/EC)and his Daughter Directives (1999/30/EC; 2000/69/EC; 2002/3/EC; 2004/107/EC) and now New Air Quality Directive (2008/50/EC) Zones and assessment regimes From: Guidance on Assessment under the EU Air Quality Directives, EEA 2002 Experiences in establishment of zones and agglomerations: Piedmont 10.

Zone, why ???

The new air quality directives oblige the Member States to divide their territories in zones

Zones are primarily units for air quality management

Municipal Administrative Boundaries

but the directives also specify assessment requirements per Zone

Homogeneous Data on Air Pollution

These requirements depend on how far air quality levels are below a limit value

Air Quality Data in All the Territory

it is clear that the collection of air quality data throughout the territory is the priority in order to properly zoning the entire region

The Piedmont region in 1999 (year of implementation in Italy) did not have an adequate network of continuous monitoring

The estimation methodology used in reference to the limits of long-term (annual average) for nitrogen dioxide and PM10 is based on the correlation between the amount of pollutant emitted annually per unit area in a given municipality, and the concentrations found in the same town stations of regional air quality monitoring network

The simplifying assumptions underlying this approach is to consider that the average concentration of a pollutant in the territory of a municipality is substantially dependent on sources of emissions within the same municipality

It should be stressed that the methodology based on regional emissions inventory provides, at present, an estimate of the average concentration of a pollutant in the territory of a municipality

So, 10 years ago, we have used one incorrect methodology for zoning an entire region, using only few monitoring data and emission data (not so bad) for every municipalities !

This is the correlation between emission NO x[t/km 2 ] and NO 2[ m g/m 3 ]

Regional Emission Inventory: yearly total emission

Experiences in establishment of zones and agglomerations: Piedmont NO x NMVOC PM 10 14.

Piemonte : Zones and agglomerations (after preliminary assessment)

Experiences in establishment of zones and agglomerations: Piedmont 15. CTM application: air quality assessment & zoning According to the European Directive 96/62/EC and the Italian Legislation (D.L. 351/1999), the Regional administrations are in charge to perform theair quality assessment (AQA)in order to divide their territories in homogeneous zones and assess air quality within them, taking into account the different pollutants limit and threshold values Since 2005, Piemonte authorities have charged the ARPA Piemonte to perform yearly AQA integrating the information provided by the regional monitoring network with concentration fields supplied by a CTM modelling system. The modelling system has been built and optimised performing year 2004 AQA and successively applied for year 2005, 2006 and 2007. The modelling system works on a computational domain covering the whole Piemonte Region with an horizontal resolution of 4 km and 12 vertical levels, spanning the lower 3500 metres of the atmosphere. Experiences in establishment of zones and agglomerations: Piedmont 16.

Air quality assessment modeling system architecture

Pollutants: SO 2 NO x NO 2 COPM 10 PM 2.5 O 3 C 6 H 6 Simulation time : 1 year Temporal resolution : 1 hour Meteorological fields reconstructed integrating analysed fields from the European Centre for Medium Range Weather Forecast (ECMWF) with observations (radiosoundings and surface data provided by the regional meteorological network) by means of a mass-consistent model Boundary conditions obtained assimilating continental runs of the CTM CHIMERE and air quality monitoring observations Post-processing tools to compute all the indicators required by EU air quality legislation.220 x 284 km 2 xy:4 km Nx = 56 Ny= 72 Vertical levels: 12 (up to 3500m agl) Simulation length: 1 year Output fields time resolution: 1hour Experiences in establishment of zones and agglomerations: Piedmont 17. Modelling system main modules Experiences in establishment of zones and agglomerations: Piedmont METEOROLOGICAL DATA MINERVESURFPROEMISSION MANAGERFARM GEOGRAPHICAL DATA EMISSION DATA METEOROLOGICAL INPUTEMISSION INPUT IC/BC 18.

FARM: (Flexible Air quality Regional Model)

3D Eulerian model, derived from STEM-II(Carmichael et al.)

Diffuse sources & LPS with plume rise

Horizontal adv.-diff.: Blackman cubic polynomials(Yamartino, 1993)

Vertical adv.-diff.: hybrid semi-implicit Crank-Nicolson / fully implicit scheme(Yamartino et al., 1992)

Actinic flux reduction effect from clouds

SOx-NOx-NH3 simplified scheme(EMEP)

Photochemistry: SAPRC-90 chemical scheme

PM: Models-3/CMAQ aero-3 module(Binkowski, 1999) ; aero-0 simplified bulk module

One- or two-way nesting

Emission data

The Emission component is prepared from IREA (Regional Emission Inventory) by EMMA code (Arianet), to obtain hourly emission rates gridded on the regional domain and nearby regions.The emissions inside IREA are defined on a polygonal bases (details on municipal or provincial boundaries) and a specific module of EMMA is able to take into account the characteristics of land-use componentsinside each cell, so that emissions associated with individual human activities (production of energy, agriculture, farming, transport, waste treatment, etc ...) or natural (biogenic, soil dust, ecc. ) are distributed only into the cells of the domain intersected by each polygon emission and the appropriate land-use type. Experiences in establishment of zones and agglomerations: Piedmont DTM 250 m CORINE LC 250 m Regional transport graph Urban area from CTR INEMARPiemonte EMEP CORINAIR Italy Valle d'Aosta RegionalInventory INEMAR Lombardia 20.

Emission data treatment

Example:

A: distribution of conifers by CORINE Land CoverB: spatialisation biogenic emissions from conifer grills on the domain

Experiences in establishment of zones and agglomerations: Piedmont A B 21. 99 total observations Observations: Wind, T, Q 30 grid points (0.5 lat lon resolution), every 6 hours ECMWF Analyses:

Meteorological data

Experiences in establishment of zones and agglomerations: Piedmont (Cuneo Levaldigi e Milano Linate) 2 Radiosoundings: VV, DV, T, RH VV, DV, T, P, RH (hourly data) 94 surface stations: 3 WMO (Synop): VV, DV, T, P, RH, (every 3 hours) 22.

Stations & fields used to define initial and boundary conditions

ChimereContinentalrun provided byPrevAir

0.5 space resolution, 8 vertical levels

Chemical mechanism: MELCHIOR + aerosol

Air Qualitymeasurementsfrom Piemonte and nearby Italian Regions networks Experiences in establishment of zones and agglomerations: Piedmont 23.

Evaluation of model performances: PM 10

2005 PM10 yearly averages 2005 PM10 daily averages Station: Torino-Consolata Experiences in establishment of zones and agglomerations: Piedmont 24.

Evaluation of model performances: NO 2

2005 NO 2daily averages Station: Torino-Rebaudengo (roadside) 2005 NO 2hourly averages Experiences in establishment of zones and agglomerations: Piedmont 25.

Evaluation of model performances: O 3

Station: Novara-Verdi(urban background) Station: Buttigliera dAsti(rural background) Experiences in establishment of zones and agglomerations: Piedmont 2005 O 3max 8-hour average 2005 O 3max 8-hour average 26.

The Piemonte Region Air Quality Assessmentmodelling system has been evaluated comparing modelled concentrations with measurements from stations located in different geographic, topographic and land cover environment.

O3 indicators shows satisfactory results, with good accuracy in reproducing measured levels, for both eight-hours and hourly averages.

NO2 concentration fields are satisfactorily reproduced, yearly averages and concentration statistics are compatible with measured values. The model accuracy remains within the requested limit for most monitoring sites. Some underestimation of peak concentration has been detected where model resolution is insufficient and during adverse meteorological conditions (severe episodes).

PM10 concentrations have been reconstructed with values within the requestedaccuracy for both daily and yearly averages at most station locations, even though a clear underestimation can be noticed for stations located outside the Torino metropolitan area.

This shortcoming can be partly ascribed to a certain underestimation of emissions outside Torino (e.g. from wood burning) and to the difficulty to simulate PM accumulation phenomena within the Po valley.

Regione Piemonte implemented the modelling system as a permanent air quality assessment tool and ARPA Piemonte is now in charge to perform modelling applications on a yearly basis.

Piemonte 2005: PM 10- Zones and agglomerations

Piemonte 2005: NO 2- Zones and agglomerations

Piemonte 2005: O 3- Zones and agglomerations

the problem is:

Experiences in establishment of zones and agglomerations: Piedmont PM 102005yearly average The AQA modelling system provides data on a regular grid We need to aggregate data at municipality scale ( change of support problem ) 31.

The output fields produced by the modeling system are superimposed (using GIS GRASS) to vector thematic maps that describe the region:

in this way are identified for each administrative unit, the cells that fall within it, and, on the basis of intersection polygon-cells, the breakdown of communal cells

Experiences in establishment of zones and agglomerations: Piedmont Note: the grid points belonging to the municipality of Alessandria However, very few municipalities in Piedmont have more than 5 data 32.

the aggregation at municipality scale can be realized solving the so-called change of support problem.

A simple solution is to integrate over the area, that means to average the field values weighted by areas over the cells belonging to a certain municipality.

Two alternatives for the aggregation are the average of the field valuesweightedby thebuilding percentagefor every cell (a point in the grid represent a cell) and the90th percentileover the cell values in a municipality.

Thebuilding percentageis an important indicator of theantropic activitywhich could generate more pollution in a municipality, whereas a wide country area could not contribute at all.

Instead the90th percentileis chosen as a measure ofextreme casesin a municipality, in a precautionary perspective.

Methodology

Experiences in establishment of zones and agglomerations: Piedmont

Weighted block average (weight = municipality area percentage)

Weighted block average (weight = municipality built area percentage)

Weighted block average (weight obtainedusing municipality area and build area percentage)

90 thpercentileor maximum over the cell valuea in a municipality

Comparison between the three algorithms (examples of Biella, where we see major differences - with built concentrated in a restricted area) and Torino (little difference)

Functional approach for land classification using data coming from AQA modelling system - Cluster Analysis

Pollutant time series can be considered as realizations of continuous processes that are recorded in discrete time. The conversion from discrete data to curves involves smoothing, using linear combinations of B-spline functions. Thus we have functional data that can be classified by clustering procedures. The non hierarchical algorithm PAM (Partioning Around Medoid) is used to cluster objects.The parameterization of functions using B-splines can replace the classical analysis by synthetic indicators (as average levels, standard deviations, quantiles) to respect the functional form of pollutant concentrations through time. Synthesizing features contained within time series in a little number of coefficients, we preserve in the zoning most information about pollutant concentration levels and time profiles.

This approach is applied to functional data of the biggest critical atmospheric pollutants in Piemonte.

Clustering results

Experiences in establishment of zones and agglomerations: Piedmont NO2 (left, 4 cluster)and PM10 (right, 3 cluster) Data coming from simulations for AQA 2005 37.

Multi-pollutant zoning

To support decisors considering jointly different pollutants, in order to get a multi-pollutant zoning, we propose to summarize them in an air quality index (1) or, alternatively, in principal components (2).

We propose to aggregate over pollutants by the maximum function, in order to keep information about critical cases, to obtain air quality index time series for all the municipalities. Considering these time series as functional data we can cluster them and obtain groups of municipalities, through a functional cluster analysis where PAM algorithmis embedded

Multi-pollutant zoning

As alternative technique, we explore the Functional Principal Component Analysis (FPCA) in its multivariate version, since we consider different pollutants taking into account their interactions. Then we apply the PAM algorithm to the scores of the principal components, obtaining groups of municipalities. In this case the medoids are scores where the temporal component is integrated out

The Difference

1999 2009 40.

Air Quality Plan

only using quality models we can take good plan with reasonable scenario.

Italy: national, regional, provincial & municipal plans

Regional decision maker: LV will be respected ?

role of local measures ?

Develop coherent evaluations, taking into account:

local features(sources, topography & land-use, meteorology)

broader context

Experiences in establishment of zones and agglomerations: Piedmont 41. Regional AQ system Boundary conditions Meteo model Emissionmanager Dispersion& chemistry Impacts ReferenceInventory Regional plan Emission scenarioAtmospheric simulation system Experiences in establishment of zones and agglomerations: Piedmont Costs Emissions Atmospheric dispersion Health & env. impacts Enargy / ind. / agric. Projections Emission control RAINS-Italy RAIL 42.

Methodology

Inventories harmonization:IREA (Regional Emission Inventory) vs. RAINS-Italy, for baseline year

Projected emission scenario:RAINS-Italy CLE, downscaled at municipality level using IREA

Regional Plan emission scenarios:translation (at municipality level) of local measures

Impacts on concentrations:simulation through regional 3D atmospheric modelling system

Ongoing process, promoted by the Ministry of the Environment, with the progressive involvement of Regional Governments Experiences in establishment of zones and agglomerations: Piedmont 43.

Piemonte AQ Plan: measures at 2010

Heating

energy efficiency (new & renovated)

boilers: efficiency & emission limits

ban of dirtier fuels (coal and distillate oil)

incentives for solar heating for sanitary water

district heating expansion

Transport

whole region: progressive ban of the most polluting vehicles

Plan Zones: restricted traffic zones in municipalities > 10000 inh.

adoption of DPF (diesel particles filters)

Piemonte Case: Effects on emissions

Variations for Plan 2010 scenarios respect to Base 2005

(all categories)

Detail on individual measures Heating

Plan 2010 vs. Reference 2010 variations,by measure (local measure)

Detail on individual measures Mobility

Plan 2010 vs. Reference 2010 variations,by type (example)

Piemonte Case: Effects on concentration

NO 2concentrations, yearly averages

Plan-R 2010 vs. Baseline 2005 (CLE + local measures)

PM 10concentrations, yearly averages

Plan-R 2010 vs. Baseline 2005 (CLE + local measures)

Room for improvement ?

(RAINS technical measures) Experiences in establishment of zones and agglomerations: Piedmont 52.

Lessons learned

implementing integrated N policies: difficulties & suggestions

There is room for local, coordinated policies

supported by regional- and urban-scale analyses(spatial details, hotspots)

linked to broader context

multiple models: need of consistent tools & data - harmonization

Long-term perspective(4/5 years)

Lack of multidisciplinary connections(e.g. mobility, agriculture)

especially on quantitative approaches

Italian peculiarity ?

Experiences in establishment of zones and agglomerations: Piedmont 53. Thank You For Your Attention Experiences in establishment of zones and agglomerations: Piedmont

the experience of piedmont in air quality

Education