Economic valuation of environmental effects of NOxemissions from air traffic at different altitudes

1
Economic valuation of environmental effects of
NOx-emissions from air traffic at different
altitudes

• Robert Bergström and Joakim Langner
• SMHI The Swedish Meteorological and
  Hydrological Institute
• Robert.Bergstrom_at_smhi.se
• Lena Nerhagen
• VTI The Swedish National Road and Transport
  Research Institute
• Bertil Forsberg
• Umeå University

2
SMHI The Swedish Meteorological and
Hydrological Institute www.smhi.se

• Meteorology, Hydrology, Oceanography, Climatology
  and related Environmental fields
• Operates under the Swedish Ministry of the
  Environment
• 550 employees
• Products and services include
• Forecasts, Analyses, Surveys, Statistics
• Research and consulting

3
SMHI Aviation

• The Environmental Safety Service Department
• Forecasting service for the aviation sector
• Climate and Environmental Studies of Airports
• Hans.Backstrom_at_smhi.se
• The Research Department
• Regional and continental scale modelling of air
  traffic emission impacts
• Robert.Bergstrom_at_smhi.se

4
Background

• Swedish transport policy implies that the
  external marginal costs of transport should be
  the basis for taxes and charges
• The external marginal costs include the cost of
  the wear and tear of infrastructure, environment,
  noise, accidents and congestion
• Environmental costs may, e.g., include damages to
  natural ecosystems, agricultural and forest
  production, human health and changes in climate

5
Background

• Air traffic emission charges usually only
  consider local air quality issues and take into
  account emissions in the Landing and Take-Off
  (LTO) cycle
• Only part of the aircraft emissions are released
  during the LTO-cycle (for Swedish air space the
  non-LTO emissions of NOx are about 5 times larger
  than the LTO-emissions)

6
Background

• If the contributions from non-LTO emissions to
  environmental problems are significant it is
  reasonable to introduce NOx emission charges also
  for these emissions
• The Swedish Civil Aviation Authority have
  therefore initiated studies aimed at providing
  the information needed for economic valuation of
  the environmental effects caused by aircraft
  emissions at different altitudes in Sweden

7
The Impact Pathway Method

• The Impact Pathway (IP) Method is a bottom-up
  approach used for assessment of external impacts
  and associated costs resulting from energy
  production or transport
• The IP Method was developed within the ExternE
  research projects
• An impact pathway is the sequence of events
  linking a burden or emission to an impact and
  subsequent valuation

8
Schematic view of the Impact Pathway approach
9
Results marginal costs of LTO NOx emissions in
Sweden
10
Results marginal costs of non-LTO NOx emissions
in Sweden
11
Results Summary

• Marginal costs for NOx emissions from aviation in
  Swedish air space are estimated to be higher for
  LTO emissions than for non-LTO emissions
• LTO-emissions ca 1.89 EUR (17 SEK) / kg NOx
• Non-LTO-emissions ca 0.91 EUR (8 SEK) / kg NOx
• The corresponding total NOx emission costs from
  emissions in Swedish air space are higher for
  non-LTO emissions than for LTO emissions
• LTO-emissions ca 1.7 million EUR
• Non-LTO-emissions ca 4.4 million EUR
• Using ExternE methodology lead to higher costs
• LTO-emissions ca 7.2 million EUR
• Non-LTO-emissions ca 13 million EUR

12
Results Summary

• The costs due to acidification and eutrophication
  of natural ecosystems dominate but they are very
  uncertain.
• If ecosystem effects are excluded the marginal
  costs become almost the same for LTO and non-LTO
  emissions of NOx.
• LTO-emissions 0.63-0.77 EUR / kg NOx
• Non-LTO-emissions 0.56-0.74 EUR / kg NOx
• Costs due to crop loss and effects on
  agricultural soils were estimated to be much
  smaller than costs due to health effects.

13
Outline

• Environmental effects studied
• Models used
• Emissions scenarios
• The results in more detail

14
Considered effects in the valuation

• Impact on human health from surface ozone and
  particulate matter (PM)
• Impact on agricultural crops from surface ozone
• Impact on agricultural soils and natural
  ecosystems from acidifying and eutrophying
  deposition of oxidised nitrogen

15
Some excluded effects in the valuation

• Impacts on climate
• Impact on forests and natural vegetation from
  surface ozone
• Impact on building materials through corrosion
  and soiling and impacts on crops through SO2

16
MATCH, 3D transport/chemistry/deposition model
3D meteorological data, HIRLAM, ECMWF, ERA40,
MM5, Aladain, RCA...
Emission data e.g. EMEP-2000
3D transport/chemistry model, MATCH
Chemical mechanism, KPP syntax
Land use - e.g. forest
17
Modelling domains
18
Models emission data

• Detailed emission data for traffic in Swedish air
  space were taken from the projectInput data
  for model studies of environmental effects of
  NOx-emissions from air traffic at different
  altitudes Commissioned byThe Swedish Civil
  Aviation Administration, Luftfart och samhälle
• Joakim Langner and Robert Bergström SMHI
• Jana Moldanová IVL, (The Swedish Environmental
  Institute)
• Anette Näs och Anders Hasselrot FOI (The Swedish
  Defence Research Agency)

19
Emission data for traffic in Swedish air space

• Emissions of nitrogen oxides (NOx), carbon
  monoxide (CO), Volatile Organic Compunds (VOC),
  and sulphur dioxide (SO2) were derived, for
  Swedish air space, for the year 2002
• Horizontal resolution 20 km
• Vertical resolution 500m
• Information about fuel consumption was also
  derived with the same geographical resolution

20
Emission data for traffic in Swedish air space

• LTO-emissions based on official statistics from
  LFV
• 19 airports, run by LFV, included
• non-LTO-emissions
• Domestic traffic
• International flights
• Overflights
• Temporal emission variations based on traffic
  statistics from Arlanda airport 2002
• Monthly variation
• Weekday variation
• Diurnal variation

21
Geographic distribution of fuel consumption 2002
22
Vertical distribution of total emissions in
Swedish airspace 2002
23
Global emission data from ANCAT/EC2 and DLR
NOx
Unburned hydrocarbons
Unburned hydrocarbon (UHC) emissions. Refined
using fuel use data from the higher resolution
ANCAT/EC2 database
The fine grid emissions of NOx from civilian
aircraft
24
Emission data

• Large scale aviation emissions from the ANCAT/EC2
  and DLR2 databases, 3D, increased by 2,6/year
  from 1992
• Swedish aircraft emissions from FOI
• Non-aircraft emissions (SO2, NOx, NH3, NMVOC and
  CO) from CLRTAP/EMEP
• NOx emissions from lightning, calculated in the
  MATCHmodel
• Biogenic emissions of VOC (isoprene), calculated
  in MATCH
• Sea salt emissions, calculated in MATCH

25
Impact of NOx-emissions from air traffic in
Swedish air spaceScenario-calculations
Model calculations using refined aircraft
emissions in Sweden Three different emission
scenarios have been investigated A. All aircraft
emissions included B. LTO emissions of NOx
excluded C. non-LTO emissions of NOx excluded The
LTO-impact is given by the difference A B The
non- LTO-impact is given by A C
26
Model results

• Model calculations have been performed for one
  year (2000)
• The following data are stored
• Hourly concentrations of surface level ozone and
  NOx
• Daily mean concentrations of fine particulate
  matter
• Daily deposition of acidifying and eutrophying
  species
• Statistics of the impacts of Swedish aviation
  emissions in different countries are calculated
  (annual averages of impacts)

27
Deposition of oxidised nitrogen

• Aircraft emit NOx mainly in the form of NO
• NO is transformed in the atmosphere to NO2 and
  other oxidised nitrogen species, such as nitric
  acid, HNO3
• The oxidised forms are finally deposited at the
  surface of plants, soils and water
• Deposition of oxidised nitrogen has several
  environmental effects
• Acidification of ecosystems and agricultural soil
• Fertilisation/Eutrophication of agricultural and
  natural ecosystems

28
Deposition of oxidised nitrogen
LTO-contribution non-LTO-contribution
Model calculated contribution to the deposition
of oxidised nitrogen from emissions of NOx in
Swedish air space. Unit mg/m2.
29
NOx-deposition effects on agriculture

• The economic effects on agriculture of
  NOx-deposition due to aircraft emissions are
  small
• Potential reduction of fertiliser use (i.e.,
  savings)
• LTO emissions 0.016 EUR / kg NOx
• Non-LTO emissions 0.008 EUR / kg NOx
• Potential increase in liming cost to counter
  acidification of agricultural soil
• LTO emissions 0.002 EUR / kg NOx
• Non-LTO emissions 0.0009 EUR / kg NOx

30
NOx-deposition valuation of ecosystem effects
based on abatement costs

• ExternE projects have used relatively high
  abatement cost estimates for acidification and
  eutrophication of ecosystems
• 176 000 / km2 for acidification
• 25 900 / km2 for eutrophication
• Vermoote and DeNocker (2003) developed a
  Standard Price approach to be compatible with
  the ExternE methodology and estimate an abatement
  cost of
• 10 000 / km2 both for acidification and
  eutrophication

31
Acidification of ecosystems

• Estimated costs due to acidification of
  ecosystems
• Using standard ExternE valuation
• LTO emissions 4.57 EUR / kg NOx
• Non-LTO emissions 1.09 EUR / kg NOx
• Using Vermoote DeNockers Standard Price
  approach
• LTO emissions 0.26 EUR / kg NOx
• Non-LTO emissions 0.06 EUR / kg NOx

32
Eutrophication of ecosystems

• Estimated costs due to eutrophication of
  ecosystems
• Using standard ExternE valuation
• LTO emissions 2.60 EUR / kg NOx
• Non-LTO emissions 0.76 EUR / kg NOx
• Using Vermoote DeNockers Standard Price
  approach
• LTO emissions 1.00 EUR / kg NOx
• Non-LTO emissions 0.29 EUR / kg NOx

33
Impact on ozone concentrations near ground

• In the presence of sunlight and organic compounds
  NOx-emissions can lead to formation of ozone, O3
• Ozone is very toxic to plants
• Ozone is harmful to humans
• Ozone is a strong greenhouse gas

34
Ozone daily max 8h-mean conc
LTO-contribution non-LTO-contribution
Calculated annual average contribution to the
daily maximum 8-hour average concentration of
surface ozone from emissions of NOx in Swedish
air space. Unit ng/m3. 1 ng/m3 0.001 mg/m3.
35
Ozone population weighted exposure
POPULATION
36
Ozone population weighted exposure
LTO-contribution non-LTO-contribution
37
Ozone - valuation of impact on mortality

• Impact on mortality 0.3 risk increase per
  10mg/m3 increase in, maximum daily 8-h mean,
  concentration (short-term exposure)
• Only applied for persons above 30years old
• Assumed 1 year lost of life
• The value of a lost year of life was set to 73
  000 EUR
• (In earlier ExternE studies the risk was set to
  0.59 and the value of a lost year of life was
  set to 160 000 EUR)

38
Ozone - valuation of impact on mortality

• The resulting estimated marginal cost for life
  years lost by ozone exposure due to aviation
  NOx-emissions in Swedish air space are
• 0.038 EUR/kg NOx for LTO emissions (or 0.16 EUR
  using ExternE valuation)
• 0.039 EUR/kg NOx for non-LTO emissions (or 0.17
  EUR using ExternE valuation)

39
Ozone and Particulate Matter- valuation of
impacts on morbidity

• Impact on morbidity considered effects
• Chronic bronchitis
• Hospital admissions
• Cerebrovascular disease
• Heart failure
• Respiratory diseases (chronic cough, restricted
  activity)

40
Monetary values for morbidity (EUR)
41
Valuation of morbidity impacts - uncertainties

• According to a recent WHO review there are few
  European ozone studies using other endpoints than
  daily number of deaths.
• A few studies on hospital admissions did not show
  a significant overall estimate in single
  pollution models, which may be a result of a
  negative correlation between ozone and primary
  combustion products.
• Neither did studies on admissions for asthma in
  children find conclusive associations, which may
  be explained by increased medication when ozone
  levels are high.
• Studies of ozone exposure and asthma incidence
  and prevalence in children and adults are not
  consistent. Available evidence suggests that
  long-term exposure possibly reduces lung function
  growth in children.

42
Valuation of morbidity impacts - uncertainties

• For the purpose of this study we have chosen not
  to update the Exposure-Response (ER) relations
  for impacts on morbidity due to ozone and PM.
• Instead we indicate that the morbidity effects
  are in the range between zero and the ExternE
  estimates.
• This is in line with the assumptions made in the
  CAFÉ (Clean Air For Europe) work where morbidity
  is excluded in the analysis of abatement costs.

43
Ozone - valuation of impact on morbidity

• The resulting estimated marginal cost for
  morbidity by ozone exposure due to aviation
  NOx-emissions in Swedish air space are
• For LTO emissions 0 0.29 EUR / kg NOx
• For non-LTO emissions 0 0.30 EUR / kg NOx

44
Crop losses due to ozone exposure

• Elevated concentrations of ozone damages plants
  and thereby may lead to crop losses
• Slightly sensitive crops include rye, oats and
  rice
• Sensitive include wheat, barley, potato and
  sunflower
• Very sensitive include tobacco
• Damage to crops are assumed to be linearly
  dependent on accumulated ozone exposure above a
  certain threshold concentration (AOT40)

45
Impact on accumulated ozone exposure - AOT40
AOT40 (Accumulated Ozone exposure over Threshold
40 ppb(v)) Definition WHO guidelines for
protection of crops are based on
AOT40 Protection of crops against 5 yield
loss AOT40 May-July (daylight hours) lt 3
ppm h Protection of crops against 10 yield
loss AOT40 May-July (daylight hours) lt 6
ppm h Protection of natural vegetation AOT40
May-July (daylight hours) lt 3 ppm
h Protection of trees (forests) AOT40
April-September (daylight hours) lt 10 ppm h
46
Impact on accumulated ozone exposure - crops and
natural vegetation - AOT40 May-July
LTO-contribution non-LTO-contribution
Calculated contribution to AOT40 for May-July,
2000, from emissions of NOx in Swedish air space.
Units ppb(v) hours.
47
Valuation of crop losses due to ozone exposure

• Valuation of crop losses was based on national
  producer prices for the year 2000 for the
  different crops considered. Data were taken from
  EUROSTAT.
• Two different sets of dose-response relationships
  for various crops were used, one compiled by
  Friedrich and Bickel 2001 (ExternE) and an
  updated set by Holland et al. (2002)

48
Valuation of crop losses due to ozone exposure

• The Holland (2002) ER relationships give the
  costs
• LTO emissions 0.03 EUR / kg NOx
• Non-LTO emissions 0.05 EUR / kg NOx
• The ExternE ER relationships give the costs
• LTO emissions 0.09 EUR / kg NOx
• Non-LTO emissions 0.12 EUR / kg NOx

49
Impact on fine particulate matter - PM2.5

• NOx is transformed into HNO3, which can form
  particles
• NO2 gas phase oxidation
• NO2 OH ? HNO3 (day time reaction)
• NO2 O3 ? NO3 ? NO3- (night time reaction)
• Ammonium chemistry
• NH3(g) HNO3(g) ? NH4NO3 (s, aq) (equilibrium
  temperature- and humidity dependent)
• Heterogeneous reactions
• HNO3(g) ? NO3- (s, aq)

50
Indirect influence of NOx-emissions on
concentrations of PM

• NOx emissions influence the concentrations of
  various oxidants in the atmosphere (O3, OH,
  H2O2). The concentration of these determine how
  fast, e.g., sulphur dioxide is transformed into
  sulphuric acid and sulphate particles
• SO2 gas phase oxidation
• SO2 OH ? SO42-
• Heterogeneous reactions
• SO2 ( H2O2 or O3, in cloud droplets) ? SO42-

51
Impact on fine particles - PM2.5
LTO-contribution non-LTO-contribution
Calculated annual average contribution to the
surface PM2.5 concentration from emissions of NOx
in Swedish air space. Unit pg/m3 (STP). 1 pg/m3
0.000001 mg/m3.
52
PM - valuation of impact on mortality

• Impact on mortality 6 risk increase per 10mg/m3
  increase in, PM10 concentration, (long-term
  exposure)
• Only applied for persons above 30years old
• Assumed 1 year lost of life
• The value of a lost year of life was set to 49
  000 EUR
• (In earlier ExternE studies the risk was set to
  2.6 and the value of a lost year of life was set
  to 93 000 EUR)

53
PM - valuation of impact on mortality

• The resulting estimated marginal cost for life
  years lost by PM exposure due to aviation
  NOx-emissions in Swedish air space are
• 0.20 EUR/kg NOx for LTO emissions (or 0.16 EUR
  using ExternE valuation)
• 0.13 EUR/kg NOx for non-LTO emissions (or 0.10
  EUR using ExternE valuation)

54
Valuation of morbidity impacts - uncertainties

• The exposure-response relationships used in the
  ExternE calculations regarding the effects of PM
  on hospital admissions can be questioned for
  several reasons
• The ER relations have been taken from a limited
  number of studies
• The base frequencies of incidence of hospital
  admissions used in the calculations were not
  taken from the nations or regions to which the
  calculations were applied.
• The scientific basis for using different
  coefficients for nitrates and PM10 and sulphates
  and PM2.5 can also be questioned.
• There are new risk factors for hospital
  admissions due to respiratory diseases and
  hospital admissions for cerebrovascular diseases
  due to PM10 available from the EU-projects Air
  Pollution and Health A European Information
  System (APHEIS) (www.apheis.net) and APHEA2
  (Short-term effects of Air Pollution on Health a
  European Approach using epidemiological
  time-series LeTertre, 2003). For chronic
  bronchitis updated risk factors are lacking.

55
Valuation of morbidity impacts - uncertainties

• For the purpose of this study we have chosen not
  to update the Exposure-Response (ER) relations
  for impacts on morbidity due to ozone and PM.
• Instead we indicate that the morbidity effects
  are in the range between zero and the ExternE
  estimates.
• This is in line with the assumptions made in the
  CAFÉ (Clean Air For Europe) work where morbidity
  is excluded in the analysis of abatement costs.

56
PM - valuation of impact on morbidity

• The resulting estimated marginal cost for
  morbidity by PM exposure due to aviation
  NOx-emissions in Swedish air space are
• For LTO emissions 0 0.08 EUR / kg NOx
• For non-LTO emissions 0 0.05 EUR / kg NOx

57
Total LTO NOx emission costs (1 SEK 0.11 EUR)
58
Total non-LTO NOx emission costs (1 SEK 0.11
EUR)
59
Conclusions

• Marginal costs for NOx emissions from aviation in
  Swedish air space are estimated to be higher for
  LTO emissions than for non-LTO emissions
• LTO-emissions ca 1.9 EUR (17 SEK) / kg NOx
• Non-LTO-emissions ca 0.9 EUR (8 SEK) / kg NOx
• The costs due to acidification and eutrophication
  of natural ecosystems dominate but they are very
  uncertain.
• If ecosystem effects are excluded the marginal
  costs become almost the same for LTO and non-LTO
  emissions of NOx.
• LTO-emissions 0.25-0.63 EUR / kg NOx
• Non-LTO-emissions 0.21-0.56 EUR / kg NOx
• Costs due to crop loss and effects on
  agricultural soils were estimated to be much
  smaller than costs due to health effects.

60
Conclusions
61
Conclusions

• The total calculated costs for NOx emissions from
  aviation in Swedish air space are estimated to be
  about two times higher for non-LTO emissions than
  for LTO emissions
• LTO-emissions ca 1.7 M EUR
• Non-LTO-emissions ca 4.4 M EUR
• Using ExternE methodology and the higher
  abatement cost for ecosystem damages lead to
  higher costs
• LTO-emissions ca 7.2 M EUR
• Non-LTO-emissions ca 13 M EUR

62
Conclusions
63
Future work

• Future work to improve the present assessment of
  environmental costs due to aviation emissions of
  NOx are needed in all links of the Impact Pathway
  chain
• A number of uncertainties have been identified.
  The following topics should be considered in
  future studies
• Use of improved ER-relations and valuation of
  morbidity due to both PM and ozone ongoing work
  for road traffic scenarios
• Grid based assessment of all the effects studied
  instead of nation averaging better population
  data available now
• Calculations for more than one year in order to
  reduce the impact of meteorological variability

64
Uncertainties

• The valuation of ecosystem effects includes a
  number of uncertainties
• The mapping of ecosystem sensitivity varies to
  some extent between different countries and the
  valuation is based on abatement costs.
• The method for calculating the change in
  unprotected ecosystem area in the present study
  is based on country average critical loads. This
  also introduces some uncertainty. A more detailed
  approach would be to make calculations on a grid
  square by grid square basis, using grid square
  specific critical loads. This is possible but was
  outside the scope of the present study.

65
Uncertainties

• We have not used updated ER relations for effects
  on morbidity in this study.
• There are new European studies for hospital
  admissions, but for chronic bronchitis updated
  risk factors are lacking, and the calculations
  are based on one old study from the USA.
• An additional problem is the valuation of
  morbidity effects. The valuation can be expected
  to vary between countries but such information is
  lacking for many health outcomes. We have used
  the ExternE ER relations and valuation to
  indicate the possible magnitude of the morbidity
  costs but the uncertainties here are substantial.

66
Uncertainties

• PM formed from aviation NOx emissions is mostly
  nitrate in this study. Is nitrate really
  dangerous?
• Studies on animals do not indicate that ammonium
  nitrate in itself is toxic in relevant
  concentrations.
• However, when the short-term effect of PM in the
  US was compared between regions, California with
  the highest nitrate proportion had an
  ER-coefficient above the average.
• A few studies of short-term effects on mortality
  have also shown that nitrate particulates seem
  important, but maybe not due to nitrate in
  itself.

67
Uncertainties

• The sensitivity of different chemical transport
  models to changes in emissions is an important
  area of uncertainty.
• This appears to be the most important factor
  explaining the differences in valuation of health
  effects due to PM in this study. Harmonisation of
  chemical transport models is certainly an
  important topic here.
• Part of the explanation could also be related to
  the choice of time period for the modelling.
  Meteorological conditions are known to be
  important and simulations for different years are
  expected to give different results.

68
Further information
contact Robert.Bergstrom_at_smhi.se