admsstar – a model of puff dispersion · admsstar – a model of ... 0 200 400 600 800 1000 1200...

26
CERC ADMSSTAR a model of CERC Presentation to UK Atmospheric Dispersion Modelling Liaison Committee (ADMLC) puff dispersion Health Protection Agency September 2008 David Carruthers and Martin Seaton Cambridge Environmental Research Consultants

Upload: lamduong

Post on 29-Aug-2018

219 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERC

ADMSSTAR – a model of

CERCPresentation to

UK Atmospheric Dispersion Modelling

Liaison Committee (ADMLC)

ADMSSTAR – a model ofpuff dispersion

Health Protection Agency

September 2008

David Carruthers and Martin Seaton

Cambridge Environmental Research Consultants

Page 2: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCContent

• Applications

• ADMS – plumes and puffs

• Requirements

• Puff dispersion model overview

• Spatially varying meteorology

• Variable terrain and surface roughness

• Observed site meteorological data

• Conclusions

Page 3: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCApplications

Currently being implemented in ADMSSTAR 2

• Produced for the Food Standards Agency­ Interested in accidental and explosive releases

• Includes existing features of ADMSSTAR 1• Includes existing features of ADMSSTAR 1­ Radioactive decay

­ Back calculation from sample strengths

­ Calculation of MPL contours

Page 4: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCADMS - Plumes

• Steady state infinite release

Elevation

• Long term results are the average over a series of independent, infinite, steady state plumes

Plan

Page 5: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCADMS - puffs

• Puffs in ADMS 4 also assume steady met conditions but release has finite duration

• Concentration depends on time since release and puff has a defined front and rear

Page 6: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCADMS - puffs

1 - - 180

200

400

600

1 - - 300

200

400

600

1 - - 420

200

400

600

0 200 400 600 800 1000 1200 1400

Metres

-600

-400

-200

0

Me

tres

0 200 400 600 800 1000 1200 1400

Metres

-600

-400

-200

0M

etr

es

0 200 400 600 800 1000 1200 1400

Metres

-600

-400

-200

0

Me

tres

180s 300s 420s

Release lasts until 300s

Page 7: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCRequirements

• For short releases temporal variation of meteorology can be important

• Ability to handle spatially varying meteorology­ flow field: Horizontal and vertical wind speed components

­ boundary layer properties: e.g. boundary layer height, surface heat ­ boundary layer properties: e.g. boundary layer height, surface heat flux

• Generate flow field from terrain data and upwind meteorological conditions

Page 8: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCMethod - overview

Multiple puff approach• Overview

­ A sequence of instantaneous puffs released­ Each puff advected on a time scale smaller than time scale on which

the meteorology changes­ Concentration field is the sum of the concentration fields of the

individual puffsindividual puffs

• However­ Time history of each puff is different, so more complex than steady

state­ Longer run time

• Advantages­ Flexible, allows concentration with time with temporally and spatially

varying meteorology­ Allows incorporation of both small scale and large scale

meteorology

Page 9: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCRelease types

• Release can either be continuous or explosive­ Continuous release has a defined start and end time, a single initial

puff and more are added as the release continues

­ Explosive release is a single event with 5 initial puffs and no more added over time

fraction of CTH

2 = 0.4

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

3

2

1

2 x = 0.2

x or y

2 z = 0.4

5

4

Page 10: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCMethod – details 1

Instantaneous puff

U

Puff properties

• Centre of mass

• Spread parameters

• Amount of material

• Deposition properties

Page 11: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCMethod – details 2

• Centre of mass advected based on wind velocity at mean puff height

)()()(

)()()(

ttUtYttY

ttUtXttX

y

x

• Takes into account plume rise and gravitational settling

• Meteorological parameters calculated at mean puff height

)( )()()(

)()()(

spr WWtUttZttZ

ttUtYttY

z

y

Page 12: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCMethod – details 3

• Spreading parameters, relative to local streamline, calculated based on change assuming current met conditions persisted since start of release e.g.

)()()()( 2222 tttttt yyyy

• With the plume spread assuming the current met conditions persisted since the start of release

calculated using the standard ADMS plume equations

)(ty

Page 13: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCMethod – details 4

• Deposition equations adapted to take into account instantaneous puff approach

• Concentration for individual puff calculated assuming a Gaussian puff (X,Y directions, Z direction like ADMS plume)

• Overall concentration calculated by summing concentrations due to individual puffsdue to individual puffs

• Total accumulated deposition calculated by summing deposition due to each puff at each time step

Page 14: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCResults - concentration

Concentration for a single puff after 15 minutes and 30 minutes. The wind direction is 180o for the first 15 minutes and 235o for the second

187000

188000

189000

187000

188000

189000

15 Minutes 30 Minutes

530000 531000 532000 533000 534000 535000 536000 537000 538000 539000 540000

179000

180000

181000

182000

183000

184000

185000

186000

187000

530000 531000 532000 533000 534000 535000 536000 537000 538000 539000 540000

179000

180000

181000

182000

183000

184000

185000

186000

Page 15: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCResults – continuous release

188000

189000

Concentration and total accumulated deposition results for a continuous release after two 15 minute met periods with wind directions of 180o and 235o

188000

189000

530000 531000 532000 533000 534000 535000 536000 537000 538000 539000 540000

179000

180000

181000

182000

183000

184000

185000

186000

187000

530000 531000 532000 533000 534000 535000 536000 537000 538000 539000 540000

179000

180000

181000

182000

183000

184000

185000

186000

187000

Concentration Deposition

Page 16: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCResults – total deposition

Total accumulated deposition for a 5 minute release over a series of met periods with differing meteorological conditions

Page 17: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCSpatially varying meteorology

• Extend this to be able to deal with spatially varying meteorology

1000

2000

3000

4000

res

1000

2000

3000

4000

res

• Meso scale meteorological data supplied by the Met Office in NetCDF format at 4km or 12km horizontal resolution

-3000 -2000 -1000 0 1000 2000 3000 4000

Metres

-3000

-2000

-1000

0

1000

Met

r

-3000 -2000 -1000 0 1000 2000 3000 4000

Metres

-3000

-2000

-1000

0

1000

Met

r

Hour 1 Hour 2

Page 18: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCSpatially varying meteorology 2

• Spatially varying meteorology includes­ 3D flow fields

­ 3D temperature field

­ 2D rain fall rate, cloud cover, surface roughness and boundary layer height

• This means ADMS met variables need to be calculated across the meteorological grid.

Page 19: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCSpatially varying meteorology 3

• ADMS met processor run at each grid point of the input meteorology grid to calculate met properties

• Dispersion calculations carried out in the same way as for spatially homogeneous met data but now met properties are a function of space and time e.g.a function of space and time e.g.

)),(()()( tttUtXttX x X

Page 20: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCResults – spatially varying meteorology

Total deposition for two runs with the same set up and for the same time period, the only difference is in the source location

760000

780000

800000

1.151.21.25

820000

840000

0 20000 40000 60000 80000 100000 120000 140000 160000 180000 200000

Metres

600000

620000

640000

660000

680000

700000

720000

740000

Met

res

0.050.10.150.20.250.30.350.40.450.50.550.60.650.70.750.80.850.90.9511.051.11.15

160000 180000 200000 220000 240000 260000 280000 300000 320000 340000

Metres

660000

680000

700000

720000

740000

760000

780000

800000

Met

res

Page 21: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCFLOWSTAR

• Use FLOWSTAR (ADMS mean flow and turbulence model) to incorporate effects of­ Variable terrain height

­ Variable surface roughness

­ Both

• This can be incorporated within the spatially varying meteorology framework with an upstream condition for FLOWSTAR determined from the spatially varying meteorology

Page 22: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCFLOWSTAR

Horizontal locations of flow field data for variable terrain embedded within the meso scale meteorology data. The extent of the terrain region is given by the green box.

Page 23: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCResults - FLOWSTAR

80000

85000

11001140 80000

85000

Without terrain With terrain

Total deposition for two runs with the same set up and for the same time period, except that one run uses variable terrain height

160000 165000 170000 175000 180000 185000 190000

Metres

55000

60000

65000

70000

75000

80000

Me

tre

s

20601001401802202603003403804204605005405806206607007407808208609009409801020106011001140

160000 165000 170000 175000 180000 185000 190000

Metres

55000

60000

65000

70000

75000

80000

Me

tre

s

Page 24: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCObserved site met data

• With spatially varying meteorology can include site observed values of­ Wind speed and direction

­ Surface temperature

­ Rain fall rate

• Used to improve calculations near to the source

• Observed site data relaxes over a radius of influence back to the meso scale meteorological data

Page 25: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERC

Total deposition for two runs with the same set up and for the same time period, except that one run has site observed data changing the wind speed near the source

Results – observed site data

No site data Wind on site 315o

186000

188000

190000

14

TotDepBq/m² Cs-137 - - -

186000

188000

190000

520000 522000 524000 526000 528000 530000 532000 534000 536000 538000 540000

Metres

170000

172000

174000

176000

178000

180000

182000

184000

Met

res

0.05

1

2

3

4

5

6

7

8

9

10

11

12

13

520000 522000 524000 526000 528000 530000 532000 534000 536000 538000 540000

Metres

170000

172000

174000

176000

178000

180000

182000

184000

Me

tre

s

Page 26: ADMSSTAR – a model of puff dispersion · ADMSSTAR – a model of ... 0 200 400 600 800 1000 1200 1400 Metres-600-400-200 0 M e t r e s ... • ADMS met processor run at each grid

CERCConclusions

• Puff dispersion model.­ Release modelled as a set of instantaneous puffs

­ Individual puffs properties updated based on local, in time and space, meteorological conditions

• Allows short term releases to be modelled in spatially and temporally varying meteorologytemporally varying meteorology

• Allows complex terrain to be modelled

• Allows observed meteorological data at the source to be incorporated

• Validation and verification of the model

• Further developments …