DLR contributions to WP 3.3.1 Katrin Obermaier, Volker Grewe, Rudolf Deckert

1
DLR contributions to WP 3.3.1Katrin Obermaier,
Volker Grewe, Rudolf Deckert
2
WP 3.3.1 Impact of future climate change

• Investigate existing simluations with regard to
  impact of climate variability on traffic induced
  ozone changes.
• Impact of chemistry-climate feed-backs on traffic
  induced ozone changes.
• Each participating group performs 4 simulations
  (two may be covered in 3.3.1)
• A Climate2000 Emissions2050 all emissions
• BR Climate2000 Emissions2050 -5 for road
  traffic (recommended by WP 6)
• C Climate2050 Emissions2050 all emissions
• DR Climate2050 Emissions2050 -5 for road
  traffic
• DLR investigates transient simulations 1960 to
  2020
• with and without climate change and the impact
  on traffic induced ozone changes
• Additionally Calculation of the effect of
  lifetime changes on
• the transient behaviour of methane changes

3
WP 3.3.1 Impact of future climate change

• Future climate change simulations yet to be done
• Investigation of existing runs
• a) Understanding the impact of changing atm.
  composition on chemistry and radiative forcing
  (Obermayer et al.)
• b) Understanding the impact of climate
  variability on regional ozone amounts (Deckert et
  al.)
• Appendix calculation of methane changes, as
  suggested in Budapest (Volker Grewe)

4
WP 3.3.1 Existing DLR simulations

• Simulations with climate change
• ensemble of three simlulations 1960-1999
• ensemble of four simulations 2000-2019
• Simulation without climate change
• single simulation 1980-2019

5
WP 3.3.1 Existing DLR simulations

• In the simulations tagging approach for O3 by
  Grewe (2004)
• tagged-O3 production rates calculated relatively
  to NOy concentrations
• four anthropogenic NOx emission sectors NOx
  natural sources
• 2004-2019 growth rate of traffic NOx emissions
  varies between regions

6
WP 3.3.1 Changing atmospheric composition

• Simulations evolution of emissions and tagged O3
  matches
• tagging method works
• Lightning and air traffic large tagged-O3 column
  despite weak NOx emissions
• Reason emissions in upper troposphere result in
  high O3 production efficiency
• high level of UV radiation
• favourable background NOx concentrations

7
WP 3.3.1 Changing atmospheric composition

• Nonlinear dependence of O3 production efficiency
  on background NOx concentrations (Grooß et al.
  1998)
• efficiency increases up to a NOx concentration of
  about 0.3 ppb
• efficiency decreases for higher NOx
  concentrations
• net O3 destruction for excessively low or high
  NOx concentrations

Grooß et al. 1998
8
WP 3.3.1 Changing atmospheric composition

• Nonlinear dependence explains
• low production efficiency of industrial emissions
• 1960-1990 increase in production efficiency of
  air traffic emissions

Obermaier et al.
Grooß et al. 1998
9
WP 3.3.1 Changing atmospheric composition

• Further effect of nonlinear dependence O3
  production efficiencies from ground sources
  decrease during 1960-2019
• due to increasing NOx background, in accordance
  with Lamarque et al. 2005
• Exception for road traffic after 2005
• emissions only increase in nonindustrial regions
  with low NOx background

Obermaier et al.
10
WP 3.3.1 Changing atmospheric composition

• Investigation of climate change effects on tagged
  O3 not so easy
• Main reason evolution of SSTs, CO2 and CH4
  different for simulations with and without
  climate change
• CH4 impacts on O3 chemistry via HO2
• Work still in progress

11
WP 3.3.1 Changing atmospheric composition

• What is the radiative impact of each tagged-O3
  species? - Consider radiative forcing
• Radiative forcing (RF) relates to surface
  temperature response (?Tsurf)
• ?Tsurf ? RF, ? climate sensitivity
  parameter
• RF calculated for each tagged-O3 species
• defined by RF(all O3 s) minus RF(all O3 species
  except species of interest)
• radiation calculations similar to Stuber et al.
  (2001)

12
WP 3.3.1 Changing atmospheric composition

• Method works similar evolution of RF and tagged
  O3
• However RF of an O3 molecule depends, among
  others, on
• surface temperature (hence on latitude)
• air temperature and pressure (hence on altitude)

13
WP 3.3.1 Changing atmospheric composition

• RF efficiency (RFE) the dependencies mentioned
  become apparent
• RFE increases towards the equator
• highest RFE for lightning (high altitude low
  latitude)
• lowest RFE for road traffic and industry (low
  altitude high latitude)
• RF efficiency decreases for many sectors since
  1975 saturation effect due to increasing
  background O3 concentrations
• O3 long-wave absorption band saturates

14
WP 3.3.1 Impact of climate variability

• Analyse transient simulation with climate change
  1960-1999, 2000-2019
• consider three industrialised regions East Coast
  USA, Central Europe, East China
• average of nine grid points for each region
• tagged-O3 mass from surface to 500hPa

15
WP 3.3.1 Impact of climate variability

• Example of tagged-O3 time series anthropogenic
  NOx emissions
• how investigate interannual variability?
• multiple regression analysis
• accounts for annual cycles and trends
• relates interannual variability to
• other simulated quantities (e.g. wind velocity)
• model boundary conditions (e.g. SSTs)

16
WP 3.3.1 Impact of climate variability

• Regression model adopted
• O3(t) a seascycle(t) b1 trend(t) b2
  wind(t ) b3 radiation(t) b4 enso(t) b5
  qbo(t) resid(t)
• predictors
• trend linear, NOx emissions, ozone production
  efficiency
• wind 700hPa horizontal-wind components and
  geopotential
• radiation absorbed surface solar radiation
• enso calculated from equatorial sea surface
  temperatures
• qbo 50hPa equatorial zonal wind
• Seasonally dependent relationships
• bi bi1 bi2 cos(bi3 tp /6) bi4
  cos(bi5 tp /12) ...
• Try to meet requirements for statistical
  inference
• resid(t) autoregressive model AR(n)
• seascycle(t) calculated prior to regression
• seasonal weighting

same region as O3 (t)
simulation boundary conditions
17
WP 3.3.1 Impact of climate variability

• Example of regression result East China,
  industrial O3
• 95 confidence band for the regression line
• regression works

Deckert et al.
18
WP 3.3.1 Impact of climate variability

• Example of regression result East China,
  industrial O3
• 95 confidence band for the regression line
• regression works
• Without annual cycle more obvious
• seasonally dependent trend
• predictors capture many prominent peaks, but not
  all

Deckert et al.
19
WP 3.3.1 Impact of climate variability

• Example of regression result without annual
  cycle
• East China and East Coast USA, industrial O3
• regression explains a noticeable fraction of
  interannual variability for both cases
• variability greater for East China

20
WP 3.3.1 Impact of climate variability

• Problem Requirements for statistical inference
  not satisfied
• residuals autocorrelated, not normally
  distributed
• despite AR(n) stochastic model
• Therefore confidence bands/intervals
  underestimated statistical tests for predictor
  significance invalid
• Hence adopt those predictors that reduce the
  residual sum of squares (SSQ) more strongly than
• same predictor with time lag
• other clearly unsuitable predictors
• Consider SSQ reduction for each predictor adopted
• SSQ difference for regression with and without
  a given predictor
• measures interannual variability captured by this
  predictor
• problem only rough measure and not strictly
  additive (due to correlated predictors?)

21
WP 3.3.1 Impact of climate variability

• East China
• large SSQ reduction for industry, soils,
  lightning
• due to large interannual variability
• large interannual variability for soils Central
  Europe not well captured

22
WP 3.3.1 Impact of climate variability

• Central Europe
• dominance of wind due to transport? (prevailing
  westerlies and nearby Atlantik)
• Noticeable ship effect for the same reason?

23
WP 3.3.1 Impact of climate variability

• Stratosphere
• qbo and enso are known to affect the
  Brewer-Dobson circulation
• dominance of radiation due to dynamics and/or
  chemistry?

Deckert et al.
24
WP 3.3.1 Impact of climate variability

• More thoughts needed to provide physical/chemical
  explanations

25
Calculation of methane changes
(GreweStenke, 2008 ACP)
Instantaneous change
Old method CCMs/CTMs ? ?t(2000) ? ?CH4
(2000) ? RF New method CCMs/CTMs ? ?t(2000) ?
?t(t)E ?t(2000) ? ?CH4 (t) ? RF
Change according to lifetime
Unperturbed case
Perturbed case
Methane change
Input Lifetime changes, emission evolution,
background methane
26
Calculated of methane changes and RF changes
Inputs
Percentage lifetime changes for 2000 road
traffic -1.69 air traffic -1.07 ship
traffic -4.27
27
Methane changes Results
RF changes are smoother Max. impact occurs later
Methane RF lasts longer Timescale of changes
decade
28
Summary

• ECHAM5/MESSY, resolution T42/L41, is being
  developed
• Existing simulations
• Katrins findings show that
• NOx emissions produce a realistic O3 response
  dependence on altitude and background NOx
• the associated RF and RFE is realistic
  dependence on altitude, latitude, and background
  O3
• climate dependency of tagged O3 more work
  necessary
• My investigations of regional ozone show that
• multiple regression explains a noticeable
  fraction of interannual variability
• predictors wind, radiation, qbo, enso
• physical/chemical explanations more work needed
• statistical inference not allowed
• Volkers consideration of methane lifetime
  changes
• smoother evolution of methane concentration and
  RF changes
• later maximum impact of
• longer lasting impact of