The seasonal burden of Dimethyl sulphide-derived aerosols in the ...

The seasonal burden of
Dimethyl sulphide-derived aerosols
in the Arctic
and the impact on global warming
Bo Qu 1 *, Albert Gabric 2 , Patricia Matrai 3 and T. Hirst 4
1. Lecturer, Nantong University, Science Faculty, China;
2. Associate Professor, School of Australian Environmental Studies, Griffith University, Australia;
3. Principal Scientist, Bigelow laboratory for Ocean Sciences, USA;
4. Principal Scientist, CSIRO atmospheric Research, Australia

Global climate changes have led to remarkable environmental
changes in the Arctic. On the other hand, Dimethyl sulphide
(DMS) emission in Arctic Ocean plays an important role for the
global warming. The ice cover as the special feature of Arctic
Ocean has significant effect on regulation of the large distribution
tion
of phytoplankton production. Chlorophyll-a a (CHL), as the primary
production of phytoplankton, has its strong relationship with DMS
derived aerosol in the ocean surface.

Arctic
Ocean
North Pole

Arctic Current
The sunlight goes some way towards heating the Arctic, but heat also
comes from the south with ocean currents and airstreams. One branch of
the Gulf Stream, called the North Atlantic Current, flows along the coast
of Norway and continues all the way to the Arctic Ocean.
The area of the Barents Sea
where the cold, relatively fresh,
Arctic water meets the warm,
saline Atlantic water is called
the polar front. The polar front
does not lie in a specific
geographical position, but may
move somewhat from year to
year.

Study region: Barents Sea, 30-35E, 70-80N
Three Cruises
collected DMSP data
from Biglow
Laboratory for Ocean
Sciences , ME, USA

MLD (m)
Monthly Mean MLD (m)
80
70
60
50
40
30
20
10
0
1 2 3 4 5 6 7 8 9 10 11 12
Month
Ice Cover (%)
100
75
50
25
0
Ice cover (%) (1988-2002)
1 2 3 4 5 6 7 8 9 10 11 12
Month
Study region:
30-35E,
70-80N
SST (Celsius)
Monthly Mean SST (1980-1993)
7
6
5
4
3
2
1
0
1 2 3 4 5 6 7 8 9 10 11 12
Cloud Cover (Octas)
8
6
4
2
0
Monthly Mean Cloud Cover (1980-1993)
1 2 3 4 5 6 7 8 9 10 11 12
Month
Month
Speed (m/s)
Monthly Mean Wind Speed (1980-1993)
12
10
8
6
4
2
0
1 2 3 4 5 6 7 8 9 10 11 12
Month
PAR (W/m^2)
1998-2002 Mean PAR (1998-2002)
120
100
80
60
40
20
0
73 103 133 163 193 223 253
Day of year

The comparisons for the calibrated model results and the original
SeaWiFS data for 1998. Due to the fitness function is calculated
by using negative value with maximum optimization rule
applied, hence, the closer to zero, the better the fitness.
Year 1998 had its best fitness of -2.6758, while the 5 years
mean (from year 1998 to 2002) also achieved quite high fitness
(-4.4098).
CHL (mg/m^3)
7
6
5
4
3
2
1
1998 GA calibration for CHL Comparison
( k4=0.00002106, k23=0.3157, Ik=17.8188, no=450.73, k19=0.0084, k20=0.012)
fitness=-2.6758
CHL(SeaWiFS)
CHL(Model)
0
3/ 7 3/ 31 4/ 24 5/ 18 6/ 11 7/ 5 7/ 29 8/ 22 9/ 15

CHL and DMS in 1999 in the study region
CHL
4
DMS
10
CHL (mg/m^3)
3
2
1
8
6
4
2
DMS (nM)
0
0
1 49 97 145 193 241 289 337
Day of year
CHL and DMS in 2000 in the study region
CHL
CHL (mg/m^3)
2.5
2
1.5
1
0.5
0
DMS
12
10
8
6
4
2
0
DMS (nM)
1 49 97 145 193 241 289 337
Day of year
CHL and DMS in 2001 in the study region
CHL
4
DMS
15
CHL (mg/m^3)
3
2
1
10
5
DMS (nM)
0
0
1 49 97 145 193 241 289 337
Day of year

CHL(mg/m^3)
6
4
2
0
CHL and DMS in 1998
(5 days later than the first DMS bloom, 45 days to the peak DMS)
CHL
DMS(nmol/l
1/1 2/18 4/7 5/25 7/12 8/29 10/16 12/3
0.12
0.08
0.04
0
DMS(nmol/L)
CHL(mg/m^3)
8
6
4
2
0
CHL and DMS in 2002
(4 days later than the first DMS bloom and 68 days to the peak DMS)
CHL
DMS(nmol/l
1/1 2/18 4/7 5/25 7/12 8/29 10/16 12/3
0.2
0.1
0
DMS(nmol/L)

The spring blooms of phytoplankton (CHL peaks) were gradually
shifted ahead from 17th May to 5th May during the 5 years. This
is due to the increased SST and the earlier ice melting each year
happened in the study region. The time lags between the CHL
blooms and DMS peak blooms are increased from 45 days to 68
days in the 5 years. It is interesting to see that year 1998 had its
DMS first blooms in spring a few days before CHL bloom while in
other years, DMS had its first bloom a few days after the CHL
spring bloom. The reason could be that the higher ice cover in
year 1998 in south part of the study region could generate more
and earlier DMS from the ice algal when ice started melting in
spring .

5
GA Calibration for 98-02 mean CHL for zonal 70-80N :
k4=0.00005501, k23=0.4964, ik=11.3997, no=536.88,
k19=0.0424, k20=0.1381, fitness=-0.7598
4
CHL (mg/m^3)
3
2
1
CHL Model
SeaWiFS
0
1 2 3 4 5 6 7 8 9 10 11 12
Month
Before the transit climate data prediction, another CHL GA
calibration was carried out using average CHL SeaWiFS data in
zonal 70-80N during the 5 years period: 1998-2002. The
excellent fitness value (-0.7598) gives better model parameter
values .
There was a second smaller bloom in September due to the
shallower MLD and increased wind speed from August to
September.

The transit climate data from GCM simulations for the period of
1960-1970 (pre-industry level: 1CO2) and 2078-2086 (tripled
equivalent CO2: 3CO2) were obtained in the zonal 70-80N.
SST (Celsius)
2.5
2
1.5
1
0.5
0
-0.5
-1
1CO2
3CO2
1 2 3 4 5 6 7 8 9 10 11 12
Cloud Cover (%)
100
80
60
40
20
0
1CO2
3CO2
1 2 3 4 5 6 7 8 9 10 11 12

The transit climate data from GCM simulations for the period of
1960-1970 (pre-industry level: 1CO2) and 2078-2086 (tripled
equivalent CO2: 3CO2) in the zonal 70-80N.
100
Ice Cover (%)
80
60
40
20
1CO2
3CO2
0
1 2 3 4 5 6 7 8 9 10 11 12
MLD (m)
60
50
40
30
20
10
0
1CO2
3CO2
1 2 3 4 5 6 7 8 9 10 11 12

Table : Mean Transit Climate Data for 1 CO2 and 3 CO2 in zonal 70-80N
SST
Wind
Speed
Cloud
Cover
Ice
Cover
MLD
DMS
DMS
Flux
1CO2
-0.24
4.2
82.1%
68.8%
41.8
12.1
0.8
3CO2
0.71
4.5
74.7%
50.3%
35.3
13.1
1.8
Increased
40%
3%
-9%
-27%
-13%
8%
117%

CHL
3
2
1
CHL(1CO2)
CHL(3CO2)
0
0 30 60 90 120 150 180 210 240 270 300 330 360
Day of year
2
Z
1
Z(1CO2)
Z(3CO2)
0
0 30 60 90 120 150 180 210 240 270 300 330 360
Day of year
DMS
32
28
24
20
16
12
8
4
0
DMS(1CO2)
DMS(3CO2)
0 30 60 90 120 150 180 210 240 270 300 330 360
Day of year
3
umole/m^2/d
2
1
1xCO2 Flux
3xCO2 Flux
0
1 2 3 4 5 6 7 8 9 10 11 12
m onth

DMS (nM)
5 years DMS (nM) comparison in the study region
1998
35
30
1999
2000
25
2001
20
2002
15
10
5
0
73 89 105 121 137 153 169 185 201 217 233 249 265
Day of year
Monthly Mean Ice Cover and DMS in 1998 in the study region
Ice Cover (%)
70
60
50
40
30
20
10
0
Ice Cover
DMS
1 2 3 4 5 6 7 8 9 10 11 12
Month
25
20
15
10
5
0
DMS (nM)
umole/d^2/d
Monthly Mean DMS Flux in Ice Free Water for Year 1998-2002 in the Study Region
18
16
1998
14
1999
12
2000
10
2001
8
2002
6
4
2
0
-2 1 2 3 4 5 6 7 8 9 10 11 12
Month

Ice Cover (%)
50
40
30
20
10
Annual mean ice cover in the study region during 1988-2002
Error bar is the Standard Error of the Mean
y = -0.565x + 1156.4
R 2 = 0.3275
0
1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Year
Average Ice cover in winter-sping (Nov.-Mar.) in southern 70-75N, 30-35E
Ice Cover (%)
8
7
6
5
4
3
2
1
0
1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Year

We can conclude that the significant decrease of
ice cover and increase of SST are the main
reason of increasing DMS flux to more than
100% by year 2086. This significant change in
the northern belt would cause large impact on
global warming.