Atmospheric Impact of the 1783-1784 Laki volcanic eruption

1
Atmospheric Impact of the 1783-1784 Laki volcanic
eruption

• David Stevenson (University of Edinburgh)
• Thanks to
• Ellie Highwood (Univ. Reading),
• Colin Johnson, Bill Collins, Dick Derwent (Met
  Office)
• Funding

2
Talk Structure

• Motivations volcanoes and the atmosphere
• Introduce
• The Laki eruption
• Atmospheric chemistry model
• Tropospheric sulphur cycle
• Model experiments results
• Radiative forcing climate impact
• Conclusions / problems

3
0530 15 June 1991
4
0630 15 June 1991
5
0730 15 June 1991
6
SO2 oxidises to H2SO4 aerosol Residence time
year (mid-stratosphere) weeks-months (UT/LS)
days-weeks (troposphere)
Ash falls out quickly (hours-days)
7
Pinatubo aerosol from the Space Shuttle
Aerosol layer 20-25 km
Tropopause 15 km
SO2 (gas) OH ? H2SO4 (aerosol)
8
Volcanoes and Climate

• Natural climate variability
• How much of this is due to volcanoes?

9
Volcanoes and Climate

• Test climate model sensitivity/response

10
Volcanoes and Climate

• Focus to date on large explosive eruptions, or
  those that leave a record in the Greenland or
  Antarctic ice.
• What about large effusive eruptions?

11
Fires of the Earth The Laki Eruption
1783-1784 Eyewitness account of the eruption by
the local vicar(Rev. Jon Steingrimsson) Recently
reprinted by the University of Iceland, and
translated into English 1/5 of Icelands
population died (10,000 people) Much better than
that cack by Tolkien
12
(No Transcript)
13
27 km long fissure
580 km2 of lava
14
1783-84 Laki eruption, Iceland

• 8 June 1783 27 km long fissure opens
• 15 km3 of basalt erupted in 8 months
• 60 Tg(S) released
• 60 in first 6 weeks
• Fire-fountaining up to 800 - 1450 m
• Eruption columns up to 6 - 13 km
• Tropopause at 10 km
• Dry fog or blue haze recorded over Europe,
  Asia, N. Atlantic, Arctic, N. America
• This appears to have been a sulphuric acid
  aerosol layer in the troposphere and/or lower
  stratosphere

15
200m Fire-fountaining at Etna, 2002
Photos from Tom Pfeiffers web-site
www.decadevolcano.net
16
(No Transcript)
17
Environmental impacts

• 20 of the Icelandic population die
• Acid deposition destroys crops grazing
• Similar impacts across Europe
• Cooling of NH (regionally extreme, e.g. Alaska)
• Cooling for several years (Franklin, 1785)
• Famine

18
Questions

• Using our best estimate of the Laki SO2
  emissions, what is the modelled impact on the
  global atmospheric composition?
• Does it agree with observations?
• Can it generate a climate impact?

19
Atmospheric model STOCHEM

• Global 3-D chemistry-transport model
• Meteorology HadAM3
• GCM grid 3.75 x 2.5 x 58 levels
• CTM 50,000 air parcels, 1 hour timestep
• CTM output 5 x 5 x 22 levels
• Detailed tropospheric chemistry
• CH4-CO-NOx-hydrocarbons
• detailed oxidant chemistry
• sulphur chemistry
• Normally used for air quality/climate studies
• This version has high resolution tropopause

20
STOCHEM framework
21
For each air parcel

• Advection step
• Interpolated winds, 4th order Runge-Kutta
• Plus small random walk component (diffusion)
• Calculate emission and deposition fluxes
• Prescribe gridded emissions for NOx, CO, SO2,
  etc.
• Integrate chemistry
• Photochemistry (sunlight, clouds, etc.)
• Gas-phase chemistry (T, P, humidity, etc.)
• Aqueous-phase chemistry (cloud water, solubility,
  etc.)
• Mixing
• With surrounding parcels
• Convective mixing (using GCM convective clouds)
• Boundary layer mixing

22
Sulphur chemistry
O3(aq)
Only deposition rates determine the SO4 lifetime
Oxidation and deposition rates determine the SO2
lifetime
23
Present-day tropospheric sulphur cycle
24
Laki sulphur emissions

• Analysis of the S-content of undegassed magma
  suggests 60 Tg(S) released by Laki (Thordarson
  et al., 1996)
• 1990 global annual anthropogenic input
• Compare to Pinatubo 20 Tg(S)
• What was the vertical profile of emissions?

25
1990 Anthropogenic SO2 emissions (annual total)
0.1
1
10
100
Tg(S)/yr/5x5
26
Model experiments

• 1990 atmosphere
• Background pre-industrial atmosphere
• Two Laki emissions cases
• lo emissions evenly distributed 0-9 km
• hi 75 emissions at 8-12 km, 25 at 0-3 km
• All runs had fixed (1996-97) meteorology
• No attempt made to simulate 1783 weather
• Run for one year following start of eruption
• Generate aerosol distributions
• No feedback between aerosols ? climate
• Calculate radiative forcings and climate effects
  later

27
Zonal mean JJA SO2 sulphate
28
Zonal JJA mean SO2
1990
1860
laki hi
laki lo
29
Zonal JJA mean SO4
1990
1860
laki hi
laki lo
30
Model present day observations
? Model ? Observations
31
July SO2 (ppbv) Laki hi
Surface 0.5 km
550 hPa 5 km
0.1
0.2
0.5
2
10
20
50
1
5
100
0.1
0.2
0.5
2
10
20
50
1
5
100
350 hPa 8 km
200 hPa 12 km
0.1
0.2
0.5
2
10
20
50
1
5
100
0.1
0.2
0.5
2
10
20
50
1
5
100
32
July SO4 (pptv) Laki hi
Surface 0.5 km
550 hPa 5 km
350 hPa 8 km
200 hPa 12 km
33
Laki SO4 evolution
Upper Trop
Lower Strat
Surface
90N
lo
Eq
90S
May 1784
June 1783
SO4 / pptv
90N
hi
Eq
90S
34
Laki sulphur budget
Hi case
22 Tg(S) or 89 Tg (H2SO4.2H2O)
SO2 gas
Emissions 61 Tg(S)
35
Impact on oxidants (JJA)
H2O2
O3
OH
36
Zonal mean JJA Laki SO2 sinks (ppbv/day)
SO2 OH
Aqueous phase oxidation
Altitude / km
Wet deposition
Dry deposition
37
Laki SO2 budget (JJA)
Lower Stratosphere
tSO2 25 days (no transport 170 days)
Upper Troposphere
tSO2 12 days (no transport 67 days)
Lower Troposphere
tSO2 5.3 days
38
Laki Sulphate budget (JJA)
LS
tSO4 67 days (no transport gt7 yrs)
UT
tSO4 10 days (no transport 32 days)
LT
tSO4 5.3 days
39
Acid deposition to Greenland
40
Greenland Ice-core data
H2SO4 deposition rates mg(S)/m2/yr
Modelled Observed
Background 5.0 2.9-8.6
Laki 63-65 18-107
41
Total atmospheric aerosol mass
42
Aerosol yield/peak loading
Total yield Peak load
This work 51-66 4.4-5.1
Clausen Hammer(1988) 280 -
Zielinski (1995) 40 -
Stothers (1996) 150 4.5
Clausen et al (1997) 100-150 -
Thordarson and Self (2002) 200 (150?)
Tg(H2SO4)
43
Radiative forcing Climate impactEllie Highwood
(Reading Univ)

• Aerosol fields inserted into Reading IGCM
• 3 experiments
• 1. Hi/long decay (10 month e-fold)
• 2. Hi/short decay (3.6 month e-fold)
• 3. Lo
• Each has a 10 member ensemble of 3 yr runs
• Compare to control run with no forcing
• Only direct aerosol effect

44
Radiative forcings
45
Climate Impact
lo
hi
46
Climate impact

• Hi runs have NH cooling of 0.21K, in good
  agreement with observations (-0.14 to 0.27K)
• Lo runs show no significant cooling
• BUT runs neglect indirect aerosol effects
• Hi runs also have cooling persisting for 3 years,
  due to feedbacks (ice/snow albedo)

47
Conclusions(1)

• 1st attempt at chemistry-climate modelling of the
  Laki eruption
• Simulated a sulphate aerosol cloud across much of
  the NH during the 8-month eruption
• Deposition to Greenland similar to ice-core
  record
• 60-70 of emitted SO2 is deposited before forming
  aerosol (previous studies assumed it all formed
  aerosol)
• Mean lifetime week
• Atmospheric loading less than previous estimates

48
Conclusions(2)

• Oxidants H2O2 OH strongly depleted
• lengthens the SO2 lifetime
• more likely to be deposited as SO2
• Climate modelling suggests Hi scenario gives
  0.2K cooling, and persists for gt2 years this
  matches observations
• But many processes missing
• For more info 2 papers in ACP
• www.copernicus.org/EGU/acp
• davids_at_met.ed.ac.uk

49
Problems(1)

• Volcanology
• Emissions uncertain
• Magnitude
• Vertical profile
• Temporal distribution episodic
• Plume processes, e.g. scavenging of SO2/SO4 by
  ash in the eruption column

50
Problems(2)

• Chemistry modelling
• Not fully coupled
• No aerosol microphysics
• No coupling of aerosol to photolysis rates
• Only 1 heterogeneous reaction (N2O5 loss)

51
Problems(3)

• Climate model
• Simplified radiation scheme more bands suggest
  forcing is smaller by factor 0.6
• Humidity assumptions 80 RH increases forcing
  by factor 2.6
• No aerosol indirect effects
• Climate sensitivity low compared to other models
  suggests duration of forcing may be underestimated

52
2 Papers in Atmospheric Physics
Chemistry www.copernicus.org/EGU/acp/acp/3/487/a
cp-3-487.pdf www.copernicus.org/EGU/acp/acp/3/117
7/acp-3-1177.pdf