Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

Forecast and Hindcast Modeling in the Grand Bay Mercury Intensive

Mark Cohen, Fantine Ngan, Roland Draxler, Winston Luke, Paul Kelley, and Richard Artz

NOAA Air Resources Lab, Silver Spring, Maryland _____________________

Grand Bay Mercury Intensive Data and Planning Meeting, Jan 24, 2011Georgia Tech UniversityAtlanta, GA

1

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

Outline

emissions of Hg(0), Hg(II), Hg(p)

Hg from other sources: local, regional & more distant

wet and dry depositionto the water surface

Enhanced oxidation of Hg(0) to RGM?

Enhanced deposition?

Reactive halogens in marine boundary layer

Measurement of wet

deposition

Measurement of ambient air

concentrations

wet and dry deposition

to the watershed

1. Forecasting for mission planning

2. High-resolution hindcast meteorological modeling (Fantine Ngan)

3. Atmospheric Mercury and Other Hindcast Modeling

2

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

Forecasting for Mission Planning During the Summer 2010 Intensive

Based on the NOAA NCEP NAM 12km weather forecast model product

This forecast starts at UTC 00 for the given date, which is 7 PM the night before in Grand Bay time (GBT).

The hours displayed in the product – 9 AM to 8 PM GBT – thus represented hours 14-25 in the 48 hr t00z forecast.

The file became available to NOAA ARL at ~2:00 AM GBT, and the processing of the data to make this product was generally done by 5:00 – 6 00 GBT.

Goal was to have product ready each day by 7:00 GBT

3

HYSPLIT-based forecast product for the Grand Bay Intensive

Trajectories (pages 2-13)

One page for each local hour from 9 AM to 8 PM

Image in upper left corner is mixing height (m)

The other five images are back-trajectory maps, each map representing a starting point at a different elevation (meters) above mean-sea-level (250, 500, 1000, 2000, and 3000)

There are nine trajectories shown on each map: one starting at the Grand Bay NERR and on a +/- 1 deg lat/long grid around the NERR

The trajectories each go back 96 hours; but the trajectories may not stay on the map for all 96 hours.

On each trajectory there is a little dot showing the location at six-hour intervals; and a larger symbol at 00 UTC each day.

In the panel below the trajectories, the height above the surface is shown for each trajectory as it goes back in time

Note that elevations in trajectory maps are shown as being “above ground level”, and the starting heights labels in the panels below each trajectory are shown for the first trajectory run. This happened to started over land where the elevation was ~80m, so, the labels happen to refer to that trajectory, so the labels say ~170, 420, 920, 1920, and 2920.

Wind Direction at Different Elevations (pages 14-25)

One page for each local hour from 9 AM to 8 PM

Each image shows a map of wind direction, at each grid point in the NAMSF 12-km forecast at a particular vertical level in the met data set

There are 6 maps, corresponding approximately to the elevations 10, 250, 500, 1000, 2000, 3000.

The met data is on “terrain following sigma levels”; so, the heights above the ground at any given location are influenced by the terrain height.

Meteorological Data Contours (pages 26-37)

One page for each local hour from 9 AM to 8 PM

Image in upper left corner and upper middle are the mixing height (same map as shown on Trajectory pages). The two maps are slightly different due do differing interpolation procedures

The map in the upper right corner is forecast precipitation, shown as a 3-hr accumulation, ending at that hour; that is the amounts shown are the total forecast precipitation over the previous 3 hrs.

Other maps shown are for the following three surface parameters: pressure, downward shortwave radiation flux, and friction velocity

Meteorological Data for Grand Bay NERR (pages 38-50) Page 38 is description of data that are shown Then, one page for each local hour from 9 AM to 8 PM Each page shows the meteorological data at the surface and at each level

of the gridded, forecast meteorological data set. Note that on these pages and throughout, times are expressed in UTC

(Universal Time Coordinate). These are currently 5 hrs ahead of Grand Bay, e.g., 9 AM Grand Bay is 2 PM UTC (or UTC 14)

RGM Plumes from Large Regional Sources (pages 51-63)

Page 51 shows a map of some of the large sources in the region; as stated on map, some of the sources are no longer emitting. These sources were not included in these simulations.

One page for each local hour from 9 AM to 8 PM; each page shows model-estimated RGM concentrations for six different vertical layers n the atmosphere.

These maps do not represent the total RGM in the atmosphere, but only the fraction contributed by large regional sources.

The maps show average concentrations for the hour leading up to the stated hour, e.g., the map for UTC 14 represents average concentrations between UTC 13 and 14 (8 – 9 AM Grand Bay)

Based on the t00z NAM-12km forecast generated by the National Weather Service

This forecast starts at UTC 00 for the given date, which is 7 PM the night before in Grand Bay time (GBT). The hours displayed in this product – 9 AM to 8 PM GBT – thus represent hours 14-25 in the 48 hr t00z forecast. The file becomes available to NOAA ARL at ~2:00 AM GBT, and the processing of the data to make this product is generally done by 5:00 – 6 00 GBT.

4

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

The first set of pages were hourly maps of mixing heights and back trajectories

5

6

7

8

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

The second set of pages were hourly maps of wind vectors at different heights

9

10

11

12

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

The third set of pages were hourly maps of mixing height, PBL-height, 3-hr precipitation, surface pressure,

downward shortwave radiation flux and friction velocity

13

14

15

16

17

18

19

20

21

22

23

24

25

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

The fourth set of pages were hourly tables of surface variables and meteorological

parameters in a vertical column above the site

26

27

28

29

30

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

The fifth set of pages were hourly maps of modeled RGM concentrations in the region,

arising from anthropogenic sources in the region

31

32

33

34

35

36

37

38

39

40

41

42

43

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

Forecasting for next intensive?

What format would be most helpful? Maybe just have one page per hour, and then one could look at

one part of the page as one scrolled through… This would mean far less information, though, as we are limited

to about six images per page… If we did it this way, what would the six images be? Or maybe this is too limited? Another issue was bandwidth for downloading – I had to

degrade the quality of the images significantly in order to keep file size as small as it was (~12 MB), and even this was perhaps too large

44

45

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

2. High Resolution Hindcast Meteorological Modeling

A 4-km meteorological data modeling analysis (with data assimilation) has been produced by Dr. Fantine Ngan, a post-doc at NOAA ARL, for the Summer 2010 Intensive

Unfortunately, Fantine couldn’t be at this meeting as she is at the AMS meeting in Seattle this week, but she prepared the slides in this section for presentation here

10 m wind fields 18 UTC August 06

46

South-west corner (km)

Number of cells Resolution (km)

Starting point relative to mother domain

X-origin Y-origin Easting Northing X-direction Y-direction

D01 -2808 -2268 157 127 36 1 1

D02 180 -1548 193 223 12 84 21

D03 708 -1260 163 151 4 45 25

Projection center: 40N, 100WStandard latitude: 30N, 60N

Layers: 43, with model top at 50 mb (1st layer thickness is 33 m and 15 layers are below 850 mb)

Domain configuration

D01

D02

D03Grand Bay– 30.4123, -88.4037, 5 m

47

Simulation period: 210/07/31 00 UTC – 08/13 12 UTC

IC/BC for D01 is from GFS data + objectively analysis (OBS2GRID)Others are nestdown from coarse domain.

3D grid nudging, SFC nudging and OBS nudging are on for all domains

The model was run in 5.5 day segments and re-initialized every 5 days. There were 12 hours overlapping between each run segment.The soil moisture and temperature in the input files were replaced with the WRF output in the previous run segment

run segment period p1 7/30 00 UTC – 8/4 12 UTC p2 8/4 00 UTC – 8/9 12 UTC p3 8/9 00 UTC – 8/13 12 UTC

Physical optionsmicrophysics: WSM 3-class schemeradiation scheme: RRTM scheme for longwave radiation

Dudhia scheme for shortwave radiationland surface model: PX LSMPBL scheme: ACM2 schemecumulus scheme: Grell-Devenyi Ensemble scheme

Model Setup

48

Temperature time series at Grand BayBlack *: observation, Red line: WRF model

Relative humidity time series at Grand Bay

Measurements are provided by Winston Luke 49

Wind speed time series at Grand BayBlack *: observation, Red line: WRF model

Wind direction time series at Grand Bay

50

Soundings at Grand Bay at 2010/08/06 16 UTC

Wind speed

Potential temperature

Measurements are provided by Winston Luke 51

2 m temperature 09 UTC August 06 2 m temperature 19 UTC August 06

PBL height 09 UTC August 06 PBL height 19 UTC August 06

Shaded background: modelColor circles: MADIS OBS

5252

10 m wind fields 12 UTC August 06 10 m wind fields 15 UTC August 06

10 m wind fields 18 UTC August 06 10 m wind fields 21 UTC August 06

Shaded : model wind speedVector: model wind vectorsColor circles: MADIS OBS

5353

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

Questions for the group

We are going to be using this dataset for modeling, but would this met data set be useful for anybody’s else’s data analysis?

Fantine (and/or Mark) might be able to help extract needed data for any particular use…

54

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

3. Atmospheric Mercury and Other Hindcast Modeling

A. Back-trajectory modeling using HYSPLIT

B. Comprehensive Fate and Transport Modeling using HYSPLIT-Hg

Note – both kinds of modeling will utilize high-resolution met data developed by Fantine to the greatest extent possible

55

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

3. Atmospheric Mercury and Other Hindcast ModelingA. Back-Trajectory Modeling Using HYSPLIT

Can model back-trajectories arriving at the Grand Bay site and try to correlate with observations at the site

Can model back-trajectories associated with any given air-craft measurement and try to correlate with these measurements

The idea is that maybe having information regarding where air masses were coming from will help with data interpretation

56

Time series of Reactive Gaseous Mercury (RGM), Fine Particulate Mercury (FPM) and Gaseous Elemental Mercury (GEM) from two co-located instruments (D1 and D2) (top graph) and of SO2, O3, NO, NOy, and CO (bottom graph) measured at the Grand Bay NERR from May 3-8, 2008

57

Time series of Reactive Gaseous Mercury (RGM), Fine Particulate Mercury (FPM) and Gaseous Elemental Mercury (GEM) from two co-located instruments (D1 and D2) at the Grand Bay NERR from May 3-8, 2008

58

Date and Time (UTC)59

Date and Time (UTC)60

Date and Time (UTC)61

Date and Time (UTC)62

Date and Time (UTC)63

Date and Time (UTC)64

Date and Time (UTC)65

Date and Time (UTC)66

Date and Time (UTC)67

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

3. Atmospheric Mercury and Other Hindcast ModelingB. Comprehensive Fate and Transport Modeling Using HYSPLIT-Hg

Model mercury from regional and global sources for the time period of the Intensive

Compare model predictions against surface measurements at the Grand Bay NERR site as well as measurements aloft during flights

One goal is help with data interpretation, i.e., help explain why we saw what we saw

Another goal of this analysis will be model sensitivity and model evaluation, i.e., Can model reproduce measurements? If not, why? Sensitivity to key uncertainties?

68

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

Emissions from sources during intensive, especially local/regional sources that particularly impacted our observations?

Some issues with mercury fate/transport modeling

69

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

Emissions data: a critical model input

In model evaluation, want to diagnose weaknesses in model without large influence of emissions errors

Need accurate, speciated emissions estimates for all sources impacting the site – for the time period of the episode

Resolution: 2.5 min Duration: 11 Days

0

2

4

6

8

10

12

25-Aug 26-Aug 27-Aug 28-Aug 29-Aug 30-Aug 31-Aug 01-Sep 02-Sep 03-Sep 04-Sep 05-Sep

Hg

- (

ug

/m3 )

HgT

Hg0

Hg2

Series 3300 CEM - Continuous Speciated Mercury Data

Pascagoula MSW Incinerator shut

down in Jan 2001, but still in 2002 NEI

CRIST: New scrubber installed Dec 2009 will dramatically reduce RGM emissions

Lowman reported dramatic drop in mercury emissions in 2008 TRI

Brewton paper mill : Hg emissions in 2002 NEI , but do not appear in 2000-2008 TRI

IPSCO Steel: significant Hg

emissions in 2002 NEI, but

negligible emissions in

2008 TRI

Large mercury emissions point sources from USEPA 2002 National Emissions Inventory (NEI)

Grand Bay NERR site

70

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

Emissions from sources during intensive, especially local/regional sources that particularly impacted our observations?

Some issues with mercury fate/transport modeling

How to incorporate bromine and other chemical measurements into the modeling? What kind of collaboration makes sense?

Model configuration for prediction and subsequent comparison at a fixed site – e.g., the Grand Bay NERR site -- is “routine”, but prediction & comparison for a moving platform will be a little more challenging

71

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

Summary of modeling questions for the group discussed in this presentation

FORECASTING: What kind of forecast product would be most useful for the next intensive?

HIGH-RES MET DATA: Does anybody need/want extracts from the high-resolution met data set, or can think of any other way that this data set might be useful for their analysis?

BACK-TRAJECTORY MODELING: We will most likely carry out some kind of back-trajectory analysis, using the high-resolution met data, but does anybody have any particular requests? The goal is to help with data interpretation.

FATE/TRANSPORT MODELING: We will definitely carry out this kind of modeling and would like to collaborate in any way that makes sense…

For both the BACK-TRAJECTORY and FATE/TRANSPORT modeling, do we have the complete set of flight locations (lat/long/elevation) and times associated with each measurement?

72

Extra Slides

73

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

color of symbol denotes type of mercury source

coal-fired power plants

other fuel combustion

waste incineration

metallurgical

manufacturing & other

size/shape of symbol denotes amount of RGM) emitted during 2002 (kg/yr)

10 - 50

50 - 100

100 – 300

5 - 10

urban areas

Grand Bay NERR monitoring site

Watson

Daniel

Barry

Lowman

CristMobile

Pensacola

MississippiAlabama

Florida

Louisiana

Pascagoula MSW incinerator **

Brewton paper mill*

* Hg emissions included in 2002 NEI , but do not appear to be in 2000-2008 TRI ** Hg emissions included in 2002 NEI but incineration ceased in January 2001*** Significant Hg emissions in 2002 NEI, but negligible emissions reported in 2008 TRI

Eaton

Gaylord Container Bogalusa

IPSCO Steel ***

Biloxi-Gulfport

New Orleans

Hattiesburg

Slidell

Pascagoula

Location of the Grand Bay NERR sampling site, along with large point sources of RGM in the region, based on the EPA’s 2002 National Emissions Inventory

74

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

Mercury Air Emissions from Charles R. Lowman Power Plant as reported to the Toxic Release Inventory

For some sources, there have been big changes since 2002, the date of the last comprehensive mercury inventory

75

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

Current Atmospheric Measurements of Ambient Air Concentrations and Meteorological Data

Elemental mercury (two instruments)

Fine particulate mercury (two instruments)

Reactive gaseous mercury (two instruments)

Sulfur dioxide

Ozone

Carbon Monoxide

Nitrogen Oxides (NO, NOy)

Aerosol Black Carbon

Wind speed, Wind Direction

Temperature, Relative Humidity

Precipitation Amount

“Speciated” Atmospheric Mercury Concentrations

Trace gases and other measurements to help understand and interpret mercury data

Meteorological Data

76

Grand Bay Mercury Intensive, Data and Planning Meeting, Jan 24, 2011, Georgia Tech University

Wet Deposition Measurements added in 2010 by the Mississippi Department

of Environmental Protection (Henry Folmar, Becky Comyns, others), with

funding from the USEPA

Precipitation Continuous digital measurement of precipitation amount

Major IonspH, SO4

-2, NO3-, PO4

-3, Cl-,

NH4+, Ca+2, Mg+2, K+, Na+

Weekly measurements of concentrations in precipitation (NADP-NTN)

Total Mercury Weekly measurements of concentration in precipitation (NADP-MDN)

Methyl Mercury Monthly measurements of concentration in precipitation (composite)

Selected Trace MetalsAs, Cd, Cr, Cu, Pb, Ni, Se, Zn

Weekly measurements of concentrations in precipitation (MDN Heavy Metal Protocol)

77