arctic climate change and the role of ozone - gruan.org · icm-10, 23-27 april 2018, potsdam,...
Post on 08-Sep-2018
216 Views
Preview:
TRANSCRIPT
ICM-10, 23-27 April 2018, Potsdam, Germany
Arctic Climate Change and the Role of Ozone
Markus Rex Alfred Wegener Institute
ICM-10, 23-27 April 2018, Potsdam, Germany
Arctic
Equator
Antarctic
Kelvin
http://data.giss.nasa.gov/gistemp/time_series.html
The Arctic is the key area for climate change
90
60
30
0
-30
-60
-90 1960 1970 1980 1990 2000 2010
La
titu
de
[d
eg
]
Year
Observed near surface temperature changes
ICM-10, 23-27 April 2018, Potsdam, Germany
Severe loss of sea ice in the Arctic
Update from Stroeve et al., GRL 2007
Observed September sea ice extent
(106 km2) 10
8
6
4
2
0 1950 1960 1970 1980 1990 2000 2010 2020
larger
uncertainty
ICM-10, 23-27 April 2018, Potsdam, Germany
Perovich et al. 2015
Transition from multi-year to first year ice
ICM-10, 23-27 April 2018, Potsdam, Germany
Sea ice thickness in various regions (mid September)
Average thickness loss: 1.6 m (or 53%)
Loss of ice volume: ~75%
Kwok & Rothrock (2009)
Large reductions in sea ice thickness
ICM-10, 23-27 April 2018, Potsdam, Germany
Aktuelle Meereisentwicklung
Situation in 2017 & 2018
ICM-10, 23-27 April 2018, Potsdam, Germany
Updated from Stroeve et al., 2007
Arctic climate change is not well represented in state-of-the-art Earth System Models
10
8
6
4
2
0 1900 1950 2000 2050 2100
September sea ice extent (106 km2)
Models versus observations
CMIP3 / AR4
1950 2000 2050 2100
Update from Stroeve et al., GRL 2012
CMIP5 / AR5
ICM-10, 23-27 April 2018, Potsdam, Germany
10
8
6
4
2
0 1900 1950 2000 2050 2100 1950 2000 2050 2100
Update from Stroeve et al., GRL 2012
CMIP3 / AR4 CMIP5 / AR5
September sea ice extent (106 km2)
Models versus observations
Arctic climate change is not well represented in state-of-the-art Earth System Models
Updated from Stroeve et al., 2007
ICM-10, 23-27 April 2018, Potsdam, Germany
[m]
September sea-ice thickness
2000-2004, CMIP5 / AR5
Maslowski et al., 2012
ICM-10, 23-27 April 2018, Potsdam, Germany
Taylor et al., based on RCP8.5
Pro
jecte
d te
mp
era
ture
ch
an
ge
[K
]
ab
ove
glo
ba
l a
ve
rag
e
(20
80
-90
min
us 2
00
0-1
0)
Latitude
Arctic is the area of largest uncertainty in climate projections
Arc
tic
Model spread
CMIP5 climate models:
Warming above the global average
ICM-10, 23-27 April 2018, Potsdam, Germany
Slide & planned investments: http://www.slideshare.net/guggenheimpartners/the-melting-arctic-business-opportunities-in-arctic-development
Investment needs: https://www.guggenheimpartners.com/firm/news/guggenheim-partners-endorses-world-economic-forums
Rapid economic development in the Arctic
Rapid development in several areas:
- Shipping
- Mining / resource extraction
- Fishing
Investments planned in Arctic
Infrastructure 2014-2024:
~100 billion US$
Investment needs over next two
decades:
~1000 billion US$
Source: Guggenheim Partners, 2014
Real GDP Growth by Region
(2002=100%)
ICM-10, 23-27 April 2018, Potsdam, Germany
Taylor et al., based on RCP8.5
Pro
jecte
d te
mp
era
ture
ch
an
ge
[K
]
ab
ove
glo
ba
l a
ve
rag
e
(20
80
-90
min
us 2
00
0-1
0)
Latitude
Arctic is the area of largest uncertainty in climate projections
Arc
tic
Model spread
CMIP5 climate models:
Warming above the global average
Understanding of key climate processes in the
Arctic is limited by lack of observations!
Many processes in the Arctic climate system
only roughly represented in Climate Models
ICM-10, 23-27 April 2018, Potsdam, Germany
A major international research initiative under IASC to improve the representation of Arctic processes in weather forecast and climate models
Multidisciplinary drifting Observatory for the
Study of Arctic Climate
ICM-10, 23-27 April 2018, Potsdam, Germany
Main scientific focus areas
Aerosols
Ozone Layer
Dynamical coupling
Boundary layer
Clouds
Sea ice
Ecosystem
Energy & momentum
fluxes
Radiation
• Formation
• Drift
• Deformation
• Melting
• Vertical fluxes through boundaries
between ocean, sea ice and atmosphere
• Meridional fluxes in atmosphere and ocean
• Thermal structure
• Small scale processes
• Chemistry in ocean, sea ice & atmosphere
• Interaction with aerosols and clouds
Vertical and meridional
dynamical coupling in the
Arctic climate system
• Radiative properties of clouds & aerosols
• Microphysical interactions between
clouds & aerosols
• Interactions with boundary layer
• Precipitation
Biogeochemical fluxes
Chemistry
Full seasonal cycle:
• Ecosystem dynamics
• Populations in ocean, sea ice
and melt ponds
ICM-10, 23-27 April 2018, Potsdam, Germany
Link between Arctic and European climate: The Arctic Oscillation (AO)
High AO Index Low AO Index
ICM-10, 23-27 April 2018, Potsdam, Germany
Link: Sea ice changes atmospheric circulation Maximum covariance analysis, ERAi 1979-2015
Jaiser et al. 2013, 2016 , Handorf et al. 2015
September sea ice anomaly
Surface Lower stratosphere
50 hPa GPH anomaly [gpm] [hPa] [%]
Atmospheric pressure anomalies in following winter:
Negative ice anomaly September low AO index in winter
ICM-10, 23-27 April 2018, Potsdam, Germany
Effect of decreasing sea ice on northern hemispheric weather patterns
Decreasing sea ice
ICM-10, 23-27 April 2018, Potsdam, Germany
Advection of warm and
humid air into the Atlantic sector of the Arctic
Decreasing sea ice
Potential for cold air outbreaks
Cold spells in Europe and US
Effect of decreasing sea ice on northern hemispheric weather patterns
ICM-10, 23-27 April 2018, Potsdam, Germany
Winter warming is most severe in the Atlantic sector of the Arctic
M. Maturilli et al.
850 hPa temperature trend
ERA-interim, 1996 - 2017
AWIPEV
research station
ICM-10, 23-27 April 2018, Potsdam, Germany
Climate Change at AWIPEV - In the Atlantic sector of the Arctic -
M. Maturilli et al.
Annual mean trend:
+1.5±0.7oC
per decade
Winter trend:
+3.0±2.0oC
per decade
ICM-10, 23-27 April 2018, Potsdam, Germany
65
o-9
0oN
Tem
pera
ture
An
om
aly
[K
]
Low ice versus high ice conditions in Climate Model
Romanowsky et al., 2017
Pre
ssu
re [
hP
a] „Observations“
ECMWF ERAi
Sep Oct Nov Dec Jan Feb Mar Apr May
10
weaker & warmer
polar vortex
ECHAM6
standard
Pre
ssu
re [
hP
a]
10
Sep Oct Nov Dec Jan Feb Mar Apr May
ICM-10, 23-27 April 2018, Potsdam, Germany
Decreasing sea ice
Effect of decreasing sea ice on northern hemispheric weather patterns
Weaker & warmer polar vortex:
• More transport of ozone into Arctic
• Less chemical loss
Radiative feedback further weakens AO
ICM-10, 23-27 April 2018, Potsdam, Germany
100
0
-100
Ch
an
ge
of to
tal o
zo
ne
du
rin
g w
inte
r [D
U]
Chemical and dynamical contribution
to the variability of the late winter Arctic ozone column
Tegtmeier et al., 2008
1.0 1.2 1.4 1.6
Winter average
EP-Flux (Fz) [105 kg s2]
dynamical
supply
chemical
loss
500
400
300
To
tal o
zo
ne
[D
U]
Winter average
EP-Flux (Fz) [105 kg s2]
1.0 1.2 1.4 1.6
late March
ozone
range of October
observations
high AO low AO high AO low AO
ICM-10, 23-27 April 2018, Potsdam, Germany
~20km altitude, 20 model days, dynamical tracer (PV), ~50km resolution run
driven by ERA interim wind/temperature
• detailed homogeneous and heterogeneous chemistry
• Lagrangian particle sedimentation scheme for aerosols
• no numerical diffusion, sophisticated 3d diffusion scheme
• Tropospheric aerosol & cloud microphysics scheme
• Parametrization for convection
ATLAS: Lagrangian Chemical Transport Model
ICM-10, 23-27 April 2018, Potsdam, Germany
Ozone 16 March 2011
46 hPa
5
3
1
Ozo
ne [
ppm
v]
5
3
1
Ozone [p
pm
v]
ATLAS MLS
4
2
4
2
0
MLS ATLAS
ICM-10, 23-27 April 2018, Potsdam, Germany
Computational effort to calculate
ozone changes
ATLAS solves a set of 49 coupled differential equations
based on 55 initial and boundary conditions.
DO3 calculated at each time step and each model point:
For a 100-year model run the system needs to be solved
~2.5 trillion (1012) times.
Computational effort is 4 years for one 100 year run
much too large to be included in climate models.
But: Virtually all of these calculations
are redundant!
ICM-10, 23-27 April 2018, Potsdam, Germany
From ATLAS to SWIFT Extra-polar approach (polar approach different)
Values of DO3 form a hypersurface in the 55 dimensional
parameter space.
Development of SWIFT:
1. Linear combinations of parameters to reduce the number of
dimensions such that DO3 still forms a compact hypersurface in
the reduced space.
2. Shape of hypersurface is characterized by full runs of ATLAS.
3. An automatic procedure constructs a closed polynomial
expression (~150 terms, 4th order) that approximates its shape.
4. SWIFT solves this expression to give results very similar to those
of ATLAS much faster!
ICM-10, 23-27 April 2018, Potsdam, Germany
SWIFT: Fast interactive Ozone for Climate Models
Numerical effort for ozone calculations
ATLAS
~2 weeks per year
on 48 cores
SWIFT
~ 2 hours per year
on one core
100 200 300
Ozone partial column [DU]
Kreyling et al, PhD, 2017
ICM-10, 23-27 April 2018, Potsdam, Germany
ATLAS & SWIFT Comparison with MLS observations
Kreyling et al, PhD, 2017
ICM-10, 23-27 April 2018, Potsdam, Germany
65
o-9
0oN
Tem
pera
ture
An
om
aly
[K
]
Low ice versus high ice conditions in Climate Model
Romanowsky et al.,
to be submitted
Pre
ssu
re [
hP
a] „Observations“
ECMWF ERAi
Sep Oct Nov Dec Jan Feb Mar Apr May
10
weaker & warmer
polar vortex
ECHAM6
standard
Pre
ssu
re [
hP
a]
10
Sep Oct Nov Dec Jan Feb Mar Apr May ECHAM6 - SWIFT
with interactive
stratospheric ozone
Sep Oct Nov Dec Jan Feb Mar Apr May
10
Pre
ssu
re [
hP
a]
ICM-10, 23-27 April 2018, Potsdam, Germany
Two Main Messages
1. Arctic sea ice decrease affects atmospheric circulation
and increases transport of warm, humid air into the
central Arctic:
positive feedback contributes to Arctic Amplification
of global warming.
2. Interactions with the stratospheric ozone layer play a role in
this feedback and need to be taken into account when
modelling it.
top related