Wet plus Dry Deposition of Atmospheric Hg in the SE U.S. NADP Technical Meeting Sept. 10-12, 2007 Boulder, CO, USA
Outline Routine Measurements and Deposition Estimates Coarse Particle Hg Further Thoughts
SEARCH Network Hg-related measurements SEARCH Sites Weekly wet dep Hourly Hg/Daily wet dep (active as of 1/1/07) Hourly Hg/Daily wet dep (inactive as of 1/1/07)
Met Yorkville, GA (YRK) Trace Gases Tekran collector
SEARCH Atmospheric Hg Measurements ~99% 1% 1% Hg 0 - Tekran 2537A 5 minute averages (every other hour) Gold trap pre-concentration CVAFS detection RGM - Tekran 1130 Module 60 minute average (every other hour) KCl thermal annular denuder Desorption at 500C Hg P - Tekran 1135 Module 60 minute average (every other hour) Impactor @ inlet removes particles >2.5 micron Quartz filter (downstream of impactor) Desorption at 800C
Tekran Configuration Hg PM2.5 on quartz filter RGM on KCl denuder (analogous to HNO 3 ) Nafion drier Air in PM 2.5 Hg 0 transmits quantitatively
Tekran Modifications Made in SEARCH Nafion drier (keep water out of detector) Soda lime trap (keep reactive gases out of detector) Log flows (confirm proper operation and calculations) Log temperatures (confirm operation)
Data Verification Activities Hg 0 injection onto 2537A traps 100 +/- 2% recovery Method of Addition Hg 0 injection at inlet 100 +/-5% recovery, no fictive RGM or Hg PM2.5 Parallel denuder/filter sampling RGM 100 +/- 10%, Hg PM2.5 88 +/- 20% Spiked denuder experiments (RGM) HgCl 2 recoveries 56-167% (n=20) Automated and Manual Data review
Raw 5-minute Time Series Zero air purges
Supporting SEARCH Measurements Discrete Particles and Gases (24-hour) PM 2.5 and PM 10 (filter-based mass, speciation) NH 3 (annular denuder) Continuous Particles (1-min to 1-hr) PM2.5, TC, BC, NH 4+, NO 3-, SO 4 2- Trace Gases (1-min) O 3, NO, NO 2, NOy, HNO 3, SO 2, CO Meteorology and Visibility (1-min) WS, WD, T, RH, BP, SR, rainfall, Bsp (dry)
Urban Mercury Measurements Jefferson Street 2/8-9/02 10 6.0 CO Hg(0) 9 8 5.0 7 CO (ppm) 6 5 4 4.0 3.0 Elem. Hg (ng/m3) 3 2 2.0 1 Background Hg(0) 0 1.0 2/8/02 0:00 2/8/02 4:00 2/8/02 8:00 2/8/02 12:00 2/8/02 16:00 2/8/02 20:00 2/9/02 0:00 2/9/02 4:00 2/9/02 8:00 2/9/02 12:00 2/9/02 16:00 2/9/02 20:00 CO and Hg 0 track each other during major excursions
Hg(0) vs. CO regression slopes Nocturnal events @ JST Month (2002) Lower 95% CI Upper 95% CI r-sq. January 0.24 0.54 0.51 February 0.47 0.58 0.62 March 0.57 0.67 0.78 April 1.47 1.69 0.68 May 1.63 1.91 0.71 June 2.36 2.71 0.73 October 0.96 1.15 0.68 November 0.37 0.46 0.66 December 0.60 0.68 0.78
60-minute SO 2 and Hg 150 4/28/06 Event - OLF 1800 SO 2 Hg P RGM Hg 0 SO 2, Hg P, RGM 125 100 75 All 4 species track during the event 1700 1600 Hg 0 50 1500 25 0 605 805 1005 1205 1405 1605 Time 1400
Summary of SEARCH 1-Hr Hg Measurements Site Year Hg 0 (pg/m 3 ) RGM (pg/m 3 ) Hg PM2.5 (pg/m 3 ) mean max. mean max. mean max. OLF, FL 2004 1490 2440 4.8 182 1.4 48 2005 1470 2850 2.7 183 1.3 38 2006 1580 2740 4.0 127 2.5 49 BHM, AL 2004 2490 35,900 56 8560 10.3 # 185 # 2005 2420 130,200 77 13,700 24.9 2410 2006 2570 165,400 82 9210 31.6 2940 YRK, GA 2002/3 1566 9470 4.5 169 nm nm JST, GA 2002/3 1917 8400 6.1 288 nm nm # partial year All 3 forms of Hg significantly elevated at BHM; Hg 0 elevated at JST
HgP Hg Species by WD at BHM - 2006 0 0 345 15 345 15 330 30 330 100 30 Hg 0 (ng/m 3 ) 285 300 315 45 60 75 285 300 315 80 60 40 20 45 60 75 Hg P (pg/m 3 ) 270 90 270 0 90 255 105 255 105 240 120 240 120 225 135 225 135 210 195 180 165 150 330 345 750 0 15 30 210 195 180 165 150 RGM (pg/m 3 ) 300 315 600 450 45 60 285 300 75 150 270 0 90 255 105 240 120 225 135 210 195 180 165 150
BHM Aerial View
Hg 0 ng/m 3 Hg Species by WD at OLF - 2006 0 0 345 15 345 15 330 30 330 30 285 300 315 2.00 1.75 1.50 1.25 45 60 75 285 300 315 5.0 4.0 3.0 2.0 1.0 45 60 75 Hg P (pg/m 3 ) 270 1.00 90 270 0.0 90 255 105 255 105 240 120 240 120 225 135 225 135 210 195 180 165 150 330 345 6.0 0 15 30 210 195 180 165 150 RGM (pg/m 3 ) 285 300 315 5.0 4.0 3.0 2.0 45 60 75 1.0 270 0.0 90 255 105 240 120 225 135 210 195 180 165 150
OLF Aerial View Gulf of Mexico - No known local sources Pensacola
Diurnal Pattern of Hg Species at OLF 1.75 2006 Hg(0) HgP RGM 10 1.65 8 Hg(0) (ng/m3) 1.55 1.45 6 4 HgP, RGM (pg/m3) 1.35 2 1.25 0 2 4 6 8 10 12 14 16 18 20 22 All Hour of Day 0
Diurnal Pattern of Hg Species at BHM 5.0 2006 Hg(0) HgP RGM 125 4.0 100 Hg(0) (ng/m3) 3.0 2.0 75 50 HgP, RGM (pg/m3) 1.0 25 0.0 0 2 4 6 8 10 12 14 16 18 20 22 All Hour of Day 0
Dry Deposition Estimates Unlike wet deposition, dry deposition is difficult to measure. However, dry deposition can be estimated based on analogy with HNO 3 (for RGM) and PM 2.5 (for Hg PM2.5 ). No simple analogy for Hg 0 Dry deposition (DD) calculation: DD = ΣC i xv di, where C i = hourly concentration and V di = hourly deposition velocity
Dry Deposition Resistance Models Flux = C HNO 3 (RGM) i /Σr i, = C i xv di PM 2.5 (Hg PM2.5 ) Hg 0, NH 3 Simple: unidirectional Complicated: bi-directional
Vd, cm/s 3.5 3.0 2.5 2.0 1.5 1.0 Hourly Deposition Velocity (V d ) cm/s HNO 3 (RGM) PM 2.5 (Hg P ) Q1 Q2 Q3 Q4 Vd, cm/s 0.40 0.35 0.30 0.25 0.20 0.15 0.10 Q1 Q2 Q3 Q4 0.5 0.05 0.0 0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0.00 0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 Meyers, T. P., P. Finkelstein, and J. Clarke, 1998: A multilayer model for inferring dry deposition using standard meteorological measurements. Journal of Geophysical Research 103 (D17):22,645-22,661. Vd exhibits strong diurnal patterns; therefore, ΣC xvd NE 0 and it is necessary to measure RGM and Hg P with high time resolution.
Measured Wet and Estimated Dry Deposition (2005) Wet (µg/m 2 ) Dry (µg/m 2 ) Site Total-Hg Hg P RGM OLF 18.5 0.1-0.2 1.1-2.2 BHM 24.7 0.8-1.6 24-48 23.8 Wet Deposition >> Dry deposition at OLF (non-industrial) but comparable at BHM (industrial) Caveats: 1) Dry Deposition estimate does not include Hg 0 2) Dry deposition estimate for Hg P includes PM 2.5
OLF Experiments Dual Tekrans operated synchronously and asynchronously with various inlet configurations. Questions addressed: 1) What is measurement precision? 2) What is the size distribution of Hg P?
Collocation Experiments at OLF P = primary Tekran, C = colo Tekran HgP RGM pg/m 3 pg/m 3 Inlets P C P C PM 2.5 /PM 2.5 1.4 1.8 5.4 4.9 PM 2.5 /TSP 2.3 6.9 4.4 5.0 PM 2.5 /PM 2.5 1.9 1.8 5.4 4.9 PM 2.5 /PM 10 1.6 4.5 3.0 3.4 1) Measurement precision +/- 10% for RGM, +/- 25% for Hg P 2) Roughly 2-3 times as much Hg P in PM 10 and TSP as PM 2.5 3) RGM may be contaminated by Hg P if PM 2.5 inlet removed
Further Thoughts Measurement technology is approaching maturity and we are rapidly moving toward a national network for measuring atmospheric Hg (and estimating dry fluxes). We should pause to: Recognize that, except for Hg 0, we are measuring forms not species (if not, we underestimate uncertainties). For example, if we believe Hg PM2.5 is important, then we should investigate size distributions (pilot studies) Think about reporting units (ng/m 3 at 0 C and 1 atm. potentially unrepresentative). Some options are: Use current units, but always report ambient T and BP Use mixing ratios Consider more conventional calibration using bottled Hg 0. Some advantages/disadvantages: Automated calibrations (displacement and method of additions) Remote actuation and adjustment Cost and development effort/time Consider whether thermography can (or should) be used to learn more about RGM and Hg PM2.5?
Thermography of Hg PM2.5 and RGM 20.0 17.5 Purge1 HgP RGM Purge 2 800 Hg, pg/m3 Denuder T Filter T 15.0 600 12.5 pg/m3 10.0 400 T, C 7.5 5.0 200 2.5 0.0 9/21/06 22:05 9/21/06 22:10 9/21/06 22:15 9/21/06 22:20 9/21/06 22:25 9/21/06 22:30 9/21/06 22:35 9/21/06 22:40 9/21/06 22:45 9/21/06 22:50 9/21/06 22:55 9/21/06 23:00 0
Summary and Conclusions SEARCH joined MDN and initiated atmospheric Hg measurements in 2000/2001. Plans are to continue through 2010. Routine, continuous measurements of atmospheric Hg 0, RGM and Hg PM2.5 are obtainable with current technology, but require significant oversight. Hg data for 4 southeastern sites shows highly elevated concentrations (all 3 forms) at an industrial site and moderately elevated concentrations (Hg 0 only) at an urban site, relative to rural/suburban sites. Based on 2005 data, estimated DD of RGM and Hg PM2.5 < WD at rural/suburban sites, but may be comparable at industrial sites (Hg 0 deposition unknown, but likely highly variable in space and time). Inlet experiments show roughly 3 times higher Hg P in PM 10 and TSP than PM 2.5. Coarse Hg P will contribute to dry deposition and may provide clues to atmospheric Hg processing. The prospect of a nationwide Hg network is exciting, but we need to continue to think through details of the measurement and reporting process.
Acknowledgements Funding for SEARCH is provided by EPRI and Southern Company Further information about SEARCH can be obtained at: www.atmospheric-research.com