Applications of Models-3

Report
Some recent studies
using Models-3
Ian Rodgers
[email protected]
Presentation to APRIL meeting London 4th March 2003
Outline
• Why Models-3?
• Studies currently in progress using Models-3
framework for U.K. air quality and deposition
• Focus on particulates:
– Particulate modelling in CMAQ
– Some initial results from a validation study
• Conclusions and issues to resolve
Why Models-3?
• JEP wants to have good understanding of effects of power
station operation on the environment
– Key issues include air quality, health effects,
ecosystem effects
• Wanted single framework to address these issues:
– Multi-issue
– Multi timescale (episode to annual)
– Flexibility of spatial scale (local – regional –
transboundary)
– Adequate treatment of large elevated point
sources
– ‘future proof’ science
• Models-3/CMAQ emerged from review of available models
as best meeting requirements
Approach to development of
Models-3 capability
• Build Models-3/CMAQ domain to include a large enough
region of Europe to provide good boundary fluxes for U.K.
• Use nested grid capability to provide 4km grid resolution for
test region in Midlands
• Build key inputs
– Emissions
– Meteorology
• ‘unsolder’ hard wired features!
• Test performance of ‘vanilla’ model
– ‘one atmosphere’ approach
Current studies
• A number of studies are currently being
performed:
– Validation of Models-3 wet deposition modelling
– Full-year simulation of air quality and deposition
for the U.K. (joint JEP/Environment Agency
study)
– Validation of Models-3 particulate modelling
– Testing of Models-3 plume rise ad plume in grid
methodologies
Some illustrative results from the particulates study
will be shown
Particulates in Models-3
• CMAQ aerosols module derived from USEPA Regional
Particulate Model
• Modal approach: PM represented by 3 lognormal
modes
– Nucleation mode
– Accumulation mode
– Coarse mode
• Simulates 3 integral properties of each mode:
– Number concentration
– Surface area concentration
– Mass concentration
Particulates in Models-3 (2)
• Particulate species:
2-
-
+
– Fine: SO4 , NO3 , NH4 , elemental C, liquid
water, organic species (primary, anthropogenic
secondary, biogenic secondary)
– Coarse: soil-derived material, sea salt,
unspecified anthrop. species
Particulates in Models-3 dynamics
Primary emissions
0.1%
99.9%
Aitken mode
nucleation
condensation
coagulation
within-mode
In-cloud
scavenging
Accumulation mode
Inter-mode
condensation
coagulation
within-mode
CCN
Coarse mode
Validation of Models-3 for
atmospheric particulate
concentrations
• Joint Environmental Programme funded study
• Comparison of Models-3 results with TEOM
measurements from JEP monitoring stations at Cliffe,
Thorney, Bottesford
• Initial simulation run for Winter (January) and Summer
(August) condition
• ‘Vanilla’ model parameterisation (no model tuning for
particulates)
• Modelled values are with water removed
• Work in progress – example results shown for Cliffe
(only site with both PM2.5 and PM10 data)
Cliffe monitoring site
location
31/01/1999
30/01/1999
29/01/1999
28/01/1999
27/01/1999
26/01/1999
25/01/1999
24/01/1999
23/01/1999
22/01/1999
21/01/1999
20/01/1999
19/01/1999
18/01/1999
17/01/1999
16/01/1999
15/01/1999
14/01/1999
13/01/1999
12/01/1999
11/01/1999
10/01/1999
09/01/1999
08/01/1999
07/01/1999
06/01/1999
05/01/1999
04/01/1999
03/01/1999
02/01/1999
01/01/1999
Particulate matter concentration (ug/m3)
Results - Cliffe, January
Figure 3. 24 hour average modelled and measured particulate
matter concentrations at Cliffe, January 1999
90
80
70
60
50
40
Models-3
PM10
PM2.5
30
20
10
0
27/08/99
26/08/99
25/08/99
24/08/99
23/08/99
22/08/99
21/08/99
20/08/99
19/08/99
18/08/99
17/08/99
16/08/99
15/08/99
14/08/99
13/08/99
12/08/99
11/08/99
10/08/99
09/08/99
08/08/99
Particulate matter concentration (ug/m3)
Results - Cliffe, August
Figure 6. 24 hour average modelled and measured particulate
matter concentrations at Cliffe, August 1999
35
30
25
20
15
Models-3
PM10
PM2.5
10
5
0
Cliffe – scatter plot, January
Figure 9. Relationship between 24 hour average modelled and
measured PM10 data at Cliffe, January 1999
Modelled concentration (ug/m3)
120
100
80
60
y = 1.081x
R2 = 0.4731
40
20
0
0
10
20
30
40
Measured concentration (ug/m3)
50
60
Cliffe – scatter plot, August
Figure 12. Relationship between 24 hour average modelled and
measured PM10 data at Cliffe, August 1999
Modelled concentration (ug/m3)
35
30
25
20
y = 0.702x
R2 = 0.2337
15
10
5
0
0
5
10
15
20
25
Measured concentration (ug/m3)
30
35
01/01/99
02/01/99
03/01/99
04/01/99
05/01/99
06/01/99
07/01/99
08/01/99
09/01/99
10/01/99
11/01/99
12/01/99
13/01/99
14/01/99
15/01/99
16/01/99
17/01/99
18/01/99
19/01/99
20/01/99
21/01/99
22/01/99
23/01/99
24/01/99
25/01/99
26/01/99
27/01/99
28/01/99
29/01/99
30/01/99
31/01/99
01/02/99
Particulate matter concentration (ug/m3)
160
140
120
2500
100
2000
80
60
1500
40
1000
20
500
0
0
Boundary layer height (m)
Figure 21. Boundary layer heights and predicted peaks in
particulate matter at Cliffe, January 1999
3500
3000
Boundary layer
Particulate matter
Discussion & requirements
for further work
• A nested-grid implementation of Models-3 for the UK
has been built & testing for AQ and deposition is under
way
• For PM:
– Generally good level of agreement
– 60% of 24 hr average results within a factor of 3 of
measurements
– Tendency to underestimate PM10 when background
levels low
– Peaks in ground level PM in model when bl low.
– Peaks are primarily associated with long-range transport
rather than local sources or groups of sources
The next steps
• Data mining of Gbytes of data generated to date
• Detailed testing of impact of choice of science
representation for JEP company priority issues
• Use of Models-3 for effects assessment and case
making
• Comparison of performance with other models
• Would welcome interaction with other users / potential
users
• (Potential framework to link European-wide issues to
national issues)

similar documents