Environmental Engineering Reference
In-Depth Information
which the plume is discharged. TAPM uses a simplified chemical scheme and the
smog parameter largely determines the oxidation rates of NO
2
and SO
2
0
,
1
. The
external emulator from 8 runs of the model generates the health impact and
uncertainty shown in
Fig. 2.
The uncertainty is large and with realistic emissions
one would anticipate much lower numbers of life years lost (of order of a few
hundred) corresponding to values close to the origin. Larger values were used in
the calculation to avoid rounding errors in the code and to investigate the functional
dependence on the smog parameter. This represents a preliminary application
illustrating how the emulator can be used to investigate the behaviour of a
complex code. In future trials it is planned to generate an emulator from runs of
the more complex chemical transport model CMAQ (Yu et al
.
, 2008).
λ λ
Uncertainty in life years lost
Life yrs lost = f(NO3 production,vg(PM))
50
25
20
25
30
700
45
600
650
750
40
15
35
30
20
25
25
20
15
20
600
650
700
750
10
15
20
10
20
15
20
650
25
25
700
750
30
35
15
700
20
30
40
750
800
50
45
35
25
Fig. 1.
Life years lost and uncertainty (standard deviation of probability distribution) as function
of oxidation rate of NOx (x-axis) and the deposition velocity of particulate matter (y-axis) for an
analytical footprint model
Lives lost =fn(total emission,smog)
0.4
0.4
4000
3000
3 10
4
2000
3 10
4
2.5 10
4
0.3
6000
0.3
2.5 10
4
3000
5000
2 10
4
2000
4000
8000
2 10
4
1000
1000
0.2
0.2
9000
1.5 10
4
1.5 10
4
2000
3000
7000
6000
4000
5000
3000
8000
0.1
1000
0.1
1 10
4
1.5 10
4
7000
1 10
4
4000
6000
5000
2000
9000
8000
5000
0
0
0
2
4
6
8
10
0
2
4
6
8
10
Fig. 2.
Life years lost and uncertainty (standard deviation of probability distribution) as a
function of emission rate (x-axis) and smog
(y-axis) for the TAPM simulation
