next
previous
items

3. General Considerations on Revised Data Model

This chapter presents a general reflection on air emission inventories corresponding to existing, short, medium or long development terms.


3.1  Substances

Substances for which emission estimations must be provided derive from those concerned by requests listed in section 1.

USER DESIGNATION POLLUTANTS
UNECE/EMEP SO2, NOx, NMVOC, CH4, CO, NH3
PARCOM-ATMOS HEAVY METALS
  As Arsenic
  Cd Cadmium
  Cr Chromium
  Cu Cupper
  Hg Mercury
  Ni Nickel
  Pb Lead
  Se Selenium
  Zn Zinc
POPs HCH hexachlorocychlohexane
  PCP pentachlorophenol
  HCB hexachlorobenzene
  TCM tetrachloromethane
  TRI trichloroethylene
  PER tetrachloroethylene
  TCB trichlorobenzene
  TCE trichloroethane
  DIOX dioxins and furans
  PAH Polycyclic Aromatic Hydrocarbons
IPCC NOx, NMVOC, CH4, CO, CO2, N2O and to be defined more precisely
  (SOx, HFCs, PFCs, SF6)
EU LCPD SO2, NOx
EU CO2 CO2

3.2  Emission generating activities (for long term development)

The reflection is based on a similar concept which has been used for the ACCOR nomenclature developed in the spirit of integrated emission inventory.

The concept of activity is based on four complementary components.

Any source of pollutant can be expressed as :

belonging to one economical sector specified by operations, using machines/technologies generally involving commodities (e.g. products/fuels).

Each of these four components may be more or less split as shown on figure 3.



This approach which includes potentially elements to satisfy a lot of needs, can be implemented practically by different ways to be examined further in details. Nevertheless in front of the increase of information to handle, it will be necessary to prioritize some of them.

Moreover, changes in emission generating activities need to be :

  • agreed on a common basis which fits as far as possible with all national specificities and practices,
  • in accordance with available data. When absolutely necessary, it should be recommended to make available some information,
  • aware of different requests which imply to consider the minimum common level of details, that is always more detailed than any of individual requests.

Consequently, it is recommended to introduce changes in emission generating activities definition and structure cautiously and progressively.

As basis for further consideration, there is the following formal proposal from the expert panel on projection.


E = S [ Ai,j,k . [ S Fj,k,l . Pi,j,k,l ] ]

i,j,k l

with E total emission for one pollutant

A activity rate

F emission factor

P fraction of sector, activity, fuel and technology

i economic sector

j source type

k fuel type

l technology type (including control device)


3.3  Types of sources

Emission generating activities can be split in different source types :

  • Large point sources (LPS) which correspond to actual or potential large emitters.

Such sources must be investigated on an individual basis in order to :

- satisfy some requests (e.g. Large Combustion Plant Directive, modellers, ...)
- increase the accuracy of inventories.

The definition of LPS is provided in section 4.2.3 and annex 3.
It is to be noticed that some LPS are sometimes extended sources (e.g. international airports).

  • Medium or small point sources which correspond to other point sources generally investigated on an extended geographical scale.
  • Extended area sources (landfill, swamps, rice cultivation, ...).
  • Mobile sources which can be split in :

- linear sources (highways, main roads, inland traffic, air or marine traffic, etc...)
- "Brownian" sources (urban/local traffic).


Data model of these types of sources :

Large point sources are treated on an individual basis.

Sometimes medium sources need to be treated individually (e.g. some Large Combustion Plants) but generally medium and minor sources are aggregated with extended area sources and "Brownian" mobile sources within so called area sources relating to more or less extended geographical zones.

Linear sources have not been considered specifically in CORINAIR until now but included in area sources.

While LPS emission are estimated either generally from specific data (measurements, mass balance, specific emission factor, ...) or sometimes with general emission factor, area source emissions are generally estimated from average emission factors.


3.4  Spatial and time resolution

The spatial resolution depends on the goals previously defined and differs considerably when interest focuses on local air pollution or deals with national figures.

With regard to needs expressed in section 1, the national level is mainly requested. Nevertheless, EMEP needs data by a 50 x 50 km grid squares. Most countries provide also periodically figures spatially disaggregated.

As previously demonstrated in CORINAIR, statistical data are generally available at administrative territorial units levels and moreover estimations on this pragmatical basis are of interest. These data can be allocated to various cells (i.e. EMEP but also GEIA, LOTOS, ...) by using relevant keys.

It is recommended to maintain the geographical resolution based on territorial units defined by Eurostat NUTS level III. They correspond fairly well to needs expressed by different users (e.g. modelling groups) at the European scale and constitute a manageable harmonious combination of about 1300 units for EU-15 and 1800 units for 28 countries providing CORINAIR 90 data.

Both levels (national and spatially disaggregated) will be considered. The spatially disaggregated data being provided periodically only in accordance with EMEP periodicity (at this time every four years).

With regard of all international requests described in chapter 1 emission need to be estimated every years at national level.


3.5  Energy balance

It is recommended to include national energy balance facilities (which was not included in previous CORINAIR 90 system) :

  • to improve the consistency of the emission estimates
  • to facilitate the achievement of inventories
  • to speed up CO2 emission estimation
  • to get a better compatibility with IPCC format and to facilitate the CORINAIR/IPCC conversion.

The energy balance can be made for more or less detailed fuel types according to available statistics. It is to be noticed that international statistics (e.g. EUROSTAT) are available for a large number of (European) countries on a consistent basis by more than 20 different fossil fuel types.

Moreover, a split by main economical sectors should be requested.


3.6  Overview of air emission inventory model

The air emission inventory model is based on the relevant arrangement of different specific modules which have to fit with different kinds of needs/qualities :

  • annual national level data
  • periodical spatial level data
  • individual large point source data
  • transparency
  • quality assurance.

Each module includes all or some of the following elements :

  • data management procedures
  • calculation procedures
  • functions insuring convergence or split of data.

As far as the geographical resolution of the inventory is on one hand the national level and on the other hand a finer territorial unit (e.g. NUTS level III), two different approaches are used :

  • some national systems or source activities only have data at national level which must be spatially allocated using surrogate data (top-down approach),
  • other national systems provide data (activity, emission factors, point source data) on a local basis which can be aggregated up to national level (bottom-up approach).

Generally, both approaches are mixed and the air emission model must allows to use them.

 

3.6.1  Description of components

The description of the model is given below, the figure 4 presents the corresponding flowsheet which is very useful to have a complete overview especially concerning the links between the different modules. Application is described in section 3.6.2.


Image3.gif (21583 bytes)



National aggregated data (NAD) (not existing in CORINAIR 90)

Function of NAD: to store aggregated data (activity rates, fuel/energy consumption, emission factors and emissions) at national level by source categories.

It is generally easier and quicker to use global figures than area and individual large point sources figures, to provide national estimations ; namely, aggregated source level instead of area and large point sources levels. Nevertheless, a detailed approach must be also practicable for national estimation.

Relation with other components:

  • with energy balance (EB)
  • with national detailed area sources (NDAS)
  • with large point sources data (LPSD)

The relevant source categories considered here correspond to "emission generating activities" (discussed in section 3.2). They are based on a more or less detailed level of each of the four components of emission generating activities, in such a way that all international needs may be satisfied. These "emission generating activities" have attached aggregated emission factors for calculating the emissions.

For energy related activities the fuel component is necessary (e.g. NAPFUE or equivalent). The fuel split has to be at least the same or possibly more detailed than the energy statistics fuel split considered in EB. In the same way the sector activity split in NAD has to be the same or to be more detailed than the sector split used for energy statistics in EB.

NAD module enables to provide national estimations for UNECE, IPCC, PARCOM/HELCOM, ...

EMEP and LCPD needs cannot be satisfied at this stage.


Energy balance (EB) (not existing in CORINAIR 90)

Function of EB:
- to carry out an energy balance between energy statistic and energy input in the system.
- to store possibly national energy statistics by fuel types and economical sectors.

Relation with other components:
with national aggregated data module (NAD).

The energy taken into account in NAD can be easily compared with reference national energy consumption.


Large Point Sources data (LPSD) (existing in CORINAIR 90)

Function of LPSD: to store specific data relating to sources considered individually.

Relation with other components:
- with national aggregated data (NAD) as detail level
- with national detailed area sources (NDAS) as complementary information
- with spatially disaggregated inventory (SDI) as one of its components

LPSD module constitutes a module oriented for requests such as LCPD, EMEP, ...

LPS have to be arbitrarily defined according to objectives of inventories (e.g. cf LPS definition of CORINAIR 90) (4).

LPSD deals with a lot of parameters concerning the identification and the location of plants. Specific information such as activity rates, capacities, energy consumption, processes, controls, emissions, working time, exhausting characteristics are generally collected.


National detailed area sources data (NDAS) (existing in CORINAIR 90)

Function of NDAS: to store area activity rates, energy consumption, area emission factors of emission generating activities according to the detailed source category split selected by the national expert. The provision of these data instead of direct emission input allows transparency.

Relation with other components:
- with national aggregated data (NAD) by the way of national aggregation function (NAF) in combination with LPSD
- with spatially disaggregated area sources data either via allocation procedures (top-down approach) or via bottom up function (BUF) (bottom-up approach).

NDAS and LPSD constitute the national level detailed reference inventory.


National aggregation function (NAF) (not existing in CORINAIR 90)

NAF function: to get a national set of national aggregated data by aggregating source categories data from national area sources and large point sources.

Relation with other components:
- with national detailed area sources data (NDAS) and large point sources data (LPSD) as input,

- with national aggregated data (NAD) as output.

NAF is part of NDAS module.

Spatially disaggregated inventory (SDI) (existing in CORINAIR 90)

Function:
- to provide emission estimations by pollutants, territorial units and emission generating activities.

- to store area sources data (activity rates, energy consumption, emission factors, surrogate data).

It proposes the allocation procedure which allows to estimate unknown activity rates from an upper level by using socio-economical parameters.

It includes large point sources data (LPSD) as complementary component.

Relatmponentsion with other co:
- with national detailed area sources data as a breakdown (top-down approach) and via bottom-up function (BUF) (bottom-up approach).

SDI allows to satisfy requests from modellers (e.g. EMEP).


Bottom-up function (BUF) : spatial aggregation (existing in CORINAIR 90 but available separately of software)

Function: to provide national detailed area data from sub-national levels

Relation with other components:
- with area sources data

- with national detailed area data (NDAS).

BUF is part of SDI module.


 

3.6.2  Implementation of the revised data model

3.6.2.1  General overview

Pragmatically, as shown by figure 5, national CORINAIR inventory can be achieved according to national preferences as follows :

  • by using the CORINAIR software (e.g. when one country is using the CORINAIR System as its national system).
  • by transferring data within the CORINAIR database.
  • In this case, there is two alternatives to provide NAD figures:

* to transfer directly area sources data and LPS data. Therefore, national aggregated data will be obtained automatically by using the CORINAIR System.

* to transfer directly national data into NAD module of EEA/CORINAIR database.


Regarding CORINAIR 90 System, this scheme included a supplementary intermediate step : transfer of DOS/CORINAIR databases into ORACLE/centralized CORINAIR database (including data dictionary transformation).

In order to avoid such an intermediate step, it is proposed to have only one single data dictionary to be used by individual CORINAIR databases and by the centralized EEA database.

Due to the use of one single data dictionary, CORINAIR databases under different platforms are equivalent by using existing ODBC drivers. Each inventory producer may be free to chose his most convenient platform.

3.6.2.2  Dynamic process

The data model is designed for a dynamic application as presented in figure 6.

NAD of the previous year (N-1) are duplicated. Major data available for the year N are taken into account (activity rates, energy consumption). Emission factors are assumed relevant as well for the year N-1 as the year N as far as there is no major changes in the source categories structures. Corrections have to be made when major changes are identified.

This process allows to provide estimations in due time.

It can be refined progressively according to specificity’s and willingness of each country.

When detailed data (NDAS and LPSD) are available (possibly the year after), the NAF procedure aggregates these data (i.e. calculates national aggregated emission factors) and replaces the preliminary and provisional NAD data with them.

It is also possible to consider an alternative way, which consists in introducing only data into NDAS and LPSD, then to aggregate them in NAD with NAF.

Periodically (e.g. every 4 years), the spatial disaggregated inventory is achieved within two years after the reference year.

3.6.2.3  Detailed implementation for one current year

The figure 4 shows the flowsheet of the model application.

National aggregated data (NAD) are available :

  • either from direct data input by duplication of NAD from the previous year and progressive replacement with data specific to the current year.
  • or from detailed data (NDAS and LPSD) aggregated by using the NAF function.

The first approach needs very few time and reporting of preliminary emission estimations can be performed within six/twelve months. Moreover, data can be progressively completed and more accurate estimations provided some months later.

Due to the very high contribution of energy related sources considering some pollutants, it is essential that inventories take into account relevant energy figures. Especially, energy data input in NAD have to be well balanced with national energy statistics in EB module. It provides feed-back information to refine energy related data within the inventory.

Large Point Sources data have to be introduced as complementary information :

  • every year when specific requests has to be complied from this inventory (e.g. LCPD inventory). In such a case the LPSD module is partially filled in for relevant combustion plants only.
  • when national detailed data have to be considered or when a spatially disaggregated inventory has to be performed.

Image4.gif (11163 bytes)


Image5.gif (16184 bytes)

The inventory producer can start without considering LPS data and, if necessary, will fill in progressively LPS figures which are necessary in further steps.

Reference specifications have to be available to define LPS.

Periodically (e.g. every 1 to 4 years) the NAD (National Aggregated Data) must be considered in more details to establish update reference by specifying area sources and large point sources. The latter is already known at least partially and can be completed according to CORINAIR definitions and national wishes. Therefore, national detailed area sources (NDAS) must be provided.

More detailed means here to split, when relevant some source sectors considered in NAD, in order to appreciate finest details (i.e. differences of technologies). Example, cement production as a whole in NAD can be split in dry, wet or semi-wet processes.

Alternatively to the previous process, it may be possible to input NDAS and LPSD data annually and to use the NAD module as an output for the annual needs.

In spite of fusion of AS + LPS data and possible more detailed split of information, there is a closed link between NDAS and NAD. Three situations are existing for emission generating activity.


AS LPS COMMENT
Yes No For most activities: NDAS=NAD for others NDAS includes more details
Yes Yes Specific area data in NDAS (i.e. emission factors are necessary)
No Yes No area sources in NDAS

The convergence process : National Aggregated Function (NAF) is used to build or to update NAD.

From the national reference inventory (NDAS + LPSD) data are aggregated according to the format considered in NAD. Especially aggregated emission factors are calculated.

NDAS and LPSD constitute the national reference inventory. From these detailed data NAD can be built or updated (aggregation) when necessary. The finest structure of information (i.e. emission factors) referring to one reference year N can be used automatically for the preliminary estimation of the year N + 1, possibly N + 2, ...

Periodically, a spatial disaggregated inventory (SDI) has to be achieved. In such a case NDAS activity rates must be distributed over different territorial units (TU). Either activity rates are known at some levels and directly introduced, or activity rates are allocated according to surrogate data and procedures defined by each expert.

When necessary TU specific emission factors can be attached to some activity rates, so weighted emission factors must be generated at intermediate levels (cf. bottom-up function).

LPSD are already available for any territorial unit.

The bottom-up function (BUF), which is a spatial aggregation function, is to be used in both following cases : activity rates aggregation and emission factors aggregation.

  • when activity rates are known, at sub-national levels, BUF aggregates activity rates at upper TU levels.
  • when TU specific area emission factors are introduced BUF is used for calculation of weighted emission factors at upper TU levels.

Example: When it is assumed by one expert that all data are known at level NUTS III, it is possible to calculate automatically activity rates and weighted emission factors for upper levels.

lijn.gif (900 bytes)

up.gif (859 bytes)
left.gif (869 bytes) right.gif (869 bytes)


</head0

Permalinks

Document Actions