Climatology of Aerosol Optical Properties over the High Arctic

Auromeet Saha 1, Norm O’Neill 1, Bob Stone 2 Ed Eloranta 3, Kim Strong 4, Camille Viatte 4, Andreas Herber 5 Ihab Abboud 6, Jim Freemantle 1

1 CARTEL, Universite de Sherbrooke, Sherbrooke, , Canada 2 CIRES/NOAA ESRL, Boulder, Colorado, USA 3 SSEC, University of Wisconsin, Madison, Wisconsin, USA 4 University of Toronto, Toronto, Ontario, Canada 5 Alfred Wegener Institute for Polar & Marine Research, Germany 6 Environment Canada, Canada

NOAA Global Monitoring Annual Conference, Boulder, May 15-17, 2012 High Arctic Sites Closer look at Eureka

MODIS-Terra (250 m) July 8, 2004

PEARL

OPAL What is PEARL?

PEARL - Polar Environmental Atmospheric Research Lab. It is operated by the Canadian Network for the Detection of Atmospheric Change (CANDAC).

PEARL is located at Eureka, (80N, 86W) on in Canada's high Arctic, 450 km north of , the most northerly permanent settlement. PEARL is 1,100 km from the North Pole.

ØPAL - Zero altitude Polar Atmosphere Laboratory. It is located about 15 km southeast of the PEARL ridge lab which is at an elevation of 610 m. Instruments & Data used in this Study (Eureka)

Two Cimel Sun Photometers AHSRL ( βββ & δδδ) (aerosol optical properties) NOAA/SEARCH Arctic High Spectral Resolution Lidar (AHSRL) for vertical profiles of backscatter coefficient ( βββ) & depolarization ratio ( δδδ)

Bruker Fourier Transform Spectrometer Brewer Spectrometer (SO 2) (CO Column concentration) Instruments & Data used in this Study (Barrow & Alert)

NOAA Sun Photometer SP02 used at Barrow & Alert (aerosol optical properties)

wavelength photometer deployed traps heat and stabilizing temp Sunphotometers installation in Eureka

PEARL Ridge Lab

OPAL runway SAFFIRE OPAL – CIMEL 327 PEARL – CIMEL 401

This dual placement was designed to study the layer between the two sites as well as provide an element of redundancy for the AOD measurements. PEARL

~0.61 km ~12 km OPAL

AEROCAN (Canadian Sunphotometer Network)

Eureka (PEARL, OPAL) h

Resolute Bay h

Iqaluit h Yellowknife h

Churchill Fort McMurray h h Kuujjuarapik h Waskesiu h Kelowna h h Bratt’s Lake Chapais Sable Island Saturna h h h h h Island Lethbridge Pickle Lake Sherbrooke h h Halifax Egbert hh Toronto AERONET vs AEROCAN

AEROCAN is a full fledged sub-network of AERONET and benefits from all services that AERONET offers.

AEROCAN has a few unique features that go beyond AERONET protocols:

- All instruments use the high frequency, 3-minute mode (O’Neill mode). - All data transmission is done using FTP. - Complementary AOD processing chain in preparation for real time ingestion into an aerosol assimilation model (currently on hold). - System wide display tools for real-time QA & troubleshooting. - Separate prototype MySQL database for data QA (and science). - Satellite calibration facility being developed (on hold with EC freezes). AEROCAN * (Canadian Starphotometer Sites)

Eureka h

Sherbrooke Egbert h h Eureka Aerosol Optical Depth Climatology 0.16 2007-2011 OPAL PEARL

0.12

0.08 AOD AOD (500 nm)

0.04

0.00 150 OPAL 100 PEARL (Days) 50 obs N 0 MAR APR MAY JUN JUL AUG SEP Mid-Latitude AOD Climatology

0.25 Egbert (1998-2011) CARTEL(1995-2010) 0.20

0.15

0.10 AOD AOD (500 nm)

0.05

0.00

300 Egbert 200 CARTEL

(Days) 100 obs N 0 JUL JUN JAN FEB SEP APR DEC AUG OCT NOV MAR MAY Seasonal and inter-annual variation of AOD over Eureka 0.20

Biomass burning Kasatochi volcanic Sarychev volcanic OPAL from Forest & aerosols aerosols Agricultural Fires in PEARL Russia & Kazakstan 0.15

0.10 AOD (500 nm) AOD(500

0.05

0.00 30 JUL JUL JUL JUL JUL JUN JUN JUN JUN JUN OPAL 20 PEARL (Days) obs

N 10

0 JUL JUL JUL JUL JUL JUN JUN JUN JUN JUN APR SEP APR SEP APR SEP APR SEP APR SEP AUG AUG AUG AUG AUG MAR MAY MAR MAY MAR MAY MAR MAY MAR MAY 2007 2008 2009 2010 2011 Biomass burning aerosols during Spring 2008 (ARCTAS)* PEARL (Level 1.0), poly. order = 2, λ = 0.5 µm 0.4 Total Fine 0.3 Coarse

τ 0.2

τττf correlates well with βββ variations 0.1 in the presence of small δδδ, while τττc values correlate with βββ in the

0.0 presence of large δδδ. 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 AHSRL aerosol backscatter cross section ( β)

Fine mode aerosols were associated with biomass burning (agricultural and forest fires) from E/SW Russia & northen Altitude (km) Altitude Kazakhstan.

β (β ( m-1 str -1) AHSRL particulate circular depolarization ratio ( δ)

Coarse ‐‐‐mode events were likely due to ice particles whose nucleation may have been associated with the presence of smoke, or

Altitude (km) Altitude possibly dust.

Day of Month (April 2008) δ (%) *Saha et al., GRL, 2010 Sarychev volcanic aerosols over the Arctic during Summer 2009*

(a)

β ( m-1 str -1) (b)

Pre- and post- SAT τττf, and reff,f for four AEROCAN / AERONET high-Arctic stations 0.15 Pre-eruption Post-eruption δ (%)

0.10 fine mode OD coarse mode OD 0.4 total OD (c) (500 nm) (500

AHSRL SOD f , , , , 0.05

0.2

0.00

Optical Depth Optical 0.4

0 0.3 m] ∗ ∗ 182.0 182.5 183.0 183.5 184.0 ∗ ∗ fine mode OD [ eff,f 10 r 0.4 Brewer #007 0.2 Brewer #021 (d) Brewer #069 0.1 [DU]

0.2 5 2 0.0 SO

Optical Depth Optical 0PAL PEARL Thule Hornsund

0.0 0 182.0 182.5 183.0 183.5 184.0 Doy.UT/24 *O’Neill et al., JGR 2012 July 1 July 2 Variations of CO Total Column over Eureka

2500 15 2000 ] x ]x 10 -2

1500 CO[mol. cm

1000 30

20 (Days) obs

N 10

10000 JUL JUL JUL JUL JUL JUN JUN JUN JUN JUN APR MAY SEP APR MAY SEP APR MAY SEP APR MAY SEP APR MAY SEP AUG AUG AUG AUG AUG MAR MAR MAR MAR MAR 2007 2008 2009 2010 2011

CO being a tracer of biomass, one would expect a degree of correlation with the fine mode aerosols that are also a primary product of biomass burning. Fine Mode AOD vs CO

0.3 0.3 2007 2009

0.2 0.2

y = 0.0099e 0.001x

(500 (500 nm) Sarychev

(500 (500 nm) 2 f

f R = 0.5884 τ τ τ τ τ τ τ τ 0.1 y = 0.0027e 0.0014x 0.1 R2 = 0.5781

0 0 1000 1500 2000 2500 3000 1000 1500 2000 2500 3000 -2 15 CO Column [mol. cm ] x 10 CO Column [mol. cm -2 ] x 10 15 0.3 2008 Shown in the figures are the variation of fine-mode

0.0026x AOD as a function of CO concentration for the 0.2 y = 0.0003e R2 = 0.7209 three years of 2007, 2008 and 2009. (500 nm)

f τ τ τ τ One can observe a moderate (2007), strong (2008) 0.1 and insignificant (2009) correlation respectively for these three years. However if we separate out the Sarychev fine-mode AODs from the 2009 2 0 regression, we find a R value that is comparable 1000 1500 2000 2500 3000 to the other two years. CO Column [mol. cm -2 ] x 10 15 Pan-Arctic sites for the AOD Climatology

Hornsund Ny-Alesund

Alert Barrow-NOAA

Barrow-AERONET PEARL

OPAL Thule

Resolute Bay

Kangerlussuaq Pan-Arctic AOD Climatology 0.20 Barrow NOAA (2001-2011) Barrow AERONET (1999-2011) Resolute Bay (2004-2011) OPAL (2007-2011) PEARL (2007-2011) 0.15 Alert (2004-2011) Thule (2007-2011) Kangerlussuaq (2008-2011) Hornsund (2006-2011) Ny-Alesund (1991-2011) 0.10 AOD (500AOD nm)

0.05

0.00 Pan-Arctic Aerosol Climatology 300 Barrow NOAA Barrow AERONET Resolute Bay OPAL 200 PEARL Alert Thule (Days) Kangerlussuaq

obs 100 Hornsund

N Ny-Alesund

0 MAR APR MAY JUN JUL AUG SEP OCT AOD monthly climatology for an enhanced number of high Arctic stations on a pan- Arctic scale. It simultaneously shows the spring to summer decrease noted for Eureka as well as a west to east AOD decrease. Retrievals of fine & coarse mode AOD and effective radius

0.15 The fine-mode AOD shows total fine mode the spring to summer coarse mode decrease (confirms that the 0.10 total AOD is fine-mode dominated). AOD (500 nm) (500 AOD 0.05 The coarse mode AOD shows apparent springtime activity that could possibly 0.00 be Asian Dust. 0.3 PEARL The fine-mode effective r ~ 0.2 µµµm eff,f radius shows a relatively 0.2 consistent variation except m) µ µ µ µ

( in 2009 where we know eff,f r there was a strong 0.1 Sarychev influence (O’Neill et al., 2012). 0.0 3 PEARL The coarse mode effective radius shows significant 2 variations that could be m) µ µ µ µ real (possibly associated ( eff,c

r with marine aerosols) but 1 which might, just as well, be associated with low- AOD retrieval artifacts. 0 MAR MAY JUL SEP MAR MAY JUL SEP MAR MAY JUL SEP MAR MAY JUL SEP MAR MAY JUL SEP 2007 2008 2009 2010 2011 Illustration of Coarse Mode Event (Asian Dust)

April 27, 2009

Dust β Precipitating Ice crystals

δ

Taklimakan Desert April 19, 2009

0.25 Total 0.20 Fine Coarse )

τ 0.15

0.10 AOT ( AOT 0.05

0.00

00 04 08 12 16 20 24 Time (GMT) Coarse Mode Event (Nucleation of Smoke & Dust) during ARCTAS*

PEARL (Level 1.0), poly. order = 2, λ = 0.5 µm AHSRL, λ = 532 nm, Fine: 0.05 < δ < 0.4, Coarse: δ > 0.4

0.8 Fine Fine Coarse Coarse

0.6

τ 0.4

0.2

0.0 18 19 20 April 2008

*Saha et al., GRL, 2010 Coarse mode event at PEARL (Artifacts)

PEARL (Level 1.0), poly. order = 2, λ = 0.5 µm 0.4 Total Suspicious coarse Fine 0.3 mode peaks Coarse possibly due to

τ 0.2 frost in the Cimel

optics. 0.1

0.0 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 April 2009

β

AHSRL does not show any coarse mode event

δ Frost Issues

Cimel Sunphotometer Cimel Sunphotometer Sept. 9, 2011 Sept. 22, 2009

Star Photometer Winter 2012 Coarse mode event at Thule (Artifacts)

April 2008

1 fine mode OD coarse mode OD total OD

0.8 2 April 15 April 16 0.6 April 14 Linear (April 16) Linear (April 15)

0.4 1.5 Linear (April 14)

y = 1.8462x + 0.348 0.2 R2 = 0.2402

1 y = 1.7528x + 0.0893 0 R2 = 0.7687 14 15 16100*[L(sky)-L(ext.)]/L(ext) at 1020, 17 870, 675 and 4 40 nm 18 19 AOT(500 nm)

50

0.5

y = 0.9659x + 0.1163 30 R2 = 0.8315

0 Almucantar % differences, 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1 / m 10

14 15 16 17 18 19 Computed AOT varies -10 day in April, 2008 suspiciously as 1 / m Almucantar radiance differences at 6° azimuth should (m – airmass) agree within 5%. Coarse Mode Retrievals vs Wind speed (Eureka)

3 PEARL

2 m) µ µ µ µ ( eff,c r 1

0 MAR MAY JUL SEP MAR MAY JUL SEP MAR MAY JUL SEP MAR MAY JUL SEP MAR MAY JUL SEP

Eureka (OPAL) Wind Speed Climatology

20 Climate Normals (1971-2000) ) -1

10 Wind Speed (km hr Speed Wind (km

0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec High Arctic Starphotometer / moonphotometer network Summary & Conclusions

A multi-year AOD and effective radius climatology for the high Arctic showed a number of consistent features:

 Spring to summer decrease of fine-mode AOD (probably attributable to biomass burning and/or anthropogenic pollution).

 Significant correlation of fine mode AOD with CO concentration (which indicates a predominance of biomass burning aerosols throughout the entire year).

 West to east decrease in AOD on a pan-Arctic scale. Acknowledgements