The Brewer Triad and Calibrations;

the Canadian Brewer

Spectrophotometer Network

Vitali Fioletov and Stoyka Netcheva

Measurements and Analysis Research Section (ARQM)

Air Quality Research Division

Atmospheric Science and Technology Directorate

COST ACTION ES1207 EUBREWNET OPEN CONGRESS /

14th WMO-GAW BREWER USERS GROUP MEETING

State Meteorological Agency of Spain, Tenerife, Spain

24th – 28th March 2014

Network Integration

• Brewer network (7 sites with 2+ instruments and 3 sites with one instrument, 42 instruments including 10 Double Brewers) and 2 sites outside Canada

• AEROCAN sunphotometer network (20 sites)

• Ozonesondes network (8 sites)

• (new) Pandora (2 instruments)

• The Canadian Air and Precipitation Monitoring Network (CAPMoN) – operated by EC

– Particles and Related Trace Gases , 24 hour integrated filter samples. particulate Cl-, NO3-, SO4=, NH4+, Na+, K+, Ca++, Mg++, gaseous HNO3 and SO2

– Ground level Ozone Measurements , hourly averages.

– Nitrogen Measurements , gaseous NO, NO2, and NOy hourly averages.

– Size Selective Particulate Matter ,24 hour integrated samples PM2.5 , PM10 and coarse fraction mass

– Precipitation Chemistry , 24 hour integrated samples. Cl-, NO3-, SO4=, NH4+, Na+, K+, Ca++, Mg++, pH

– …

• National Air Pollution Surveillance (NAPS) – SO2, NO2, O3, PM2.5, PM10 - continuously monitored

AR

QM

Team members & support

• Stoyka Netcheva: network management

• Vitali Fioletov: science, algorithms, data analysis, QA/QC

• Tom Grajnar: network operation, calibrations, instrument maintenance

• Michael Brohart: instrument maintenance, repairs, upgrades and calibrations

• Akira Oguy: computer support, web, BPS, data processing, QA/QC

• Other ARQM staff providing support as needed

• Volodya Savastiouk - regular contracts for calibrations, instrument repairs, upgrade and maintenance of software for Brewer operation, calibrations, and diagnostics …

Support from other branches of EC (infrastructure, instrument operations, basic maintenance, and inspections) at 7 stations CANDAC occasional support at Eureka Pearl Observatory NOAA at Mauna Loa Observatory NOAA at South Pole Observatory

Network

New developments 2012-2014:

•Double Brewers at Churchill,

Goose Bay, and Stony Plain: no

more data gaps in winter

•Calibration of 4+2 Brewers at

Mauna Loa in October 2013

•The new triad of double Brewers

(#145, #187, #191) established in

October 2013

•Two new Brewer instruments

(#223 and #224) purchased in

2014

Brewers applications: Ozone change estimates Ozone changes on global scale from Dobson, Brewers and Filter Ozonometers

Deseasonalized, area-weighted total ozone deviations for 90°S–90°N, adjusted for solar, volcanic, and QBO effects. The yellow line represents the component of ozone variability due to changes in the equivalent effective stratospheric chlorine (EESC) based on a fit to data from 1964–2012 (updated from WMO (2007))

90°S–90°N

0

20

40

60

80

100

1978-1985 1986-1990 1991-1995 1996-2000 2001-2005 2006-2010

Dobson Brewer Filter

Quantity

The total number of sites reported data to the WOUDC. Stations submitted less than 100 DS observations per bin are not included

Quality

Relative number (in % of the total

number) of sites with "no issues" in the

record in 6 bins for Dobson, Brewer,

and filter instrument sites located

between 60ºS and 60ºN

Brewers applications: Brewers and others

Ground-Based total ozone data available from the World Ozone and UV Data Centre (WOUDC). In the most recent 5-year interval, the number of Brewer sites is almost equal to the number of Dobson sites.

Updated in 2013 from Fioletov et al., JGR, 2008

EC Brewer at NOAA at SPO vs. other instruments

WMO Antarctic Ozone Bulletin No 6/2013

Seasonal variation of the noontime UV Index at Eureka. The

top (first) panel compares noontime UVI measurements

performed in 2011 (red dots) with the average noontime UVI

(blue line), the range between the 10th and 90th percentile

(dark shading), and the range of historical minima and maxima

(light shading). The second panel shows the 2011 UVI

anomaly in absolute terms, calculated as the difference

between measurements and the average. The third panel

shows the relative UVI anomaly calculated as the percentage

departure from the climatological mean. The fourth panel

shows a similar anomaly analysis for total ozone derived from

satellite measurements. Vertical broken lines indicate the

times of the vernal equinox, summer solstice, and autumnal

equinox, respectively.

Form Bernhard et al., ACP, 2013

60% above the normal

Brewer applications: UV in the Arctic

Brewer UV from 3 Arctic Brewer sites (Resolute,

Eureka, Alert) are used in NOAA Arctic Report Card,

BAMS State of Climate report, and other publications

In the spring of 2011, they reported UV levels that are

60% higher than normal

Brewers application: UV Index forecast validation

Scatterplots and linear regression fits of EC and NOAA forecast methods (FM) plotted against collocated Brewer measurements. Blue open circles show EC forecasts, and red plus signs show NOAA forecasts. A small amount of random noise is added to the UV index data for better visualization. Four major weather types are represented: cloud-free sky (type 1), light clouds (type 2), heavy clouds (type 3), and rain (type 4). The correlation coefficients R between the measured and forecast values are also shown.

From He et al., JAMC, 2013

Brewers are used EC and

NOAA UV Index forecast

validation

The Brewer Triads and Calibrations

Toronto Brewers in March 2014

The triad update:

long-term record

Deviations of ozone values of

individual triad Brewers from the

mean of the three instruments. Each point on the graph represents a 3-month

average. Plots for ozone values from b-files

(top) and after re-processing (bottom) are

shown. The vertical line indicates the end of the

period used in the original triad paper.

Updated from Fioletov et al., GRL, 2005 (the triad paper)

Devia

tion

(%)

-8

-6

-4

-2

0

2

4

6

8

1985 1990 1995 2000 2005 2010 2015

Devia

tion

(%)

-8

-6

-4

-2

0

2

4

6

8

1985 1990 1995 2000 2005 2010 2015

Ozone from b-files

Reprocessed ozone with SL-corrections applied

Brewer 008

Brewer 014

Brewer 015

Brewers 008 and

014 were calibrated

in late 2008 σ = 0.5%

The main challenge in maintaining

the triad’s instruments is their

stability.

Calibrations and service every 2

years are necessary.

Toronto triad record vs. satellites Difference between triad data (available from the WOUDC) and various satellite instruments

-15

-10

-5

0

5

10

15

1978 1983 1988 1993 1998 2003 2008 2013

065 TORONTO, 44N, 79W, Brewer

Direct Sun Statistics 1978-1985 1986-1990 1991-1995 1996-2000 2001-2005 2006-2010 Mean Range

Number of days 368 1259 822 1217 1337 1437 1073.33 1069

Bias 0.16 0.05 -0.01 0.87 0.25 0.04 0.23 0.88

Daily sigma 2.33 2.4 2.4 2.16 2.03 1.73 2.17 0.67

Monthly sigma 1.13 1.03 1.05 0.9 0.75 0.69 0.92 0.44

Seasonal ampl. 0.56 0.15 0.33 0.24 0.32 0.32 0.32 0.41

Annual range 0.59 1.06 1.14 1.52 0.83 1.01 1.03 0.93

Seasonal mean differences between triad Brewer DS measurements and Nimbus 7 TOMS, Meteor 3 TOMS, EP TOMS, OMI, GOME, GOME-2, SCIAMACHY satellite total ozone data.

The whiskers represent 95% confidence limits for seasonal mean differences Table of the difference statistics. Note that the standard deviation of monthly mean differences is 0.7% in the most recent 5-year bin. updated from Fioletov et al., JGR, 2008.

Dif

fere

nce

in %

Pandora

Spectrometer

• Sun and sky spectrometer – measures solar spectra

• Can also work in ZS and MAX-DOAS modes

• Under development since 2006 (Jay Herman et al., NASA)

• Designed for satellite validation and pollution monitoring

• Operation and software design are similar to these for the

Brewer spectrophotometer (commands, schedules)

• Automated, established algorithms, data available in real

time

• Specifications:

– Czerny-Turner spectrometer with backthinned CCD detector

(Avantes)

– 270-530 nm at 0.6 nm spectral resolution, 4 pixels oversampling

– Wavelength independent FOV of 1.5 (FWHM)

– S/N of 400 at 400 nm

– T stabilized spectrometer ( enclose insulation under

improvement)

– High temporal resolution (<30 seconds per measurement)

– Simultaneous measurements of various trace gases incl. O3,

NO2, SO2, BrO, HCHO, water vapour

– Small size and portability (20 kg)

– Cost: ~$40k

• Two instruments were deployed in 2013 (McKay and

Toronto)

• Continue comparison with Triad instruments in Toronto.

The Pandora-Brewer difference

There is a 0% to 4% systematic difference between

Brewer and Pandora total ozone caused likely by the

difference in ozone absorption coefficients and their

temperature dependence.

Bre

wer

-Pan

dora

Diff

eren

cein

%

-2

-1

0

1

2

3

4

5

01OCT13 01NOV13 01DEC13 01JAN14 01FEB14 01MAR14 01APR14

5-day averages

Brewer 008 Brewer 014 Brewer 015 Brewer 145 Brewer 187 Brewer 191

The triad update - 2013

Tota

lO

zone

(DU

)

200

250

300

350

400

450

500

550

EST6:00:00 8:00:00 10:00:00 12:00:00 14:00:00 16:00:00 18:00:00

Brewer 008 Brewer 014 Brewer 015 Brewer 145Brewer 187 Brewer 191 Pandora 103 Pandora 104

An example: Feb 22, 2014

Diffe

rence

in%

-6

-5

-4

-3

-2

-1

0

1

2

Airmass value

1 2 3 4 5 6 7

Diffe

rence

in%

-4

-3

-2

-1

0

1

2

3

4

01OCT13 01NOV13 01DEC13 01JAN14 01FEB14 01MAR14 01APR14

5-day Statistics, mu<3

Brewer 008 Brewer 014 Brewer 015 Brewer 145 Brewer 187 Brewer 191

Brewers 014, 015, 191 returned from Mauna Loa

Old Triad

New Triad

Hg bulb was replaced in Brewer 008

Brewer and Pandora measurements at Toronto. Pandora measurements with 1.5 minute frequency were used to account for the difference in ozone due to difference in observation time between individual Brewer measurements.

Stray light causes an underestimation of ozone values by the single Brewer. Pandora measurements adjusted for the bias were used as a reference.

The whiskers represent 5st and 95st percentiles, the box

edges represent the 25th and 75th percentiles, the centre is

drawn at the median value

The old and new triads agree within 1%, but there are some systematic differences likely due to stray light. The long-term instrument stability is a challenge.

The biases between individual Brewer triad instruments and the “baseline.” The baseline was established using high-frequency Pandora measurements adjusted for the Pandora-Brewer bias:.The biases between Pandora and the new triad due to different ozone cross-sections were removed on a daily basis. The error bars represent 95% confidence limits for 5-day mean differences

New Triad Old Triad

Brewer UV measurements were normalized to modeled UV at 324 nm for sea level clear sky, no aerosol conditions. Noon values were

calculated as averages between 11:30 and 12:30. The whiskers represent 1st and 99st percentiles, the box edges represent the 25th and 75th

percentiles, the centre is drawn at the median value.

MLO Brewer observations: Based on cloud cover information, June and September are the best months for calibration, while March, April, and November are the worst ones

0

10

20

30

40

50

60

70

80

90

100

110

120

130

140

150

160

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Noon Values (11:30 - 12:30 LST)

Brewer 009 Brewer 119

0

10

20

30

40

50

60

70

80

90

100

110

120

130

140

150

160

7 8 9 10 11 12 13 14 15 16 17

Brewer 009 Brewer 119

Noon normalized UV at 324 nm values in different months Normalized UV at 324 nm values as a function of solar time

0

3

6

9

12

15

18

21

24

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Noon Values (11:30 - 12:30 LST)

Brewer 009 Brewer 119

Noon UV index values in different months

EC operates 2 Brewers at the NOAA Mauna Loa Observatory. Brewers 009 and 119 were installed in 1997. The most recent calibration of both instruments was in October 2013

Brewer 009 data for 1997-2012 and Brewer 119 data for 2003-2012 were used for this slide

Sun seen through fume

cloud Kilauea (photo by J.D. Griggs,

http://pubs.usgs.gov/of/2007/1114/)

MLO Brewer observations:

Volcanic SO2

The activity of Kilauea volcano near Mauna

Loa has increased after 2007

This may affect ozone calibration activities, particularly for Dobsons.

SO2 values as high as 50 DU under clear sky and as high as 400 DU

under cloudy conditions (with the multiple scattering effect) were seen.

Note that the Observatory (3400 m) is located above the volcano (1222

m) otherwise column SO2 would be even larger.

-1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 DU

OMI mean total column SO2 over Hawaii in 2005-2007 and 2008-2010 Mauna Loa UV spectra, April 26, 2008, 22:53 UTC (12:33 LST), Brewer #009, SO2=400 DU due to multiple scattering within the cloud.

290 300 310 320Wavelength (nm)

1.0E-6

1.0E-5

1.0E-4

1.0E-3

1.0E-2

1.0E-1

Irra

dia

nce (

W/m

2/n

m)

Statistical model estimation

3 std. deviations envelope

Measurement

290 295 300 305 310 315 320 325Wavelength (nm)

1

10

100

1000

10000

Ra

tio

(expe

cte

d/m

easure

d)

0

5

10

15

20

25

SO

2 a

bso

rption

(cm

-1 b

ase e

)

Global UV spectrum expected in absence of SO2

Measured spectrum

Ratio: expected/measured

MLO Brewer and Dobson observations Brewer ozone retrievals are not sensitive to SO2, while Dobson overestimates ozone in presence of SO2

Tota

lO

zone

(DU

)

220

230

240

250

260

270

280

16OCT13 01NOV13 16NOV13 01DEC13 16DEC13 01JAN14

Brewer 009 Brewer 014 Brewer 015 Brewer 017 Brewer 119 Brewer 191 Dobson

Large Dobson-Brewer differences

correspond to high SO2 values

The last

day of the

calibration

trip

The whiskers represent 10st and 90st percentiles, the box edges represent the 25th and 75th

percentiles, the centre is drawn at the median value. Dobson measurements (direct sun or blue

zenith) are taken once a day near local noon. Dobson data were provided by NOAA.

SO2=2 DU

SO2=8 DU

SO2=5 DU

Tota

lS

O2

(DU

)

-5

0

5

10

15

20

25

30

35

10:00:00 12:00:00 14:00:00 16:00:00 18:00:00

Brewer 009 Brewer 119

Total column SO2 measured by

Brewers 009 and 119 measured

on November 20, 2013.

Dobson and Brewer

ozone measurements

in October-December

2013.

Dobson measurement was at 13:00 HAST

New Brewer ZS algorithm: The same parameterization, but only 2 parameters instead of 9 are estimated from measurements Fioletov, V. E., C. A. McLinden, C. T. McElroy, and V. Savastiouk (2011), New method for deriving total ozone from Brewer zenith

sky observations, J. Geophys. Res., 116, D08301, doi:10.1029/2010JD015399.

The Brewer ZS algorithm does really not work when

μ>4 and ozone>300: the Brewer wavelengths and

weighting coefficients produce values that are not

sensitive to ozone.

The plot shows weighted sums of logarithms of zenith

irradiance intensities (G) as a function of total ozone for

various values of slant path (μ) for high latitudes. The

model calculations were done for albedo = 0.80. G shows

little sensitivity to total ozone when μ >4 and total

ozone is greater than 400.

The shape of the vertical ozone profile affect the ZS

retrievals when when μ>3.

Difference between DS ozone daily means and ZS ozone

means for different µ values for Churchill (59°N). ZS data were

processed using models based on midlatitudinal (black) and

high‐latitudinal (gray) ozone profiles. Calculations were done

using 0.8 albedo for December–April and 0.05 for the rest of

the year. The mean values (horizontal lines) with 2 standard

errors of the mean (vertical bars) are shown.

UV calibration: comparison between 40 and 50

cm calibration systems (4 identical UV lamps

and data logger on both systems)

The combined effect of polar ozone depletion and high tropopause is seen below for March 17, with

a large ozone anomaly (30-35% below normal) centered over the most populated areas of Europe.

Work with Kipp and Zonen to facilitate service

and maintenance and improve instrument

reliability • Replace NFS80 with NFS110 power supply -more current on 24V line for more reliable micrometer motor

movement

• Higher power level applied to standard lamp - improves stability of dead time and run stop test results

• Ball-end grating arm and micrometer points-eliminates potential source of hysteresis in micrometer movement

• Rotary connector installed in trackers along with heater system - cables do not rotate with tracker during day, eliminates data and power cable movement stress and breakage, internal shut off switch not required, no accidental breakage of shut-off string due to accidental over-rotation during drive plate cleaning, tracker heater helps to ensure better tracker movement in cold environments

• Use 296.728 nm hg single peak line for hg test instead of 302.1nm doublet peak line - reduce hg lamp warm-up to 2 minutes from 5 and reduce micrometer positioning changes since doublet line peak can change with lamp age, reduces hg lamp aging, saves time for more ds, uv, etc.

• Neutral density filters with tighter tolerances on specified attenuation values-prevents artificial spikes in ozone measurements when successive neutral density filters are introduced into the optical path

• Magnetic window in back tracker cover with chain link to tracker cover bolt - facilitates visual inspection and tool-free cleaning of the drive plate edge, eliminates tracker seal issues due to cover bolt spacer positioning errors

• 1/8” Hole in center of blocked filter - to determine and confirm the UV focusing of the Brewer

• Micrometer movement referenced to spectrometer LEDs - allows for reproducible micrometer reset even after reinstallation/replacement of the micrometer and/or pushrod replacement

• Power and data cables rated to -70 C - More reliable in colder environments

Brewer operational issues that still need to be

addressed

• Long-term stability for micrometer motor movement—mostly resolved by new NFS110 power supply but still have instances of poor micrometer movement and jamming

• Standard lamp power—needs to be adjustable as it was in single board Brewers so that standard lamp intensity can be optimized for each lamp as required in order for stable and acceptable diagnostic test results

• Sensitivity to power issues—single board Brewers appear to be sensitive to power failures or brownouts and sometimes require config and/or firmware to be re-installed

• PMT sensitivity decline over time-- Brewer PMT’s have been originally set with a relatively low high voltage setting and the NFS80 power supply may not have been providing the necessary current at 24V

• PMT acceptance testing prior to installation into Brewers--one of our PMTs recovers slowly after being exposed to light and residual signal can bleed over into the next measurement

• Noisy negative voltages—understand the cause of the variability to confirm no effect on the instrument

Strategy

• Maintain smaller network but with assured stream of high

quality data with 2+ Brewers at each site plus other

instruments

• Let the Brewer do what it can do best, i.e., to produce

high quality and stable over time total ozone and UV data

and use other instruments for other tasks

• Focus on implementing what has been developed by

other, rather than developing everything ourselves

• Strongly support unification and standardization of

measurements and data processing

• Open to suggestions